Remote monitoring data processing method and processing system for heavy vehicles
By adopting a remote monitoring data processing method that classifies vehicle speeds in three stages and adapts to vehicle usage, the problems of high cost, incomplete coverage, and poor stability in heavy vehicle emission monitoring have been solved, achieving low-cost and efficient emission assessment and dynamic supervision.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 天津仁爱学院
- Filing Date
- 2026-05-07
- Publication Date
- 2026-06-05
AI Technical Summary
Existing technologies for monitoring emissions from heavy-duty vehicles suffer from problems such as high offline testing costs, limited coverage of operating conditions, unstable online assessment results, and low data processing efficiency. In particular, they cannot accurately cover the complex operating conditions and dynamic changes of vehicles in actual operation.
A three-stage vehicle speed classification method combined with vehicle application adaptation is adopted. Vehicle information is obtained through OBD remote monitoring data, effective power window filtering and multiple random combinations are performed, and median statistics are combined to achieve refined coverage of operating conditions and stability of evaluation results.
It achieves low-cost, large-scale batch supervision, and the assessment results are highly consistent with offline tests. It can accurately capture changes in operating conditions, significantly improve data processing efficiency and assessment stability, and support dynamic emission performance tracking and early warning.
Smart Images
Figure CN122153605A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle emission monitoring technology, and in particular to a method and system for processing remote monitoring data of in-use heavy-duty vehicles. Background Technology
[0002] Currently, the mainstream emission assessment technologies in the industry mainly rely on two types of methods: one is offline testing based on bench, rotating drum and actual vehicle PEMS (Portable Emission Measurement System) (such as the test methods specified in GB 17691-2018 and GB / T 27840-2011), and the other is online assessment using vehicle remote monitoring data to perform PEMS cyclic reconstruction (such as the technical solution proposed in "Data Processing Method for Monitoring NOx Emissions of In-Use Vehicles Using Network Transmitted Data").
[0003] While the aforementioned existing technologies meet emission regulatory requirements to some extent, they have significant limitations, mainly in the following aspects: 1. Inherent defects of offline testing: Bench and slewing tests require fixed sites and special equipment, with a single test costing up to hundreds of thousands of yuan, and cannot cover the complex working conditions of actual vehicle operation; PEMS real vehicle testing is close to reality, but the equipment rental cost is expensive (the cost of a single test exceeds 10,000 yuan) and the test cycle is long (usually several days), making it difficult to achieve large-scale batch supervision of vehicles in use, and even more unable to meet the needs of dynamic tracking.
[0004] 2. Shortcomings of existing online data processing technologies: The PEMS cyclic reconstruction technology proposed in "Data Processing Method for Monitoring NOx Emissions from In-Use Vehicles Using Online Data" solves the problem of "no additional equipment," but it has two core shortcomings: First, the operating condition coverage is inaccurate. It only divides urban / suburban / highway operating conditions by "55km / h speed boundary" and "30-minute duration segmentation," without considering the dynamic fluctuations of vehicle speed under the same operating condition (e.g., the "urban operating condition" may be a short-term low-speed congestion in the highway or suburbs), resulting in a large deviation between the reconstructed cycle and the actual driving conditions. Second, the evaluation stability is poor. It relies on "random sampling splicing + single cycle calculation," and there may be huge changes in operating conditions within the reconstructed cycle. It is easily affected by extreme data interference and cannot objectively reflect the vehicle emission level.
[0005] Meanwhile, in-use vehicle remote monitoring systems have accumulated massive amounts of operational data, including key parameters such as vehicle speed, engine power, NOx sensor readings, and intake airflow. However, existing technologies cannot efficiently extract the value from this data. Either the "coarse data filtering" leads to redundant calculations (e.g., failure to exclude invalid low-power data), or the "single operating condition classification" results in inaccurate assessments (e.g., failure to differentiate between urban and non-urban vehicle operating conditions). Therefore, there is an urgent need for a NOx emission compliance assessment method that can be based on existing online data, adapt to multiple operating conditions, and provide stable and reliable assessment results. Summary of the Invention
[0006] Therefore, the purpose of this invention is to provide a method and system for processing remote monitoring data of in-use heavy-duty vehicles, employing an effective power window to address the problem of "single operating condition classification and poor adaptability." Existing technologies only classify operating conditions into three categories based on "vehicle speed threshold + duration," failing to cover scenarios of "dynamic speed changes under the same road conditions" in actual driving (such as fluctuations between "low speed-medium speed-low speed" in urban roads). This solution, through "three-segment vehicle speed classification + vehicle usage adaptation," classifies non-urban vehicles into 27 operating conditions and urban vehicles into 8, achieving refined coverage of operating conditions. It also addresses the problem of "poor stability of evaluation results and susceptibility to extreme data interference." Existing technologies rely on "single PEMS cycle reconstruction" to calculate emission results, and the randomness of the data can easily lead to evaluation bias (e.g., occasional high-speed operating conditions may inflate emission values). This solution, through "multiple random combinations + median statistics," eliminates the influence of extreme values, significantly improving the stability and objectivity of the evaluation results.
[0007] To achieve the above objectives, the present invention provides a method for processing remote monitoring data of in-use heavy vehicles, comprising the following steps: S1. Obtain OBD remote monitoring data of in-use heavy vehicles; S2. Preprocess the remote monitoring data; S3. Use the effective power window to perform coarse filtering on the preprocessed monitoring data; S4. Classify and label the monitoring data after coarse screening using the window classification method to obtain the corresponding category; the window classification method includes: dividing each effective power window selected into three segments according to time, calculating the average vehicle speed of each segment, and labeling the vehicle speed of each segment according to the preset vehicle use and speed range mapping relationship, and determining the working condition category of the power window by the combination of the three speed labeling types. S5. From each power window after classification in S4, perform multiple random samplings without replacement according to the preset baseline sampling number. Each sampled window is combined into a feature window combination, forming a total of N independent feature window sets, where N is an integer greater than 1. S6. Calculate the NOx emission characterization value for each group of feature window combinations, compare the NOx emission characterization value with the preset compliance threshold, obtain the compliance judgment result for each group of window combinations, and statistically analyze the judgment results of N groups of feature windows, and use the statistical results as the final judgment result of vehicle NOx emission compliance.
[0008] More preferably, in S4, the mapping relationship between vehicle use and speed range includes: classifying vehicles into urban vehicles and non-urban vehicles; for urban vehicles, the speed range includes a low-speed range and a medium-speed range; for non-urban vehicles, the speed range includes a low-speed range, a medium-speed range, and a high-speed range.
[0009] More preferably, for non-urban vehicles, the low-speed range is 15~40 km / h, the medium-speed range is 40~70 km / h, and the high-speed range is greater than 70 km / h; For urban vehicles, the low speed range is 15-40 km / h, and the medium speed range is 40-70 km / h.
[0010] More preferably, for non-urban vehicles belonging to category M1 or N1 in the GB17691-2018 national standard, the medium speed range is 40~90 km / h, and the high speed range is greater than 90 km / h.
[0011] More preferably, in S1, the OBD remote monitoring data includes: vehicle speed, engine speed, torque, NOx sensor reading, intake air flow, fuel injection quantity, atmospheric pressure, and SCR temperature.
[0012] Further preferably, in S2, the remote monitoring data is preprocessed, including: removing sensor fault data from the remote monitoring data, the sensor fault data including data where the vehicle speed is negative and the NOx reading exceeds the sensor range of 0~5000 ppm; removing non-driving state data, the non-driving state data including zero value data when the engine is off and data where the vehicle has been stationary for more than 5 minutes.
[0013] More preferably, in S3, the coarse screening of the preprocessed monitoring data using an effective power window includes: The effective power window is a power window whose average power within the window is greater than or equal to the power threshold. The power threshold is set to 10% to 15% of the engine's rated power. The window length is determined based on the cumulative power within the window reaching the engine's WHTC cycle power. The average power of each window is calculated by sliding calculation, and power windows with average power greater than or equal to the power threshold are selected as effective power windows, thus completing the coarse screening process for effective power windows.
[0014] Further preferably, in S6, calculating the NOx emission characterization value for each group of feature window combinations includes: calculating the NOx ratio emission for each window; sorting the NOx ratio emission values of all windows in a set of windows in ascending order, removing the highest 10%, and taking the maximum value from the remaining 90% of the data as the NOx emission characterization value for that set of windows.
[0015] Further preferred, S1 to S4 are repeated at preset time periods to form a dynamic change curve of vehicle NOx emissions for long-term tracking and early warning of emission performance.
[0016] More preferably, the preset benchmark sampling quantity is determined based on the frequency of occurrence of various operating condition windows in actual road emission statistics, and the benchmark sampling quantity corresponding to the operating condition category with a high occurrence frequency is greater than that of the operating condition category with a low occurrence frequency.
[0017] More preferably, the statistical analysis of the determination results of the N sets of feature windows specifically involves taking the median of the N sets of determination results as the final determination result of the vehicle's NOx emission compliance.
[0018] More preferably, the calculation of the NOx emission characterization value is achieved through the following model: defining the relationship between the NOx emission rate R_NOx and the NOx concentration C_NOx in the exhaust gas and the exhaust gas mass flow rate F_exh as follows: R_NOx = k × C_NOx × F_exh Where k is the conversion coefficient derived from the gas law; the window effective work is calculated based on the window average power and window duration; the window NOx emission intensity = (R_NOx × window duration) / window effective work.
[0019] The present invention also provides a remote monitoring data processing system for in-use heavy-duty vehicles, which is based on the steps of the above-described remote monitoring data processing method for in-use heavy-duty vehicles, including: The data acquisition module acquires remote OBD monitoring data from in-use heavy vehicles. The data preprocessing module preprocesses the remote monitoring data; and performs coarse filtering on the preprocessed monitoring data using an effective power window. The window classification processing module classifies and labels the coarsely filtered monitoring data using the window classification method to obtain the corresponding categories. The window classification method includes: dividing each effective power window into three segments according to time, calculating the average vehicle speed of each segment, and labeling the vehicle speed of each segment according to the preset vehicle use and speed interval mapping relationship, and determining the working condition category of the power window by the combination of the three speed labeling types. The feature window combination module randomly selects multiple times without replacement from each power window after classification, according to a preset baseline sampling number. Each selected window is combined into a feature window set, forming a total of N independent feature window sets, where N is an integer greater than 1. The emission compliance determination module calculates the NOx emission characterization value for each group of feature windows, compares the NOx emission characterization value with a preset compliance threshold, obtains the compliance determination result for each group of window combinations, and statistically analyzes the determination results for N groups of feature windows, using the statistical results as the final determination result for vehicle NOx emission compliance.
[0020] This application directly utilizes existing remote monitoring data of vehicles, eliminating the need for additional hardware such as PEMS. The cost of a single assessment is less than 1% of that of offline testing, and a single server can process data from 1,000 vehicles simultaneously, achieving large-scale batch monitoring with zero additional equipment investment. Furthermore, this solution uses power threshold screening to eliminate approximately 10% to 30% of low-power invalid windows (such as idling and low-speed creep conditions), significantly reducing the data processing volume for subsequent condition classification and combination calculations. Compared to traditional network transmission data processing technologies, it significantly improves data processing efficiency while ensuring assessment accuracy, better meeting the actual needs of large-scale emission monitoring of in-use vehicles.
[0021] This application addresses the shortcomings of existing technologies that only classify operating conditions into three categories based on "vehicle speed threshold + duration," failing to capture dynamic changes in operating conditions. This solution proposes a refined classification mechanism of "vehicle usage adaptation + three-segment vehicle speed classification." For urban vehicles, it classifies them into eight operating conditions; for non-urban vehicles, it classifies them into 27 operating conditions. This accurately depicts complex operating condition changes such as "low speed - medium speed - high speed," effectively avoiding inaccurate assessments caused by coarse operating condition classification.
[0022] This application also enables repeated assessments on a daily or weekly basis, generating a dynamic curve of vehicle NOx emissions. When a vehicle's assessment results change from the median "compliant" to "non-compliant" over several consecutive days, a timely warning of declining emissions performance can be issued, prompting regulatory authorities to conduct on-site inspections or guiding automakers to perform maintenance (such as replacing SCR catalysts). This represents an upgrade from a "single-point sampling inspection" to a "continuous monitoring" regulatory model, demonstrating significant social and economic value. Attached Figure Description
[0023] Figure 1 This is a flowchart illustrating a remote monitoring data processing method for heavy-duty vehicles according to the present invention.
[0024] Figure 2 This is a schematic diagram of the effective power window classification method of the present invention.
[0025] Figure 3This is a graph showing the percentage of various effective power windows for different vehicles in multiple real-world emission tests.
[0026] Figure 4 A graph showing the differences in NOx emission ratios across various effective power windows.
[0027] Figure 5 This is a comparison chart of the evaluation results and the experimental results. Detailed Implementation
[0028] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0029] like Figure 1 As shown in the figure, the remote monitoring data processing method for heavy-duty vehicles provided in one embodiment of the present invention includes the following steps: The figure illustrates the complete process of this solution from "data input" to "compliance judgment": The first step is data acquisition; the second step is data preprocessing (importing data transmitted from the network → removing abnormal values from sensors); the third step is preliminary screening of power windows (excluding low-power windows according to the effective power window method of GB17691-2018); the fourth step is window classification and marking (classifying according to vehicle use + three speed segments); the fifth step is feature window combination (multiple random sampling, without replacement after use); the sixth step is emission calculation and judgment (comparing with the threshold of GB 17691-2018). The logical relationship of each step is marked with arrows in the figure, clearly showing the operation process of this solution: The detailed process is as follows: S1. Obtain OBD remote monitoring data of the heavy-duty vehicle in use; the OBD remote monitoring data includes: vehicle speed (km / h), engine speed (r / min), torque (N·m), NOx sensor reading (ppm), intake air flow (kg / h), fuel flow (kg / h), atmospheric pressure (kPa), and SCR temperature (°C).
[0030] S2. Preprocess the remote monitoring data; the preprocessing process includes: removing two types of invalid data: (1) sensor fault data (such as negative vehicle speed, NOx reading exceeding the sensor range of 0~5000 ppm); (2) non-driving state data (such as zero value data when the engine is off, data of the vehicle being stationary for more than 5 minutes). This step is consistent with the "invalid data removal" logic in the "Data Processing Method for Monitoring NOx Emissions of In-Use Vehicles Using Network Transmitted Data", but adds "sensor range verification" to further ensure the validity of the data.
[0031] S3. The preprocessed monitoring data is coarsely screened using an effective power window. This coarse screening of the preprocessed monitoring data using an effective power window includes: the effective power window is a power window whose average power within the window is greater than or equal to the power threshold. The power threshold is set to 10% of the engine's rated power. The window length is determined based on the cumulative power within the window reaching the engine's WHTC cycle power. The average power of each window is calculated using a sliding calculation. Power windows with an average power greater than or equal to the power threshold are selected as effective power windows, completing the coarse screening process for effective power windows. Existing technologies do not distinguish between "power correlation" and include low-power operating conditions (such as idling and low-speed creep) in the calculation, resulting in emission results that are not of reference value. This solution uses a power threshold for screening, retaining only windows with high power and high emission contributions, significantly reducing the subsequent calculation workload (experiments have verified that approximately 10% to 30% of invalid windows can be eliminated).
[0032] S4. Classify and label the coarsely filtered monitoring data using a window classification method to obtain corresponding categories; the window classification method includes: dividing each selected effective power window into three segments according to time, calculating the average vehicle speed of each segment, and labeling each segment's speed according to a preset mapping relationship between vehicle use and speed range, and determining the combination of the three segment speed label types as the operating condition category of the power window; for example... Figure 2 As shown in the figure, the classification and labeling process of a single power window is as follows: the window is divided into 3 segments according to time (allowing ±1 second deviation to avoid data breakage caused by overly fine time division), and each segment is labeled with the working condition type according to the interval to which the average vehicle speed belongs. Finally, the segments are combined to form the category of the window (such as "low speed-medium speed-high speed" corresponding to a specific category). The mapping relationship between vehicle purpose and speed range includes: classifying vehicles into urban vehicles and non-urban vehicles; For urban vehicles, the speed range includes a low-speed range and a medium-speed range; for urban vehicles, the low-speed range is 15~40 km / h, and the medium-speed range is 40~70 km / h. For non-urban vehicles, the speed range includes low speed range, medium speed range and high speed range.
[0033] For non-urban vehicles, the low-speed range is 15-40 km / h, the medium-speed range is 40-70 km / h, and the high-speed range is greater than 70 km / h. For M1 or N1 category non-urban vehicles, the medium-speed range is 40-90 km / h, and the high-speed range is greater than 90 km / h. M1 category vehicles are passenger vehicles with no more than nine seats (including the driver's seat) as defined in the national standard GB17691-2018, and whose maximum design gross vehicle weight exceeds 3500 kg. N1 category vehicles are freight vehicles with a maximum design gross vehicle weight not exceeding 3500 kg according to the same standard.
[0034] The three-segment classification method divides each power window (within which the cumulative power reaches the engine's WHTC cycle power) after screening in step 1 into three segments based on time (allowing a deviation of ±1 second). The average vehicle speed of each segment is calculated, and the segment is labeled "low / medium / high" according to its speed range. The final combination forms the window category. For example, for non-urban vehicles, the three-segment combinations of the three speed ranges generate 3×3×3=27 window categories (e.g., "low-low-low" is category 1, "low-medium-high" is category 6), as shown in Table 1. For urban vehicles, the three-segment combinations of the two speed ranges generate 2×2×2=8 window categories (e.g., "low-low-medium" is category 2, "medium-medium-medium" is category 8), as shown in Table 1.
[0035] Table 1. Effective Power Window Classification Rules .
[0036] S5. From each power window after classification in S4, perform multiple random samplings without replacement according to the preset baseline sampling number. Each sampled window is combined into a feature window set, forming a total of N independent feature window sets, where N is an integer greater than 1. The number of baseline windows is determined based on actual road emissions test statistics (such as...). Figure 3 As shown in the figure, this figure uses four non-urban heavy-duty diesel vehicles as samples to show the proportion of 27 effective power windows appearing in 3-4 actual road emission tests for each vehicle. It can be observed from the figure that: (1) the number of window categories in a single test is generally no more than 10; (2) some categories (such as categories 1, 4, and 7) did not appear in all tests, indicating that there are "atypical operating conditions" in actual driving; (3) the proportion of windows in different tests for the same vehicle is different. This figure is used to illustrate the limitations of "existing single operating condition classification" and highlight the necessity of "multi-category coverage" in this scheme. ), determine the "baseline sampling quantity" for each type of window: for example, the proportion of categories 5, 14, and 18 of non-urban vehicles in actual road emission tests is relatively high, so the corresponding sampling quantity is also relatively large.
[0037] Multiple random combinations and sampling without replacement are performed from each class of windows marked in step 2. Windows are randomly selected according to a baseline quantity and combined into a "feature window set". The selected windows are not replaced to avoid bias caused by repeated calculations. This process is repeated N times to form N independent window sets. The selection of "N times" is based on the fact that when the number of combinations is ≥ N, the fluctuation range of the evaluation results is < 5%, balancing stability and computational efficiency.
[0038] S6. Calculate the NOx emission ratio characterization value for each group of feature window combinations, compare the NOx emission ratio characterization value with a preset compliance threshold to obtain the compliance judgment result for each group of window combinations, and statistically analyze the judgment results for N groups of feature windows, using the statistical result as the final judgment result for vehicle NOx emission compliance. Calculate the NOx emission ratio for each window; for example... Figure 4 This figure uses vehicles 1-4 as samples to show the NOx specific emissions (mg / kWh) for different power windows. For example, vehicle 1 has higher NOx specific emissions in windows 20-22 (starting at high speed) than adjacent windows because the SCR system temperature drops and conversion efficiency decreases during the "high speed → low speed" switch. The figure visually demonstrates the significant differences in emissions from the same vehicle under different operating conditions, illustrating that using only a single window for evaluation can easily lead to inaccurate results due to operating condition selection bias, thus proving the rationality of the proposed "multi-window combined evaluation" approach.
[0039] Sort the NOx emission ratio values of all windows in a set of windows from smallest to largest, remove the top 10%, and take the maximum value of the remaining 90% of data as the NOx emission ratio characterization value of the set of windows.
[0040] NOx emission intensity calculation is performed for each set of windows, calculating the NOx emission intensity (mg / kWh) for each window. This embodiment uses the following calculation model: Step 1: Define the relationship between the NOx emission rate R_NOx, the NOx concentration C_NOx in the exhaust gas, and the exhaust gas mass flow rate F_exh as follows: R_NOx = k × C_NOx × F_exh Where k is the conversion coefficient derived from the gas law, which can be obtained through experimental calibration.
[0041] Step 2: Calculate the effective work (kWh) within the window. Effective work = average power (kW) × window duration (h); where average power is calculated using "speed × torque × π / 30000".
[0042] Step 3: Calculate NOx specific emissions (g / kWh). NOx specific emissions = NOx emission rate × window duration / effective work.
[0043] Step 4: Calculate the NOx emission ratio for each window set, sort them from smallest to largest, remove the top 10%, and take the maximum value of the remaining 90% as the NOx emission characterization value for that group. This step refers to the "Statistical Method of Emission Results" in actual road emission tests, which can effectively eliminate the interference of extremely high or extremely low emission data.
[0044] Step 5: Compliance Judgment and Result Statistics. According to GB 17691-2018, if the NOx emission characterization value of a certain window set is <0.69 g / kWh, then that set is judged as "compliant"; if the NOx emission characterization value is ≥0.69 g / kWh, then it is judged as "non-compliant". The median of the judgment results for N window sets is taken: if the median is "compliant", then the vehicle's NOx emissions are ultimately judged as compliant; if the median is "non-compliant", then it is ultimately judged as non-compliant. The reason for taking the median is that the distribution of compliant and non-compliant numbers in N sets of results usually follows a "normal distribution," and the median can reflect the emission level in most scenarios, avoiding misjudgments caused by single-set bias. Figure 5 The figure uses vehicles 1-4 as samples, with the horizontal axis representing the evaluation date. Each evaluation date generates 3-4 evaluation results (corresponding to the number of tests, e.g., vehicle 1 has 4 sets of baseline data). The vertical axis represents the "evaluation results," and the range of actual road emission test results is also marked (magenta crosses). From the figure, it can be observed that: (1) regardless of which set of baseline data is used, the evaluation results are highly concentrated; (2) the median of the evaluation results falls within the range of the actual test results. This figure is used to verify the accuracy of this scheme and prove the consistency between its evaluation results and offline tests.
[0045] It also includes S7, which repeats S1 to S4 at a preset time period to form a dynamic change curve of vehicle NOx emissions for long-term tracking and early warning of emission performance. The evaluation can be repeated on a "daily" or "weekly" basis to form a "dynamic change curve" of vehicle NOx emissions. For example, if a vehicle's evaluation result changes from the median "compliant" to "non-compliant" for three consecutive days, an early warning of "deteriorating emission performance" can be issued, prompting regulatory authorities to conduct on-site verification or the automaker to perform maintenance (such as replacing the SCR catalyst).
[0046] Compared to existing technologies (offline testing, traditional online data processing), it has the following significant advantages: 1. Cost and efficiency advantages: zero additional equipment investment, enabling large-scale batch monitoring.
[0047] Offline PEMS testing equipment often costs over a million yuan, and calibration of the measuring instruments is required before and after each measurement. This solution directly utilizes existing remote monitoring data, requiring no additional equipment installation. Data processing only requires a conventional server (a single server can process data from 1000 vehicles simultaneously), with a total cost less than 1% of that of offline testing. Compared to traditional network-transmitted data processing techniques (such as PEMS cyclic reconstruction), this solution eliminates 10% to 30% of invalid data through "power threshold filtering," shortening data processing time and better meeting the needs of large-scale supervision.
[0048] 2. The assessment has an advantage in accuracy, is highly consistent with offline tests, and has more comprehensive coverage of operating conditions.
[0049] This scheme covers 8 to 27 operating conditions, capturing emission differences between these conditions and avoiding inaccurate assessments due to variations in operating conditions. Four China VI heavy-duty diesel vehicles (vehicles 1-4) were selected, and their emissions were evaluated using both this scheme and actual road tests. The results are shown in Table 2. It can be seen that the proportion of assessment results falling within the range of the test results is higher than 78%, and three vehicles achieved 100% compliance, meeting emission regulatory requirements.
[0050] Table 2 Comparison of the results of this plan and actual road emission tests. .
[0051] 3. Stability advantage: multi-combination statistical analysis eliminates interference from extreme data. Traditional technologies rely on single-cycle calculations, which are susceptible to interference from extreme data. This solution uses a triple mechanism of "N combinations + median statistics and NOx emission characterization values" to eliminate extreme values and ensure that the assessment results reflect the vehicle's "normal emission level".
[0052] The present invention also provides a remote monitoring data processing system for in-use heavy-duty vehicles, which is based on the steps of the above-described remote monitoring data processing method for in-use heavy-duty vehicles, including: The data acquisition module acquires remote OBD monitoring data from in-use heavy vehicles. The data preprocessing module preprocesses the remote monitoring data; and performs coarse filtering on the preprocessed monitoring data using an effective power window. The window classification processing module classifies and labels the coarsely filtered monitoring data using the window classification method to obtain the corresponding categories. The window classification method includes: dividing each effective power window into three segments according to time, calculating the average vehicle speed of each segment, and labeling the vehicle speed of each segment according to the preset vehicle use and speed interval mapping relationship, and determining the working condition category of the power window by the combination of the three speed labeling types. The feature window combination module randomly selects multiple times without replacement from each power window after classification, according to a preset baseline sampling number. Each window combination is a feature window combination, forming a total of N independent feature window sets, where N is an integer greater than 1. The emission compliance determination module calculates the NOx emission ratio characterization value for each group of feature window combinations, compares the NOx emission ratio characterization value with a preset compliance threshold to obtain the compliance determination result for each group of window combinations, and statistically analyzes the determination results for N groups of feature windows, using the statistical result as the final determination result for vehicle NOx emission compliance. The implementation process is the same as described in the above embodiment and will not be repeated here.
[0053] In summary, this solution addresses the problems of "high cost, low efficiency, and weak stability" in existing technologies. It can be widely applied to scenarios such as emission supervision of in-use vehicles by environmental protection authorities, after-sales emission optimization by automakers, and vehicle operation and maintenance management by logistics companies, and has significant economic and social value.
[0054] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.
Claims
1. A method for processing remote monitoring data of in-use heavy vehicles, characterized in that, Includes the following steps: S1. Obtain OBD remote monitoring data of in-use heavy vehicles; S2. Preprocess the remote monitoring data; S3. Use the effective power window to perform coarse filtering on the preprocessed monitoring data; S4. Classify and label the monitoring data after coarse screening using the window classification method; Get the corresponding category; The window classification method includes: dividing each selected effective power window into three segments according to time, calculating the average vehicle speed of each segment, and labeling the vehicle speed of each segment according to the preset mapping relationship between vehicle use and vehicle speed range, and determining the working condition category of the power window by the combination of the three segment vehicle speed label types. S5. From each power window after classification in S4, perform multiple random samplings without replacement according to the preset baseline sampling number. Each sampled window is combined into a feature window combination, forming a total of N independent feature window sets, where N is an integer greater than 1. S6. Calculate the NOx emission ratio characterization value for each group of feature window combinations, compare the NOx emission ratio characterization value with a preset compliance threshold to obtain the compliance judgment result for each group of window combinations, and statistically analyze the judgment results of N groups of feature windows, using the statistical results as the final judgment result for vehicle NOx emission compliance.
2. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, In S4, the mapping relationship between vehicle purpose and speed range includes: classifying vehicles into urban vehicles and non-urban vehicles; For urban vehicles, the speed range includes a low-speed range and a medium-speed range; For non-urban vehicles, the speed range includes low speed range, medium speed range and high speed range.
3. The method for processing remote monitoring data of in-use heavy vehicles according to claim 2, characterized in that, For non-urban vehicles, the low speed range is 15-40 km / h, the medium speed range is 40-70 km / h, and the high speed range is greater than 70 km / h. For urban vehicles, the low-speed range is 15-40 km / h, and the medium-speed range is 40-70 km / h.
4. The method for processing remote monitoring data of in-use heavy vehicles according to claim 3, characterized in that, For non-urban vehicles classified as M1 or N1 in the national standard, the medium speed range is 40-90 km / h, and the high speed range is greater than 90 km / h.
5. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, In S1, the OBD remote monitoring data includes: vehicle speed, engine speed, torque, NOx sensor reading, intake air flow, fuel injection quantity, atmospheric pressure, and SCR temperature.
6. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, In S2, the remote monitoring data is preprocessed, including: removing sensor fault data from the remote monitoring data, the sensor fault data including data where the vehicle speed is negative and the NOx reading exceeds the sensor range of 0~5000 ppm; removing non-driving state data, the non-driving state data including zero value data when the engine is off and data where the vehicle has been stationary for more than 5 minutes.
7. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, In S3, the coarse screening of the preprocessed monitoring data using an effective power window includes: the effective power window is a power window whose average power within the window is greater than or equal to a power threshold. The power threshold is set to 10% of the engine's rated power. The window length is determined based on the cumulative power within the window reaching the engine's WHTC cycle power. The average power of each window is calculated by sliding calculation. Power windows with an average power greater than or equal to the power threshold are selected as effective power windows, thus completing the coarse screening process of the effective power window.
8. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, In S6, the NOx emission ratio characterization value for each group of the aforementioned feature window combinations is calculated, including: Calculate the NOx emission ratio for each window; Sort the NOx emission ratio values of all windows in a set of windows from smallest to largest, remove the top 10%, and take the maximum value of the remaining 90% of data as the NOx emission characterization value of the set of windows.
9. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, It also includes repeating S1 to S4 at preset time periods to form a dynamic change curve of vehicle NOx emissions for long-term tracking and early warning of emission performance.
10. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, In S5, the preset baseline sampling quantity is determined based on the frequency of occurrence of various operating condition windows in actual road emission statistics. The baseline sampling quantity corresponding to the operating condition category with a high frequency of occurrence is greater than that of the operating condition category with a low frequency of occurrence.
11. The method for processing remote monitoring data of in-use heavy vehicles according to claim 1, characterized in that, The statistical analysis of the judgment results of N sets of feature windows is specifically as follows: the median of the N sets of judgment results is taken as the final judgment result of the vehicle's NOx emission compliance.
12. A remote monitoring data processing system for in-use heavy vehicles, characterized in that, The steps of the remote monitoring data processing method for in-use heavy vehicles according to any one of claims 1-11 include: The data acquisition module acquires remote OBD monitoring data from in-use heavy vehicles; The data preprocessing module preprocesses the remote monitoring data; and performs coarse filtering on the preprocessed monitoring data using an effective power window. The window classification processing module classifies and labels the coarsely filtered monitoring data using the window classification method to obtain the corresponding categories. The window classification method includes: dividing each effective power window into three segments according to time, calculating the average vehicle speed of each segment, and labeling the vehicle speed of each segment according to the preset vehicle use and speed interval mapping relationship, and determining the working condition category of the power window by the combination of the three speed labeling types. The feature window combination module randomly selects multiple times without replacement from each power window after classification, according to a preset baseline sampling number. Each window combination is a feature window combination, forming a total of N independent feature window sets, where N is an integer greater than 1. The emission compliance determination module calculates the NOx emission characterization value for each group of feature windows, compares the NOx emission characterization value with a preset compliance threshold, obtains the compliance determination result for each group of window combinations, and statistically analyzes the determination results for N groups of feature windows, using the statistical results as the final determination result for vehicle NOx emission compliance.