To: Members of the Technical Advisory Subcommittee Modeling Work Group

DRAFT

October 5, 1995

Attached is a draft listing of potential MOBILE6-related projects prepared by the Emission Inventory Group (EIG), Assessment and Modeling Division (AMD), Office of Mobile Sources. We are distributing this document in draft form to enlist your assistance in helping EPA to focus on the most important issues and analyses for the development of MOBILE6, the next major update to the highway vehicle emission factor model.

The listing is divided into sections by topic: Exhaust Emission Estimates, Evaporative Emission Estimates, Inspection/Maintenance Program Issues, and so on. Please review this list and provide (a) any additions or other ideas for related work, and (b) your perception of the three highest priority issues in each topical listing. Please provide this information to either Lois Platte or Terry Newell of the EIG. (Fax number 313/668-4368; mailing address U. S. EPA NVFEL, 2565 Plymouth Road, Ann Arbor MI 48105 Attn: EIG/AMD) This process will be further discussed at the next conference call. A suggested format for your response is provided below.

The next conference call is scheduled for Friday, October 13, 1995. The Technical Advisory Subcommittee meeting that had been scheduled for that date has been cancelled due to government travel restrictions during the period of the continuing resolution. This will permit us to have the modeling work group conference call at 1:00 p.m. EDT on October 13. Further information regarding this call will follow.

Suggested Response Format

For each of the topical headings in the attached list, list your selection of the top three issues/projects in terms of priority and any suggestions for additions to the list. For example:

Exhaust Emission Estimates

1.Issue No. 3, Interaction of Correction Factors

2.Issue No. 7, Refinement of Tier 1 Estimates, etc.

3.Issue No. 1, Modeling of Non-FTP Emission Effects

Additions: None.

Evaporative Emission Estimates

and so on.

DRAFT10/05/95

List of projects/analyses for MOBILE6

This list was developed using the results of the June 23, 1995 "brainstorming" session and the "States' Needs" memo prepared by Dale Aspy/Region 4 which summarized the results of the "MOBILE user workshop" held in June 1994. Higher priority items, as viewed by the Emission Inventory Group (EIG) of the Assessment and Modeling Division (AMD), are listed first, along with brief descriptions; all but a few of the specific items mentioned at the brainstorming session and by the States at the 6/94 user workshop are listed.

The list consists of several sub-lists related by topic (exhaust emission estimates, evaporative emission estimates, and so forth). Relative rankings are within sub-lists only (not overall).

EXHAUST EMISSION ESTIMATES

1.Modeling of non-FTP (off-cycle) emissions impacts:

(i) Incorporate more "real-world" (non-FTP) driving effects in all vehicle activities; (ii) Evaluation of the emissions impacts of (a) non-standard loads, (b) driver behaviors, (c) defeat devices, (d) road grade, and (e) air conditioner usage.

(i) The fact that the Federal Test Procedure (FTP) and its driving cycle are not representative of all real-world vehicle operation has been widely recognized for some time. The 1990 Clean Air Act Amendments instructed EPA to study the extent to which real-world vehicle operation differs from that included in the FTP, and to revise the FTP as needed to ensure better in-use emission control. Primary among these driving aspects not reflected in the FTP are speeds (top speed in the existing FTP is < 60 mph, while significant VMT is accumulated at speeds up to at least 80 mph) and accelerations (maximum acceleration rate in existing FTP is 3.4 mph/s, while studies and surveys have shown in-use acceleration rates at least as high as 8 mph/s are not uncommon in some circumstances).

(ii) The specific items listed here are among the most important aspects of real-world driving not currently captured by the FTP. With the exception of (c), these generally can be described in terms of non-standard loading on the vehicle. "Non-standard loads" can include high acceleration rates and towing (trailers, boats, etc.). "Driver behaviors" can also include high acceleration rates and high travel speeds, but is more commonly used to describe "throttle flutter," the minor variations in speed/throttle position which have been found to result in significant emission increases relative to true steady-state operation. "Defeat devices" refer to situations where the vehicle/engine/emission control system are controlled in such a way as to effectively bypass emission controls when necessary (for example, hard accelerations may lead to open-loop highly enriched operation in order to provide the power demanded by the driver, resulting in large emission spikes). "Road grade" is simply another type of non-standard load; all current and past emission testing, driving cycles, and so on have inherently assumed level roadways, while tunnel studies among other evidence have shown that emissions can be much higher while climbing grades. "Air conditioner usage" has been simulated as part of the FTP for years, but the recent "Non-FTP Study" provided evidence that at least for some vehicles, actually operating the AC while driving has much bigger emission impacts (especially for NOx) than does simulating AC use in a dynamometer test.

Need: Data collection on these items has been in progress for some time, but substantially more data is needed to adequately characterize the emissions impacts of all of these items as well as their frequency of occurrence. The MOBILE model needs to be revised to account for these items and the associated increases in emissions. Guidance on determining how frequently each of these items occur will also be needed by States and others in order to properly account for these effects.

2.Reevaluate Warm/Cold Start Assumptions

New pollution control equipment, including improved catalytic converters and engine management systems, have reduced the time needed to warm up the vehicle to operating temperatures. Another factor is the amount of time between starts that will either keep the vehicle warm or to allow it to cool down. The issue of open loop operation impacts overall mobile emissions due to its large contribution to the total. Currently, it is assumed 508 seconds must pass for closed loop operation. However, SAI has conducted research indicating the time should be closer to one minute. This can significantly impact emission, especially on short trips.

Need: Data concerning actual cold/hot start operation needs to be collected, analyzed, and the results incorporated into the next version of the model.

3.Interaction of various correction factors used in MOBILE

The basic emission rate equations for LDGVs and LDGTs are developed from testing over the FTP (and in MOBILE5, from IM240 tests and the correlation of IM240 and FTP test results). In order to estimate in-use emission factors for other conditions (temperatures other than 75¡F, diurnal temperature range other than 60¡-84¡F, average speed other than 19.6 mph, operating mode fractions other than 20.6/27.3/20.6, fuel volatility other than 9.0 psi RVP, and so forth), correction factors are applied to the basic emission rate estimates. Most such correction factors are multiplicative, and almost no data on the interaction between them is available for analysis. Thus, for example, if emissions at 32¡F are twice emissions at 75¡F and emissions at 10 mph are twice emissions at 19.6 mph, MOBILE5 estimates emissions at 32¡F and 10 mph as being four times emissions at 75¡F and 19.6 mph.

Need: Test programs and resulting data to examine the interactions of the many variables affecting in-use emissions levels. Practically speaking, there is no way that OMS can ever hope to have sufficient resources (money, time, testing capability) to perform the amount of testing that would be required to address all of the interactions. The interactions of average speed and ambient temperature would be the most important to investigate.

Note: This issue is intimately related to the issue of bag-specific (operating mode-specific) emission factors and correction factors. Beyond the interactions of all of these different correction factors are the problems that some correction factors are bag-specific while others are not, and that most of correction factors are developed from data on hot stabilized operating mode vehicle tests.

4.Changes to the modeling of effect of speed on emissions:

(i) Change "speed" option to include both average and cruise speeds

(ii) provide options for assuming various driving cycles (variable, California, etc.)

(iii) revisit the speed correction factor for NOx.

(i) Speed is a major factor in estimating emission factors. Currently, the speed used in modeling is an average speed. However, such average speeds may not be representative of travel on a given type of roadway. On free flowing expressways there is no need to incorporate acceleration and deceleration from low speeds into the average speed. In this case the average speed should be a weighted average speed of vehicles at normal conditions. Emission characteristics of vehicles change dramatically under such conditions. Factors such as cold and hot start are very different for speeds typical of expressway versus local roads.

(ii) There has been much discussion on the use of the current FTP to model actual driving conditions. Various areas may experience different driving cycles. California is investigating the way people drive under various conditions and roadway types. Small changes in speed are not as important on expressways as they are on local roads. A change of 10 mi/hour has a much larger effect on the emissions of a vehicle traveling at 15 mi/hour than it does on a vehicle at 60 mi/hour. The inclusion of off cycle emissions may play a significant role in addressing underestimation of emissions by the model.

(iii) NOx emissions are speed dependent. The average speed used in the model may not accurately reflect the actual average speed on a given link or roadway. (OMS feels that this issue is raised by the States in part because NOx emission factors increase with increasing speed, creating problems in situations where increased average speed/reduced congestion, which reduces estimated VOC emissions, also increases estimated NOx emissions.)

Need: (i) Provide options for the use of speed data in the next MOBILE model. Differentiate between average speed and speed in cruising conditions. Incorporate the work California is doing in various speed distributions. Consider multiple speed algorithms, i. e., different factors would be used in different speed ranges.

(ii) Provide an option for the use of variable driving cycles in the estimation of emission factors. Include off cycle emission factors.

(iii) Provide for a NOx speed correction factor. (This is from the "States' Needs" memo list of suggestions.)

OMS Note: MOBILE5a does include NOx speed corrections; as noted above, NOx increases with increases in speed over the same range of average speeds in which VOC emissions decrease with increases in average speed.

5.Revisit use of IM240 data for modeling of exhaust HC/CO/NOx emissions; otherwise account for age effects. Review of high mileage (> 50K) deterioration rates

MOBILE5a, with its basic emission rate equations for 1981 and later model year LDGVs revised on the basis of IM240 data, has substantially higher in-use deterioration rates than did previous versions of the model. This is especially true with respect to "DR2," the deterioration rates applied for accumulated mileage in excess of 50,000 miles. There are some indications that the DR2s used in MOBILE5a overstate actual typical in-use emissions deterioration after 50,000 miles. In addition, the available data from high-mileage vehicles is not sufficient to distinguish the relative contributions of increased mileage and increased age to in-use deterioration of emissions performance. Finally, the issue of where the DR "kink" is best modeled should be further investigated; there are indications that 50,000 miles (the current DR flex point) may not be the best point for changing the DR.

Need: Significantly more data on in-use emissions performance on high-mileage (over 50,000 miles) vehicles; reanalysis of issue using new and existing data; potential revisions to the DRs (particularly DR2s for higher mileages) used in the model and to the mileage point at which the DR increases.

6.Provide for idle emission factors output; reevaluate idle data and emission factors.

MOBILE5 does not provide idle emission factor output as did previous versions did. This is a valuable output that is widely used. State users also requested that currently available idle data be reevaluated.

Note: OMS has issued guidance to model users that idle emission factors be estimated for a given set of conditions (temperature, fuel RVP, operating mode) by modeling 2.5 mph emissions for the desired conditions and multiplying that emission factor in g/mi by 2.5 mi/hr to yield a g/hr estimate of idle.

Need: Provide for idle emission factor output, reexamine the current idle data used in MOBILE5.

------

7.Refinement of Tier 1 emission standard impacts; new evaporative emissions test data.

New exhaust (tailpipe) emission standards under the 1990 Clean Air Act Amendments, known as "Federal Tier 1" standards, are phased in for LDGVs and LDGTs over the 1994-98 model years. A new evaporative emissions test procedure, intended to control evaporative emissions under more realistic conditions (i.e., temperature profiles and time durations) is also phased in over the 1996-99 model
years. Current estimates of the effects of these new standards and test procedure changes on in-use emission levels are of necessity based on past performance and engineering judgment. As vehicles manufactured in compliance with these standards and test procedure changes enter the in-use fleet and accumulate mileage, data will become available to allow refinement of the estimated effects in the MOBILE model.

Need: Test programs, data collection, and data analysis focused on new vehicles certified to Tier 1 standards and under the new evaporative emission test procedure are required as early as feasible so as to better estimate the actual in-use emissions impacts of these standards and test revisions.

8.Improvements to estimates of in-use emission factors for heavy-duty vehicles (HDVs):

(i) Review need for HDV data collection and analyses

(ii) Development of HDV chassis test cycle(s)

(iii) Review of HDV data collected in I/M programs

(iv) Is there a "real-world" way of measuring HDV emissions?

(v) Coordination with clean-fleet programs (EPCD).

(vi) Division of HDVs (especially HDDVs) into subclasses, analogous to PART5.

(i) Historically, there are few heavy-duty vehicle (HDV) analogues to the in-use emission factor data collection program that forms the basis of the emission factor estimates for light-duty vehicles (LDVs) and trucks (LDTs). The emission factors for HDVs produced by MOBILE are developed on the basis of new engine certification data and limited data from 1980s technology HDVs (zero mile levels from cert data, DRs from 1980s technology engines), engineering judgment in the extension of LDV assumptions and methodologies to HDVs, and the use of conversion factors (necessitated by the regulation of emissions from such engines in terms of grams per brake horsepower-hour and the need for in-use emission factor estimates in terms of grams per mile).

(ii) The current approach of regulating HDE emissions in terms of mass pollutant-per-unit work performed is based on testing of HDEs over engine duty cycles on engine dynamometers; even if chassis dynamometer testing of in-use HDVs were less costly and difficult to implement, there is currently no defined HDV driving cycle.

(iii) Given the above, the one source of data on in-use emissions from HDVs would be those relatively few inspection and maintenance (I/M) programs that include some subset of HDVs in their coverage requirements. This would be limited to the lighter portion of the HDGV fleet (generally Class IIb-III vehicles, 8501-14,000 lbs GVW), which represent the majority of all HDGVs. Such data will generally be limited to idle emission measurements until such time (if ever) as some HDVs have been tested in IM240 programs.

(iv) An alternative approach would be to determine if there is any feasible way of obtaining "real-world" in-use emissions measurements from such vehicles (i.e., some type of instrumented vehicles).

(v) Much of the focus of clean-fleet programs will be on HDGVs. There are currently no in-use data on the performance of fleets under this program.

(vi) Some modelers particularly want "urban buses" to be a separate category with its own emission factor estimates. More generally, there is demand on the part of modelers that the HDV categories be divided into subclasses, along the lines of the division in the PART5 particulate model. It is reasonable to assume that given sufficient data to work with, such subdividing would improve the overall modeling of emissions estimates for HDVs.

Need: Only substantial collection and analysis of in-use data, preferably from chassis dynamometer testing of HDVs (as vs engine dynamometer testing of heavy-duty engines (HDEs)), can clearly address the extent to which current approaches fail to accurately characterize in-use emissions from these vehicles. Considering the logistics and expense of doing so, and the lack of any agreed upon representative chassis cycle(s) for HDVs, this is unlikely to occur in the short term. Thus, other approaches must be developed and evaluated for assessment of the accuracy of in-use emission factor estimates for HDVs. Collecting all available HDV emissions data from operating I/M programs for analysis represents one alternative approach which can at least be initiated in the short term. Development and funding of real-world measurements, such as might be obtained from adequately instrumented HDVs, represents another alternative worth considering. EPA's Office of Research and Development (ORD) has funded work in this area recently. The division of HDGVs and HDDVs into subcategories for emission factor estimates requires not only substantial data collection and analysis, but also significant programming effort.