





Site Visits Related to Combustible Dust
Synthesis Report




                                                                               
                                                                 	Prepared for:
U.S. Department of Labor
                                               	Occupational Safety and Health 
                                                                 Administration
                                         	Directorate of Standards and Guidance
                                                                               
                                                                               
                                                                              	
                                                                               
                                                                  	Prepared by:
                                                  	Eastern Research Group, Inc.
Lexington, MA 02421
                                                                               
                                                                               

March 23, 2011
                               Table of Contents
                                       
Introduction and Project Overview	1
1	Industry and Facility Background	2
2	Combustible Dust Definitions Used	3
3	Hazard Recognition	5
4	Hazard Assessment	6
5	Hazard Communication and Training	8
5.1.	Content of Material Safety Data Sheets (MSDSs)	9
5.2.	Training	12
6	Consensus, Industry, and Insurance Standards Used	12
7	State and Local Code Enforcement	14
8	Engineering Controls	15
8.1.	Primary Engineering Controls	15
8.2.	Secondary Engineering Controls	19
9	Administrative Controls	24
9.1.	Housekeeping Procedures	24
9.2.	Other Administrative Controls	26
10	Emergency Response	29
11	Investigation of Incidents	29
12	Regulatory Approaches	30
13	Economic Impacts and Benefits	34
14	Impacts on Small Entities	39
15	Compliance Assistance	39
16	References	40



Abbreviations

ERG		Eastern Research Group, Inc.
FIBC		flexible intermediate bulk containers
Kst			deflagration index
LOC		limiting oxygen concentration
MEC	 	minimum explosible concentration
MIE		minimum ignition energy
MSDS		Material Safety Data Sheet
NFPA		National Fire Protection Association
OSHA		Occupational Safety and Health Administration
pmax			maximum explosion overpressure 
PPE		personal protective equipment
PSM		process safety management
μm			micron (or micrometer)

 Introduction and Project Overview 
Under contract to the Occupational Safety and Health Administration (OSHA), Eastern Research Group, Inc. (ERG) conducted site visits to 11 industrial facilities to gather information on the development and implementation of combustible dust recognition, prevention, and protection programs. This report documents common themes and important observations from the 11 site visits. More detailed insights on the individual visits can be found in the facility-specific site visit reports.
ERG implemented this project according to specifications in a Project Protocol (ERG, 2008), which OSHA reviewed and approved in December, 2008. As the protocol describes, ERG recruited facilities to host site visits through telephone calls, focusing on industrial sectors identified by OSHA. The 11 site visits were conducted between April, 2009 and December, 2010. During that time, ERG visited two food manufacturing facilities, two metals processing facilities, a paper mill, a pharmaceutical manufacturing facility, a furniture manufacturing facility, a coal-fired power plant, two facilities with wet corn milling operations, and a sulfur processing facility. Each visit lasted 1 to 3 days and was conducted by chemical engineers from ERG and a fire protection engineer, a consultant to ERG. OSHA personnel attended three of the site visits. 
Information on the facilities' combustible dust safety programs was obtained during guided facility tours; interviews with operators, process engineers, safety personnel, and facility managers; and document reviews. Although the site visits were not designed to assess facility adherence to National Fire Protection Association (NFPA) consensus standards or compliance with OSHA regulations, this report identifies certain instances where facilities apparently may not be meeting requirements of these guidelines and regulations. During some visits, samples of selected materials were collected and sent to OSHA's laboratory in Salt Lake City and analyzed for explosibility and combustibility parameters. 
The remainder of this report documents common themes and salient findings from the 11 site visits, based on the judgment of the site visitors. While this report presents summary information on selected facilities' perspectives and experiences with combustible dust, the report does not document every observation made during the site visits. Readers should refer to the individual site visit reports for a more complete account of observations made throughout this project. Moreover, because the project is based on visits to only 11 industrial facilities, the report should not be viewed as a comprehensive nationwide evaluation of current combustible dust safety programs. 
The remainder of this report organizes the site visitors' observations into 15 different topics. These topics were selected because they parallel the organization of information in OSHA's Advanced Notice of Proposed Rulemaking (ANPR) for combustible dusts (OSHA, 2009b). Some observations are relevant to multiple topics. Rather than repeating these observations throughout the report, such observations are described in detail once, and then referenced in other sections as appropriate. 

Industry and Facility Background
Site visits were conducted at two facilities in the food manufacturing industry, two facilities in the metals processing industry, one paper mill, one pharmaceutical manufacturer, one furniture manufacturing facility, one coal-fired power plant, two facilities with wet corn milling operations, and one sulfur processing facility. Facilities were visited at locations throughout the United States. The facilities varied greatly in terms of many factors: the number of employees at each facility ranged from less than 20 to more than 2,000; some facilities had only a few production lines while others had more than 10; and some production processes were more than 30 years old while others were constructed within the last 5 years. The site visit reports include more detailed information on each facility visited, and this section describes the facilities' in-house and corporate resources for addressing combustible dust safety issues: 
   # The facilities had varying in-house capacity for conducting technical analyses and evaluations of combustible dust safety issues. Specifically, all but the smallest facility had between one and five employees who worked nearly full-time on occupational safety issues, and these individuals were responsible for implementing and overseeing a broad range of safety (and environmental) programs. At the smallest facility visited, only one person worked part-time on health and safety issues. The individuals who worked on combustible dust issues had varying academic backgrounds and professional experience. 
   # Every facility visited was part of a larger company; these parent companies had varying levels of influence on the individual facilities' combustible dust safety programs. At one extreme, a corporation had its own "process safety" staff that was dedicated to providing technical support on combustible dust (and many other safety issues) to the company's individual facilities. This corporation was very actively engaged with facility-specific safety issues and issued corporate-wide policies on many topics related to combustible dust (see Section 6 for further information). At the other extreme, several facilities received minimal support from corporate officials on combustible dust safety issues. 
   # When characterizing combustible dusts or constructing or implementing major modifications to processes involving combustible dust, the facilities relied to varying degrees on external engineering and design firms and consulting companies. Some facilities relied entirely on these external resources, while others sought external engineering and design support only for specialized tasks not adequately covered by in-house expertise. 
   # The facilities' experiences with insurance underwriters varied. Each facility's insurance underwriter reportedly conducted at least annual visits to evaluate fire protection measures and other safety issues. One facility whose parent company is self-insured is audited annually by an outside loss prevention consultant. Some insurance underwriters played a very active role in informing and educating facilities about combustible dust safety issues, while other insurance underwriters offered limited to no insight on the matter. One insurance underwriter was noted as being particularly engaged in facilities' combustible dust safety programs (see Section 6 for more information). 

Combustible Dust Definitions Used
The facilities visited for this project typically deferred to NFPA definitions for identifying combustible dusts. However, the various definitions published in current and former NFPA standards are not consistent and may change in future editions of these standards. Specific definitions cited during the site visits included "combustible metal dust" as defined in NFPA 484; "combustible dust" as defined in NFPA 499 and NFPA 654; and "agricultural dust" as defined in NFPA 61. These definitions differ primarily in terms of whether they specify particle size cut-offs. 
Given the lack of a universally accepted definition of combustible dust, nearly every facility representative expressed interest in precisely how OSHA's standard will define the term. Some facility representatives preferred a definition based on objective criteria (e.g., specified ranges of particle sizes, deflagration indices, and so on) while others facility representatives advocated for more open-ended definitions (e.g., "combustible dust" as currently defined in NFPA 654). See Section 12 for additional discussion of this issue.
The majority of facilities visited in this project had some laboratory analytical data that characterized materials that could be classified as combustible dusts. Detailed information on these testing data can be found in the facility-specific site visit reports. The testing data that ERG reviewed quantified various parameters, including particle size distribution, deflagration index (Kst), moisture content, maximum explosion overpressure (pmax), minimum explosive concentration (MEC), and minimum ignition energy (MIE). Facility representatives raised the following issues when discussing how laboratory analytical data might be used to help define which materials are combustible dusts:
   # While facilities recognized the need to characterize their materials, facility representatives also noted the need for direction and guidance from OSHA on many aspects of testing combustible dusts: What materials must be tested (e.g., raw materials, intermediates, products, settled dusts, material from dust collectors)? How often must this testing occur? What parameters must be measured? Will facilities or their suppliers be expected to provide testing data? Will OSHA publish cut-offs for particle size, moisture content, and other properties when defining combustible dusts? What laboratory analytical methods must be used? Under what circumstances is it acceptable for facilities to refer to published data for comparable materials? 
   # Facility representatives generally acknowledged the importance of having extensive testing data for their materials. However, some facilities -- particularly the larger facilities that used many different combustible materials in their processes -- expressed concern about the costs associated with running a full battery of laboratory tests on all materials that are, or may generate, combustible dusts. (Refer to Attachment 1 of this report for laboratory testing costs that commercial laboratories quoted for two facilities.) To illustrate this concern, one of the food manufacturers processed dozens of different dry ingredients, which were subsequently milled, mixed, and blended into many mixtures; and these various materials were found throughout the facility in different particle size fractions (e.g., finer dusts settled on surfaces, coarser material captured in dust collectors). According to representatives at this facility, hundreds of laboratory tests would be required to fully characterize these materials.  
   # Most facilities shared testing data with the site visitors. The facilities generally tested bulk raw materials and products believed to pose the greatest explosion hazards, but these materials may not have reflected hazards associated with settled dusts in the workplace or material from dust collectors. Even though the settled dusts and collected dusts are chemically similar to their respective raw materials and products, these dusts can have considerably varied particle size distributions and therefore would be expected to exhibit different explosion hazards. Facilities would likely benefit from guidance on how to select materials for testing, especially when limited numbers of samples will be collected (e.g., testing for milled products instead of un-milled products, consideration for testing of settled dusts or material collected in dust collectors). Some facility contacts stated that they were not sure what parameters should be measured by laboratories. 
   # One facility representative said suppliers, under OSHA's Hazard Communication standard, should be obligated to provide facilities a broad range of testing data for combustible dusts and materials that can generate combustible dusts under normal processing and use. Section 5 of this report expands on these concerns. 

Hazard Recognition
Employees at every facility visited exhibited strong awareness of which materials are combustible dusts. Facilities generally did not assign specific employees responsibility for making this determination. Rather, identification of combustible dusts occurred through various means: experience from process operators, engineers, and safety staff; input from corporate officials and insurance underwriters; "common knowledge" (e.g., facilities that handled certain pyrophoric metallic dusts did not need to conduct research to determine that these materials were combustible); and rules of thumb (e.g., finely milled grain products and starches are likely combustible dusts). 
While facilities were generally aware that Material Safety Data Sheets (MSDSs) contain information about potential combustible dust hazards, MSDSs did not seem to be a primary source of information for identifying combustible dusts. Refer to Section 5.1 for facility feedback on MSDSs and the role that suppliers play in providing quantitative information on potential combustible dust hazards. Representatives from all 11 facilities said that local fire authorities offered no useful insights on determining which materials are combustible dusts. 
When seeking quantitative information on combustibility and explosibility, facilities tended to have samples tested by laboratories for various parameters or seek reported values from the literature. In general, most facilities had materials tested only once for explosibility parameters and did not re-test these materials, presumably because the material composition did not change considerably with time. One facility had extremely extensive testing data available, both in terms of the number of materials that have been tested and the number of parameters that were measured. On the other hand, at least one facility did not have any combustibility or explosibility testing performed on any of its combustible dusts. 
Some facilities apparently assumed that testing data for a given material, whether reported by a laboratory or taken from a publication, apply wherever that material is found in the same chemical form, without regard for changes in physical properties (e.g., particle size distribution, moisture content). This assumption may have led to facilities incorrectly concluding that certain materials were not combustible dusts. For example, one facility used testing data for its bulk product to conclude that a sugar compound was not a combustible dust. However, the mean particle size for the bulk product was considerably greater than that of settled dusts and one of the facility's milled products; and these finer materials -- though chemically identical to the bulk product -- would be expected to exhibit different explosibility properties. As another example, multiple facilities referred to published parameters (e.g., deflagration indexes, minimum ignition energies) for certain materials when evaluating potential combustible dust hazards, even though those published accounts did not document the particle size of the tested material. 
Note: 	ERG contacted dozens of facilities before identifying the 11 facilities that agreed to host site visits. When other facilities were asked why they chose not to participate, the facility representatives often asserted that their operations did not generate combustible dusts, even though some worked in industries known to have more frequent combustible dust fires or explosions (e.g., aluminum foundries). Thus, it is possible, if not likely, that facility recognition of combustible dust hazards is much poorer than implied by the facilities that were visited in this project. 

Hazard Assessment
ERG asked representatives from all 11 facilities if they have conducted hazard analyses on processes involving combustible dusts. Examples of some facility practices follow:
   # One facility followed its corporate policy of applying process safety management (PSM) requirements to every chemical production operation, regardless of whether the process included chemical usage that would actually trigger PSM applicability. Due in part to this policy, this facility conducted detailed hazard analyses on every process involving combustible dusts. Groups of up to 12 employees -- typically safety professionals, engineers, operators, and maintenance personnel -- conducted the hazard analyses and addressed many topics included in PSM hazard analyses: human factors, siting, operating procedures, process safety information, mechanical integrity, and incident investigations. 
   # One facility developed a spreadsheet-based approach to prioritize its resources for retrofitting existing processes with controls to minimize combustible dust hazards. This facility had hundreds of unit operations that processed materials containing combustible dust and developed a scoring scheme to determine which unit operations should receive immediate attention. In this scheme, each unit operation was ranked according to several factors: fuel properties (e.g., type, quantity, particle size, potential for suspension); operating temperatures; potential ignition sources (e.g., static, friction, hot work); proximity to employees; and history of past fires and explosions. The outcome of this scoring scheme was a composite rank that was qualitatively indicative of each unit operation's combustible dust hazard. ERG site visitors found this "risk assessment" approach to be a useful tool in ranking hazards and prioritizing control upgrades for individual process.   
   # Another facility established "Dust Explosion Protection Hazard Analysis Criteria" to determine what specific controls may be warranted on all newly installed equipment. The criteria were organized into a decision tree that asked a series of questions (e.g., Is dust combustible? Is equipment confined? Is the particle size less than 420 microns? Are ignition sources present? Has explosibility testing been conducted?). Based on answers to these and other questions, the decision tree would indicate whether equipment explosion protection is required, whether fire protection options are necessary, and other design considerations. A second tier of questions then provided additional specificity on the required controls. For instance, if the first tier of questions indicates that equipment explosion protection is needed, the second tier of questions would help determine what type of protection (e.g., venting, flame quench, suppression) is most appropriate. In short, this facility's hazard analysis criteria walked process safety engineers through the process of determining whether protection is required and, if so, what type of protection is preferred. 
   # One facility follows corporate guidelines that present a series of technical evaluations that should be conducted for any process with a dust explosion potential. These evaluations begin with a preliminary hazard assessment, in which material-specific combustibility and explosibility parameters are collected and evaluated. The guidelines recommend extensive testing and explain how particle size and moisture content should factor into the testing strategy, but they also list several references with compilations of combustibility and explosibility parameters for dusts commonly encountered in their industry. A flow chart is used to recommend specific hazard mitigation strategies based on material properties (e.g., MIE, minimum ignition temperature of a dust cloud, Kst, thermal instability temperature) and production operations. These parameters are presented in a matrix to assign a "low," "medium," or "high" explosion rank risk. These combustible dust evaluations are in addition to the facility's more generic process hazard analyses.
   # The parent company of another facility issued a written procedure on "Combustible Dust Management" after OSHA launched its National Emphasis Program for combustible dust. The procedure, which is designed to identify and control potential combustible dust hazards, requires facilities to use a compliance checklist to conduct "initial dust assessments." The checklist was designed to help facility representatives identify action items for further reducing potential safety hazards. Questions on the checklist fell into several categories, including electrical classification designations, ignition control, dust collectors, size reduction equipment, fire protection, hazard communication, and housekeeping. At the time of the site visit, the facility had completed the assessment for the first time. The assessment is supposed to be revisited annually or whenever the facility undergoes any major changes to its equipment, materials, or processes.
      
Hazard Communication and Training
Though the site visits were not designed to assess the overall effectiveness of facilities' hazard communication programs, this section documents several observations pertaining to hazard communication in the context of combustible dusts. Observations are organized into two topics: information documented on Material Safety Data Sheets (Section 5.1) and information conveyed during employee training programs (Section 5.2). 
Content of Material Safety Data Sheets (MSDSs)
During site visits, ERG reviewed approximately 150 Material Safety Data Sheets (MSDSs) for both (a) materials suspected to be combustible dusts and (b) materials that could generate combustible dusts under normal conditions of use. The extent of documentation of potential combustible dust safety hazards varied greatly across these MSDSs. In cases where MSDSs acknowledged and characterized combustible dust safety hazards, relevant information was typically found in sections on firefighting measures, accidental release measures, handling and storage, and stability and reactivity. Following are several general observations about the MSDSs that the site visitors reviewed:
   # Quantitative information. MSDSs exhibit great variability in terms of quantifying flammability and explosibility: most MSDSs presented only qualitative descriptors of hazards, and a relatively small fraction included quantitative data on various parameters. Of the MSDSs reviewed, the parameters that were most frequently quantified were minimum explosible concentration (MEC), lower explosive limit, and cloud ignition temperature. Following is an excerpt from the MSDS that included the most detailed quantitative information (in this case, for a modified food starch):
         Lower explosive limit: 	60 g/m[3].
         Starch is a class St1 dust at normal moisture level.
         Minimum ignition energy (MIE): > 30 mJ at normal moisture level.
         Pmax: 9.5 bar
         Kst: 170 bar-m/s
         Layer ignition temperature: > 450 [o]C.
         Auto ignition temperature: 170 [o]C. (Above this temperature starch will self-heat.)
      Facility representatives also exhibited different expectations regarding the anticipated level of quantitative information in MSDSs. One facility contact noted that his company processes more than 150 raw materials that are likely considered to be combustible dusts; however, the facility received complete testing data for only 35% of these materials. The specific testing data requested from suppliers and manufacturers, in this case, were MEC, Kst, MIE, and pmax. This facility recommended that OSHA clarify whether the supplier or facility is obligated to test for these parameters. One facility was able to comment on the consistency of information found on MSDSs with information generated by a testing laboratory. The facility found that testing data occasionally conflicted with data provided on MSDSs; for example, for one tested dust, the supplier's MSDS reports a lower dust explosion class (St1) than the facility's testing data (St2).
   # Qualitative information. Most of the MSDSs that ERG reviewed had some qualitative characterization of potential fire and explosion hazards; however, the level of detail widely varied. At one extreme, certain MSDSs included very generic information (e.g., "avoid dusty conditions") without acknowledging the full range of hazards and how they can be minimized or controlled. At the other extreme, some MSDSs offered much more detailed and thorough characterization of the potential hazards and means for controlling them.  For example, one MSDS included the following:
            Sources of ignition can start a fire on fine particle sizes... Layers of 3-micron diameter dust are susceptible to spontaneous ignition. [Material] burns with very little flame but with large quantity of heat... Do not spray water on burning powder, fines, chips, or dust as violent explosion may occur. This hazard increases with finer particles. If a fire starts in a mass of wet metal fines, such as a drum of damp machining chips, an explosion and a very high temperature flash radiation may follow the initial fire. Therefore, when in doubt, individuals should leave and not attempt to extinguish the fire but let it burn out. The explosion characteristics of such material are caused by the hydrogen and steam generated by the burning mass... Carbon dioxide is not effective in extinguishing burning [material] and will accelerate combustion. Fire extinguishers should not be pointed directly at the burning material, so as not to stir up and aggravate the fire. 
   # Inclusion of hazards for "normal conditions of use." According to OSHA's recent guidance on hazard communication for combustible dusts (OSHA, 2009c), suppliers and importers must consider the potential hazards of materials "that may occur under normal conditions of use" and address known hazards on MSDSs. The MSDSs that ERG reviewed exhibited considerable variability in meeting this requirement, particularly in terms of various food items sold in bulk but subsequently milled at other facilities and metal products sold as "articles" (e.g., billets, ingots, rods, plates) but subsequently machined at other facilities. 
      To illustrate this point, ERG reviewed MSDSs for two bulk metal "articles" that facilities sold to their customers, who then machined the products. Both of these articles were pure forms of combustible metals listed in NFPA 484. One MSDS provided no information on potential hazards associated with machining the object, but the other MSDS included the following disclaimer: 
            This metal in its present state is an "article" as defined in 29 CFR 1910.1200. However, changing its shape or form may promote a chemical or fire hazard... Shiny, odorless, metal that presents little or no unusual hazard if involved in a fire in its solid mass state. When in a particle state, it can be flammable to pyrophoric depending on its size, surface area, and if wet or damp.
      ERG also notes that one supplier completely avoided the issue of hazards associated with normal conditions of use by claiming on an MSDS: "...we cannot predict the uses of [wheat flour] and deny any liability for injuries or illnesses that might be occasioned by unanticipated or non-normal uses." As a result, the supplier provided limited information on this material's MSDS, and the facility that purchased and subsequently processed the wheat flour may not have been fully informed of potential hazards.
   # Regulatory interpretations. In some cases, suppliers of specific food items provided facilities letters or other communications claiming that certain materials were exempted from OSHA's Hazard Communication standard. The rationale provided by these suppliers follows:
            (1) Food grade products intended for edible uses do not require Material Safety Data Sheets (MSDS) to comply with OSHA regulation 29 CFR 1910.1200 (The Hazard Communication standard). Under the standard, MSDSs are only required for hazardous chemicals. Hazardous chemicals are defined as any chemical which is a physical or health hazard. Physical or health hazards are further defined and Appendix B of the standard describes criteria used to determine whether or not a chemical is to be considered hazardous. Foods regulated under The Food, Drug, and Cosmetic Act (which all of our foods and food ingredients are) do not meet any of the criteria which require them to be listed as hazardous. Therefore, our rices and rice by-products do not require MSDSs.
            
            (2) [Company] believes its wheat flour and other milled wheat products, for human consumption, are not hazardous materials according to the OSHA Hazard Communication standard (29 CFR 1910.1200). These products are labeled according to the Federal Food, Drug, and Cosmetic Act and are exempt from further OSHA or EPA labeling requirements.
      Similarly, some suppliers are under the impression that MSDSs are not required for "natural products," like coal, and therefore do not provide MSDSs to their customers. 
      The previous accounts are included in this report to inform OSHA of circumstances by which facilities may not be fully informed of potential combustible dust hazards. It is outside the scope of this project to comment on whether these suppliers' regulatory interpretations are correct. 
         #       MSDSs from Foreign Suppliers. One facility noted that they have encountered some difficulty obtaining material-specific hazard information from foreign suppliers, who might not be subject to minimum disclosure and reporting requirements set forth in OSHA's Hazard Communication Standard. In cases where suppliers translate hazard information into English, facility representatives voiced concern about the quality of translations in some instances, particularly with regards to highly technical detail.
Training
Every facility visited (for which information on training was available) offered numerous different initial and refresher training programs to employees, contractors, and visitors. Training topics varied across the facilities, but typically addressed plant security, personal protective equipment, emergency response, standard operating procedures, lockout/tagout, confined space entry, hot work permits, smoking, and chemical safety. Most facilities had at least one course that addressed combustible dust as part of the training on another topic, but few facilities developed courses devoted entirely to combustible dust safety issues. 
Multiple different delivery methods were used for the training, including classroom settings, videos, online courses, interactive courses on CD-ROM, and on-the-job training. A few facilities use computerized systems to manage their training programs. These programs track training requirements and ensure that employees are up-to-date on all required training. As time permitted, site visitors viewed the various training materials, some of which did not mention unique hazards of combustible dusts, while others offered detailed information on this topic. 

Consensus, Industry, and Insurance Standards Used
The facilities listed various standards used for identifying combustible dusts and controlling and mitigating their hazards. Specific examples follow:
   # NFPA. Nearly every facility that ERG visited was aware of NFPA standards on combustible dusts (e.g., NFPA 61, 484, 654), though the awareness of specific requirements varied greatly across these facilities. Some facilities had employees who are actively engaged in the development of NFPA standards, either by serving as Technical Committee members or by commenting on proposed revisions to the standards. In one case, the facility representatives that ERG met during a site visit exhibited close to no awareness of potentially applicable NFPA standards, including the standard developed specifically for that industry. 
   # Industry. The 11 facilities that ERG visited were not aware of any industry standards specific to combustible dusts. However, most of these facilities are members of industry-specific trade associations. Though most of these associations have not published combustible dust safety standards, they offer varying levels of engagement on combustible dust safety issues. Most trade associations identified during site visits were actively tracking OSHA's rulemaking process. Some trade associations had participated in the agency's combustible dust stakeholder meetings, and others had commented on OSHA's advanced notice of proposed rulemaking (ANPR) or encouraged their members to do so. 
      Site visitors learned of specific examples of certain trade associations providing facilities extensive technical information on combustible dusts. For example, the PRB Coal Users Group is currently drafting a combustible dust best practices guideline that will address various related safety topics, such as firefighting considerations for dust collectors and inerting systems for pulverizers. The PRB Coal Users Group has also conducted a webinar on managing combustible dust and developed a 70-minute presentation on combustible dust hazards (distributed on compact disk). The association further maintains a website with extensive resources accessible by member facilities and an online discussion forum where members can share experience on combustible dust safety issues. As another example, a facility informed site visitors that the International Titanium Association may soon be developing a manual to describe safe practices for using and handling titanium (including titanium dusts). 
   # Insurance. The 11 facilities that ERG visited were insured by different underwriters, including one facility that was self-insured under its parent company. Of these, only one (Factory Mutual) was identified as having developed extensive property loss prevention guidelines specific to combustible dust (FM, 2009) and requiring its policyholders to adhere to these guidelines, though exceptions can be negotiated. All 11 facilities noted that their insurance underwriters conducted inspections at least annually (at the self-insured facility an outside loss prevention consultant conducts annual audits), but the extent to which those inspections addressed combustible dust varied. 
      Some facilities identified their insurance underwriter as a trusted resource for current, technical information on combustible dust safety issues, while others noted that they tended to rely on other resources because their insurance underwriter did not seem to offer strong expertise on combustible dusts. In many cases, facilities carefully considered and adopted recommendations made by their insurance underwriters. For example, an insurance underwriter made recommendations that recently led a facility to install spark/ember detection and suppression systems and high-speed abort gates in selected ductwork locations. Following these improvements, the facility's annual premiums decreased by approximately $21,000; the installation of dust controls were believed to be one of many factors that led to these reduced premiums. 
   # Food and Drug Administration (FDA). One facility is typically inspected annually by the FDA to evaluate whether the facility is following "Current Good Manufacturing Practices" (CGMP) to ensure product quality and avoid cross-contamination in products. To comply with these procedures, production equipment is thoroughly cleaned and tested for cleanliness before switching production campaigns. Even though FDA regulations were not promulgated to protect employees from combustible dust safety hazards, their emphasis on product purity has the secondary benefit of minimizing combustible dust accumulations.
   # Other. One facility was part of a corporation that had recently prepared a process safety technical guideline on combustible dusts that applied to all of the corporation's research and manufacturing facilities. These written guidelines were prepared by corporate process safety officials, with input and comment received from individual facilities. Many references were considered when developing these guidelines (e.g., NFPA standards; Eckhoff, 2003; Bartnecht, 1981; Barbauskas, 2003). The guidelines are not publicly available information. 

State and Local Code Enforcement
For all facilities visited, the local fire marshals and fire departments conducted periodic inspections of the facilities, but these inspections focused primarily on general fire detection and suppression principles. The local fire authorities provided virtually no specialized expertise in combustible dust safety issues, particularly for the facilities that were located in sparsely populated areas and served by volunteer fire departments. In most cases, the industrial facilities actually educated the fire marshals (rather than fire marshals educating the facilities) about the unique and specific combustible dust safety issues found at their sites.
Facility representatives also noted that their local fire marshals did not suggest adherence to NFPA standards specific to combustible dust. Therefore, either (1) compliance with various NFPA standards was not required in their jurisdictions or (2) compliance was required but the local fire marshals exhibited limited to no awareness of the potentially applicable NFPA standards and how to evaluate facility compliance. 

Engineering Controls
ERG site visitors observed a broad array of industrial processes, which included many unit operations that potentially posed combustible dust hazards (e.g., dryers, ovens, dust collectors, hammer mills, grinders, pulverizers). This section reviews ERG's observations regarding the various engineering controls facilities employed to prevent an incident from occurring (Section 8.1) or to minimize the effects of an incident after it had been initiated (Section 8.2). The need for engineering controls ultimately depends on the potential hazards associated with the materials being processed. In many cases, lack of extensive testing data complicated efforts to assess the effectiveness of facilities' engineering controls. For instance, because some facilities lacked MIE data for their main products, site visitors could not assess whether sufficient electrostatic prevention measures were in place. Many of the engineering controls that site visitors observed were installed more than 10 years ago, and original design specifications, hazard analyses, and other documentation were no longer available for review. 
Primary Engineering Controls
This section summarizes site visitors' observations regarding primary controls designed to prevent hazards from occurring. Observations are organized according to specific types of engineering controls listed in the combustible dust ANPR. This section covers only those controls that were observed at the 11 facilities that participated in this program:
   # Features to prevent accumulation of dust on surfaces. Nearly every facility faced challenges of preventing dust accumulations on surfaces, particularly overhead horizontal surfaces. One facility addressed this by installing overhead oscillating fans that continuously blew air over and near surfaces where dusts previously accumulated. The fans proved highly effective at reducing accumulations within a 35-foot radius and minimized the need to periodically clean these overhead structures, which was previously conducted manually with explosion-proof vacuum cleaners. Further, emergency stop switches were installed in the working area, which employees were instructed to activate in the event of a fire; this measure was implemented to ensure that fans do not suspend smoldering or burning material and possibly exacerbate a dangerous condition. Industrial hygiene evaluations verified that operation of the fans did not lead to excess noise exposure, and the fans did not adversely affect employee comfort. At another facility, representatives are testing the use of downward-sloping metal coverings for the structural steel on the building walls. Described as "purlins" by the facility, the sloped coverings are intended to reduce dust accumulations due to their steep angle. The coverings were designed by facility employees and have been installed in a small section of the production room. According to facility representatives, vibrations from production machinery are sufficient to prevent dust from accumulating on top of the coverings. The facility is planning to extend their use throughout the entire production area.
   # Oxygen concentration reduction. Six facilities operated certain production processes under inert atmospheres: two facilities used argon, one used nitrogen, two used combustion gases (i.e., primarily carbon dioxide [CO2]), and one used steam. All but one of the processes was equipped with oxygen monitoring devices to ensure that oxidant concentrations remained at safe levels. Processes for which inerting was used include dryers, solids charging in reactors, and milling or pulverizing operations. Original design specifications were not available at these facilities, so site reviewers commented primarily on their potential compliance with specifications in NFPA 69. While process engineers generally knew the oxygen concentration that would trigger process shutdown and other measures, they had less awareness for how that oxygen concentration was derived. One facility was unaware of the limiting oxygen concentration (LOC) for their dusts, which made it impossible to assess the effectiveness of the inerting system. 
      Site visitors had some concerns about the reliability of sampling and monitoring instrumentation used to measure oxygen concentrations -- a critical component of this engineering control. One facility commented on technical difficulties associated with installing continuous oxygen monitors in unit operations that were not stationary, such as rotating blenders. Another facility distrusted the sensor responsible for monitoring oxygen concentration due to the high rate of corrosion of electronic components in the plant atmosphere. As a result, employees used a hand-held gas analyzer to measure instantaneous CO2 concentration in the mill prior to startup of the production equipment and every 15 minutes during production. A measure of CO2 can be used as a proxy for oxygen in this case because the quantitative relationship between CO2 and oxygen concentrations for natural gas combustion gases has been documented. 
   # Foreign material separation devices. Six facilities used equipment (e.g., scalping screens, magnetic separators) to remove metallic scrap in process locations where tramp metal can introduce serious spark-producing hazards, such as upstream of milling operations. Use of these devices in food manufacturing facilities was motivated largely by product safety concerns, with the combustible dust safety viewed as an added benefit. The design of the magnetic separators varied, with only some facilities using devices with self-cleaning designs. At one facility, for example, the magnetic separators were automatically energized whenever the upstream belt conveyor began operating. Whenever the conveyor stopped, diversion gates beneath the magnets were activated and the magnets de-energized, allowing for collected material to drop by gravity into designated tramp metal collection areas.
   # Monitoring and alarms for abnormal conditions. Facilities employed various monitoring systems to detect abnormal operating conditions. Some of these systems monitored parameters for production quality control purposes, while others were implemented specifically to detect potentially hazardous situations associated with combustible dusts. Examples of monitoring systems observed included temperature sensors in dryers and dust collectors, oxygen analyzers in processes required to operate under inert atmospheres, low-level water sensors in wet dust collectors, and carbon monoxide monitors to detect spontaneous coal combustion. In several cases, alarm conditions triggered control systems such as abort gates or sprinklers, or shut down production. Very few instances were observed where facilities operated monitoring devices (e.g., flame sensors, spark/ember detection systems) that might detect the presence of a fire that outlet temperature monitoring equipment might otherwise fail to detect. 
   # Automatic interlocks, shutoffs, or overflow systems. Many examples of automatic interlocks and shutoffs were observed at the 11 facilities. At one facility, for instance, metal cutting operations could not be operated unless the water reservoir in the wet dust collectors was at a certain level; and processes operating under inert atmospheres could not be started until measured oxygen concentrations were safely below the limiting oxygen concentration (LOC) for a specified time frame. At another facility, woodworking equipment was interlocked with the dust control system such that employees could not conduct any woodworking operations unless the dust control system was fully operational. 
      Additionally, site visitors also noted several instances where the absence of interlocks was problematic. At one facility, for example, multiple dust-generating operations vented to a single wet dust collector. Before engaging production activity, operators visually verified that the proper ductwork connections were in place and also checked to see that the wet dust collector was ready for operation (e.g., sufficient water reservoir). In cases like these, engineering controls (interlocks) would help guarantee that dust collection systems were functioning properly before dust-generating operations could be initiated, without introducing the possibility of human error. 
   # Manual emergency controls. Several facilities had manual emergency controls that, when activated by employees, would trigger various instantaneous actions. Examples included employee-activated steam suppression in dryers, argon quenches in metal milling and grinding operations, and fuel cut-offs for gas-fired dryers and roasters.
   # Class II electrical equipment and wiring. The facilities exhibited widely varying appreciation for designating hazardous locations and ensuring use of equipment rated for those designations. Some facilities conducted very thorough process evaluations to identify and delineate hazardous locations, and they also ensured that all equipment operating in Class II locations was approved for the hazard classification. However, other facilities were not as thorough in this regard, whether completely failing to designate Class II locations in the first place (as suggested by sample results collected by site visitors) or using equipment (e.g., propane-powered forklifts, vacuum cleaners) not rated for combustible dust atmospheres in their designated Class II locations. At one facility which did provide a listing of Class II hazardous locations, representatives acknowledged that some motors located in those hazardous locations were not rated for Class II environments in part due to the costs of replacing the motors. Further, multiple facilities struggled with dust accumulations inside electrical equipment enclosures, even when the enclosures were reportedly "dust-tight."  
   # Grounding, bonding, and other electrostatic controls. Facilities adopted different strategies for controlling electrostatic discharge. In some cases, the adequacy of these controls could not be assessed due to the lack of sufficiently informative testing data (e.g., MIE) for certain materials. In other cases, extensive electrostatic controls were applied in some parts of facilities, but not in others. For instance, one facility grounded and bonded its pneumatic conveying lines in some but not all of the production areas involving the same combustible dust materials; further, one facility that processed a highly sensitive material (MIE < 4 mJ) had extensive grounding and bonding of facility equipment, but did not have comparable controls (e.g., use of grounding straps, conductive shoes, static-dissipative clothing) to ensure that employees themselves did not contribute to electrostatic hazards. 
      Other facilities shared challenges associated with electrostatic discharge related to packaging. One facility that is obliged to comply with the FDA's CGMPs has had difficulty identifying an adequately protective plastic lining for holding powder material in fiber-board drums and plastic drums. The facility had considered using dissipative plastic lining bags made from carbon impregnated polymers, but did not pursue this after finding evidence of carbon leaching into the bags' contents. At another facility, packaging operators are trained to always attach two grounding clamps to flexible intermediate bulk containers (FIBCs) prior to filling in order to ground the bag and avoid static discharges. However, the facility only used type A and type B FIBCs, whereas the type C FIBC is the only type that is specifically designed to be grounded.
Secondary Engineering Controls
This section summarizes site visitors' observations regarding secondary controls designed to minimize the effects of an incident after it has been initiated. The observations are organized according to specific secondary engineering controls listed in the combustible dust ANPR. This section covers only those controls that were observed at the 11 facilities that participated in this program:
   # Dust collectors. Every facility that ERG visited had dust collectors, and these varied tremendously in terms of placement, design, operation, and controls. Some facilities had "dry" dust collectors (e.g., baghouses) in which collected material was separated from an air stream and later removed from the device, others had "wet" dust collectors in which collected material was collected in a water reservoir and eventually removed as sludge, and some facilities used both wet and dry collectors onsite. Given the commonplace use of this technology to collect combustible dusts and the frequency with which fires and explosions originate in these devices, the following paragraphs present detailed observations about the facilities' dust collectors. However, most facilities no longer had extensive documentation of the original engineering evaluations and design specifications for their dust collectors.
      A large number of dust collectors that ERG observed were located inside production buildings. Although most dust collectors inside buildings were placed alongside or in close proximity to external walls, many exceptions were observed, primarily for the older facilities and processes. Later sections of this report comment on the potential regulatory implications of the placement of these devices. 
      Many dust collectors observed during the site visits vented exhaust air to the outdoor, ambient environment. However, numerous exceptions were observed where dust collector exhaust air was returned into the workplace, sometimes after passing through one or multiple particulate filters. Two reasons were cited for preferring to return dust collectors' exhaust air to workplaces: 1) environmental permitting regulations in some jurisdictions require facilities to obtain air permits before constructing any new vent from a production area to outdoor air -- a situation that presented a disincentive for venting dust collector exhaust to the ambient environment; and 2) venting exhaust air to the ambient environment increases demand for air conditioning in the summer and heating in the winter. In almost every instance where facilities returned exhaust air from dust collectors to the workplace, no explosion isolation devices were installed at the dust collectors or the ductwork, which raised concern about deflagrations or explosions starting in a dust collector and propagating throughout a facility. 
      The most commonly used explosion protection technology in dust collectors was explosion panels designed to vent material during an explosion through ductwork to the outdoors; however, numerous dust collectors were operating without any such controls. The design of the explosion venting varied greatly, with some examples listed here:
         o The preferred placement and size of explosion panels depends on the shape of the dust collector and the explosibility of the dusts being controlled. ERG did not evaluate these design parameters during the site visits. However, site visitors noted that some panels were placed very high on the dust collectors' walls, raising questions about their ability to release and vent deflagrations before the dust collector itself explodes. 
         o In most indoor dust collectors, the ductwork for explosion vents was relatively short (10 feet or less) and directed outdoors, but in one case, ductwork for one explosion vent was approximately 50 feet long, including a 90-degree bend, and site visitors questioned the effectiveness of this design. In this particular case, construction of a shorter explosion vent was not feasible without completely redesigning an entire process; however, other forms of controls could have been employed to prevent or suppress explosions in the dust collector. 
         o For a few indoor dust collectors, an explosion panel was installed on a dust collector inside a production building, but the panel was not connected to any ductwork: thus, an explosion in this dust collector would vent directly into the workplace. 
         o In a few outdoor installations, dust collectors that were placed alongside a building, and the explosion panels were placed on the side facing the building, rather than the side that faced away from the building. Similarly, another dust collector was placed outdoors but adjacent to a flammable chemical storage area. Site visitors noted that major fires or explosions that start in the baghouse could rapidly spread to the chemical storage area.
      Although many dust collectors that ERG observed were not equipped with measures (e.g., metal detection systems at inlets, spark/ember detection and suppression systems) that would help detect ignition sources or fires inside equipment, several facilities did install such detection systems and equip them with measures that would extinguish fires before they could propagate into larger events. Several examples follow:
         o One facility had linear heat detectors installed on some of its dust collectors, and these detectors activated an argon quench when measurements exceeded a trigger level. 
         o At another facility, most of the dust collectors were equipped with smoke detectors. Alarm conditions would typically activate abort gates in the exhaust ductwork and shutdown airflow through the dust collection system, but would not trigger fire suppression. In some instances, smoke detectors were interlocked directly with sprinklers inside the dust collectors. 
         o One facility took a preventive approach and equipped the inlet duct to the dust collectors with a spark/ember detection and suppression system, which is intended to prevent sparks or smoldering material from entering the dust collector. 
         o Another facility retrofitted their baghouses with continuous carbon monoxide (CO) monitoring devices, which can detect signs of fires or smoldering coal before. These devices output trends in CO concentrations to the control room, where operators determine the proper course of action if "hot spots" are detected. 
         o One facility relied on drawing process air from production equipment operating in an inert atmosphere to prevent fires in its dust collector. However, site visitors suspected that the introduction of ambient air from the compressed air used to clear the filters and from openings in the dust pickup duct connections reduced the efficacy of the system. 
      In terms of dust collector types, two facilities reported that they were in the process of replacing "dry" dust collectors with "wet" dust collectors. Several reasons were provided for making this change: 1) for materials prone to self heating and spontaneous combustion, the wet systems immerse the combustible dust in scrubber water, greatly reducing fire and explosion concerns; 2) the wet systems are viewed as being less hazardous to maintain because employees do not need to enter enclosed spaces where dry combustible dusts are present; and 3) the wet systems tend to be smaller than dry systems with comparable airflow capacities. However, maintenance of the wet dust collectors posed several challenges for one facility. The slurry discharge flow reportedly can become plugged and ductwork was highly susceptible to corrosion. In this particular case, scrubber malfunctions went undetected because the devices had no monitoring systems (e.g., level alarms on the slurry reservoir) and because the equipment was not routinely inspected. Greater oversight of the operation and maintenance of the scrubbers would help ensure that they operate correctly.
      Individual site visit reports document other observations in which dust collector design and operation were apparently not compliant with NFPA standards. For instance, some facilities vented exhaust from multiple processes through manifolded ductwork into a single dust collector, without having conducted a hazard evaluation or installing appropriate explosion isolation devices. This practice raised concerns of deflagrations spreading from a dust collector through manifolded ductwork to multiple facility locations. 
      Despite the various controls listed above, dust collectors (and interconnecting ductwork) at most facilities were generally not equipped with explosion isolation devices, raising concern about explosions propagating from the collector back to the process equipment. This issue is discussed further later in this section. 
   # Deflagration suppression systems. Multiple facilities employed deflagration suppression technology to control potential explosion hazards. At one facility, this technology was installed at raw material storage silos. These systems were designed to inject large volumes of sodium bicarbonate into the silos upon detection of elevated pressures. Though facility representatives had confidence in the systems' effectiveness, they also had reservations about using this technology. For example, the facility currently pays approximately $8,000 a year to train employees to inspect and maintain these explosion suppression systems. While these costs are substantially lower than the initial purchase and installation costs of the deflagration suppression systems, the facility noted that the cumulative long-term maintenance costs can be significant. Further, the facility was concerned about the possibility of "false trips" that can contaminate storage vessels and large quantities of raw materials. (Note: The particular suppression agent used in this case was compatible with U.S. Food and Drug Administration requirements for food additives, which helped address concerns for food manufacturing facilities.) 
   # Manual fire suppression equipment. Facilities had varying types of manual fire suppression systems, but specific challenges associated with manual fire suppression equipment were observed primarily at the metals processing facilities, where Class D extinguishing agents were required. One facility researched numerous candidate Class D agents and found sodium chloride ("table salt") to be most effective. At this facility, employees extinguished smoldering metal fires by scooping or shoveling salt from barrels and pouring the salt directly on and around fires. This approach raised concern to site visitors because employees had to come in very close proximity to fires in order to suppress them. The facility had considered other means for delivering the table salt to the fires, but most extinguishing systems result in a high pressure stream of the suppression agent, which has the potential for spreading a combustible metal fire instead of extinguishing it. The facility was actively researching various combinations of Class D extinguishing agents and delivery systems that would allow employees to effectively suppress fires from a safe distance. 
   # Isolation devices to preclude deflagration propagation. Many of the facilities that ERG visited had unit operations equipped with devices that could isolate operations from other parts of the plant. These include abort gates, rotary valves beneath dust collectors, and mechanical quick acting valves separating dryers from other processes. However, most of the installations viewed do not meet the explosion isolation design specifications in NFPA 69. In most cases, they would not activate quickly enough to prevent deflagrations from propagating through ductwork from one unit operation to another. It should be noted that one facility stated that they had considered such devices, but claimed that the devices would not be effective due to the rate at which their material of interest (a pyrophoric metal dust) would be expected to react. 
   # Spark/ember detection systems. Several facilities installed infrared spark/ember detection and suppression systems on different ductwork segments to detect and promptly suppress smoldering material before it could become the source of a fire or explosion. One such system is currently programmed to have two alarm levels: The low-level alarm, which is issued frequently, is triggered when sensors inside the ductwork detect a preset sparking threshold. It causes an audible alarm to sound, and the system releases a water spray for 5 seconds. The high-level alarm is triggered by an extended period of sparking that exceeds the programmed threshold and activates abort gates in the downstream ductwork in addition to the water spray and audible alarm. The high-level alarm has never been triggered in the system's history. Proper maintenance of spark/ember detection and suppression systems is important for successful operation, because detection sensors can lose effectiveness if they become coated with dusts and suppression systems can become plugged. More information on maintenance consideration has been published in the trade literature (e.g., Barnum, 2007). 
   # Deflagration venting. As noted previously, the vast majority of dust collectors are equipped with deflagration venting as their primary explosion protection technology. In addition, one facility has equipped most of their production rooms with building explosion panels. At this facility, at least one wall in most production areas had panels that were designed to release when interior pressures exceed 65 pounds per square foot, or approximately 0.5 pounds per square inch.
   
Administrative Controls
ERG reviewed various types of administrative controls and work practices that facilities implemented to control or mitigate hazards. This section presents detailed information on housekeeping programs and procedures (Section 9.1), given their significance in reducing the likelihood and severity of secondary explosions, and summary information for other administrative controls (Section 9.2) for combustible dusts. 
Housekeeping Procedures
The majority of facilities that ERG visited had documented housekeeping and equipment cleaning procedures to minimize accumulations of combustible dust. One facility was part of a corporation that recently developed a "Combustible Dust Housekeeping Guidance Document" that was distributed to all manufacturing facilities in the company, while other facilities developed their own internal guidance, typically in the form of standard operating procedures. At those facilities that did not have a written housekeeping program, interviews with multiple employees painted a consistent picture of housekeeping practices. The extent and effectiveness of the facility-specific housekeeping procedures varied. Some general observations about housekeeping and equipment cleaning follow: 
   # While all 11 facilities recognized the role that housekeeping plays in minimizing combustible dust safety hazards, facilities' housekeeping efforts were also motivated by other factors. These include the need for product purity, the potential for customer visits, pest control, and compliance with FDA regulations. The significance of these other factors was most evident at the pharmaceutical manufacturer and, to a lesser extent, at the food manufacturing facilities. Metal production facilities also had concerns about contamination, in part because trace amounts of certain impurities were known to compromise the performance of certain specialty products. 
   # The facilities implemented various types of housekeeping procedures, which typically included routine removal of dust accumulations in visible areas during process operations, immediate removal of localized accumulations following process upsets, and periodic "top-to-bottom" cleaning of entire production areas during process shutdowns or between production campaigns. The frequency of the routine and periodic cleaning activities varied across the facilities. For instance, "top-to-bottom" cleaning campaigns were conducted at frequencies ranging from every two weeks to every three years.
   # The approaches used to remove dust accumulations were determined by the individual materials' hazards and properties. The most common approaches to removing dusts included sweeping, mopping, hosing down equipment, vacuuming, and using compressed air. Water wash-downs were commonplace for soluble materials (e.g., sugars, paper dust, pharmaceuticals) but generally not for metals. At facilities in colder climates, water wash-downs in unheated production areas were not an option during the winter months. While water wash-downs have the benefit of placing combustible dust into a physical state that does not pose a hazard, the techniques used in the wash-downs were sometimes of concern. In one instance, where water wash-down was performed in a confined space (underground tunnels), the use of high pressure water spray had the undesired effect of re-suspending some of the settled dust, forming dense floor-to-ceiling dust clouds. 
      Compressed air was used to remove dust accumulations at multiple facilities. In general, site visitors were told that this practice occurred only when process operations were "shut down," only after bulk accumulations had been removed to the extent possible using other means, and only to remove accumulations from crevices, elevated areas, and other parts of operating equipment that were difficult to reach by other means. However, employees were observed using compressed air to remove dust that routinely Compressed air is often utilized during "top-to-bottom" cleaning campaigns to blow down dust on the rafters, equipment, walls, and stop equipment where it accumulates on the floor and can be easily swept up. Facility representatives often viewed vacuuming as more time-consuming than blowing dusts with compressed air and noted that vacuums are difficult to use when removing accumulations from areas with limited accessibility (e.g., interior workings of some machinery, overhead rafters). To facilitate clean-up following the use of compressed air, one facility reported using an oil-based sweeping compound to prevent re-suspension of dislodged dusts. In one industry, the compressed air blow-downs were conducted a few times daily and while process equipment was still activated and energized, and this activity generated massive dust clouds. This practice concerned site visitors, especially considering that operators used compressed air to remove combustible dust from heated surfaces -- a practice that could suspend smoldering or burning material in a dusty atmosphere. The facility cited many extenuating circumstances for why its use of compressed air was believed to be necessary. Section 12 of this report further discusses recommendations to OSHA regarding use of compressed air to remove accumulations of combustible dust. 
      Additionally, while most of the dust eventually settles back to the ground where employees can then wash it away, some dust settles on top of motors and other locations where accumulations should be avoided. Removal of dust accumulations in Class II locations was typically conducted using either (1) portable vacuum cleaners certified for use in Class II locations or (2) connections to central vacuum systems. However, one facility operated a gasoline-powered riding sweeper to remove dusts that had either settled onto the floor or had been swept into piles. It was unclear if the powered sweeper was approved for use in Class II, Division 1, Group G locations.
      
   # At two facilities, dust waste from housekeeping was collected and removed to centralized storage locations (e.g., wooden totes, outdoor shed). These storage locations were explosion hazards due to the large quantities of combustible dust that they contain. At one facility, the front-end loader used to move the dust, which was not rated for explosible atmospheres, was a possible ignition source.
   # Some facilities sought guidance from OSHA on quantitative measures for acceptable dust accumulations. For instance, facilities wondered if OSHA's standard will establish objective, quantitative metrics that facilities can use as goals for housekeeping efforts. Examples include specifying a maximum dust thickness that is considered acceptable (similar to the Agricultural Grain Dust standard) or a maximum allowable mass accumulation over a given area. Section 12 presents additional information on this issue.
   # At some facilities, site visitors identified inconsistent messages in housekeeping procedures. For instance, at one facility, instructions given in employee training differed from specifications in written operating procedures. At other facilities, housekeeping procedures described by safety and health personnel differed from the procedures employees actually followed. 
   # Several facilities had developed checklists for routine housekeeping activities that operations and maintenance staff were instructed to complete. The facilities also used periodic inspections to verify the effectiveness of their housekeeping programs. 
Other Administrative Controls
Site visitors identified several other types of administrative controls that facilities employed to control or mitigate combustible dust hazards. Examples follow:
   # Personal protective equipment (PPE). All facilities required employees in production areas to wear PPE, with the specific requirements varying across different processes and for different reasons. Most facilities required employees to wear hard hats, some form of safety shoes, and protective eyewear in all production areas. Additional requirements were observed in the food and pharmaceutical manufacturing facilities, but primarily in the interest of product purity. Multiple facilities also issued respiratory protection (e.g., air purifying respirators) for employees performing certain dusty tasks, such as changing out filter media in baghouses. Only one facility implemented additional PPE requirements specifically to address hazards posed by combustible dusts: a metals processing facility required its operators to wear flame-retardant clothing, consistent with specifications in NFPA 484 (NFPA, 2008). However, another facility was in the process of switching to fire-resistant garments when the site visit took place, and had issued all of its workers the new fire-resistant uniforms within a few months of the visit.
   # Confined space entry. Most facilities had a confined space entry program, through which employees obtained permits before entering or working in confined spaces. ERG's observations focused on the extent to which these programs acknowledged the unique hazards posed by combustible dusts. At some facilities, the confined space entry program did not allow employees to enter confined spaces if visible dust clouds were present. One facility, for example, had procedures denying entry in cases where dust levels impaired visibility of objects at a distance of 10 feet. Another facility established a similar condition of entry: "...a concentration of combustible dust that obscures vision at a distance of five feet or less." The origin of this rule of thumb was not known, but may be based on a previous scientific publication (Eckhoff, 2003). Finally, one facility explicitly considered the explosibility characteristics of the dust in addition to following the five-foot rule. The facility defined a hazardous atmosphere within a confined space when "airborne combustible dust is at a concentration that meets or exceeds its lower explosive limit." However, it is unclear if and how the facility actually measured the instantaneous dust concentration.
      Some facilities noted that OSHA could provide more concrete guidance on what specific dust concentrations must be avoided for confined space entry. These facilities noted that they have no difficulty establishing safe guidelines for gases (e.g., oxygen levels, explosive gases) given the availability of continuous monitoring devices and well-established safe thresholds (e.g., lower- and upper-explosive limits), but encouraged OSHA to develop guidance to inform facilities of similar strategies that can be implemented for detecting and avoiding potentially hazardous combustible dust atmospheres.
   # "Hot work" permits. Most facilities had a "hot work" permit program that both defined "hot work" activities and specified what steps employees must take to obtain a "hot work" permit. Most of these programs included general information about ensuring shutdown of all nearby equipment before "hot work" could commence. However, one facility specifically addressed combustible dust in their permit program by requiring employees to "...clean built-up sugar dust from I-beams and floors" before a permit can be issued. Another facility had multiple criteria that must be met before authorized personnel can issue hot work permits, such as removing coal dust from a "safe radius" around the hot work location (the exact radius is unspecified), venting airborne dust or allowing airborne dust to settle prior to the hot work activity, and ensuring that ventilation systems are operating prior to and during hot work activity.
   # Management of change. Most facilities, particularly those with processes subject to OSHA's PSM standard, had written management of change procedures that outlined special requirements for managing various types of process changes beyond replacements-in-kind. These written procedures were fairly generic, but did set a reasonable framework for ensuring that potential health and safety hazards are evaluated before changes are made. During employee interviews, however, site visitors found examples of process operators implementing changes without going through the facilities' written management of change procedures. 
   # Maintenance and repair. Maintenance and repair is crucial for ensuring that explosion protection systems operate as they are intended and for preventing process equipment from becoming an ignition source. Most facilities performed some type of regularly scheduled preventive maintenance -- for example, inspecting dust collectors for leaking filter bags on a monthly basis or greasing mill bearings each day -- although many did not have a written program. Some facilities contracted out specialized maintenance activities to third-party vendors. However, site visitors observed many instances where additional regularly scheduled maintenance activities would reduce fire and explosion hazards, such as periodic inspection of electrical control panels for dust accumulations and regular removal of tramp metal from magnetic separators. Another facility was struggling to keep up with needed repairs. A scrubber was not operating due to a broken fan shaft that employees identified more than a week prior to the site visit. Although a work order had been replace the fan shaft, facility representatives did not know when the repairs would take place. Other scrubbers plagued by corrosion had received patchwork repairs, including using tape to seal the inlet ductwork.

Emergency Response
Emergency response procedures varied across the facilities. Six of the facilities visited (including the smallest facility visited) had their own fire brigades, ranging in size from 4 to 75 employees. At the smallest facility, however, the four-man fire brigade typically only handled small fires, such as small amounts of smoldering material, and called the local fire department in the event of uncontrolled fires. Based on limited employee interviews, fire brigade members appeared to be well trained on the unique hazards posed by working with the specific combustible dusts at their sites. Time constraints prevented site visitors from conducting thorough evaluations of the facilities' emergency response measures, though some potential concerns were identified. For example, at facilities where argon deluge systems are used to suppress fires, site visitors encouraged the facilities to ensure that first responders were equipped with oxygen monitors to ensure that they do not become accidentally submersed in an argon atmosphere, which can quickly overcome employees who are not wearing suitable respiratory protection. Similarly, site visitors recommended that another facility develop specific guidelines on when the fire brigade should don SCBA when approaching different size sulfur fires, given that burning sulfur can emit highly toxic and irritating gases. Examples of other concerns are noted in the individual site visit reports. 

Investigation of Incidents
At every facility, site visitors inquired about past "incidents" involving combustible dusts. For purposes of this report, an "incident" includes fires and explosions (regardless of the extent of damage or injury) and near misses during which a serious fire or explosion appeared imminent but was avoided. The individual site visit reports document the range of incidents that occurred at the 11 facilities, and the site visitors made the following general observations pertaining to combustible dust incident investigations:
   # Some facilities had very detailed, written accounts of past incidents. For example, two facilities shared spreadsheets documenting every fire that occurred over a recent multi-year time frame. For each fire, the spreadsheets included information on the nature and extent of the event, possible causes, consequences, and action items; at one facility, the spreadsheet also assigned different employees responsibility for ensuring that recommended corrective measures were implemented. Similarly, other facilities shared copies of hazard analyses that were conducted, as per company policy, following specific incidents (e.g., fires in dryers). 
   # Several facilities relied entirely on its employees' recollection of past events for insights on incidents. This facility did not have a systematic approach for identifying and investigating incidents, which raised concern about the potential for incidents to recur in the future. 
   # At one facility, employees frequently encountered smoldering material or small fires that had to be extinguished -- this occurred several times a day in a particular operation. While the employees in these cases were trained on how to control these fires and felt comfortable doing so, and none of the fires propagated to major events according to site representatives, site visitors were concerned about the sense of complacency regarding a potentially hazardous occurrence (e.g., "...we see smoldering material all the time") and encouraged facilities to avoid downplaying the significance of these events. Two other facilities had long histories of minor deflagrations (referred to by the facilities as "puffs" or "pops"). Although both of these facilities have implemented engineering controls to reduce the frequency of these events, the facilities still experience small deflagrations on an intermittent basis.
   # The most serious incident that a facility shared with site visitors occurred when an employee responding to evidence of fire in a dust collector opened the access door to the dust collector. The influx of air dispersed the dust and caused the dust collector to explode, leading to one fatality and one injury. Two other facilities have experienced large fires leading to extensive property damage (estimated at $400,000 by one facility), but these fires reportedly did not cause any injuries.
   # NEAR MISSES

Regulatory Approaches
At the end of every site visit, ERG asked facility representatives to share any specific questions or concerns regarding OSHA's combustible dust rulemaking. While facility representatives seemed to appreciate the need for a standard focusing specifically on combustible dust, they relayed several concerns about the rulemaking and its potential implications on industry: 
   # Applicability. Almost every facility inquired about how OSHA's standard will define "combustible dusts." Some facility representatives emphasized that hazard potential varies greatly with the material of interest and its properties (e.g., density, moisture content, potential for agglomerating). One facility had internal combustible dust guidelines (see Section 6) that apply as follows: "A dust explosion potential is considered to exist where 1 pound or more of combustible dust per 1,000 cubic feet of volume is normally in suspension or could be put in suspension in an enclosure or inside pieces of equipment." Information on the scientific basis for this guideline was not provided. 
   # Retroactivity. Most facilities that hosted site visits operated processes that were constructed more than 20 years ago and therefore predate requirements set forth in recent editions of the NFPA standards. These facilities wondered if OSHA's combustible dust standard would apply retroactively to all existing processes, regardless of their construction date, or if it would include "grandfather" clauses for older processes. These representatives encouraged OSHA to consider how potential compliance costs can vary with facility age.
   # Cost. Almost every facility that ERG visited had concerns about the potential costs associated with a new combustible dust standard. Section 13 of this report provides more detail on the cost implications of a new combustible dust standard, based on limited observations collected from the 11 facilities that were visited. 
   # Timing. Concerned about their ability to quickly comply with a new safety standard, especially one that might require engineering evaluations and retrofits to processes with combustible dusts, two facilities suggested that OSHA consider a standard with "phased-in" compliance periods. For example, the standard might require a certain percentage of highest priority work be completed within 1 year; another percentage of work must be finished within 2 years; and so on. One facility was particularly concerned about timing given their large number of production lines (more than 10) and dust collectors that do not fully comply with NFPA standards (more than 150). 
   # Limited expertise. Facilities expressed concern about having limited internal resources and expertise to address combustible dust safety issues, due to the highly specialized nature of the topic. Although every facility had highly proficient on-site safety staff, not every facility had safety personnel with a strong command of NFPA standards and current administrative and engineering controls for addressing combustible dust hazards. Some facilities noted that they would likely have to hire consultants and engineering firms to implement additional engineering controls that a new combustible dust standard may require. Moreover, these facilities feared that the most qualified consultants and engineering firms would not be able to handle the amount of business that would be generated once OSHA promulgates a combustible dust safety standard. In contrast, one facility voiced concern about not being able to utilize the expertise that they do have on-staff. In particular, representatives opposed specifying minimum educational and professional certification requirements for individuals who develop and implement combustible dust safety programs. One representative noted that numerous employees do not have college degrees but have many years of professional experience that qualifies them to help develop dust safety programs. Minimum educational requirements could force the facility to hire external consultants to accomplish certain tasks that qualified facility personnel could handle on their own.
   # Acceptable dust accumulations. Facilities had many comments about whether the standard should include prescriptive limits for allowable dust accumulations, whether in terms of thicknesses or area of coverage. Some facilities expressed concern about establishing objective dust accumulation criteria, because a given accumulation (e.g., 1/32″) can pose dramatically different hazard potentials depending on the material's chemical and physical properties). While facility representatives appreciated the scientific basis for a proposed new approach in NFPA 654 to specify maximum allowable accumulations (i.e., in terms of mass per area), some people questioned the practicality of this approach and argued that measuring an accumulation depth is far easier than determining a mass per unit area. However, it should be noted that the next NFPA 654 edition will likely have options for relating accumulated mass to layer thickness, so that facilities that want to continue using thickness measurements will have a way to do so. 
   # Prescriptive or performance-based. Facility representatives offered conflicting opinions regarding the preferred approach of the pending combustible dust standard. Some representatives advocated for a highly prescriptive standard that would leave little ambiguity about exactly what facilities must do to comply. Other representatives supported development of a performance-based standard (e.g., similar to the PSM standard), which outlines general principles that facilities can then adopt in a process-specific manner. One suggestion was to incorporate combustible dust into the PSM standard. In contrast, representatives at another facility encouraged OSHA to not model its combustible dust standard after the PSM standard because they found it onerous to implement. 
      One facility offered a specific example arguing against the use of prescriptive requirements. The facility questioned the current NFPA 654 allowance for dust collectors with volumes less than 8 cubic feet to operate indoors. The facility found this threshold somewhat arbitrary (e.g., would it be sufficiently protective to have a dust collector located indoors for a highly explosive material?) and noted that the correct limiting size of dust collector volumes is best determined by engineering calculations and process hazard analyses. (Note: Future versions of NFPA 654 may no longer include the size cutoff in the requirement for dust collectors being placed outdoors; these specifications may be replaced with a requirement for some type of hazard or risk evaluation regarding when dust collectors can operate indoors.) 
   # Material-specific requirements. Representatives from the metal processing facilities advocated a rulemaking format consistent with NFPA's combustible metals standard (NFPA, 2008), which presents separate requirements for different metals (e.g., aluminum, magnesium, titanium). These individuals explained that having material-specific requirements greatly simplified compliance, because the facilities only needed to focus on certain sections of the standard. They also supported this approach given that the combustion and explosion properties vary considerably from one metal to the next. However, recognizing the broad range of materials (and mixtures) that OSHA's combustible dust standard might address, these facility representatives also acknowledged that including material-specific requirements might not be feasible. One facility generally recommended that OSHA be sensitive to dust-specific challenges (e.g., sulfur dust, metal dust): controls that make sense for one type of dust may not be appropriate for another type of dust.
   # Consistency with other standards. Some facilities voiced concern about having to understand and comply with multiple consensus and regulatory standards all pertaining to combustible dust (e.g., NFPA standards, insurance guidelines, OSHA regulations). One suggested approach for alleviating this situation was for OSHA to consider the use of multiple compliance options in its combustible dust standard, with one option being strict compliance with all applicable NFPA standards. By this approach, facilities that already have invested considerable resources to meet applicable NFPA standards would not have to invest additional resources to evaluate the specific requirements in OSHA's pending rulemaking. This commenter acknowledged, however, that simply referencing the NFPA standards would also present compliance and enforcement obstacles, given that these standards can change with every new edition issued.
   # Streamlining requirements of multiple agencies. Facilities in industries regulated by multiple entities (e.g., the Food and Drug Administration) expressed concern about OSHA's upcoming combustible dust standard conflicting with requirements under other statutes. One facility subject to regulation by FDA encouraged OSHA to ensure that any requirements in its proposed rule leave ample compliance options for the facility while not complicating compliance with FDA regulations. For instance, FDA regulations already limit the types of plastic bags that the facility can use, and additional OSHA requirements (e.g., use of material with a low charge retention time) might leave the facility with extremely limited options -- and possibly no options -- for purchasing compliant plastic bags. Another facility urged OSHA to ensure that its new combustible dust standard does not conflict with environmental regulations, which typically govern how dust control systems are designed, with certain operating parameters required to fall within permitted limits.
   # Other concerns about restrictive standards. Given the many different industry-specific nuances associated with controlling dust hazards, many facilities expressed concern about OSHA developing a standard with overly restrictive requirements. For instance, one facility feared that OSHA's combustible dust standard would prevent the use of compressed air to remove dust accumulations or would severely limit its use (e.g., limiting the air pressures allowed). This facility argued that the practice not only was commonplace throughout this industry, but was also the only feasible approach to removing settled dusts from crevices and other inaccessible areas in large production equipment. As another example, multiple facilities expressed concern about OSHA possibly requiring all dust collectors larger than a certain volume to be placed outside, as is outlined in some NFPA standards. One facility representative noted that, in certain scenarios, dust collectors may be operated safely within buildings. For example, he explained that wet dust collectors are inherently safer than dry dust collectors, because they capture and extinguish any ember or spark entrained in dust-laden process air streams; facilities also expressed concern about the economic impacts of being required to reconfigure processes to move dust collectors outdoors (see Section 13). 

Economic Impacts and Benefits
While this project was not designed to provide detailed economic analyses of combustible dust control options, site visitors were provided with costs associated with implementing various combustible dust control measures at several of the facilities visited. A summary of relevant observations made during the site visits follows: 
   # Costs of complying with NFPA standards. One facility had previously conducted an internal assessment of the estimated costs to bring the entire facility into compliance with NFPA standards pertaining to combustible dust. This facility, which operated more than 160 dust collectors (most of which had no explosion protection features), estimated that the up-front costs for upgrading these operations would range from $10 million to $28 million. These costs focused exclusively on retrofit activities for dust collectors, and did not consider costs for upgrading other devices that may be regulated under a new combustible dust standard (e.g., dryers, mixers, commodity bins). These projected costs also do not reflect a variety of additional costs that the facility might incur due to a new combustible dust standard. Examples of such additional costs include those associated with conducting hazard analyses, developing and implementing specialized training, and testing of new materials. Another facility performed significant retrofitting in the early 1990s, installing enhanced engineering controls for some of its 21 baghouses. Although the exact nature of the improvements was unspecified, the facility invested approximately $650,000 in the retrofits.
   # Industry-specific considerations. When conducting its regulatory impact analysis, facilities encouraged OSHA to consider several factors that may be unique to certain industries. For example, some industries rely upon production technologies (e.g., electron beam furnaces in metals production, large paper machines at paper mills) that cost more than $200 million to purchase and install; representatives from these industries said they would bear unusually high costs to replace such operations or retrofit them with dust controls. As another example, equipment replacement costs can vary greatly due to required materials of construction; this is of particular concern in the food and pharmaceutical manufacturing facilities, which must use materials that do not introduce contaminants into their products. Finally, compliance with NFPA standards has cost implications that vary across industries: for instance, NFPA 484 does not allow for titanium manufacturing facilities to operate dry dust collectors inside buildings, and these facilities would either have to reconfigure their processes to move devices outdoors or install different controls to comply with the standard; however, food manufacturing facilities can leave such devices indoors and still meet NFPA 61 requirements, though certain design criteria must be met.
   # Consideration of up-front and ongoing costs. Regarding OSHA's economic analyses for the combustible dust standard, facilities noted that compliance costs are not limited to up-front capital investments for purchasing new fire and explosion equipment, but also include costs associated with engineering and design evaluations, installation, and ongoing operations and maintenance. In some cases, these other costs outweigh the costs of purchasing new equipment. For instance, one facility provided itemized costs for installing a new wet dust collection system at a metal saw operation: the dust collector itself reportedly cost $15,000, but the facility spent an additional $45,000 on ductwork and installation costs. Some examples of ongoing costs include:
         o One facility pays approximately $8,000 a year to have employees attend training courses on inspection and maintenance of explosion suppression systems.
         o One facility hires a third-party vendor to maintain some of its scrubbers for $150 per month per unit ($1,800 per year per unit). This monthly servicing fee does not include costs for purchasing replacement parts that need to be installed or travel costs.
         o Another facility has a maintenance contract for its spark/ember detection and suppression systems. The contract, which covers quarterly servicing and annual inspections, costs approximately $3,000 per year. Roughly $2,000 of this expense is paid to the vendor for its labor, and the remaining $1,000 is paid to a separate company to rent a powered lift to help the vendor access the equipment.
   # Detailed cost information obtained during site visits. Several facilities provided site visitors information on the costs of various activities associated with combustible dust safety programs. For some combustible dust control measures, facilities made records available that clearly documented associated costs; but in other cases, facilities only estimated what they had paid. Most cost data provided by the facilities was gathered between the late 1990s and the present. 
         o Attachment 1 to this report lists laboratory analytical costs that two facilities shared for different types of testing. As the table shows, these costs can vary significantly from one laboratory to the next. It is possible that some quotes shown in the table represent discounted costs due to the large number of samples that certain facilities had analyzed. Moreover, in cases where combustible dusts are extremely valuable materials (particularly for pharmaceuticals), the value of the material "lost" in collecting a 1-liter sample can exceed the laboratory analytical costs shown in Attachment 1. 
         o According to the one facility that used oscillating ceiling fans to prevent dust accumulations on overhead structures, the cost to purchase and install the fans was approximately $8,000 per fan, though the facility expected this cost to decrease when purchasing large numbers of these controls. Costs associated with operating the fans were not available. Note that the facility believed the purchase and operation costs for the fans may have been less than the costs of having two employees work full-time vacuuming dust accumulations from overhead structures throughout the production building. The oscillating fans were also preferred given the potential safety hazards associated with employees operating vacuum cleaners while on scissor lifts.
         o Costs for installing new dust collectors can vary widely, depending upon design specifications. One facility noted that it cost approximately $60,000 to purchase and install a "small" dust collector dedicated to a band saw operation. Another facility stated that the overall cost of replacing a dry dust collector with a wet scrubber designed to control 26,500 cubic feet per minute air flow of a dust-laden process stream was approximately $550,000. An additional facility had also recently began replacing dry dust collectors with wet dust collectors. According to facility representatives, the scrubbers cost between $200,000 and $500,000 to design, purchase, and install. This cost range applies to systems with capacities ranging from 1,300 to 35,000 cubic feet per minute. The facility is also paying an additional $150 monthly maintenance fee (described earlier) for each scrubber. Another facility noted that its new dust collector (containing 36 cartridge filters) cost $21,000 to purchase and an additional $5,000 to $6,000 to install (including setting up the ductwork). 
         o One facility was in the process of transitioning from regular uniforms to fire resistant garments for all of its employees (approximately 20 people). Representatives noted that providing fire-resistant uniforms will approximately double the current annual budget for uniforms from $10,400 to $21,000.
         o One facility was considering purchasing dust ignition proof, static dissipative portable vacuums to assist with housekeeping based on their successful use by two of the facility's sister companies. The facility shared records showing that the vacuum system purchased by one sister plant cost close to $14,000 (excluding freight, installation, and startup assistance).
         o Another facility installed two infrared spark/ember detection and suppression systems and two abort gates on different ductwork segments. These systems were certified by Factory Mutual and installed at the recommendation of their insurance underwriter. The total purchase and installation cost for the spark/ember detection and suppression systems was approximately $130,000, and the two abort gates cost approximately $60,000 to purchase, install, and interface with the spark/ember detection and suppression systems. In addition, the facility pays an additional $3,000 per year in maintenance costs (described earlier). Although the facility incurred substantial costs to purchase, install, and maintain these systems, it also received substantial reductions in its insurance premium (a decrease of approximately $21,000), which partially offset these costs.
         o One facility installed building explosion panels in various production rooms. The most recent installation involved adding 13 panels (ranging in size from roughly 3 x 4 feet to 11 x 11 feet) to two production rooms. Total project costs were approximately $77,000 -- a cost that included purchase, shipping, installation, and taxes, but did not include engineering design. The costs for the engineering analysis (e.g., drawings, panel selection, panel sizing) were not available during the site visit.
As stated earlier, the previous information should be viewed only as a small collection of observations regarding costs of implementing combustible dust control strategies. A more extensive review would be needed to more fully evaluate the potential range of costs for identifying and mitigating combustible dust hazards. 

Impacts on Small Entities
This project was not designed to characterize potential impacts of a combustible dust rulemaking on small entities. The facilities that ERG visited were a self-selected subset of companies within the industries of interest identified by OSHA. All but one facility visited during this project had at least 80 employees. The smallest facility visited had less than 20 employees. This facility was well-informed about its combustible dust hazards, had an inerting system in place, recently replaced its dust collector, and was undertaking other facility improvements to reduce fire and explosion hazards (e.g., replacing a wooden storage structure with a concrete storage structure). Site visitors saw no evidence that this facility would encounter more difficulty than a large entity in controlling its combustible dust. However, this single, self-selected facility is likely not representative of other small entities that would be affected by an OSHA combustible dust standard.

Compliance Assistance
Every facility that ERG visited provided feedback on the compliance assistance resources they have accessed for information on combustible dust and the quality of technical support offered by those resources. A summary of this information follows: 
   # When asked about the possibility of a new combustible dust standard, some facility representatives recommended that OSHA place lesser emphasis on regulation and enforcement, and greater emphasis on guidance, compliance assistance, and other proactive measures to help facilities identify and implement best practices for controlling combustible dust hazards. One suggestion was for OSHA technical staff to work with facilities in more of a cooperative and consultative manner (e.g., through voluntary site visit programs), rather than interact with facilities largely through inspections. Another suggestion was for OSHA to develop industry- or material-specific guidance on combustible dust safety issues, similar to U.S. Environmental Protection Agency guidance for Risk Management Plan requirements (see: http://www.epa.gov/emergencies/guidance.htm#rmp). 
   # Facilities' awareness of OSHA publications on combustible dust also varied. On one hand, some facility representatives were aware of the National Emphasis Program, had read various OSHA documents on combustible dust posted to the agency's Web site (http://www.osha.gov/dsg/combustibledust), and had attended OSHA-hosted seminars on combustible dust safety issues. Other facility representatives were not aware of these and other OSHA resources. Some facilities subscribed to regulatory summary services, which informed designated facility contacts of ongoing activity pertaining to safety and health, environmental, food safety, and other applicable regulations. 
   # The facilities also had differing experiences with personal interactions with OSHA on combustible dust safety issues. Representatives from one facility commented that state OSHA officials seemed relatively experienced with agricultural dusts, but had limited expertise with combustible metals. (Note: Most of the facilities that ERG visited are located in states with approved State Plans.)
   # Site visitors asked facility representatives to identify other potential sources of compliance assistance information, beyond those listed in above. Not every facility tapped into these other sources, and those cited as being most informative were: a combustible dust course offered by a state Safety Council, an NFPA-sponsored conference on combustible dust, a facility-sponsored symposium on combustible dust hazards posed by a unique material, and information provided by trade associations (discussed earlier).
   # Facility representatives specified several topics for which technical guidance would be beneficial. As Section 2 notes, facility representatives had numerous technical questions about how best to use laboratory analyses and other methodologies to characterize their materials. They encouraged OSHA to issue guidance on testing (e.g., which materials to test, what methods to use, how to interpret results). Further, the facilities requested clarification from OSHA on exactly what parameters suppliers are expected to include on MSDSs versus what parameters facilities would be expected to determine on their own. Other suggested guidance topics included a compilation of explosibility data for dusts commonly encountered in industry, housekeeping procedures, and information on continuous air monitoring systems.
   # Many different compliance assistance resources are currently available on combustible dusts and their associated hazards, but the resources are not always consistent. Facilities noted the need to have a single resource of trusted information that they can access to ensure compliance with OSHA's upcoming standard. 

References

Barbauskas, 2003. Ignition Handbook. Vytenis Babrauskas. Fire Science Publishers: Issaquah, WA.

Barnum, B, 2007. Spark Detection: Maintain Spark Detection/Extinguishment Systems to Ensure Effectiveness. Bob Barnum, GreCon. In: Modern Woodworking. September, 1997. Document available online at: http://www.grecon-us.com/pdfs/maintain.pdf. 

Bartnecht, 1981. Explosions, Course, Prevention, Protection. W. Bartnecht. Springer-Verlag: New York, NY. 

Eckhoff, 2003. Dust Explosions in the Process Industries (3[rd] Edition). R.K. Eckhoff. Gulf Professional Publishing, Elsevier Science: Burlington, MA. 

ERG, 2008. Site Visits Related to Combustible Dust: Project Protocol. Prepared by Eastern Research Group, Inc. Prepared for OSHA Directorate of Standards and Guidance. December 12, 2008. 

FM, 2009. FM (Factory Mutual) Global Data Sheet 7-76: Prevention and Mitigation of Combustible Dust Explosion and Fire. March, 2009.

NFPA, 2008. NFPA 484: Standard for Combustible Metals. 2009 Edition. National Fire Protection Association. October 10, 2008.

OSHA, 2009a. U.S. Department of Labor's OSHA announces rulemaking on combustible dust hazards. U.S. Department of Labor, OSHA, Office of Communications. National News Release: 09-475-NAT. April 29, 2009. 

OSHA, 2009b. Advanced Notice of Proposed Rulemaking for Combustible Dusts. Federal Register, 74(202):54334-54347. October 21, 2009. 

OSHA, 2009c. Hazard Communication Guidance for Combustible Dusts. OSHA 3371-08. 2009. < http://www.osha.gov/Publications/3371combustible-dust.html >
Attachment 1: Laboratory Analytical Cost Data Provided by Selected Facilities

Multiple facilities informed site visitors of the prices commercial laboratories charge for analyzing samples for explosibility parameters. In most cases, however, the information provided was either incomplete or did not itemize costs by the specific analyses conducted. The following table provides the only itemized costs that ERG received, and these costs are based on recent testing quotes provided by two different commercial analytical laboratories. 

              Reported Laboratory Costs for Analyzing One Sample
                                       
                                Test Parameter
                               Analytical Method
                               Analytical Costs
                                       
                                       
                                 Laboratory #1
                                 Laboratory #2
Minimum explosible concentration
                                  ASTM E 1515
                                     $520
                                    $1,495
Minimum ignition energy
                                  ASTM E 2019
                                     $695
                                    No data
Full particle size distribution
                                  ASTM D 1921
                                     $220
                                    No data
Minimum auto-ignition temperature of dust clouds
                                  ASTM E 1491
                                     $680
                                    $1,150
Explosion severity test (Pmax and Kst)
                                  ASTM E 1266
                                    $1,000
                                    $2,095
Limiting oxidant concentration
                                 Not specified
                                    $1,200
                                    No data

Note: 	These laboratories also had supplemental charges for their analyses not specified here. Additionally, the above costs are based on analyses of a single sample. These laboratories offered their customers volume discounts for having multiple samples analyzed. It is possible that the costs quoted for "Laboratory #1" are biased low, because the facility that provided these costs is a frequent customer of the laboratory and therefore might be offered more competitive rates.
