                       UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                      NATIONAL VEHICLE AND FUEL EMISSIONS LABORATORY
                                   2000 TRAVERWOOD ROAD
                                   ANN ARBOR, MI  48105




                                                                                                                 	      			                     
May 4, 2012
                                                         OFFICE OF 
                                                      AIR AND RADIATION

MEMORANDUM

SUBJECT:	Diesel Particulate Filter Regeneration

FROM:	Lauren Steele, Office of Transportation and Air Quality

--------------------------------------------------------------------------------
TO:		Docket EPA-HQ-OAR-2011-1032
Engine Combustion Characteristics and Emissions Formation
      Particulate matter (PM) consists of many components.  Carbonaceous soot and solid-phase hydrocarbons originate from incomplete combustion of fuel and lubricating oil.  Metal and metal oxide PM (i.e. metallic ash) originates from engine wear and/or lubricating oil additives.  Sulfur-containing compounds, including solid-phase sulfate particles, are formed when other compounds react with the sulfur present in the fuel and lubricating oil.
      Within the combustion chambers of the cylinders in an operating engine, there exists an important trade-off between PM (soot and hydrocarbon) formation and nitrogen oxides (NOX) formation.  To explain this trade-off, we describe the combustion as taking place in sequential stages.  The initial formation of soot and hydrocarbon PM occurs during the early stages of combustion.  However, most of the initially formed soot and hydrocarbon PM is actually oxidized in-cylinder, late within the combustion process, so it is never seen at an engine's tailpipe.  In contrast, NOX is formed in the middle stage of the combustion process when the highest combustion temperatures occur, and it is not largely eliminated late in the combustion process as is soot and hydrocarbon PM.  This is significant because in-cylinder NOX control is achieved by lowering peak combustion temperatures, but these lower temperatures in turn decrease oxidation of in-cylinder soot and hydrocarbon PM; thereby increasing soot and hydrocarbon PM at the tailpipe.  Thus we have a trade-off between engine combustion conditions that either optimize NOX control or PM control, but not both.  This requires crucial decision-making when designing an engine to meet EPA emissions standards.  As is explained in the preamble to the proposed rule, this trade-off becomes even more important when designing an engine that will be successfully integrated with an emissions control system equipped with a diesel particulate filter (DPF).
      DPFs typically use a porous ceramic or cordierite substrate or metallic filter to physically trap particulate matter (PM) and remove it from the exhaust stream.  In some non-road applications such as in underground gassy mines, disposable paper-based or fiberglass-based filters may be employed downstream of an exhaust cooling system (e.g. a "wet scrubber").
       In contrast with soot and hydrocarbon PM, which is periodically oxidized from a DPF, most metallic ash and other noncombustible materials remain on the filter. This accumulated ash must be cleaned offline periodically by removing the DPF and sending it to a service center.  Regarding sulfate PM, it should be noted that DPFs must be used in conjunction with ultra-low sulfur diesel fuel, which has a sulfur content of less than 15 parts per million.  Higher levels of fuel sulfur adversely affect sulfate formation downstream of the DPF, and sulfur also interferes with DPF regeneration.  Sulfur can deactivate the catalytic materials within the DPF, which in turn can inhibit the oxidization of trapped soot and hydrocarbon PM at lower exhaust temperatures.
      Engines operating over relatively low-temperature, low-speed duty cycles can have more difficulty achieving conditions sufficient for DPF regeneration, and may require more frequent maintenance and driver interaction.  When an engine manufacturer certifies a new engine for application in vocational vehicles, it is tested using EPA's standard federal test procedure (FTP) which simulates urban stop-and-go driving.  It is not the purpose or intent of the FTP to simulate conditions under which a DPF may regenerate.  Indeed, it is rare that an engine will experience DPF regeneration during an FTP test cycle.  Instead, manufacturers typically collect data in other ways to determine whether their control algorithms are appropriate for the engine and its intended applications.
      When a truck engine operates at low speeds and loads, its temperature remains relatively low and this is typically a point of less efficiency on its power curve.  When a diesel engine operates at higher loads, its exhaust temperature increases and it generally shifts to a place of higher efficiency on its power curve.  This higher, more optimal temperature not only enables greater fuel efficiency and the activation of catalysts within the emission control system, it also can lead to the generation of less soot and more NOX emissions.  Because the principal NOX formation mechanism in an engine is thermal formation within the combustion chambers, many NOX reduction technologies have focused on ways to decrease combustion temperatures.  For successful operation of a DPF, however, the highly reactive NOX compounds, especially NO2, are useful, acting as powerful oxidizers.  When NO-rich engine exhaust is passed over a DOC, a large fraction of NO2 is formed, which increases the rate of oxidation of the deposited soot in the cells of the DPF, thus cleaning the filter through passive regeneration.  When combustion temperatures are lower, NOX content tends to be lower, decreasing the rate of oxidation in the DPF, leading to an increased need for active regeneration.
      Furthermore, the NOX-PM trade-off of most diesel combustion processes (homogeneous charge compression ignition (HCCI) being an exception) dictates that when combustion temperatures are lower, in-cylinder soot oxidation is lower, and this increases the PM loading of the DPF.  Engines operating at lower temperatures may also experience higher lubricating oil consumption than engines operating over duty cycles with higher operating temperatures, leading to higher rates of ash accumulation and eventual DPF removal for ash cleaning.  Therefore, both the engine's duty cycle and the overall control strategy of the engine's emissions control system play a large role in the success of integrating a DPF with an engine to control PM emissions.
Types of Regeneration
Passive Regeneration
      Passive regeneration refers to methods that rely strictly on the available exhaust temperatures and exhaust constituents to oxidize PM from a DPF in a given vehicle application.  Passive regeneration is an automatic process that occurs without the intervention of an engine's on-board diagnostic and control systems, and often without any operator alerts.  Passive regeneration is often a continuous process, because of which, it is sometimes referred to as continuous regeneration.
      DPFs that regenerate in this fashion cannot be used in all situations, primarily due to insufficient exhaust gas temperatures associated with some types of diesel engines, the level of PM generated by a specific engine, and/or application operating experience.  When a vehicle with only a passively regenerating DPF experiences unacceptable engine backpressure conditions, the driver may be alerted to take corrective action.  Operation at higher vehicle speeds and loads typically leads to successful passive regeneration of a DPF.  When driving the vehicle in this manner does not achieve passive regeneration, the vehicle must be taken to a service facility for maintenance, where the DPF is typically removed for offline regeneration.
      Many DPF's today are designed with a diesel oxidation catalyst (DOC) directly upstream.  The presence of oxidized compounds such as NO2 in the exhaust stream, in combination with sufficient temperatures, can enable the burn-off (oxidizing) of accumulated soot on a continuous basis while those conditions exist.
      In designing a system to enable passive regeneration, engineers seek to achieve a rate of soot oxidation that is in equilibrium with the rate of soot deposition. When deposition and oxidation rates are in equilibrium, the pressure drop due to PM accumulated on the filter will remain constant.  When the deposition rate is higher than the oxidation rate, particles continue to accumulate and the pressure drop rises.  For a given DPF configuration, catalyst, fuel, and engine, the temperature at which this equilibrium is attained is called the balance point temperature.  An ideal DPF system will be optimized for its engine so that the balance point temperature is attained or exceeded frequently throughout the duty cycle.
      Scientists have studied the effects of modifying DPF systems, including alternative catalysts, to facilitate oxidation and reduce the balance point temperature.  While most systems in use today do not incorporate the use of fuel borne catalysts, the study conducted by Jelles et al in 1999 illustrates the steady-state pressure drop that can be achieved.  Figure 1 depicts results of a test with dosing of a fuel-borne catalyst to determine the lowest temperature at which equilibrium is attained. 
                                       
Figure 1 Illustration of Equilibrium with Passive Regeneration 
      The experiment was conducted with a Corning EX80 wall-flow monolith filter, where the lowest balance point temperature for 100 ppm cerium catalyst additive was found to be between 430C and 440C. The curves are typical of a balance point temperature experiment.  The positive slopes indicate normal operation with soot deposition.  The horizontal lines at approximately five and 10 hours indicate when equilibrium is achieved.  In the case of an in-use engine passively regenerating, theoretically these horizontal slopes would continue until operating conditions caused PM accumulation (and pressure drop) to increase.  The periodic spikes in temperature and steep negative slopes in pressure drop shown in Figure 1 indicate active regenerations as part of the experiment, and would not be part of an in-use passive system operating profile.
Automatic Active Regeneration
      In a vehicle whose normal operation does not generate temperatures needed for passive DPF regeneration, the system needs a little help to clean itself.  This process is called active regeneration, and supplemental heat inputs to the exhaust are provided to initiate soot oxidation.  The frequency of active regeneration depends on many parameters, including engine emissions profile, engine loads, vehicle duty cycle, and ambient temperature.  There are two types of active regeneration:  those that may occur automatically either while the vehicle is in motion, while idling, or while operating in PTO mode, and those that must be driver-initiated and occur only while the vehicle is stationary and out-of-service.  The first type is discussed here, while driver-initiated parked regeneration is discussed below.  During both of these types of active regeneration, NOX emissions may increase, especially when not otherwise controlled by a NOX aftertreatment device.  An estimate of emissions during regeneration is presented in a memorandum to the docket entitled "NOX Emissions from DPF Regeneration."
      It is important to note that all DPF systems with active regeneration components also have the capability to passively oxidize soot accumulated on the filter, though some may rarely achieve a level of equilibrium and will continue to experience rising backpressures until an active regeneration is completed.  Vehicle operating conditions or duty cycles determine whether passive oxidation may occur to the degree necessary to maintain engine exhaust backpressures within acceptable limits.  In some cases, the small amount of passive oxidation merely serves to lengthen the time interval between active regenerations.  In some extreme cases where the DPF system almost never achieves its balance point temperature, passive oxidation may be negligible.  In those cases the system would rely almost exclusively on active regenerations to maintain a clean PM filter.
      With actively regenerating DPF systems, the engine's electronic control module (ECM) is programmed to detect operating conditions and produce driver signals that provide information about DPF conditions related to regeneration.  When conditions indicate a need for active regeneration, and thus more heat, a number of things can occur.  A variety of manufacturer approaches can be taken to produce the supplemental heat needed for active regeneration.  Diesel engines with MY 2007 or newer emission control systems often incorporate approaches ranging from air intake throttling to fuel dosing. Some of these approaches are described further in Section IV.C of the preamble to the proposed rule, and estimates of the additional fuel use by a vehicle whose DPF regeneration system employs fuel dosing are described in a memorandum to the docket entitled "Fuel Use with Dosing for DPF Regeneration."
      Actively regenerating DPF systems typically require sufficient air flow, temperature and soot accumulation before an automatic active regeneration will be initiated.  As mentioned above, this may occur either while the vehicle is in motion or parked, if pre-set engine operating conditions are met.  When the engine's ECM signals the initiation of an automatic active regeneration and the extra heat is generated, an ideal DPF system accomplishes this as a transparent process, with no effects perceivable by the driver.  As a safety precaution, some vehicles will illuminate a dash lamp during this time, indicating when exhaust temperatures are higher than normal.
      Vehicles with automatic active regeneration systems require operators to be alert to dash lamps and indicators.  Written instructions are provided that explain what each lamp means and what action is needed.  Some vehicle signals instruct drivers to change driving conditions to allow automatic active regenerations to occur.  As an example, an illustration of a typical DPF dash lamp is provided in Figure 2.  The left image indicates a steady illumination (usually amber), while the right image indicates the lamp is flashing.  
                                       
Figure 2 Aftertreatment Diesel Particulate Filter Lamp 
      The instructions corresponding to these lamps, as provided in a Cummins Engine Company operator manual, are shown in Table 1.
Table 1 Aftertreatment Diesel Particulate Filter Lamp Instructions
The AFTERTREATMENT DIESEL PARTICULATE FILTER lamp indicates, when illuminated or flashing, that the aftertreatment diesel particulate filter requires regeneration.
An illuminated AFTERTREATMENT DIESEL PARTICULATE FILTER lamp indicates that the aftertreatment diesel particulate filter needs to be regenerated at the next changing opportunity. This can be accomplished by:
Changing to a more challenging duty cycle, such as highway driving, for at least 20 minutes
Performing a stationary regeneration
A flashing AFTERTREATMENT DIESEL PARTICULATE FILTER lamp indicates that the aftertreatment diesel particulate filter needs to be regenerated at the next possible opportunity. Engine power may be reduced automatically.

When this lamp is flashing, the operator should:
1. Change to a more challenging duty cycle, such as highway driving, for at least 20 minutes
Performing a stationary regeneration
      Because EPA emissions standards are performance based; and therefore, do not dictate any required emission control system technologies or configurations, each manufacturer has the discretion to program the timing and sequence of lamps as needed to inform drivers of the condition of its emission control systems.  It is not uncommon in today's heavy-duty fleet for an engine's ECM to limit its maximum speed, torque or power when an overloaded DPF is detected.  These engine and emission control system protection measures can alert drivers to the need to change driving conditions to enable active mobile regeneration or to make plans to allow for a manual active regeneration, described below.  If a driver disregards such warnings, the risk of uncontrolled engine shutdown or catastrophic DPF failure increases.  The preamble to the proposed rule includes additional discussion of DPF failure.
Manual Active Regeneration
      When a vehicle's driving cycle does not present conditions for regular passive or automatic active regeneration (e.g. due to insufficient temperature or exhaust flow), a manual active regeneration becomes necessary.  In contrast to the practice for an overloaded passively regenerating DPF system, filters with heavy soot buildup on vehicles with actively regenerating systems rarely need to be removed and regenerated offline.  Instead most vehicles are equipped with switches and instructions for an operator to initiate a manual active regeneration while the truck is parked in a safe location and continuously monitored by the driver or a service technician.
      A manual active regeneration allows the engine's ECM to increase engine speed and exhaust temperature to a greater extent than what is typically allowed during an automatic active regeneration.  Because the ECM takes full control of an engine during a manual active regeneration, the vehicle must remained parked and not used for other purposes, such as pumping water in PTO mode.  While this type of regeneration does temporarily take a vehicle out of service, the duration to complete a parked manual active regeneration is typically less than 40 minutes, and the process may be interrupted at any time.  This is important, because there are often cases where a vehicle, particularly an emergency vehicle, must be returned to service prior to completion of a manual active regeneration.  As more time elapses before interrupting a manual active regeneration, the progression of warning lamps will resume earlier in the warning/DPF loading process, providing more opportunities for other types of regeneration to occur, or at least providing more operating time before the next manual active regeneration is demanded.  EPA has information from one engine manufacturer indicating that interrupting a manual active regeneration (starting with a fully loaded filter) after as little as 10 minutes may enable a vehicle to resume normal driving, which could then lead to the occurrence of an automatic active regeneration or at least to arrive safely at a local garage before re-illumination of the warning lamp indicating a need for a manual active regeneration.
      In the case where a manual active regeneration is not initiated when requested by the engine's controls, the DPF may become so overloaded that a manual active regeneration is no longer an available option.  The reason for this is that an extreme level of soot loading requires a very carefully managed regeneration, under the supervision of a highly trained technician.  A discussion of DPF failure is presented in Section IV.C of the preamble to the proposed rule, including a description of catastrophic failure, which is an uncontrolled, "runaway" regeneration.  Technician-controlled regenerations may require towing the vehicle to a special service center, and may occur while the DPF is on the vehicle, or offline with the DPF removed from the vehicle.  In such cases, if a spare DPF is not available, the vehicle could be out of service overnight.
