Sections 1065.1131 through 1065.1145 appear at 88 FR 4697, Jan. 24, 2023, unless otherwise noted.
Sections 1065.1131 through 1065.1145 specify procedures for aging compression-ignition engine aftertreatment systems in an accelerated fashion to produce an aged aftertreatment system for durability demonstration. Determine the target number of hours that represents useful life for an engine family as described in the standard setting part. The method described is a procedure for translating field data that represents a given application into an accelerated aging cycle for that specific application, as well as methods for carrying out aging using that cycle. The procedure is intended to be representative of field aging, includes exposure to elements of both thermal and chemical aging, and is designed to achieve an acceleration of aging that is ten times a dynamometer or field test (1,000 hours of accelerated aging is equivalent to 10,000 hours of standard aging).
(a) Development of an application-specific accelerated aging cycle generally consists of the following steps:
(1) Gathering and analysis of input field data.
(2) Determination of key components for aging.
(3) Determination of a thermal deactivation coefficient for each key component.
(4) Determination of potential aging modes using clustering analysis.
(5) Down-selection of final aging modes.
(6) Incorporation of regeneration modes (if necessary).
(7) Cycle generation.
(8) Calculation of thermal deactivation.
(9) Cycle scaling to reach thermal deactivation.
(10) Determination of oil exposure rates.
(11) Determination of sulfur exposure rates.
(b) There are two methods for using field data to develop aging cycles, as described in § 1065.1139(b)(1) and (2). Method selection depends on the type of field data available. Method 1 directly uses field data to generate aging modes, while Method 2 uses field data to weight appropriate regulatory duty cycles that are used for emissions certification.
(c) Carry out accelerated aging on either a modified engine platform or a reactor-based burner platform. The requirements for these platforms are described in § 1065.1141 for engine bench aging and § 1065.1143 for burner-based bench aging.
This section describes the gathering and analysis of the field generated data that is required for generation of the data cycle. Gather data for the determination of aftertreatment exposure to thermal, lubricating oil, and sulfur related aging factors. You are not required to submit this data as part of your application, but you must make this data available if we request it.
(a) Field data target selection. Use good engineering judgment to select one or more target applications for gathering of input field data for the accelerated aging cycle generation that represent a greater than average exposure to potential field aging factors. It should be noted that the same application may not necessarily represent the worst case for all aging factors. If sufficient data is not available to make this determination with multiple applications, you may select the application that is expected to have the highest sales volume for a given engine family.
(1) Thermal exposure. We recommend that you select applications for a given engine family that represent the 90th percentile of exposure to thermal aging. For example, if a given engine family incorporates a periodic infrequent regeneration event that involves exposure to higher temperatures than are observed during normal (non-regeneration) operation, we recommend that you select an application wherein the total duration of the cumulative regeneration events is at the 90th percentile of expected applications for that family. For an engine that does not incorporate a distinct regeneration event, we recommend selecting an application that represents the 90th percentile in terms of the overall average temperature.
(2) Oil exposure. Use a combination of field and laboratory measurements to determine an average rate of oil consumption in grams per hour that reaches the exhaust. You may use the average total oil consumption rate of the engine if you are unable to determine what portion of the oil consumed reaches the exhaust aftertreatment.
(3) Sulfur exposure. The total sulfur exposure is the sum of fuel- and oil-related sulfur. Oil-related sulfur will be accounted for in the acceleration of oil exposure directly. We recommend that you determine fuel-related sulfur exposure by selecting an application that represents the 90th percentile of fuel consumption. Use good engineering judgment to determine that average rate of fuel consumption for the target application. You may use a combination of field and laboratory measurements to make this determination. Calculate the average rate of fuel-related sulfur exposure in grams per hour from the average rate of fuel consumption assuming a fuel sulfur level of 10 ppm by weight.
(b) Application data gathering. Use good engineering judgment to gather data from one or more field vehicles to support the accelerated aging cycle generation. We recommend that you gather data at a recording frequency of 1 Hz. The type of data that you gather will depend on the method you plan to use for cycle generation. Record both the data and the number of engine operating hours which that data represents regardless of method, as this information will be used to scale the cycle calculations. Use good engineering judgment to ensure that the amount of data recorded provides an accurate representation of field operation for the target application. If your application includes a periodic regeneration event, you must record multiple events to ensure that you have accurately captured the variation of those events. We recommend that you record at least 300 hours of field operation, and at least 3 different regeneration events if applicable.
(1) When using Method 1, direct field data use, as described in § 1065.1139(b)(1), record data for exhaust flow rate and at least one representative inlet temperature for each major aftertreatment system catalyst component, such as a diesel oxidation catalyst (DOC), diesel particulate filter (DPF), or selective catalytic reduction (SCR) catalyst. If a given catalyst component has multiple substrates installed directly in sequence, it is sufficient to record only the inlet temperature for the first catalyst substrate in the sequence. It is not necessary to record separate temperatures for substrates that are “zone-coated” with multiple catalyst functions. Record a representative outlet temperature for any major catalyst component that is used to elevate the temperature of downstream components. This could be the inlet of the next major component if that would be representative. We recommend that you record engine fuel rate to assist in the determination of sulfur exposure rates, but you may use other data for this purpose.
(2) When using Method 2, weighting of certification cycles, as described § 1065.1139(b)(2), record data for engine speed and engine load. Record sufficient ECM load parameters to determine a torque value that can be compared directly to engine torque as measured in the laboratory. You may optionally use ECM fuel rate measurements to determine load, but only if the same measurements can also be performed during laboratory testing on certification test cycles using sensors with comparable response characteristics. For example, you could use ECM fuel consumption rates for both field data and during laboratory tests.
(i) Optionally, as an alternative to the parameters required in this paragraph (b)(2), you may use a system exhaust temperature measurement to represent load. This requires one recorded temperature that represents the aftertreatment system. We recommend that you use a temperature recorded at the outlet of the first major catalyst component. If you choose to use this option, you must use the same temperature sensor for both field and laboratory measurements. Do not compare measurements between on-engine production temperature sensors with laboratory temperature sensors.
(ii) Optionally, as an alternative to the parameters required in this paragraph (b)(2), you may use exhaust flow and temperature measurements recorded in the field to support Method 2 calculations. Only one recorded temperature that represents the aftertreatment system is needed in this case. We recommend that you use a temperature recorded at the outlet of the first major catalyst component. Do not compare measurements between on-engine production temperature sensors with laboratory temperature sensors.
(3) If you have an aftertreatment system which involves periodic regeneration events where the temperature is raised above levels observed during normal operation, you must record data to characterize each such event. Data must be recorded at a frequency of at least 1 Hz, and you must record the exhaust flow rate and inlet temperature of each key catalyst component that will experience elevated temperatures during the regeneration. In addition, record a flag or variable that can be used to determine the beginning and end of a regeneration event. You must record at least three such events to allow determination of the average regeneration profile. If you have multiple types of regeneration events which influence different catalyst components in the system, you must record this data for each type of event separately. Use good engineering judgment to determine the average duration of each type of regeneration event, and the average interval of time between successive regeneration events of that type. You may use the data recorded for this cycle determination, or any other representative data to determine average regeneration duration or regeneration interval. These values may be determined from the analysis used to determine emission adjustments to account for infrequent regeneration of aftertreatment devices in § 1065.680.
Most compression-ignition engine aftertreatment systems contain multiple catalysts, each with their own aging characteristics. However, in the accelerated aging protocol the system will be aged as a whole. Therefore, it is necessary to determine which catalyst components are the key components that will be used for deriving and scaling the aging cycle.
(a) The primary aging catalyst in an aftertreatment system is the catalyst that is directly responsible for the majority of NOX reduction, such as a urea SCR catalyst in a compression ignition aftertreatment system. This catalyst will be used as the basis for cycle generation. If a system contains multiple SCR catalysts that are separated by other heat generating components that would result in a different rate of heat exposure, then each SCR catalyst must be tracked separately. Use good engineering judgment to determine when there are multiple primary catalyst components. An example of this would be a light-off SCR catalyst placed upstream of a DOC which is used to generate heat for regeneration and is followed by a DPF and a second downstream SCR catalyst. In this case, both the light-off SCR and the downstream SCR would have very different thermal history, and therefore must be tracked separately. In applications where there is no SCR catalyst in the aftertreatment system, the primary catalyst is the first oxidizing catalyst component in the system which is typically a DOC or catalyzed DPF.
(b) The secondary aging catalyst in an aftertreatment system is the catalyst that is intended to either alter exhaust characteristics or generate elevated temperature upstream of the primary catalyst. An example of a secondary component catalyst would be a DOC placed upstream of an SCR catalyst, with or without a DPF in between.
This section describes the method for determining the thermal reactivity coefficient(s) used for thermal heat load calculation in the accelerated aging protocol.
(a) The calculations for thermal degradation are based on the use of an Arrhenius rate law function to model cumulative thermal degradation due to heat exposure. Under this model, the thermal aging rate constant, k, is an exponential function of temperature which takes the form shown in the following equation:
Eq. 1065.1137-1
Where:
A = frequency factor or pre-exponential factor.
Ea = thermal reactivity coefficient.
R = molar gas constant.
T = catalyst temperature.
(b) The process of determining Ea begins with determining what catalyst characteristic will be tracked as the basis for measuring thermal deactivation. This metric varies for each type of catalyst and may be determined from the experimental data using good engineering judgment. We recommend the following metrics; however, you may also use a different metric based on good engineering judgment:
(1) Copper-based zeolite SCR. Total ammonia (NH3) storage capacity is a key aging metric for copper-zeolite SCR catalysts, and they typically contain multiple types of storage sites. It is typical to model these catalysts using two different storage sites, one of which is more active for NOX reduction, as this has been shown to be an effective metric for tracking thermal aging. In this case, there are two recommended aging metrics:
(i) The ratio between the storage capacity of the two sites, with more active site being in the denominator.
(ii) Storage capacity of the more active site.
(2) Iron-based zeolite SCR. Total NH3 storage capacity is a key aging metric for iron-zeolite SCR catalysts. Using a single storage site is the recommended metric for tracking thermal aging.
(3) Vanadium SCR. Brunauer-Emmett-Teller (BET) theory for determination of surface area is a key aging metric for vanadium-based SCR catalysts. Total NH3 storage capacity may also be used as a surrogate to probe the surface area. If you use NH3 storage to probe surface area, using a single storage site is the recommended metric for tracking thermal aging. You may also use low temperature NOX conversion as a metric. If you choose this option, you may be limited in your choice of temperatures for the experiment described in paragraph (c)(1) of this section due to vanadium volatility. In that case, it is possible that you may need to run a longer experimental duration than the recommended 64 hours to reach reliably measurable changes in NOX conversion.
(4) Zone-coated zeolite SCR. This type of catalyst is zone coated with both copper- and iron-based zeolite. As noted in paragraphs (b)(1) and (2) of this section, total NH3 storage capacity is a key aging metric, and each zone must be evaluated separately.
(5) Diesel oxidation catalysts. The key aging metric for tracking thermal aging for DOCs which are used to optimize exhaust characteristics for a downstream SCR system is the conversion rate of NO to NO2. Select a conversion rate temperature less than or equal to 200 °C using good engineering judgement. The key aging metric for DOCs, which are part of a system that does not contain an SCR catalyst for NOX reduction, is the HC reduction efficiency (as measured using ethylene). Select a conversion rate temperature less than or equal to 200 °C using good engineering judgement. This same guidance applies to an oxidation catalyst coated onto the surface of a DPF, if there is no other DOC in the system.
(c)
(1) Use good engineering judgment to select at least three different temperatures to complete the degradation experiments. We recommend selecting these temperatures to accelerate thermal deactivation such that measurable changes in the aging metric can be observed at multiple time points over the course of no more than 64 hours. Avoid temperatures that are too high to prevent rapid catalyst failure by a mechanism that does not represent normal aging. An example of temperatures to run the degradation experiment at for a small-pore copper zeolite SCR catalyst is 600 °C, 650 °C, and 725 °C.
(2) For each aging temperature selected, perform testing to assess the aging metric at different times. These time intervals do not need to be evenly spaced and it is typical to complete these experiments using increasing time intervals (e.g., after 2, 4, 8, 16, and 32 hours). Use good engineering judgment to stop each temperature experiment after sufficient data has been generated to characterize the shape of the deactivation behavior at a given temperature.
(i) For SCR-based NH3 storage capacity testing, perform a Temperature Programmed Desorption (TPD) following NH3 saturation of the catalyst (i.e., ramping gas temperature from 200 to 550 °C) to quantify total NH3 released during the TPD.
(ii) For DOC formulations, conduct an NO Reverse Light Off (RLO) to quantify oxidation conversion efficiency of NO to NO2 (i.e., ramping gas temperature from 500 to 150 °C).
(d) Generate a fit of the deactivation data generated in paragraph (b) of this section at each temperature.
(1) Copper-based zeolite SCR. Process all NH3 TPD data from each aging condition using an algorithm to fit the NH3 desorption data.
(i) We recommend that you use the Temkin adsorption model to quantify the NH3 TPD at each site to determine the desorption peaks of individual storage sites. The adsorption model is adapted from “Adsorption of Nitrogen and the Mechanism of Ammonia Decomposition Over Iron Catalysts” (Brunauer, S. et al, Journal of the American Chemical Society, 1942, 64 (4), 751-758) and “On Kinetic Modeling of Change in Active Sites upon Hydrothermal Aging of Cu-SSZ-13” (Daya, R. et al, Applied Catalysis B: Environmental, 2020, 263, 118368-118380). It is generalized using the following equation (assuming a two-site model):
Eq. 1065.1137-2
Where:
k = e−Ea(1−αθ)/RT
Ea = thermal reactivity coefficient of ammonia desorption.
α = Temkin constant.
θ = fraction of adsorption sites currently occupied (initial θ is assumed to be 1).
R = molar gas constant.
T = aging temperature.
(A) Use Eq. 1065.1137-2 to express the NH3 storage site desorption peaks as follows:
Eq. 1065.1137-3
Where:
N1 = moles of NH3 desorbed from Site 1.
A1 = pre-exponential factor associated with Site 1.
Ea,T1 = thermal reactivity coefficient of ammonia desorption for Site 1.
N2 = moles of NH3 desorbed from Site 2.
A2 = pre-exponential factor associated with Site 2.
Ea,T2 = thermal reactivity coefficient of ammonia desorption for Site 2.
(B) Optimize Ea,T1, α1, A1, Ea,T2, α2, and A2 to fit each NH3 TPD peak to give the best fit. The moles of NH3 (N1 and N2) may vary for each individual TPD data set.
(ii) Use one of the following modeling approaches to derive the thermal reactivity coefficient, Ea,D. We recommend that you use both models to fit the data and check that the resulting Ea,D values for the two methods are within 3% of each other.
(A) General Power Law Expression (GPLE). Generate a fit of the deactivation data from paragraph (d)(1)(i) of this section for each aging temperature using the following expression:
Eq. 1065.1137-4
Where:
kD = the thermal aging rate constant.
Eq. 1065.1137-5
A = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
Ω = N2/N1 or = N2 (normalizing Ω to the degreened Ω value for each new catalyst component prior to aging is recommended (i.e., Ω = 1 at t = 0 for each aging temperature).
Ωeq = aging metric at equilibrium (set = 0 unless there is a known activity minimum).
m = model order (assumed to be 2 for copper-based zeolite SCR).
(1) Solve Eq. 1065.1137-4 for Ω to yield the following expression:
Eq. 1065.1137-6
Where:
Ω0 = 1 (assumes that N2/N1 or = N2 values were normalized to the degreened value for each aging temperature).
A = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
t = aging time.
(2) Use a global fitting approach to solve for Ea,D and AD by applying a generalized reduced gradient (GRG) nonlinear minimization algorithm, or equivalent. For the global fitting approach, optimize the model by minimizing the Global Sum of Square Errors (SSEGlobal) between the experimental Ω and model Ω while only allowing Ea,D and AD to vary. Global SSE is defined as the summed total SSE for all aging temperatures evaluated.
Eq. 1065.1137-7
Where:
n = total number of aging temperatures.
i = an indexing variable that represents one aging temperature.
SEET = sum of square errors (SSE) for a single aging temperature, T, (see Eq. 1065.1137-8).
Eq. 1065.1137-8
Where:
n = total number of aging intervals for a single aging temperature.
i = an indexing variable that represents one aging interval for a single aging temperature.
ΩExp = experimentally derived aging metric for aging temperature, T.
Ωmodel = aging metric calculated from Eq. 1065.1137-6 for aging temperature, T.
(B) Arrhenius approach. In the Arrhenius approach, the deactivation rate constant, kD, of the aging metric, Ω, is calculated at each aging temperature.
(1) Generate a fit of the deactivation data in paragraph (d)(1)(i) of this section at each aging temperature using the following linear expression:
Eq. 1065.1137-9
Where:
Ω = N2/N1 or = N2 (Ω is to be normalized to the degreened Ω value for each new catalyst component prior to aging, i.e., Ω = 1 at t = 0 for each aging temperature).
(Eq. 1065.1137-5)
A = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
(2) Generate a plot of 1/Ω versus t for each aging temperature evaluated in paragraph (c)(1) in this section. The slope of each line is equal to the thermal aging rate, kD, at a given aging temperature. Using the data pairs of aging temperature and thermal aging rate constant, kD, determine the thermal reactivity coefficient, Ea, by performing a regression analysis of the natural log of kD versus the inverse of temperature, T, in Kelvin. Determine Ea,D from the slope of the resulting regression line, mdeactivation, using the following equation:
Ea,D = −mdeactivation · R
Eq. 1065.1137-10
Where:
mdeactivation = the slope of the regression line of ln(kD) versus 1/T.
R = molar gas constant.
(2) Iron-based zeolite or vanadium SCR. Process all NH3 TPD data from each aging condition using a GPLE to fit the NH3 desorption data (or BTE surface area data for vanadium SCR). Note that this expression is different from the one used in paragraph (d)(1)(ii)(A) of this section because the model order m is allowed to vary. This general expression takes the following form:
Eq. 1065.1137-11
Where:
Ω = total NH3 (or BET surface area) normalized to the degreened value for each new catalyst component prior to aging (i.e., Ω = 1 at t = 0 for each aging temperature).
A = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
t = time.
Ωeq = aging metric at equilibrium (set to 0 unless there is a known activity minimum).
m = model order.
(i) Solve Eq. 1065.1137-10 for Ω to yield the following expression:
Eq. 1065.1137-12
Where:
Ω0 = 1 (assumes total NH3 storage, or BET surface area, was normalized to the degreened value for each aging temperature).
A = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
t = aging time.
m = model order (to be varied from 1 to 8 using whole numbers).
(ii) Global fitting is to be used to solve for Ea,D and AD by applying a GRG nonlinear minimization algorithm, as described in paragraph (d)(1)(ii)(A) of this section. Minimize the SSEGlobal for each model order, m, while only allowing Ea,D and AD to vary. The optimal solution is determined by selecting the model order, m, that yields the lowest global fit SSE. If you have a range of model order solutions where the SSEGlobal does not vary substantially, use good engineering judgement to choose the lowest m for this range.
(3) Zone-coated zeolite SCR. Derive the thermal reactivity coefficient, Ea,D, for each zone of the SCR, based on the guidance provided in paragraphs (d)(1) and (2) of this section. The zone that yields the lowest Ea,D shall be used for calculating the target cumulative thermal load, as outlined in § 1065.1139.
(4) Diesel oxidation catalyst.
(i) The catalyst monolith is modeled as a plug flow reactor with first order reaction rate:
Eq. 1065.1137-13
Where:
v = velocity.
X = conversion (NO to NO2) in %/100.
V = volume of reactor.
Eq. 1065.1137-14
AD = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
(ii) For a diesel oxidation catalyst, the preexponential term AD is proportional to the number of active sites and is the desired aging metric. Solving Eq. 1065.1137-13 for kD, substituting it for kD in Eq. 1065.1137-5, and then solving for AD yields Eq. 1065.1137-15:
Eq. 1065.1137-15
Where:
SV = space velocity used during RLO testing.
X= conversion (NO to NO2).
Ea,D = thermal reactivity coefficient.
T = temperature where X was measured.
R = molar gas constant.
(iii) Process all NO to NO2 oxidation RLO data for each aging condition by determining the average oxidation conversion efficiency, X, at the temperature determined in paragraph (b)(5) of this section. We recommend maintaining the target oxidation conversion temperature to ±5 °C. For each aging condition (aging temperature, T and aging time, t), calculate the aging metric, Ω, by normalizing AD to the degreened AD value for each new catalyst component prior to aging (i.e., Ω = 1 at t = 0 for each aging temperature).
(A) Use the GPLE to fit the NO to NO2 conversion data, X, at each aging temperature. The GPLE takes the following form:
Eq. 1065.1137-16
Where:
Ω = aging metric for diesel oxidation catalysts.
R = molar gas constant.
T = aging temperature.
t = aging time.
Ωeq = aging metric at equilibrium (set to 0 unless there is a known activity minimum).
m = model order.
(B) Solve Eq. 1065.1137-12 for to yield the following expression:
Eq. 1065.1137-17
Where:
Ωeq = 1 (assumes the oxidation efficiency, X, was normalized to the degreened value for each aging temperature).
A = pre-exponential factor.
Ea,D = thermal reactivity coefficient.
R = molar gas constant.
T = aging temperature.
t = aging time.
m = model order (to be varied from 1 to 8 using whole numbers)
(iv) Use global fitting to solve for Ea,D and A by applying a GRG nonlinear minimization algorithm, as described in paragraph (d)(1)(ii)(A) of this section. Minimize the SSEGlobal for each model order, m, while only allowing Ea,D and A to vary. The optimal solution is determined by selecting the model order, m, that yields the lowest global fit SSE. If you have a range of model order solutions where the SSEGlobal does not vary substantially, use good engineering judgement to choose the lowest m for this range.
[89 FR 29827, Apr. 22, 2024]
Generation of the accelerated aging cycle for a given application involves analysis of the field data to determine a set of aging modes that will represent that field operation. There are two methods of cycle generation, each of which is described separately below. Method 1 involves the direct application of field data and is used when the recorded data includes sufficient exhaust flow and temperature data to allow for determination of aging conditions directly from the field data set and must be available for all of the key components. Method 2 is meant to be used when insufficient flow and temperature data is available from the field data. In Method 2, the field data is used to weight a set of modes derived from the laboratory certification cycles for a given application. These weighted modes are then combined with laboratory recorded flow and temperatures on the certification cycles to derive aging modes. There are two different cases to consider for aging cycle generation, depending on whether or not a given aftertreatment system incorporates the use of a periodic regeneration event. For the purposes of this section, a “regeneration” is any event where the operating temperature of some part of the aftertreatment system is raised beyond levels that are observed during normal (non-regeneration) operation. The analysis of regeneration data is considered separately from normal operating data.
(a) Cycle generation process overview. The process of cycle generation begins with the determination of the number of bench aging hours. The input into this calculation is the number of real or field hours that represent the useful life for the target application. This could be given as a number of hours or miles, and for miles, the manufacturer must use field data and good engineering judgment to translate this to an equivalent number of operating hours for the target application. The target for the accelerated aging protocol is a 10-time acceleration of the aging process, therefore the total number of aging hours is always set at useful life hours divided by 10. For example, if an on-highway heavy duty engine has a full useful life of 750,000 miles and this is determined to be represented by 24,150 field hours, the target duration for the DAAAC protocol for this application would be 2,415 bench-aging hours. The 2,415 hours will then be divided among different operating modes that will be arranged to result in repetitive temperature cycling over that period. For systems that incorporate periodic regeneration, the total duration will be split between regeneration and normal (non-regeneration) operation. The analysis of normal operation data is given in paragraph (b) of this section. The analysis of regeneration data is given in paragraph (c) of this section.
(b) Analysis of normal (non-regeneration) operating data. This analysis develops a reduced set of aging modes that represent normal operation. As noted earlier, there are two methods for conducting this analysis, based on the data available.
(1) Method 1—Direct clustering. Use Method 1 when sufficient exhaust flow and temperature data are available directly from the field data. The data requirements for Method 1 are described in § 1065.1133(b)(1). The method involves three steps: clustering analysis, mode consolidation, and cycle building.
(i) The primary method for determining modes from a field data set involves the use of k-means clustering. K-means clustering is a method where a series of observations is partitioned into set of clusters of “similar” data points, where every observation is a member of a cluster with the nearest mean, which is referred to as the centroid of that cluster. The number of clusters is a parameter of the analysis, and the k-means algorithm generally seeks an optimal number of clusters to minimize the least-squares distance of all points to their respective centroids. There are a number of different commercially available software programs to perform k-means clustering, as well as freely available algorithm codes. K-means clustering can arrive at many different solutions, and we are providing the following guidance to help select the optimal solution for use in accelerated aging cycle generation. The process involves analyzing the data multiple time using an increasing number of clusters for each analysis. Use at least 5 clusters, and we recommend developing solutions for the range between 5 and 8 clusters, although you may use more if desired. Each cluster is a potential aging mode with a temperature and flow rate defined by the centroid. More clusters result in more aging modes, although this number may be reduced later via model consolidation.
(ii) The cubic clustering criteria (CCC) is a metric calculated for each solution having a different number of clusters. The computation of CCC is complex and described in more detail in the following reference. The CCC computation is normally available as one of the metrics in commercially available software packages that can be used for k-means clustering. The optimal solution is typically the one with the number of clusters corresponding to the highest CCC.
(iii) Check each solution, starting with the one with the highest CCC to determine if it satisfies the following requirements:
(A) No more than one cluster contains fewer than 3% of the data points.
(B) The temperature ratio between the centroid with the maximum temperature and the centroid with the minimum temperature is at least 1.6 for clusters containing more than 3% of the data points.
(C) If that solution does not satisfy these requirements move to the solution with the next highest CCC.
(iv) The process described in paragraph (c)(1)(iii) of this section generally works well for most data sets, but if you have difficulty with the CCC metric in a particular data set, use good engineering judgment to leverage additional criteria to help the down-selection process. Examples of alternate clustering metrics include a Davies-Bouldin Index (optimizing on the minimum value) or a Calinski-Harabasz Index (optimize on the maximum value).
(v) The initial candidate mode conditions are temperature and flow rate combinations that are the centroids for each cluster from the analysis in paragraph (c)(1)(iii) of this section. As part of the analysis, you must also determine the 10th percentile and 90th percentile temperatures for each cluster. These additional values may be needed later for the cycle heat load tuning process described in § 1065.1143.
(vi) The mode weight factor for a given cluster is the fraction data points contained within that cluster.
(2) Method 2—Cluster-based weighting of certification cycle modes. Use Method 2 if there is insufficient exhaust flow and temperature data from the field at the time the cycle is being developed. The data requirements for Method 2 are described in § 1065.1133(b)(2). You also need laboratory data recorded in the form of 1 Hz data sets for the regulatory duty cycles you are certifying to for your application as described in the standard setting part. Include exhaust flow rate and the inlet temperature for each key catalyst component in the laboratory data sets, as described in paragraph (e) of this section. The laboratory data sets must also include parameters that match the field data as described in § 1065.1133(b)(2), which will be used to facilitate the clustering analysis.
(i) Perform k-means clustering is described in § 1065.1133(b)(1) but using data sets containing the two parameters recorded in the field data sets. For example, you might use speed and torque, as recorded both in the field and the laboratory for Method 2 clustering.
(ii) Determine the fraction of points from each of the regulatory laboratory duty-cycles that are within each cluster, in addition to the overall fraction of points from the entire data set.
(iii) For each cycle, calculate a square sum error, SSE, as follows:
Where:
i = an indexing variable that represents one cluster.
N = total number of clusters.
Cycleprob = the fraction of points in a given cluster, i, for the regulatory duty-cycle of interest.
RefDataprob = the fraction of points in a given cluster, i, for the full data set.
(iv) For each cycle, calculate a dissimilarity index as follows:
(v) If you have more than one regulatory duty cycle, weight the regulatory cycles.
(A) Determine the weighting factors for a given regulatory cycle, wi, by solving a system of equations:
Where:
di = dissimilarity for a given regulatory cycle, i.
dj = dissimilarity for a given regulatory cycle, j.
(B) For example, for three duty cycles, calculate w1 as follows:
(C) Calculate subsequent wi values after calculating w1 as follows:
(D) Calculate the sum of the weighting factors to verify that they are equal to one.
Where:
n = number of regulatory cycles for the application.
(vi) For each regulatory cycle determine the average exhaust flow and the average inlet temperature for each key catalyst. Determine the 25th and 90th percentile inlet temperatures for the primary catalyst and the respective associated exhaust flow rate for each data point.
(vii) Use the cycle weights from paragraph (b)(2)(v) of this section and the mode conditions from paragraph (b)(2)(vi) of this section to generate a set of candidate aging modes by multiplying the cycle weight factor, w[cycle] by 0.25 for the 25th percentile temperature mode, 0.65 for the 50th percentile temperature mode, and by 0.10 for the 90th percentile temperature mode. This will generate a weighted set of mode numbers three times the number of regulatory cycles for the target application. Each mode will have a target temperature and exhaust flow rate.
(viii) If you have only one regulatory cycle for your application, use the cycle modes and weighting factors as they are given in the standard setting part.
(3) Determination of mode total durations. The output for either method will be a set of mode exhaust conditions, with an associated weighting factor for each mode. Multiply the mode weight factors by the total number of normal operating (non-regenerating) hours, to get a target mode duration for each mode. This will be used in the heat load calculations.
(c) Mode consolidation. Sometimes the clustering analysis process will generate multiple modes that are very similar to each other in temperature, such that although they are distinct modes they will not have a significantly different impact on aftertreatment aging. To reduce the complexity of the aging cycle, you may consolidate modes that are similar into a single mode as described below.
(1) Consolidate any two or more modes which have a target temperature within 10 °C into a single mode. If you choose to do this, the target temperature of the single consolidated mode is the temperature associated with the highest weight factor mode before consolidation. If the modes being consolidated all have weighting factors within 0.05 of each other, use the highest temperature among the modes.
(2) Use the highest exhaust flow target among the modes being combined as the target exhaust flow for new consolidate mode.
(3) Use the combined sum of the weighting factors for all modes being consolidate as the weighting factor for the new consolidated mode. Similarly, the total duration of the new consolidated mode is the sum of the durations of the modes being consolidated.
(d) Analysis of regeneration data. Regeneration data is treated separately from the normal operating mode data. Generally, the target for accelerated aging cycle operation is to run all of the regenerations that would be expected over the course of useful life. If multiple types of regeneration are conducted on different system components, each type of regeneration must be analyzed separately using the steps in this paragraph (d). The data requirements for input into this process are described in § 1065.1133(b)(3). The process described below is meant to determine a representative regeneration profile that will be used during aging. You may also ask us to allow the use of other engineering data or analysis to determine a representative regeneration profile.
(1) The total number of regenerations that will be run during the accelerated aging process will be the same as the total number of regenerations over useful life. Calculate this number by dividing the total number of useful life hours by the interval between regenerations as determined in § 1065.1133(b)(3).
(2) Use the 1 Hz regeneration data to determine an appropriate regeneration profile. The recorded regeneration event begins when the engine indicates it has started regeneration using the recorded regeneration indicator and ends when the aftertreatment has returned back to the normal operating temperature after the flag indicates the regeneration is complete.
(3) For each recorded regeneration, calculate the cumulative deactivation, Dt, using the equations in paragraph (e) of this section.
(4) If you have a large number of recorded regenerations in your data set, select a regeneration event with a cumulative deactivation representing the 75th percentile of the distribution of heat loads in your recorded data set. If you have a smaller number of recorded regenerations, such that you cannot clearly identify the real distribution, select the recorded regeneration with the highest recorded cumulative deactivation.
(5) This regeneration event will be used as the regeneration profile for that type of event during aging. The profile should include the entire event, include the temperature ramp and cool-down period.
(6) The regeneration must be conducted in the same manner as it is run in the field. For instance, if the regeneration temperature is generated from an exothermic reaction by injecting fuel in front of a DOC, this methodology should also be used during bench aging.
(7) If part of the system is at a lower temperature during regeneration because it is upstream of the temperature generating component, the set the target temperature for the aftertreatment system inlet to be equivalent to the system inlet temperature used during the highest duration non-regeneration mode, or 350 °C, whichever is lower.
(e) Heat load calculation and tuning for systems that have regeneration events. Perform this procedure after the preliminary cycles are completed for both normal and regeneration operation. The target cumulative deactivation is determined from the input field data, and then a similar calculation is performed for the preliminary aging cycle. If the cumulative deactivation for the preliminary cycle does not match cumulative deactivation from the field data, then the cycle is tuned over a series of steps until the target is matched.
(1) The deactivation for a given catalyst is calculated for each time step as follows:
Where:
Di = incremental deactivation for time step i.
Ea = thermal reactivity coefficient for the catalyst as determined in § 1065.1137.
R = molar gas constant in kJ/mol·K.
Tstd = standard temperature = 293.15 K.
T = catalyst temperature in K.
(2) Calculate the cumulative deactivation, Dt, for a given catalyst over a series of time steps, N, using the following equation:
Where:
i = an indexing variable that represents one time step.
N = total number of cumulative deactivation time steps in the data set.
Di = incremental deactivation for each time step.
(3) Calculate the cumulative deactivation, Dt, for the input field data set. The time step for the calculations should be 1 second for 1-Hz input data.
(i) First calculate Dt for the non-regeneration portion of the field data set. For Method 2 use the 1-Hz data from the regulatory cycles as the field data set.
(ii) Divide the calculate field Dt by the number of hours represented in the field data set.
(iii) Multiply the hourly Dt by the number of hours required to reach full useful life. This is the target Dt,field-normi.
(iv) Multiply the total number of regenerations for full useful life by the cumulative deactivation Dt for the target regeneration profile determined in paragraph (d)(4) of this section. This is the target Dt,field-regen.
(v) The total target cumulative deactivation for the field data, Dt,field, is the sum of Dt,field-normi and Dt,field-regen.
(4) Calculate the cumulative deactivation for the candidate aging cycle generated under paragraphs (c) and (d) of this section as follows:
(i) Using the modes and mode durations for normal operation generated in paragraph (c) of this section, calculate the cumulative deactivation, Dt,cycle-norm, using the method given in paragraph (e)(2) of this section.
(ii) The total cumulative deactivation for the candidate aging cycle, Dt, is the sum of Dt,cycle-norm and Dt,field-regen.
(5) If Dt,cycle is within ±1% of Dt,field, the candidate cycle is deemed representative and may be used for aging.
(6) If Dt,cycle is not within ±1% of Dt,field, the candidate cycle must be adjusted to meet this criterion using the following steps. It should be noted that if the Dt,cycle is outside of the criteria it will usually be lower than the Dt,field.
(i) Increase the duration of the stable portion of the regeneration profile, which is defined as the portion of the regeneration profile where the temperature has completed ramping and is being controlled to a stationary target temperature. Note that this will increase the number of hours of regeneration time. You must compensate for this by decreasing the total number of normal operation (non-regeneration) hours in the cycle. Recalculate the duration of all the normal operation modes. You may not increase the duration of the stable portion of the regeneration profile by more than a factor of 2. If you reach this limit and you still do not meet the criteria in paragraph (e)(5) of this section, proceed to the next step.
(ii) Increase the target temperature of the stable portion of the regeneration profile by the amount necessary to reach the target criteria. You may not increase this temperature higher than the temperature observed in the regeneration profile with the highest Dt observed in the field. If you reach this limit and you still do not meet the criteria in paragraph (e)(5) of this section, proceed to the next step.
(iii) Increase the target temperature of the highest temperature normal operation mode. You may not increase this temperature above the 90th percentile determined in paragraph (b)(1)(v) of this section for Method 1, or above the maximum temperature for the regulatory cycle from which the mode was derived for Method 2. If you reach this limit and you still do not meet the criteria in paragraph (e)(5) of this section, you may repeat this step using the next highest temperature mode, until you reach the target, or all modes have been adjusted.
(iv) If you are unable to reach the target deactivation by following paragraphs (e)(6)(i) through (iii) of this section, use good engineering judgment to increase the number of regenerations to meet the criteria in paragraph (e)(5) of this section. Note that this will increase the total regeneration hours, therefore you must decrease the number of normal operation hours and re-calculate mode durations for the normal operation modes.
(v) If you are not able to achieve the target Dt,field using the steps in paragraphs (e)(6)(i) through (iv) of this section without exceeding catalyst temperature limits, use good engineering judgement to reduce the acceleration factor from 10 to a lower number. If you reduce the acceleration factor you must re-calculate the number of hours determine in paragraph (a) of this section and re-run the process in this paragraph (e). Note that if you reduce the acceleration factor you must use the same lower acceleration factor in the chemical exposure calculations in paragraph (h) of this section, instead of 10.
(f) Heat load calculation and tuning for systems that do not have regeneration events. Follow the steps described for systems with regeneration events to calculate Dt,field and Dt,cycle, omitting the steps related to regeneration events. The Dt,cycle will be well below the Dt,field. Follow the steps given below to adjust the cycle until you meet the criteria in paragraph (e)(5) of this section.
(1) Increase the temperature of the highest temperature mode. Use good engineering judgment to ensure that this temperature does not exceed the limits of the catalyst in a way that might cause rapid deactivation or failure via a mechanism that is not considered normal degradation.
(2) Increase the duration of the highest temperature mode and decrease the duration of the other modes in proportion. You may not increase the duration highest temperature mode by more than a factor of 2.
(3) If you are not able to achieve the target Dt,field using the steps in paragraphs (f)(1) and (2) of this section without exceeding catalyst temperature limits, use good engineering judgement to reduce the acceleration factor from 10 to a lower number. If you reduce the acceleration factor you must re-calculate the number of hours determine in paragraph (a) of this section and re-run the process in this paragraph (f). Note that if you reduce the acceleration factor you must use the same lower acceleration factor in the chemical exposure calculations in paragraph (h) of this section, instead of 10.
(g) Final aging cycle assembly. The final step of aging cycle development is the assembly of the actual cycle based on the mode data from either paragraph (e) of this section for systems with infrequent regeneration, or paragraph (f) of this section for systems that do not incorporate infrequent regeneration. This cycle will repeat a number of times until the total target aging duration has been reached.
(1) Cycle assembly with infrequent regenerations. For systems that use infrequent regenerations, the number of cycle repeats is equal to the number of regeneration events that happen over full useful life. The total cycle duration of the aging cycle is calculated as the total aging duration in hours divided by the number of infrequent regeneration events. In the case of systems with multiple types of infrequent regenerations, use the regeneration with the lowest frequency to calculate the cycle duration.
(i) If you have multiple types of infrequent regenerations, arrange the more frequent regenerations such that they are spaced evenly throughout the cycle.
(ii) Determine the length of the normal (non-regeneration) part of the cycle by subtracting the regeneration duration, including any regeneration extension determined as part of cycle tuning from paragraph (e) of this section, from the total cycle duration. If you have multiple types of regeneration, then the combined total duration of regeneration events performed in the cycle must be subtracted from the total. For example, if you have one type of regeneration that is performed for 30 minutes every 30 cycle hours, and a second type that is performed for 30 minutes every 10 cycle hours (such that 3 of these secondary events will happen during each cycle), then you would subtract a total of 2 hours of regeneration time from the total cycle duration considering all 4 of these events.
(iii) Divide the duration of the normal part of the cycle into modes based on the final weighting factors determined in paragraph (c) of this section following any mode consolidation.
(iv) Place the mode with the lowest temperature first, then move to the highest temperature mode, followed by the next lowest temperature mode, and then the next highest mode, continuing in this alternating pattern until all modes are included.
(v) Transition between normal modes within (60 to 300) seconds. The transition period is considered complete when you are within ±5 °C of the target temperature for the primary key component. Transitions may follow any pattern of flow and temperature to reach this target within the required 300 seconds.
(vi) For normal modes longer than 30 minutes, you may count the transition time as time in mode. Account for the transition time for modes shorter than 30 minutes by shortening the duration of the longest mode by an equivalent amount of time.
(vii) If the shortest normal operating mode is longer than 60 minutes, you must divide the normal cycle into shorter sub-cycles with the same pattern in paragraph (g)(1)(iii) of this section, but with shorter durations, so that the pattern repeats two or more times. You must divide the cycle into sub-cycles until the duration of the shortest mode in each sub-cycle is no longer than 30 minutes. No mode may have a duration shorter than 15 minutes, not including transition time.
(viii) If a regeneration event is scheduled to occur during a normal mode, shift the start of regeneration to the end of the nearest normal mode.
(2) Cycle assembly without infrequent regenerations. For systems that do not use infrequent regenerations, the cycle will be arranged to achieve as much thermal cycling as possible using the following steps.
(i) Assign a duration of 15 minutes to the mode with the lowest weight factor. Calculate the duration of the remaining modes in proportion to the final weight factors after mode durations have been adjusted during heat load tuning in paragraph (f) of this section.
(ii) Place the mode with the lowest temperature first, then move to the highest temperature mode, followed by the next lowest temperature mode, and then the next highest mode, continuing in this alternating pattern until all modes are included.
(iii) Transition between normal modes within (60 to 300) seconds. The transition period is considered complete when you are within ±5 °C of the target temperature for the primary key component. Transitions may follow any pattern of flow and temperature to reach this target within the required 300 seconds.
(iv) For normal modes longer than 30 minutes, you may count the transition time as time in mode. Account for the transition time for modes shorter than 30 minutes by shortening the duration of the longest mode by an equivalent amount of time.
(v) This cycle will be repeated the number of times necessary to reach the target aging duration.
(h) Chemical exposure targets. Determine targets for accelerated oil and fuel sulfur exposure as follows:
(1) Oil exposure targets. The target oil exposure rate during accelerated aging is 10 times the field average oil consumption rate determined in § 1065.1133(a)(2). You must achieve this target exposure rate on a cycle average basis during aging. Use good engineering judgment to determine the oil exposure rates for individual operating modes that will achieve this cycle average target. For engine-based aging stands you will likely have different oil consumption rates for different modes depending on the speed and load conditions you set. For burner-based aging stands, you may find that you have to limit oil exposure rates at low exhaust flow or low temperature modes to ensure good atomization of injected oil. On a cycle average basis, the portion of oil exposure from the volatile introduction pathway (i.e., oil doped in the burner or engine fuel) must be between (10 to 30) % of the total. The remainder of oil exposure must be introduced through bulk pathway.
(2) Fuel sulfur exposure targets. The target sulfur exposure rate for fuel-related sulfur is determined by utilizing the field mean fuel rate data for the engine determined in § 1065.1133(a)(3). Calculate the total sulfur exposure mass using this mean fuel rate, the total number of non-accelerated hours to reach full useful life, and a fuel sulfur level of 10 ppmw.
(i) For an engine-based aging stand, if you perform accelerated sulfur exposure by additizing engine fuel to a higher sulfur level, determine the accelerated aging target additized fuel sulfur mass fraction, wS, as follows:
Eq. 1065.1139-9
Where:
ṁ̅fuel,field = field mean fuel flow rate.
ṁ̅fuel,cycle = accelerated aging cylce mean fuel low rate.
mSfuel,ref = reference mass of sulfur per mass of fuel = 0.00001 kg/kg.
Sacc,rate = sulfur acceleration rate = 10.
Example:
m ̇̅fuel,field= 54.3 kg/hr
ṁ̅fuel,cycle = 34.1 kg/hr
mSfuel,ref = 0.00001 kg/kg.
Sacc,rate = 10
(ii) If you use gaseous SO2 to perform accelerated sulfur exposure, such as on a burner-based stand, calculate the target SO2 concentration to be introduced, xSO2,target, as follows:
Eq. 1065.1139-10
Where:
ṁ̅fuel,field = field mean fuel flow rate.
ṁ̅exhaust,cycle = mean exhaust flow rate during the burner aging cycle.
xSfuel,ref = reference mol fraction of sulfur in fuel = 10 µmol/mol.
Sacc,rate = sulfur acceleration rate = 10.
Mexh = molar mass of exhaust = molar mass of air.
MS = molar mass of sulfur.
Example:
ṁ̅fuel,field = 54.3 kg/hr
ṁ̅exhaust,cycle = 1000.8 kg/hr
xSfuel,ref = 10 µmol/mol
Sacc,rate = 10
Mexh = 28.96559 g/mol
MS = 32.065 g/mol
(iii) You may choose to turn off gaseous sulfur injection during infrequent regeneration modes, but if you do you must increase the target SO2 concentration by the ratio of total aging time to total normal (non-regeneration) aging time.
[79 FR 23820, Apr. 28, 2014, as amended at 89 FR 29829, Apr. 22, 2024]
An engine-based accelerated aging platform is built around the use of a compression-ignition engine for generation of heat and flow. You are not required to use the same engine as the target application that is being aged. You may use any compression-ignition engine as a bench aging engine, and the engine may be modified as needed to support meeting the aging procedure requirements. You may use the same bench aging engine for deterioration factor determination from multiple engine families. The engine must be capable of reaching the combination of temperature, flow, NOX, and oil consumption targets required. We recommend using an engine platform larger than the target application for a given aftertreatment system to provide more flexibility to achieve the target conditions and oil consumption rates. You may modify the bench aging engine controls in any manner necessary to help reach aging conditions. You may bypass some of the bench aging engine exhaust around the aftertreatment system being aged to reach targets, but you must account for this in all calculations and monitoring to ensure that the correct amount of oil and sulfur are reaching the aftertreatment system. If you bypass some of the engine exhaust around the aftertreatment system, you must directly measure exhaust flow rate through the aftertreatment system. You may dilute bench aging engine exhaust prior to introduction to the aftertreatment system, but you must account for this in all calculations and monitoring to ensure that the correct engine conditions and the correct amount of oil and sulfur are reaching the aftertreatment system. Your engine-based aging stand must incorporate the following capabilities:
(a) Use good engineering judgment to incorporate a means of controlling temperature independent of the engine. An example of such a temperature control would be an air-to-air heat exchanger. The temperature control system must be designed to prevent condensation in the exhaust upstream of the aftertreatment system. This independent temperature control is necessary to provide the flexibility required to reach temperature, flow, oil consumption targets, and NOX targets.
(b) Use good engineering judgment to modify the engine to increase oil consumption rates to levels required for accelerated aging. These increased oil consumption levels must be sufficient to reach the bulk pathway exposure targets determined in § 1065.1139(h). A combination of engine modifications and careful operating mode selection will be used to reach the final bulk pathway oil exposure target on a cycle average. You must modify the engine in a fashion that will increase oil consumption in a manner such that the oil consumption is still generally representative of oil passing the piston rings into the cylinder. Use good engineering judgment to break in the modified engine to stabilize oil consumption rates. We recommend the following methods of modification (in order of preference):
(1) Install the second compression ring inverted (upside down) on one or more of the cylinders of the bench aging engine. This is most effective on rings that feature a sloped design to promote oil control when normally installed.
(2) If the approach in paragraph (b)(1) of this section is insufficient to reach the targets, modify the oil control rings in one or more cylinders to reduce the spring tension on the oil control ring. It should be noted that this is likely to be an iterative process until the correct modification has been determined.
(3) If the approach in paragraph (b)(2) of this section is insufficient to reach the targets, modify the oil control rings in one or more cylinders to create small notches or gaps (usually no more than 2 per cylinder) in the top portion of the oil control rings that contact the cylinder liner (care must be taken to avoid compromising the structural integrity of the ring itself).
(c) We recommend that the engine-aging stand include a constant volume oil system with a sufficiently large oil reservoir to avoid oil “top-offs” between oil change intervals.
(d) If the engine-aging stand will be used for aging of systems that perform infrequent regenerations, the aging stand must incorporate a means of increasing temperature representative of the target application. For example, if the target application increases temperature for regeneration by introducing fuel into the exhaust upstream of an oxidation catalyst, the aging stand must incorporate a similar method of introducing fuel into the exhaust.
(e) If the engine-aging stand will be used for aging systems that incorporate SCR-based NOX reduction, the aging stand must incorporate a representative means of introducing DEF at the appropriate location(s).
(f) Use good engineering judgment to incorporate a means of monitoring oil consumption on a periodic basis. You may use a periodic drain and weigh approach to quantify oil consumption. We recommend that you incorporate a method of continuous oil consumption monitoring, but you must validate that method with periodic draining and weighing of the engine oil. You must validate that the aging stand reaches oil consumption targets prior to the start of aging. You must verify oil consumption during aging prior to each emission testing point, and at each oil change interval. Validate or verify oil consumption over a running period of at least 72 hours to obtain a valid measurement. If you do not include the constant volume oil system recommended in paragraph (c) of this section, you must account for all oil additions.
(g) Use good engin eering judgment to establish an oil change interval that allows you to maintain relatively stable oil consumption rates over the aging process. Note that this interval may be shorter than the normal recommended interval for the engine due to the modifications that have been made.
(h) If the engine-aging stand will be used for aging of systems that incorporate a diesel particulate filter (DPF), we recommend you perform secondary tracking of oil exposure by using clean (soot free) DPF weights to track ash loading and compare this mass of ash to the amount predicted using the measured oil consumption mass and the oil ash concentration. The mass of ash found by DPF weight should fall within (55 to 70)% of the of mass predicted from oil consumption measurements.
(i) Incorporate a means of introducing lubricating oil into the engine fuel to enable the volatile pathway of oil exposure. You must introduce sufficient oil to reach the volatile pathway oil exposure targets determined in paragraph (h) of this section. You must measure the rate of volatile pathway oil introduction on a continuous basis.
(j) If you perform sulfur acceleration by increasing the sulfur level of the engine fuel, you must meet the target sulfur level within ±5 ppmw. Verify the sulfur level of the fuel prior to starting aging, or whenever a new batch of aging fuel is acquired.
(k) If you use gaseous SO2 for sulfur acceleration, you must incorporate a means to introduce the gaseous SO2 upstream of the aftertreatment system. Use good engineering judgment to ensure that gaseous SO2 is well mixed prior to entering the aftertreatment system. You must monitor the rate of gaseous SO2 introduction on a continuous basis.
[79 FR 23820, Apr. 28, 2014, as amended at 89 FR 29831, Apr. 22, 2024]
A burner-based aging platform is built using a fuel-fired burner as the primary heat generation mechanism. The burner must utilize diesel fuel and it must produce a lean exhaust gas mixture. You must configure the burner system to be capable of controlling temperature, exhaust flow rate, NOX, oxygen, and water to produce a representative exhaust mixture that meets the accelerated aging cycle targets for the aftertreatment system to be aged. You may bypass some of the bench aging exhaust around the aftertreatment system being aged to reach targets, but you must account for this in all calculations and monitoring to ensure that the correct amount of oil and sulfur are reaching the aftertreatment system. The burner system must incorporate the following capabilities:
(a) Directly measure the exhaust flow through the aftertreatment system being aged.
(b) Ensure transient response of the system is sufficient to meet the cycle transition time targets for all parameters.
(c) Incorporate a means of oxygen and water control such that the burner system is able to generate oxygen and water levels representative of compression-ignition engine exhaust.
(d) Incorporate a means of oil introduction for the bulk pathway. You must implement a method that introduces lubricating oil in a region of the burner that does not result in complete combustion of the oil, but at the same time is hot enough to oxidize oil and oil additives in a manner similar to what occurs when oil enters the cylinder of an engine past the piston rings. Care must be taken to ensure the oil is properly atomized and mixed into the post-combustion burner gases before they have cooled to normal exhaust temperatures, to insure proper digestion and oxidation of the oil constituents. You must measure the bulk pathway oil injection rate on a continuous basis. You must validate that this method produces representative oil products using the secondary method in § 1065.1141(h) regardless of whether you will use the burner-based aging stand to age systems which include a DPF. Use good engineering judgment to select a DPF for the initial validation of the system. Perform this validation when the burner-based aging stand is first commissioned or if any system modifications are made that affect the oil consumption introduction method. We also recommend that you examine ash distribution on the validation DPF in comparison to a representative engine aged DPF.
(e) Incorporate a means of introducing lubricating oil into the burner fuel to enable the volatile pathway of oil exposure. You must introduce sufficient oil to reach the volatile pathway oil exposure targets determined in § 1065.1139(h). You must measure the rate of volatile pathway oil introduction on a continuous basis.
(f) If the burner-based aging stand will be used for aging of systems that perform infrequent regenerations, the aging stand must incorporate a means of increasing temperature representative of the target application. For example, if the target application increases temperature for regeneration by introducing fuel into the exhaust upstream of an oxidation catalyst, the aging stand must incorporate a similar method of introducing fuel into the exhaust.
(g) If the burner-based aging stand will be used for aging of systems that incorporate SCR-based NOX reduction, the aging stand must incorporate a representative means of introducing DEF at the appropriate location(s).
(h) If the burner-based aging stand will be used for aging of systems that incorporate a diesel particulate filter (DPF), we recommend you perform secondary tracking of oil exposure by using clean (soot free) DPF weights to track ash loading and compare this mass of ash to the amount predicted using the measured oil consumption mass and the oil ash concentration. The mass of ash found by DPF weight should fall within (55 to 70)% of the of mass predicted from oil consumption measurements.
(i) You must incorporate a means to introduce the gaseous SO2 upstream of the aftertreatment system. Use good engineering judgment to ensure that gaseous SO2 is well mixed prior to entering the aftertreatment system. You must monitor the rate of gaseous SO2 introduction on a continuous basis.
The aging cycle generally consists first of practice runs to validate and tune the final cycle, followed by the actual running of the repeat cycles needed to accumulate field equivalent hours to reach full useful life. During the course of the aging run, various aging parameters are tracked to allow verification of proper cycle execution, as well as to allow for correction of the aging parameters to stay within the target limits.
(a) Preliminary cycle validation runs. Prior to the start of aging, conduct a number of practice runs to tune the cycle parameters. It is recommended that initial practice runs be conducted without the aftertreatment installed, but with the backpressure of the aftertreatment simulated to help ensure that the tuned cycle is representative. For final cycle tuning, including regenerations, it is recommended to use a duplicate or spare aftertreatment system of similar design to the target system, to avoid damage or excessive initial aging during the tuning. However, it is permissible to conduct final tuning using the target system being aged, but you must limit the total duration to no more than 100 field equivalent hours (10 hours of accelerated aging), including both thermal and chemical components. The process followed for these initial runs will vary depending on whether you are using an engine-based platform or a burner-based platform.
(1) Engine-based platform.
(i) Initial cycle development. It will be necessary to determine a set of engine modes that will generate the required combinations of temperature, exhaust flow, oil consumption, and NOX to meet the target aging requirements. The development of these modes will be an iterative process using the engine and independent temperature control features of the aging stand. This process assumes that you have already implemented the oil consumption increase modifications, and that these have already been stabilized and validated to reach the necessary levels of bulk oil exposure. In general, we recommend the use of higher engine speeds and loads to generate the desired oil consumption, leveraging the temperature controls as needed to lower temperature to the targets. Several iterations will likely be needed to reach all targets. Note that during transitions you may utilize any combination of conditions necessary to help primary component catalysts reach the target temperature and flow conditions within no more than 5 minutes. For example, you may use a higher exhaust flow rate and lower temperature to rapidly cool the aftertreatment system to the next temperature. NOX targets do not need to be met during transitions. It is permissible to deviate from engine-out NOX emission targets if needed to reach the temperature, exhaust flow, and oil consumption targets. We recommend that you maintain a NOX level that is at the target level or higher, but you may lower NOX by up to 25%, if necessary, on some modes. Note that validation of oil consumption requires at least 72 hours of operation. Tune the parameters for infrequent regeneration towards then end of this initial development process (such as hydrocarbon injection schedules and temperature ramp rates).
(ii) Final cycle validation. Once the cycle is tuned, conduct a final run using the target aftertreatment system to verify conditions and log temperatures for heat load calculation. Using the recorded cycle data, calculate Dt for all primary component catalysts to ensure that you are matching the desired Dt,cycle targets. If you are not within ±3% of the target Dt,cycle, adjust the cycle accordingly. Calculate Dt for any secondary catalyst components to verify that they are within ±3% of either the target Dt or the target aging metric. Note that the accelerated aging methodology assumes that the relationship between the temperature of the primary and secondary catalyst components will the be same as the field observations. If this relationship deviates in the lab by having more or less heat transfer through the system, it may be necessary to modify that relationship on the aging stand. You may need to take measures such as adding or removing insulation or utilize external cooling fans to help these parameters match more closely.
(2) Burner-based platform.
(i) Cycle development. The burner-based platform will be able to meet the exhaust flow, temperature, NOX, and oil consumption targets directly without the need for additional cycle development. This process assumes that you have already implemented and validated your oil consumption exposure methods to reach the necessary levels of bulk oil exposure. In addition, you must meet the oxygen and water targets during aging modes within ±2% for oxygen and ±2% for water. Note that during transitions you may utilize any combination of conditions necessary to help primary component catalysts reach the target temperature and flow conditions within no more than 5 minutes. For example, you may use a higher exhaust flow rate and lower temperature to rapidly cool the aftertreatment system to the next temperature. NOX, oxygen, and water targets do not need to be met during transitions.
(ii) Final cycle validation. Once the cycle is tuned, conduct a final run using the target aftertreatment system to verify conditions and log temperatures for heat load calculation. Using the recorded cycle data, calculate Dt for all primary components catalysts to ensure that you are matching the desired Dt,cycle targets. If you are not within ±3% of the target Dt,cycle, adjust the cycle accordingly. Calculate Dt for any secondary catalyst components to check that they are within ±3% of either the target Dt or the target aging metric. Note that the accelerated aging methodology assumes that the relationship between the temperature of the primary and secondary catalyst components will the be same as that observed in the field. If this relationship deviates in the lab by having more or less heat transfer through the system, it may be necessary to modify that relationship on the aging stand. You may need to take measures such as adding or removing insulation or utilize external cooling fans to help these parameters match more closely.
(b) Aftertreatment break in. Break in the emission-data engine and aftertreatment prior to the initial zero-hour test by running both on an engine dynamometer as described in subpart E of this part. Use good engineering judgment to develop a representative cycle that represents the field data. You may use the same data used for accelerated aging cycle development or other data. If your system utilizes infrequent regeneration, include at least one complete regeneration event, but we recommend that you include at least two such events to stabilize emissions performance. Your break in process must include at least 125 hours of engine operation with the aftertreatment system. You may ask to use a longer break in duration based on good engineering judgment, to ensure that emission performance is stabilized prior to the zero-hour testing.
(c) Initial emission testing. Prior to the start of accelerated aging conduct the initial zero-hour emission test and any required engine dynamometer aging following the requirements of the standard setting part for your engine. Dynaometer aging hours count toward the total aging hours.
(d) Accelerated aging. Following zero-hour emission testing and any engine dynamometer aging, perform accelerated aging using the cycle validated in either paragraph (a)(1) or (2) of this section. Repeat the cycle the number of times required to reach full useful life equivalent aging. Interrupt the aging cycle as needed to conduct any scheduled intermediate emission tests, clean the DPF of accumulated ash, and for any facility-related reasons. We recommended you interrupt aging at the end of a given aging cycle, following the completion of any scheduled infrequent regeneration event. If an aging cycle is paused for any reason, we recommended that you resume the aging cycle at the same point in the cycle where it stopped to ensure consistent thermal and chemical exposure of the aftertreatment system.
(e) QA tracking and validation. During aging, track a number of aging parameters to ensure that fall within the required limits. Correct aging parameters as need to remain within the required control limits.
(1) Thermal load tracking. For each primary catalyst component, generate a target line which describes the relationship between aging hours on the cycle and cumulative deactivation, Dt. Generate control limit lines that are ±3% of the target line. You must remain within these control limits over the course of aging. Adjust aging parameters as needed to remain within these limits for the primary catalyst components. For each secondary catalyst component, generate both a target Dt line and a line describing the target behavior of the aging metric directly. You must remain within either ±10% of either the Dt line or ±3% of the aging metric target line for any secondary catalyst component. Adjust aging parameters as needed to remain within these limits noting that you must remain within limits for the primary components. Adjusting the secondary catalyst aging may require altering heat transfer through the system to make it more representative of the field aging.
(2) Oil consumption tracking. Generate a target oil consumption line for both the bulk and volatile pathway which describes the relationship between oil exposure and aging hours on the cycle. For the engine-based stand the control limits are ±10% for total oil consumption, noting that the volatile pathway must not exceed 30% of the total. For the burner-based stand, the controls limits are ±5% for both pathways, which are tracked separately.
(i) Changing engine oil. For an engine-based platform, periodically change engine oil to maintain stable oil consumption rates and maintain the health of the aging engine. Interrupt aging as needed to perform oil changes. Perform a drain-and-weigh measurement. If you see a sudden change in oil consumption it may be necessary to stop aging and either change oil or correct an issue with the accelerated oil consumption. If the aging engine requires repairs to correct an oil consumption issue in the middle of aging, you must re-validate the oil consumption rate for 72 hours before you continue aging. The engine exhaust should be left bypassing the aftertreatment system until the repaired engine has been validated.
(ii) Secondary oil consumption validation. If your aftertreatment includes a diesel particulate filter, we recommend that you perform secondary validation of oil consumption by using clean (soot free) DPF weights to track ash loading and compare this mass of ash to the amount predicted using the measured oil consumption mass and the oil ash concentration. The mass of ash found by DPF weight should fall within a range of (55 to 70)% of the of mass predicted from oil consumption measurements. Perform this validation at the end of aging, at any intermediate emission test points, and at any point where you need to clean the DPF of accumulated ash in according with recommended maintenance.
(iii) Sulfur tracking. Generate a fuel sulfur exposure line describing the relationship between aging hours and cumulative target sulfur exposure mass. The control limits for sulfur exposure are ±3%. Log actual fuel consumption and the measured fuel sulfur level of the current batch of fuel (if you are doping fuel to accelerate sulfur exposure) for engine stand aging. Use these measurements to ensure that sulfur exposure remains within the control limits. Adjust sulfur doping levels in the fuel from batch to batch as needed to stay within limits. If you use gaseous SO2 for sulfur acceleration, monitor the mass flow rate of the gaseous sulfur. Use these measurements to calculate total sulfur mass exposure, and correct SO2 gas flow rates as needed to stay within the control limits.
(f) Emission testing at intermediate and final test points. Conduct emission testing at the end of aging and at any intermediate emission test points as described in the standard setting part. Following installation of the aged aftertreatment system on the emission-data engine at intermediate or final test points, prior to the start of emission testing, use good engineering judgment to operate the engine and aftertreatment system for a number of hours to stabilize emission controls and to allow any adaptive controls to update. Declare the number of stabilization hours prior to the start of the accelerated aging program.
[79 FR 23820, Apr. 28, 2014, as amended at 89 FR 29831, Apr. 22, 2024]