A method and system for predicting energy efficiency degradation trend of a new energy vehicle compressor

Abstract

The application relates to the technical field of new energy automobile compressor energy efficiency data analysis, in particular to a new energy automobile compressor energy efficiency degradation trend prediction method and system. The method comprises the following steps: collecting compressor operation data, screening steady-state or quasi-steady-state driving condition segments, generating equivalent working condition characteristics, comparing the equivalent working condition characteristics with a health benchmark energy efficiency surface to obtain a driving condition degradation evaluation value; when a triggering condition is met, performing a parked standard working condition test, calculating a parked working condition degradation evaluation value; fusing and calibrating the two to generate a compressor body energy efficiency degradation index, and predicting the energy efficiency degradation trend based on the index time sequence and cumulative operation characteristics. The application can effectively separate the compressor body energy efficiency degradation and external working condition disturbance, improve the prediction accuracy and stability through parked standard working condition calibration, and realize the prediction of the energy efficiency degradation trend and the advance planning of the maintenance window.

Description

Technical Field

[0001] This application relates to the field of energy efficiency data analysis technology for new energy vehicle compressors, and in particular to a method and system for predicting the energy efficiency degradation trend of new energy vehicle compressors. Background Technology

[0002] The electric compressor in new energy vehicles is a core component of the vehicle's thermal management system, widely involved in various tasks such as cabin cooling, heat pump heating, battery cooling, and motor thermal management. The compressor's energy efficiency directly affects the vehicle's range and user experience. Its operating status is influenced by multiple internal factors, including motor efficiency, mechanical friction, sealing and lubrication conditions, and refrigerant compression efficiency. It is also closely related to external operating conditions such as ambient temperature, vehicle speed, heat exchanger airflow, electronic expansion valve opening, fan speed, water pump flow rate, and various heat load demands. Existing technologies for monitoring the thermal management system of new energy vehicles mainly include power consumption prediction based on overall vehicle energy consumption, control optimization based on operating parameters, fault diagnosis based on abnormal signals such as current or vibration, and lifespan estimation based on historical data. These solutions typically predict the overall power consumption of the air conditioning system by collecting data such as ambient temperature, vehicle speed, air conditioning set temperature, and passenger compartment heat load. Alternatively, they may adjust control parameters such as compressor speed and electronic expansion valve opening based on the prediction results, or identify compressor malfunctions by monitoring abnormal fluctuations in parameters such as pressure and temperature. Some solutions also use long-term operating data to estimate the remaining lifespan of the compressor or air conditioning system and provide maintenance reminders.

[0003] However, in existing technologies, due to the complex and varied operating conditions of vehicles, the energy efficiency performance of the same compressor varies significantly under different ambient temperatures, vehicle speeds, battery cooling requirements, and valve control states. Directly comparing energy efficiency indicators at different times can easily lead to misinterpreting drastic fluctuations in external operating conditions as performance degradation of the compressor itself. Existing energy consumption predictions mostly focus on the air conditioning system or thermal management system as a whole, making it difficult to effectively distinguish between the energy efficiency degradation of the compressor itself and external disturbances such as heat exchanger blockage, changes in refrigerant volume, fan efficiency decline, water pump flow fluctuations, and changes in cabin heat load. Furthermore, real-world vehicle driving data is usually in a non-steady-state process, with frequent changes in compressor speed, pressure ratio, and heat exchange air volume. This makes it difficult to use the collected energy efficiency data as a reliable basis for judging long-term degradation trends. Relying solely on low-frequency bench tests or after-sales inspections cannot detect the compressor entering an accelerated degradation phase in a timely manner, making it difficult to make continuous and accurate predictions of future energy efficiency decline trends. Summary of the Invention

[0004] This application provides a method and system for predicting the energy efficiency degradation trend of compressors in new energy vehicles to solve the above-mentioned problems.

[0005] In a first aspect, this application provides a method for predicting the energy efficiency degradation trend of a new energy vehicle compressor, the method comprising: S1. Collect operating data during the operation of the compressor in the new energy vehicle. The operating data includes compressor operating parameters, refrigerant side status parameters, thermal management execution component parameters, and vehicle thermal load parameters. S2. Divide the operating data during vehicle operation into segments, and based on compressor speed fluctuations, intake and exhaust pressure fluctuations, electronic expansion valve opening changes, and thermal management mode switching status, filter out steady-state or quasi-steady-state driving condition segments. S3. Generate equivalent operating condition features based on the driving condition segment, and call the reference energy efficiency surface under the compressor's healthy state to obtain the reference equivalent energy efficiency value corresponding to the driving condition segment. S4. Calculate the measured equivalent energy efficiency value based on the compressor input power and heat output characterization value in the driving condition segment, compare the measured equivalent energy efficiency value with the benchmark equivalent energy efficiency value, and generate a driving condition degradation evaluation value. S5. When the vehicle meets the parking self-diagnosis triggering conditions, control the compressor and thermal management execution components to enter the parking standard working condition test combination, and collect parking standard working condition data after the parking standard working condition test combination reaches a stable state. S6. Calculate the measured standard energy efficiency value based on the parking standard operating condition data, and compare the measured standard energy efficiency value with the health standard energy efficiency value of the corresponding parking standard operating condition to generate a parking condition degradation evaluation value. S7. Based on the driving condition degradation evaluation value and the parking condition degradation evaluation value, perform fusion calibration to generate the compressor body energy efficiency degradation index; S8. Based on the time series of the compressor body energy efficiency degradation index and the cumulative operating characteristics of the compressor, predict the energy efficiency degradation trend of the compressor and output the energy efficiency degradation prediction result.

[0006] Through the above technical solutions, a complete closed loop was constructed, consisting of equivalent evaluation of driving conditions, calibration of standard parking conditions, and prediction of degradation trends. Multi-source data acquisition and steady-state screening under driving conditions resolved the problem of interference from non-steady-state data, providing a reliable foundation for evaluation. Equivalent operating condition feature generation and benchmark surface comparison enabled lateral comparability of compressor energy efficiency under different external conditions, separating operating condition disturbances from actual degradation. Active triggering and testing of standard parking conditions eliminated the influence of uncontrollable variables, providing a high-precision calibration benchmark. The fusion calibration mechanism dynamically balanced real-time performance and accuracy, using parking results to correct long-term drift in the driving model. Finally, trend prediction combining time series and cumulative operating characteristics endowed the solution with forward-looking maintenance guidance capabilities. This dual-path collaborative modeling approach effectively overcomes the shortcomings of relying solely on driving data, which is susceptible to interference, or relying solely on bench testing, which has a low frequency. It significantly improves the accuracy, stability, and interpretability of predicting the energy efficiency degradation trends of new energy vehicle compressors.

[0007] Optionally, before performing the segmentation, the collected operating data is synchronized in time and abnormal data is removed. The continuous operating data is divided into several segments to be evaluated, using the compressor start / stop status, thermal management mode status, and data continuity as segmentation boundaries. The segments to be evaluated that exhibit sensor jumps, communication interruptions, or sudden changes in compressor control commands are subject to weight reduction or removal.

[0008] By employing the aforementioned technical solutions, a robust data quality gating system is constructed through the synergistic effects of time synchronization, anomaly removal, multiple boundary segmentation, and segment-level weighted removal. Time synchronization eliminates phase errors caused by inconsistent sampling clocks from multi-source sensors, enabling instantaneous matching of speed and pressure. The anomaly removal mechanism blocks the upward propagation of noise generated by sensor faults. The triple segmentation logic based on start / stop, mode, and continuity ensures the consistency of operating conditions within each analysis unit, avoiding energy efficiency assessment distortion caused by cross-operating condition mixed calculations. The weighting and removal strategies further enhance the model's adaptability to the complex environment of real vehicles. Based on this, the cleaned and optimized segments to be evaluated more realistically reflect the actual operating state of the compressor, making the equivalent operating condition features generated subsequently based on these segments more accurate. This ensures that the calculation results of the driving condition degradation evaluation value are not affected by external data noise, ultimately achieving stable and reliable prediction of the compressor's energy efficiency degradation trend.

[0009] Optionally, the steady-state or quasi-steady-state driving condition segment is determined by a steady-state score, which is calculated based on the compressor speed change, intake and exhaust pressure change, electronic expansion valve opening change, and thermal management mode duration. When the steady-state score meets the set scoring conditions, the corresponding segment is determined as a steady-state driving condition segment. When the steady-state score does not meet the set scoring conditions but the change trend of each state quantity in the segment meets the monotonic or slowly changing conditions, the corresponding segment is determined as a quasi-steady-state driving condition segment, and the quasi-steady-state driving condition segment is dynamically corrected.

[0010] The above technical solution achieves a shift from coarse static threshold screening to refined dynamic evaluation. By jointly constraining four dimensions—compressor speed variation, intake and exhaust pressure variation, electronic expansion valve opening variation, and thermal management mode duration—a physically complete stability evaluation system is constructed, ensuring that only data truly reflecting the compressor's inherent characteristics are included in the analysis. Based on this, a logic for identifying monotonic or slowly varying conditions is introduced, coupled with a dynamic correction algorithm, successfully transforming potentially discarded quasi-steady-state operating data into high-value evaluation samples. This combination of scoring screening, trend identification, and dynamic compensation not only improves the acquisition rate of effective data segments under driving conditions but also reduces energy efficiency evaluation noise caused by slight fluctuations in operating conditions. This allows the generated driving condition degradation evaluation values ​​to more accurately reflect the compressor's energy efficiency decline trend, providing high-quality data input for subsequent fusion calibration and trend prediction.

[0011] Optionally, the equivalent operating condition characteristics are generated in the following manner: First, the driving condition segments are divided into cabin cooling, heat pump heating, battery cooling, or combined thermal management conditions according to the thermal management mode. Then, equivalent operating points under the corresponding conditions are generated according to the compressor load status, refrigerant pressure status, and heat exchange side load status, so that the compressor operating segments under different ambient temperatures, different vehicle speeds, and different heat loads are mapped to a comparable equivalent operating condition space.

[0012] The above technical solution achieves a unified standard for compressor energy efficiency evaluation under complex and variable operating conditions. First, coarse-grained partitioning based on thermal management modes resolves the fundamental contradiction of incomparable energy efficiency indicators between different thermodynamic cycles such as refrigeration, heating, and battery cooling, preventing misjudgments caused by confused evaluation benchmarks. Second, it further introduces three-dimensional state variables—compressor load, refrigerant pressure, and heat exchanger load—to construct equivalent operating points, characterizing the actual work intensity and thermodynamic boundary conditions of the compressor under specific modes, ensuring the physical meaning and representativeness of the equivalent points. Finally, using this mapping mechanism, real-vehicle operating segments affected by drastic fluctuations in ambient temperature, vehicle speed, and heat load are transformed into stable, comparable data points in the equivalent operating condition space. This not only provides a precise index for accessing the healthy benchmark energy efficiency surface but also ensures that the energy efficiency deviation calculated based on driving data truly reflects the performance degradation of the compressor itself, rather than random disturbances in the external environment, significantly improving the accuracy and robustness of energy efficiency degradation trend prediction.

[0013] Optionally, the benchmark energy efficiency surface is constructed using health samples, which are derived from one or more of the following: compressor bench health data, vehicle initial operating data after leaving the factory, initial operating data after compressor replacement, and health data of a group of vehicles of the same model. When constructing the benchmark energy efficiency surface, the healthy samples are divided into zones and clustered according to the characteristics of equivalent operating conditions, and the health benchmark energy efficiency value corresponding to each equivalent operating condition zone is determined, so that the benchmark energy efficiency surface can output a benchmark equivalent energy efficiency value that matches the driving condition segment to be evaluated.

[0014] The above technical solutions enhance the robustness and accuracy of the benchmark energy efficiency surface. On one hand, the complementary integration of bench data, initial factory data, initial post-replacement data, and population health data constructs a comprehensive data foundation encompassing laboratory calibration, individual measurements, and population statistics, effectively addressing the issues of weak generalization and poor individual adaptability caused by single data sources. On the other hand, a partitioned clustering strategy based on equivalent operating condition characteristics discretizes the continuous operating condition space into multiple local subspaces, independently establishing a health benchmark within each subspace. This ensures that the benchmark value can precisely respond to minute changes in local operating conditions. This dual-enhancement mechanism of diverse data sources and spatial localization enables the generated benchmark equivalent energy efficiency value to truly reflect the theoretical health level of a specific compressor under specific operating conditions. Consequently, in subsequent energy efficiency deviation calculations, the influence of external operating condition disturbances can be accurately isolated, truly attributing the difference between measured values ​​and benchmark values ​​to the energy efficiency degradation of the compressor itself, laying a solid data foundation for generating accurate driving condition degradation evaluation values.

[0015] Optionally, the driving condition degradation evaluation value is generated in the following manner: The measured equivalent energy efficiency value of each driving condition segment is compared with the corresponding benchmark equivalent energy efficiency value to obtain the energy efficiency deviation of that segment. The segment weight is determined based on the segment's steady-state score, equivalent operating condition coverage, and degree of matching with the benchmark energy efficiency surface. The energy efficiency deviations of multiple segments are weighted and fused, and the reliability of the degradation evaluation of the corresponding statistical period is improved when the energy efficiency deviations change continuously in the same direction.

[0016] Through the above technical solutions, a complete chain was constructed, from single-segment quality assessment to multi-segment weighted fusion, and then to trend-enhanced credibility determination. The steady-state score, equivalent operating condition coverage, and benchmark matching work synergistically to achieve refined screening and weighting of the original driving data, effectively removing false degradation signals caused by unstable operating conditions or model mismatch. The weighted fusion mechanism further smooths out random disturbances and improves the robustness of the evaluation values. Meanwhile, the continuous unidirectional change mechanism utilizes the trend characteristics of the time series to actively identify and reinforce genuine degradation signals, significantly improving the accuracy and credibility of the driving condition degradation evaluation values, laying a solid foundation for the final generation of a stable compressor body energy efficiency degradation index.

[0017] Optionally, the parking self-diagnosis triggering conditions include periodic triggering conditions and abnormal triggering conditions; The periodic triggering conditions are determined based on the time interval from the last parking self-diagnosis or the cumulative running time of the compressor, and the abnormal triggering conditions are determined based on the abnormal increase of the driving condition degradation evaluation value, the excessive fluctuation of the driving condition degradation evaluation value, or the insufficient number of driving condition segments. When the periodic triggering condition or abnormal triggering condition is met, and the vehicle is in a parking state that allows the compressor to run for a short period of time, parking self-diagnosis is triggered.

[0018] Through the above technical solutions, a flexible and robust parking self-diagnostic triggering mechanism was constructed. Periodic triggering conditions and abnormal triggering conditions work together; the former provides basic time or mileage coverage, while the latter provides data-driven dynamic response. The two are logically connected via OR to ensure timely diagnostics in any risk scenario. Furthermore, the triggering conditions are interlocked with the vehicle's parking status, leveraging environmental characteristics such as a stationary vehicle and sufficient power to eliminate interference from external variables like vibration and wind resistance during driving on standard operating condition tests. Based on this, the insufficient number of driving condition segments in the abnormal triggering conditions not only supplements the degradation evaluation but also triggers self-checks on sensors or data links, forming a complete closed loop of driving monitoring, anomaly identification, parking verification, and data correction. Ultimately, this dual-trigger strategy enables the compressor energy efficiency degradation trend prediction method to adapt to different vehicle usage scenarios, avoiding the rigidity and lag of fixed-period diagnostics and preventing false alarms caused by a single data-driven approach, significantly improving the accuracy of compressor energy efficiency degradation assessment and the timeliness of maintenance recommendations.

[0019] Optionally, the parking standard operating condition test combination may be performed in sequence from low load to high load, including multiple standard test phases. In each standard test phase, the compressor and thermal management actuators are first controlled to enter the corresponding standard control state, and then it is determined whether the compressor speed, suction and discharge pressure and discharge temperature have reached the stable conditions. Once stable conditions are reached, standard operating condition data from the standard test phase are collected, and corresponding phase degradation evaluation values ​​are generated based on the measured standard energy efficiency values, pressure establishment response, and exhaust temperature response from the standard test phase. These are then integrated to obtain the parking condition degradation evaluation value.

[0020] Through the above technical solutions, a rigorous standardized testing and evaluation system for parking was constructed. By employing a sequential execution strategy from low load to high load, thermal and pressure interference during operating condition switching was eliminated, ensuring the independence of data from each test point. The optimal time window for data acquisition was locked through a three-dimensional steady-state criterion of speed, pressure, and temperature, avoiding noise introduced by non-steady-state data. The introduction of pressure establishment response and exhaust temperature response as supplementary evaluation dimensions beyond energy efficiency values ​​compensates for the inability of a single COP index to identify specific degradation modes (such as micro-leakage and lubrication failure). These features work together to ensure that the generated parking condition degradation evaluation value not only reflects the compressor's current energy efficiency level but also reveals its potential mechanical health status. This provides high-confidence benchmark data for subsequent fusion and calibration with driving condition degradation evaluation values, thus solving the problem of distorted evaluation results caused by arbitrary parking test load points and ambiguous stability criteria in existing technologies.

[0021] Optionally, the fusion calibration is performed in the following manner: The reliability of the driving condition degradation evaluation value is determined based on the data coverage, segment stability, and trend consistency. The reliability of the parking condition degradation evaluation value is determined based on the test completion time, test stability, and phase consistency. The fusion weight is dynamically determined based on the driving-side credibility and the parking-side credibility. When the deviation between the driving condition degradation evaluation value and the parking condition degradation evaluation value exceeds the consistency threshold, the weight of the side with lower credibility is reduced or a re-evaluation is triggered. A time series is generated based on the compressor body energy efficiency degradation index after fusion calibration, and the energy efficiency degradation trend, degradation level and maintenance window are output according to the change slope, duration and inflection point characteristics of the time series.

[0022] Through the above technical solution, the organic integration of driving condition data and parking standard condition data is achieved. The data coverage, segment stability, and trend consistency on the driving side, in conjunction with the test completion time, test stability, and stages on the parking side, together form the decision basis for dynamic fusion weights. Through this collaborative mechanism, the system can quickly correct accumulated errors using its high-precision characteristics when new parking data is available, and smoothly transition using the high-frequency characteristics of driving data when parking data is missing, ensuring the continuity and accuracy of the compressor's energy efficiency degradation index. Furthermore, the consistency threshold mechanism acts as a safety barrier, automatically reducing the weight of the less reliable side or triggering a retest when significant conflicts occur between the two data sources, effectively preventing index distortion caused by sensor drift or extreme operating conditions. Finally, based on the slope, duration, and inflection point characteristics extracted from this highly robust index time series, the output energy efficiency degradation trend, degradation level, and maintenance window have high predictive accuracy and interpretability, thus solving the problem of inaccurate degradation prediction caused by the inability to dynamically adapt the reliability of dual-source data in existing technologies, and significantly improving the reliability of the entire life cycle management of new energy vehicle compressors.

[0023] Secondly, this application provides a system for predicting the energy efficiency degradation trend of a new energy vehicle compressor, the system comprising: The data acquisition module is used to collect operating data during the operation of the compressor in new energy vehicles. The operating data includes compressor operating parameters, refrigerant side status parameters, thermal management execution component parameters, and vehicle thermal load parameters. The operating condition analysis module is used to divide the operating data during vehicle driving into segments, and based on compressor speed fluctuations, intake and exhaust pressure fluctuations, electronic expansion valve opening changes, and thermal management mode switching status, to filter out steady-state or quasi-steady-state driving condition segments. The benchmark analysis module is used to generate equivalent operating condition features based on the driving condition segment, and call the benchmark energy efficiency surface under the compressor's healthy state to obtain the benchmark equivalent energy efficiency value corresponding to the driving condition segment. The degradation analysis module is used to calculate the measured equivalent energy efficiency value based on the compressor input power and heat output characterization value in the driving condition segment, compare the measured equivalent energy efficiency value with the benchmark equivalent energy efficiency value, and generate a driving condition degradation evaluation value. The parking condition analysis module is used to control the compressor and thermal management execution components to enter the parking standard condition test combination when the vehicle meets the parking self-diagnosis trigger conditions, and to collect parking standard condition data after the parking standard condition test combination reaches a stable state. The parking degradation analysis module is used to calculate the measured standard energy efficiency value based on the parking standard operating condition data, and compare the measured standard energy efficiency value with the health standard energy efficiency value of the corresponding parking standard operating condition to generate a parking condition degradation evaluation value. The comprehensive analysis module is used to perform fusion calibration based on the driving condition degradation evaluation value and the parking condition degradation evaluation value to generate the compressor body energy efficiency degradation index; The degradation prediction module is used to predict the energy efficiency degradation trend of the compressor based on the time series of the compressor body energy efficiency degradation index and the cumulative operating characteristics of the compressor, and output the energy efficiency degradation prediction result. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating a method for predicting the energy efficiency degradation trend of a new energy vehicle compressor, provided in one embodiment of this application; Figure 2 This is a schematic diagram of a system for predicting the energy efficiency degradation trend of a new energy vehicle compressor, provided as an embodiment of this application. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0027] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0028] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0029] Example 1 In the thermal management system of new energy vehicles, electric compressors typically participate in multiple tasks simultaneously, including cabin cooling, heat pump heating, battery cooling, and motor thermal management. The compressor's energy efficiency is not only affected by its own motor efficiency, mechanical friction, sealing condition, and lubrication condition, but also significantly influenced by a combination of external factors such as ambient temperature, vehicle speed, heat exchanger airflow, electronic expansion valve opening, fan speed, water pump flow rate, and battery and cabin heat load. Current technologies often directly compare the compressor's coefficient of efficiency (COP) under different time periods or operating conditions, which can easily misjudge normal fluctuations in energy efficiency caused by changes in the external environment or heat load as performance degradation of the compressor itself. Alternatively, relying solely on vehicle system energy consumption prediction makes it difficult to isolate interfering factors such as heat exchanger blockage and refrigerant quantity changes, resulting in an inability to accurately assess the compressor's true health status. Furthermore, real-world vehicle driving data is often in a non-steady-state process, lacking controllable standard operating conditions for calibration, leading to significant drift and uncertainty in lifespan predictions based on historical data.

[0030] Based on the above issues, please refer to Figure 1 , Figure 1 A flowchart of a method for predicting the energy efficiency degradation trend of a new energy vehicle compressor, provided in this application embodiment, is included. The method includes the following steps: Step 1: Collect operating data during the operation of the new energy vehicle compressor. The operating data includes compressor operating parameters, refrigerant side status parameters, thermal management actuator parameters, and vehicle thermal load parameters. The operational data refers to a multi-source heterogeneous data set used to characterize the compressor's operating status and its external environmental conditions. Specifically, compressor operating parameters include, but are not limited to, compressor speed, motor phase current, bus voltage, input power, drive frequency, speed following error, and compressor start / stop status. These data directly reflect the compressor's power input and mechanical response. Refrigerant-side status parameters include suction pressure, discharge pressure, suction temperature, discharge temperature, suction superheat, discharge superheat, condensing pressure, and evaporating pressure, used to characterize the refrigerant's thermodynamic state during circulation. Thermal management actuator parameters cover electronic expansion valve opening, condenser fan speed, blower speed, water pump speed, three-way valve position, and thermal management valve group status, reflecting the system's regulation of refrigerant flow and heat exchange capacity. Vehicle thermal load parameters include vehicle speed, ambient temperature, ambient humidity, sunlight intensity, vehicle driving mode, battery SOC, battery temperature, battery cooling request, motor and electronic control cooling request, cabin set temperature, actual cabin temperature, internal and external circulation status, and passenger compartment temperature drop or rise rate. The aforementioned data is collected in real-time from the vehicle's thermal management controller, vehicle controller, or cloud-based data processing unit. For example, input electrical power can be calculated by collecting bus voltage and motor phase current, or the power value calculated internally by the compressor controller can be directly read; the cabin heat load characterization value can be estimated based on the difference between the cabin set temperature and the actual temperature, the blower airflow, and the ambient temperature. Through the synchronous collection of multi-dimensional data, the overall operation of the compressor in a complex real-vehicle environment can be fully restored, providing a complete data foundation for subsequent elimination of operating condition disturbances and separation of degradation characteristics. This step aims to obtain high-granularity, comprehensive raw operating information to ensure that the subsequent energy efficiency evaluation model has sufficient input variables to decouple external interference.

[0031] Step 2: Divide the vehicle's operating data into segments and, based on compressor speed fluctuations, intake and exhaust pressure fluctuations, electronic expansion valve opening changes, and thermal management mode switching status, filter out steady-state or quasi-steady-state driving condition segments. Segment segmentation refers to dividing continuous time-series data into several evaluation windows with specific time spans. Steady-state or quasi-steady-state driving condition segments refer to operating intervals within a certain time window where the changes in key state quantities meet preset thresholds or exhibit a gradual trend. Specifically, the system uses a sliding time window mechanism to traverse the operating data, with the time window length set from 30 to 90 seconds. During the screening process, the fluctuation rates of compressor speed, intake pressure, and exhaust pressure, as well as the change in electronic expansion valve opening, are first calculated within the time window; simultaneously, the switching of the thermal management mode is monitored. When the compressor speed fluctuation rate does not exceed 3%–8%, the intake and exhaust pressure fluctuation rates do not exceed 3%–8%, the change in electronic expansion valve opening does not exceed 5%–15%, and the thermal management mode has not switched, and the compressor is not in a transient process of just starting or stopping, the segment is determined to be a steady-state driving condition segment. For segments that do not fully meet the aforementioned strict steady-state conditions, but whose state variables exhibit monotonic or slowly changing trends, they can be identified as quasi-steady-state driving condition segments and dynamically corrected in subsequent processing, such as by deducting the effects of the pressure build-up phase or the temperature response lag phase. For example, if the compressor speed remains at 3000rpm±50rpm within a 60s time window, the suction pressure fluctuation is less than 0.05MPa, and there is no mode switching command, then this segment is retained. Through this screening mechanism based on multi-dimensional index joint judgment, it is possible to effectively eliminate drastic disturbances such as vehicle acceleration / deceleration, rapid frequency conversion, and defrosting switching, avoiding misinterpreting energy efficiency fluctuations under non-steady-state conditions as compressor performance degradation, thereby ensuring the reliability and consistency of energy efficiency evaluation data.

[0032] Step 3: Generate equivalent operating condition features based on the driving condition segment, and call the benchmark energy efficiency surface under the compressor's healthy state to obtain the benchmark equivalent energy efficiency value corresponding to the driving condition segment; Here, equivalent operating condition characteristics can refer to feature vectors that map the compressor's operating state under different ambient temperatures, vehicle speeds, and heat loads to a unified comparable space. The benchmark energy efficiency surface can refer to a mathematical model characterizing the mapping relationship between different equivalent operating condition characteristics and the ideal energy efficiency value of the compressor under healthy conditions. Specifically, the equivalent operating condition feature vector... Including compressor speed Pressure ratio (Calculated from exhaust pressure and intake pressure), intake temperature Exhaust temperature Ambient temperature Electronic expansion valve opening Equivalent heat exchange air volume Equivalent coolant flow rate Cabin heat load characterization value Battery thermal load characterization value and thermal management mode indicator To improve comparability, the above characteristics can be normalized, such as mapping the rotational speed to a proportional ratio of the rated rotational speed and dividing the ambient temperature into low, medium, and high temperature ranges. (Reference energy efficiency surface) It is derived from one or more of the following: compressor bench health data, initial vehicle operating data after delivery, initial operating data after compressor replacement, or health data of a group of vehicles of the same model, and constructed through partitioned clustering. For any equivalent operating condition feature By calling this surface, the corresponding health benchmark energy efficiency value can be output. For example, when the current segment is detected to be characterized by a medium-temperature environment, medium-speed operation, and cabin cooling mode, the system interpolates from the reference surface to obtain a COP value of 3.5 that a healthy compressor should achieve under this specific operating condition. This step achieves horizontal comparability of the compressor's energy efficiency level under different external conditions by standardizing the ever-changing real-vehicle operating conditions into equivalent characteristics and using a pre-built health benchmark for comparison, thus identifying potential degradation signals that deviate from a healthy state.

[0033] Step 4: Calculate the measured equivalent energy efficiency value based on the compressor input power and heat output characterization values ​​in the driving condition segment, compare the measured equivalent energy efficiency value with the benchmark equivalent energy efficiency value, and generate the driving condition degradation evaluation value. Among them, the measured equivalent energy efficiency value can refer to the compressor energy efficiency index calculated based on actual operating data, and the driving condition degradation evaluation value can refer to the quantitative value of the deviation of the measured energy efficiency from the healthy benchmark. Specifically, the measured equivalent energy efficiency value (such as the measured equivalent COP) is expressed by the formula... Calculation, where Input electrical power to the compressor as collected. This is the estimated equivalent heat output (cooling or heating capacity) based on parameters such as temperature response, refrigerant state, heat exchanger temperature difference, and fluid flow rate. Subsequently, the energy efficiency deviation is calculated. To eliminate the interference of individual abnormal segments, the system performs weighted fusion of energy efficiency deviations from multiple driving condition segments within the same statistical period. The weights are determined based on the segment's steady-state score, equivalent driving condition coverage, and degree of matching with the reference surface. For example, if a segment has a high steady-state score and strong driving condition representativeness, its weight will be higher. Larger values ​​indicate a higher degradation rate, while smaller values ​​indicate a lower degradation rate. The final generated driving condition degradation evaluation value is... This is the result after weighted averaging. An increase in the value indicates that, under the same equivalent operating conditions, the actual energy efficiency of the compressor has significantly decreased compared to its healthy state. This step utilizes data generated at high frequencies during operation to achieve broad coverage and lightweight real-time sensing of the compressor's energy efficiency status, enabling timely detection of early signs of energy efficiency changes.

[0034] Step 5: When the vehicle meets the parking self-diagnosis trigger conditions, control the compressor and thermal management actuator to enter the parking standard operating condition test combination, and collect parking standard operating condition data after the parking standard operating condition test combination reaches a stable state. The parking self-diagnostic trigger conditions include periodic trigger conditions and abnormal trigger conditions. The parking standard operating condition test combination can refer to a series of preset, controlled operating state sequences of the compressor and thermal management components. Specifically, periodic trigger conditions can be determined based on the time interval since the last self-diagnosis (e.g., 7 days) or the cumulative operating time of the compressor (e.g., 50 hours); abnormal trigger conditions are determined based on abnormal increases, excessive fluctuations, or insufficient number of valid segments in the driving condition degradation evaluation value. When the vehicle is parked (e.g., during charging, before remote air conditioning is started, or when the vehicle is powered off but thermal management allows operation), and the battery SOC and ambient temperature meet safety requirements, self-diagnosis is triggered. After triggering, the thermal management controller controls the compressor, electronic expansion valve, fan, and water pump to sequentially execute multiple standard test stages from low load to high load. For example, in the first stage, the compressor is controlled to run at 2000 rpm, the valve opening is set to a fixed value, and 30 seconds are waited until the speed, pressure, and temperature stabilize; in the second stage, the speed is increased to 4000 rpm, and stabilization is waited again. After reaching a stable condition, standard parking operating condition data such as input electrical power, intake and exhaust pressure, temperature, pressure settling time, and response slope are collected. This step, by actively performing standardized testing in a controlled environment, eliminates interference from uncontrollable environmental and load variables during driving, obtaining a high-confidence snapshot of the compressor's performance, and providing an accurate benchmark anchor point for correcting deviations in driving data.

[0035] Step 6: Calculate the measured standard energy efficiency value based on the parking standard operating condition data, and compare the measured standard energy efficiency value with the corresponding health standard energy efficiency value of the parking standard operating condition to generate the parking condition degradation evaluation value; The measured standard energy efficiency value refers to the energy efficiency index measured during the parking standard operating condition test phase, while the healthy standard energy efficiency value refers to the theoretical or calibrated energy efficiency value of the compressor in a healthy state corresponding to the standard test phase. Specifically, for each standard test phase (such as low speed, medium speed, and high speed), the measured COP value is calculated and compared with the pre-stored healthy baseline COP value to obtain the parking energy efficiency deviation for each phase. In addition, pressure build-up time deviation and exhaust temperature residual can be introduced as auxiliary evaluation indicators. For example, if the measured pressure build-up time is significantly longer than the healthy baseline or the exhaust temperature is higher than expected during the high-load test phase, it is considered a degradation signal. The system integrates the deviations from each phase according to preset weights (the weight for the high-load phase is usually higher) to generate the final parking condition degradation evaluation value. For example, if the deviations in the three testing phases are 2%, 5%, and 8% respectively, the weighted fusion yields... The figure is 6%. This step utilizes high-precision controlled test data to quantify the actual degradation level of the compressor at the current moment. The results are minimally affected by external random factors, exhibiting high robustness and accuracy. It is mainly used to verify and correct the evaluation results of driving conditions.

[0036] Step 7: Perform fusion calibration based on the driving condition degradation evaluation value and the parking condition degradation evaluation value to generate the compressor body energy efficiency degradation index; The fusion calibration refers to the process of dynamically adjusting the weights of driving data and parking data to generate a comprehensive degradation index, combining the broad coverage of driving data and the high precision of parking data. The compressor's energy efficiency degradation index is a single quantitative indicator characterizing the compressor's current health status. Specifically, the system first evaluates the driving-side reliability (based on data coverage, segment stability, and trend consistency) and the parking-side reliability (based on test completion time, stability, and phase consistency). Then, the fusion weights are dynamically determined based on the reliability of both sides. The fusion formula can be expressed as: When the amount of driving data is large and the quality is high, improve... When a high-quality parking self-diagnosis is recently completed, it improves... If the deviation between the two exceeds the consistency threshold, the weight of the less reliable side is automatically reduced, or even a re-evaluation is triggered. For example, if driving data shows severe degradation but parking test results are normal, the system may determine that this is due to sensor drift or operational condition estimation error, thereby reducing the weight of driving data and avoiding false alarms. Furthermore, parking results can be used to reverse-correct the health baseline energy efficiency surface, allowing the model to adapt to the individualized aging characteristics of vehicles. This step, through dual-path collaborative modeling, preserves the real-time nature of driving data while using parking data to suppress long-term prediction drift, significantly improving the accuracy and stability of the degradation index.

[0037] Step 8: Based on the time series of the compressor's energy efficiency degradation index and the compressor's cumulative operating characteristics, predict the compressor's energy efficiency degradation trend and output the energy efficiency degradation prediction results.

[0038] Here, the time series refers to a set of historical compressor energy efficiency degradation indices arranged chronologically. The cumulative operating characteristics of the compressor include physical quantities reflecting cumulative damage, such as cumulative operating hours, cumulative start-stop cycles, cumulative high pressure ratio operating time, and cumulative high discharge temperature operating time. Specifically, the system construction includes historical degradation indices. The system uses an input vector containing cumulative operating characteristics and time series models or machine learning algorithms (such as piecewise linear models, exponential degradation models, or neural networks) to predict future degradation trajectories. The predictions cover the rate of energy efficiency decline within a set future period, the inflection point of accelerated energy efficiency degradation, and the expected time or mileage to enter an inefficient operating state. Based on this, degradation levels are categorized (e.g., normal, slight degradation, significant degradation, rapid degradation, and maintenance recommended). For example, if the degradation index slope exceeds a threshold for three consecutive periods, a maintenance recommendation warning is issued even if the absolute value does not reach the repair threshold. Finally, the energy efficiency degradation prediction results, including degradation trends, levels, and maintenance recommendations, are output to the vehicle controller, in-vehicle infotainment system, or cloud platform. This step transforms static health status assessment into dynamic trend prediction, introducing a physical cumulative damage dimension, making the prediction results interpretable. It guides users to perform maintenance before serious compressor failures or a sharp drop in energy efficiency occur, thereby optimizing vehicle energy consumption and extending system lifespan.

[0039] Example 2 In another optional embodiment, before performing segment division, the collected operating data is synchronized in time and abnormal data is removed. The continuous operating data is divided into several segments to be evaluated, using the compressor start / stop status, thermal management mode status, and data continuity as segmentation boundaries. The segments to be evaluated that have sensor jumps, communication interruptions, or compressor control command mutations are downweighted or removed.

[0040] Step 1: Synchronize the collected operating data with time and remove abnormal data, and divide the continuous operating data into several segments to be evaluated, using compressor start / stop status, thermal management mode status and data continuity as segment boundaries. Time synchronization refers to the process of unifying heterogeneous data from different controllers or sensors to a single time reference. Since parameters such as compressor speed, refrigerant pressure, and electronic expansion valve opening may be collected and transmitted at different sampling frequencies by the motor controller, thermal management controller, or vehicle controller, direct merging would lead to transient characteristic misalignment. Specifically, the system uses linear interpolation or nearest-neighbor interpolation algorithms to map all parameters to a unified timestamp sequence, for example, using a 100ms interval, ensuring a strict correspondence between speed and pressure data at the same moment. Outlier removal refers to identifying and removing invalid values ​​caused by sensor malfunctions, electromagnetic interference, or communication packet loss. This includes: using a sliding window to calculate the mean and standard deviation of each parameter; identifying outliers when data deviates from the mean by more than three times the standard deviation at a given moment; or detecting whether data exceeds physical limits (e.g., suction pressure less than 0 MPa or greater than 5 MPa), marking such data as invalid and removing or filling it in.

[0041] Segmentation boundaries are logical breakpoints for dividing the segments to be evaluated, specifically including three dimensions: compressor start-stop state, thermal management mode state, and data continuity. Compressor start-stop state as a boundary means dividing the transition process from shutdown to startup, the stable operation process, and the transition process from operation to shutdown into independent segments, avoiding interference from startup current surges or shutdown inertia on steady-state energy efficiency calculations. Thermal management mode state as a boundary can refer to forcibly cutting off the current segment when the thermal management system switches between modes such as cabin cooling, heat pump heating, battery cooling, or combined thermal management, because the heat exchange path and load characteristics fundamentally change under different modes and cannot be evaluated together. Data continuity as a boundary can refer to dividing the data before and after the interruption into two independent segments when a communication interruption causes data loss for more than a set threshold (e.g., 2 seconds). Through the synergistic effect of these three boundaries, the continuous raw data stream can be segmented into several semantically complete and single-condition segments to be evaluated. For example, a raw data segment containing the process of compressor startup, stable operation in cooling mode, and mode switching to heating will be automatically segmented into startup segment, cooling steady-state segment, and heating transition segment, with only the segments that meet the requirements of subsequent analysis being retained for the screening process.

[0042] This pre-cleaning and refined segmentation can eliminate transient misalignments caused by multi-source asynchronous acquisition, avoid single-point mutations from polluting the entire segment statistics, and ensure that each segment to be evaluated truly reflects a physically meaningful operating state, providing a high-quality data foundation for subsequent steady-state scoring and equivalent operating condition mapping.

[0043] Step 2: Reduce the weight or remove segments to be evaluated that exhibit sensor jumps, communication interruptions, or sudden changes in compressor control commands; The downweighting or elimination process is a secondary filtering mechanism applied to segments that, although they have passed the initial segmentation, still contain local quality defects. Sensor jumps can refer to non-physical, drastic fluctuations in parameters within a short period of time within a segment, such as a sudden change of more than 20% in intake and exhaust pressure within 100ms without corresponding speed changes. Communication interruptions can refer to brief data packet loss within a segment, causing discontinuities in the sequence of key parameters. Sudden changes in compressor control commands can refer to a step adjustment in target speed or target opening degree within a short period, indicating that the segment is in dynamic adjustment rather than steady-state operation. The system first calculates a quality score for each segment to be evaluated, and generates a confidence coefficient based on the proportion of outliers within the segment, the magnitude of jumps, and the frequency of command changes.

[0044] For segments with confidence coefficients below the first threshold, a direct removal process is performed, completely eliminating them from the subsequent evaluation set to prevent inferior data from misleading the assessment of degradation trends. For segments with confidence coefficients between the first and second thresholds, a weight reduction process is implemented, meaning that the weight coefficient of these segments is reduced when calculating the degradation evaluation value for driving conditions. For example, the weight of normal steady-state segments is set to 1.0, while the weight of quasi-steady-state segments with slight command fluctuations is reduced to 0.6. This differentiated processing strategy balances data robustness and information retention, preventing the use of sparse usable data due to overly stringent filtering conditions and effectively suppressing the negative impact of low-quality data on the overall degradation index calculation. By combining this step with the time synchronization and boundary delineation of the previous step, a dual guarantee mechanism from source cleaning to segment-level quality control is formed, ensuring that the data ultimately used to generate equivalent operating condition characteristics and calculate energy efficiency deviations are highly reliable, thereby significantly improving the accuracy of compressor energy efficiency degradation trend prediction.

[0045] Example 3 One possible implementation also includes determining steady-state or quasi-steady-state driving condition segments based on steady-state scores and making dynamic corrections, including the following steps: Step 1: The steady-state or quasi-steady-state driving condition segment is determined by steady-state scoring. The steady-state score is calculated based on the changes in compressor speed, intake and exhaust pressure, electronic expansion valve opening, and thermal management mode duration. The steady-state score is a comprehensive indicator used to quantify the stability of the compressor's operating state during vehicle operation. This score is obtained by real-time acquisition of compressor speed, suction pressure, discharge pressure, electronic expansion valve opening, and thermal management mode switching signals within the evaluated segment, and by calculating the amplitude or duration of change of each parameter within a preset time window. Specifically, the amplitude of compressor speed change reflects the fluctuation of the compressor motor's drive state, the amplitude of suction and discharge pressure change characterizes the stability of refrigerant-side fluid dynamics, the amplitude of electronic expansion valve opening change reflects the degree of adjustment of the throttling element, and the duration of the thermal management mode is used to determine whether the system is in a transient disturbance period during mode switching. These four dimensions of parameters together constitute the basis for the steady-state score calculation, enabling the score results to comprehensively cover multiple levels such as mechanical motion, fluid transmission, actuator response, and system control strategies.

[0046] For example, a sliding time window of 60 seconds is set, and the ratio of the standard deviation to the average value of the compressor speed within this window is calculated as the speed change rate. If the ratio is 2%, it corresponds to a higher speed stability score. Simultaneously, the percentage of time during which the thermal management mode does not switch within this window is counted; if there is no switching throughout, this item receives a perfect score. The normalized scores of the above four dimensions are then weighted and summed according to preset weights (e.g., speed weight 0.3, pressure weight 0.3, valve opening weight 0.2, mode duration weight 0.2) to obtain the final steady-state score for this segment. This multi-dimensional quantitative scoring mechanism effectively distinguishes between absolutely stable operating states and operating states with slight disturbances, providing accurate data support for subsequent classification processing.

[0047] This step aims to replace the traditional binary threshold judgment with a continuous scoring mechanism, thereby more precisely identifying effective data segments that can be used for energy efficiency evaluation in complex real-world driving environments. By comprehensively considering engine speed, pressure, valve opening, and mode duration, it avoids erroneously eliminating effective segments due to instantaneous fluctuations in a single parameter, and also prevents the inclusion of drastically changing transient processes in the evaluation scope, ensuring the reliability and representativeness of the data source.

[0048] Step 2: When the steady-state score meets the set scoring conditions, the corresponding segment is identified as a steady-state driving condition segment. When the steady-state score does not meet the set scoring conditions but the trend of each state variable in the segment meets the monotonic or slow-change conditions, the corresponding segment is identified as a quasi-steady-state driving condition segment, and dynamic correction is performed on the quasi-steady-state driving condition segment.

[0049] The scoring criteria can refer to pre-defined steady-state scoring thresholds, used to distinguish between steady-state and unsteady-state conditions. Steady-state driving condition segments refer to data segments where operating parameters fluctuate minimally and the system is in thermodynamic equilibrium; these can be directly used for benchmark energy efficiency comparisons. Quasi-steady-state driving condition segments refer to segments where, although overall fluctuations are slightly higher than the steady-state threshold, internal state variables (such as speed and pressure) exhibit monotonically increasing, monotonically decreasing, or linearly gradually changing trends. These segments typically occur in scenarios such as slow vehicle acceleration, gradual changes in ambient temperature, or gradual temperature increases in heat pump systems. Dynamic correction refers to mathematical compensation or trend fitting of measured data in quasi-steady-state segments to eliminate energy efficiency calculation biases caused by unsteady-state characteristics.

[0050] For example, if the scoring criteria are set at 85 points (out of 100), and a segment scores 92 points, it is directly marked as a steady-state driving condition segment. If another segment scores 78 points, failing to meet the steady-state standard, but analysis reveals that its compressor speed linearly increases from 2000 rpm to 2200 rpm within 60 seconds, and the intake and exhaust pressures show a monotonically increasing trend with a constant slope, then this segment is classified as a quasi-steady-state driving condition segment. For this quasi-steady-state segment, the dynamic correction process can use linear regression to fit the baseline of each parameter's change, and subtract the hysteresis component or transient response component caused by the rate of change from the measured values, thereby restoring the equivalent energy efficiency value close to steady-state conditions.

[0051] This step, through a combination of dual thresholds and trend analysis, enables the tiered utilization of operational data. For steady-state segments, data is directly accepted to ensure maximum accuracy; for quasi-steady-state segments, dynamic corrections are used to uncover their potential value, significantly expanding the data coverage available for degradation assessment. This approach effectively addresses the problem of insufficient effective samples caused by the over-reliance on strict steady-state conditions in traditional methods. Simultaneously, the correction algorithm ensures the accuracy of quasi-steady-state data in energy efficiency calculations, making the generation of the compressor's energy efficiency degradation index more continuous and robust.

[0052] Example 4 In another embodiment, the method further includes a process for generating equivalent operating condition features, specifically comprising the following steps: Step 1: First, divide the driving condition segment into cabin cooling, heat pump heating, battery cooling, or combined thermal management conditions according to the thermal management mode. Thermal management mode refers to the core thermodynamic cycle type executed by the thermal management system of a new energy vehicle at any given moment. Its classification is primarily based on the mode command signals issued by the vehicle controller or thermal management controller, as well as the state combinations of key execution components. Specifically, the cabin cooling mode corresponds to the operating state where the evaporator absorbs cabin heat and the compressor lowers the interior temperature. In this state, the electronic expansion valve controls the evaporator outlet superheat, and the condenser dissipates heat to the environment. The heat pump heating mode corresponds to the operating state where, through a four-way valve or refrigerant flow switching, the external heat exchanger acts as the evaporator to absorb heat, and the internal heat exchanger acts as the condenser to release heat. The battery cooling mode corresponds to the refrigerant circuit directly or through a plate heat exchanger exchanging heat with the battery coolant circuit to maintain the battery pack within a suitable temperature range. The combined thermal management mode refers to the operating state where the system simultaneously responds to cabin temperature control requests and battery temperature control requests, with refrigerant flow dynamically distributed among multiple branches. For example, when the three-way valve is detected pointing towards the battery cooling branch and the cabin blower is at low speed or off, the current condition is determined to be battery cooling. When the temperature difference between the inlet and outlet of the in-vehicle heat exchanger is positive and the temperature difference between the inlet and outlet of the external heat exchanger is negative, and the compressor is running continuously at high frequency, the condition is determined to be heat pump heating. This primary classification based on thermal management modes can isolate operating data with drastically different physical meanings and inconsistent energy efficiency evaluation metrics (such as the difference in the definition of cooling COP and heating COP) into different logical partitions, avoiding evaluation distortion caused by direct cross-mode comparisons and laying the classification foundation for the subsequent construction of targeted equivalent operating points.

[0053] Step 2: Then, based on the compressor load status, refrigerant pressure status, and heat exchanger load status, generate equivalent operating points under corresponding operating conditions, so that compressor operating segments under different ambient temperatures, different vehicle speeds, and different heat loads are mapped to comparable equivalent operating condition spaces. The equivalent operating point refers to a low-dimensional feature vector that uniquely characterizes the compressor's work intensity and boundary conditions within a specific thermal management mode partition. The compressor load state reflects the compressor's mechanical work level, specifically characterized by the ratio of compressor speed to rated speed, motor input torque, or refrigerant mass flow rate per unit time. The refrigerant pressure state reflects the compressor's thermodynamic cycle pressure ratio, specifically characterized by the pressure ratio calculated from the suction and discharge pressures, or the compression termination temperature calculated by combining suction and discharge temperatures. The heat exchanger load state reflects the intensity of external thermal disturbances faced by the compressor, specifically determined by a combination of ambient temperature, the windward heat transfer coefficient corresponding to vehicle speed, estimated battery heat generation power, or estimated cabin heat load. The specific process for generating the equivalent operating point includes: firstly, extracting the statistical feature values ​​(such as mean and weighted average) of the above three state quantities within the driving condition segment; then, using a preset normalization mapping function, transforming these physical quantities with different dimensions and orders of magnitude to a unified dimensionless equivalent spatial coordinate system. For example, in cabin cooling mode, if one segment has an ambient temperature of 35℃, a vehicle speed of 80km / h, and a compressor speed of 4000rpm, while another segment has an ambient temperature of 25℃, a vehicle speed of 40km / h, and a compressor speed of 3000rpm, although their original operating parameters are completely different, after mapping calculation, if the power consumption and pressure ratio per unit refrigerant mass flow rate generated by both segments are consistent, they will be mapped to the same equivalent operating point with the same coordinates in the equivalent operating space. Similarly, for battery cooling mode, the ratio of the current battery heat generation rate to the coolant flow rate is mapped to an equivalent heat load coefficient, which, combined with the current refrigerant pressure ratio, determines the equivalent position of that segment in the battery cooling subspace. Through this mode-state dual-driven mapping mechanism, the high-dimensional heterogeneous operating data, which was originally scattered under different ambient temperatures, vehicle speeds, and heat loads, is projected into a low-dimensional isomorphic equivalent operating condition space. This makes the compressor operating segments under different physical scenarios mathematically comparable, thereby ensuring that when the benchmark energy efficiency surface is called later, the health benchmark value corresponding to the physical essence of the segment can be accurately matched, effectively eliminating the interference of external operating condition fluctuations on the evaluation of compressor energy efficiency degradation.

[0054] Example 5 In one possible implementation, this application also provides a specific method for constructing a benchmark energy efficiency surface.

[0055] Step 1: The baseline energy efficiency surface is constructed using healthy samples. These healthy samples are derived from one or more of the following: compressor bench health data, initial vehicle operating data after factory delivery, initial operating data after compressor replacement, and health data of a group of vehicles of the same model. In this context, "health samples" refers to datasets characterizing the energy efficiency of compressors under conditions of no or slight degradation. The sources of health samples are diverse, aiming to cover the health status characteristics of compressors from different dimensions. Specifically, compressor bench health data comes from standardized performance tests conducted on compressors in a laboratory environment. This data eliminates interference from the vehicle's thermal management system and provides a theoretically optimal energy efficiency benchmark. Initial vehicle operating data comes from real-world driving records within a preset mileage or time period after new vehicle delivery, reflecting the initial individual state of a specific vehicle after assembly and eliminating the impact of individual manufacturing tolerances. Initial operating data after compressor replacement comes from operating records after replacing the compressor in after-sales maintenance scenarios, used to re-anchor the health benchmark midway through the compressor's lifecycle. Group health data for vehicles of the same model comes from aggregated statistical results of a large number of vehicles of the same model in a cloud database, used to compensate for the sparsity of single-vehicle data and provide group statistical regularities. In practical applications, the system can flexibly select one or more of the above data sources for fusion based on data availability. For example, for new vehicles, initial vehicle operating data combined with group health data for vehicles of the same model is prioritized to construct a benchmark; for vehicles with replaced compressors, initial operating data after compressor replacement is introduced to correct the original benchmark. This multi-source data fusion mechanism ensures that the benchmark energy efficiency surface has both theoretical accuracy and individual adaptability for specific vehicles.

[0056] Step 2: When constructing the baseline energy efficiency surface, the healthy samples are divided into zones and clustered according to the characteristics of equivalent operating conditions, and the health baseline energy efficiency value corresponding to each equivalent operating condition zone is determined, so that the baseline energy efficiency surface can output the baseline equivalent energy efficiency value that matches the driving condition segment to be evaluated.

[0057] Partition clustering refers to the process of dividing a continuous healthy sample space into several discrete local regions based on equivalent operating condition feature vectors. Equivalent operating condition features include, but are not limited to, compressor speed, pressure ratio, intake temperature, exhaust temperature, ambient temperature, electronic expansion valve opening, and thermal management mode identifiers. The system first extracts these feature vectors from the healthy samples and uses clustering algorithms (such as K-Means clustering, DBSCAN density clustering, or grid-based clustering methods) to group samples with similar operating condition features into the same partition. Subsequently, for each equivalent operating condition partition, the energy efficiency values ​​(such as COP, isentropic efficiency, etc.) of all healthy samples within that region are statistically analyzed. The baseline energy efficiency value for that partition is determined by calculating the average, median, or fitting a local response surface. This partitioning approach avoids the local distortion problem caused by using a single global model to fit all operating conditions. For example, within a specific equivalent operating condition partition of high speed-high pressure ratio-high temperature environment, the system only uses bench data and fleet data falling within that region to calculate the baseline value, thereby accurately reflecting the compressor's healthy energy efficiency level under this extreme condition. When it is necessary to evaluate a certain driving condition segment in the future, the system only needs to extract the equivalent operating condition features of the segment, locate the equivalent operating condition partition to which it belongs, and then directly call the pre-determined health benchmark energy efficiency value of the partition, thus achieving the matching of the benchmark value with the operating condition to be evaluated.

[0058] Example 6 One embodiment further includes generating a driving condition degradation evaluation value and performing weighted fusion and confidence adjustment, including the following steps: Step 1: Compare the measured equivalent energy efficiency value of each driving condition segment with the corresponding benchmark equivalent energy efficiency value to obtain the energy efficiency deviation of that segment; The measured equivalent energy efficiency value is an instantaneous or average energy efficiency index, such as equivalent COP or isentropic efficiency, calculated based on the compressor input power and heat output characterization values ​​in the previously selected steady-state or quasi-steady-state operating condition segments. The benchmark equivalent energy efficiency value is the theoretical healthy energy efficiency value obtained by calling the benchmark energy efficiency surface under the compressor's healthy state and querying it according to the equivalent operating condition characteristics of the segment. The energy efficiency deviation is used to quantify the degree of deviation of the compressor's actual performance from its healthy state under the current segment. Specifically, the energy efficiency deviation can be determined by calculating the difference between the measured value and the benchmark value, or by calculating the proportion of the difference to the benchmark value. For example, if the measured equivalent COP of a certain segment is 3.2, while the corresponding healthy benchmark COP under the equivalent operating condition is 4.0, then the energy efficiency deviation can be expressed as (4.0-3.2) / 4.0=0.2, that is, the energy efficiency has decreased by 20%. This deviation serves as the basis data for subsequent weighted fusion and directly reflects the signs of energy efficiency degradation within a single segment.

[0059] Step 2: Determine the segment weight based on the segment's steady-state score, equivalent operating condition coverage, and degree of matching with the benchmark energy efficiency surface; The segment weight is a coefficient used to measure the reliability and importance of the current driving condition segment in the overall degradation evaluation. The steady-state score is calculated based on the stability of parameters such as compressor speed fluctuations, intake and exhaust pressure fluctuations, and electronic expansion valve opening changes within the segment. A higher steady-state score indicates less transient interference and higher data quality, thus assigning a greater weight. The equivalent operating condition coverage refers to the distribution density or representativeness of the equivalent operating condition characteristics of the segment in the benchmark energy efficiency surface space. If the segment's operating point is located in a dense sample area or typical operating area, the coverage is high, and the weight increases accordingly; conversely, if it is in a sparse edge area, the weight is reduced to avoid extrapolation errors. The matching degree with the benchmark energy efficiency surface is obtained by calculating the similarity between the segment's feature vector and the feature distribution of the healthy samples used to construct the benchmark surface. A higher matching degree indicates stronger applicability of the benchmark value to the segment. For example, if a segment has a steady-state score of 0.9 (out of 1.0), and its operating point is located within the core clustering region of the reference surface with a matching coefficient of 0.95, then the comprehensive weight of that segment can be set to a higher value (e.g., 0.85). Conversely, for a segment that has just undergone a thermal management mode switch and has a steady-state score of only 0.4, even if its energy efficiency deviation is large, its weight will be significantly reduced (e.g., 0.1), thereby preventing abnormal fluctuations from misleading the overall evaluation. Through this three-dimensional collaborative evaluation, the system can automatically filter low-quality data, ensuring that the segments participating in the fusion have high reliability.

[0060] Step 3: Weighted fusion of energy efficiency deviations from multiple segments, and improved reliability of degradation assessment for the corresponding statistical period when energy efficiency deviations continue to change in the same direction; The weighted fusion method involves multiplying the energy efficiency deviation of all valid segments within a statistical period by their corresponding segment weights, summing the results, and then dividing by the total weights to obtain the driving condition degradation evaluation value for that statistical period. This process eliminates the influence of random noise from individual abnormal segments, making the evaluation value more reflective of the compressor's true condition. Based on this, a continuous unidirectional change mechanism is introduced to dynamically adjust the reliability of the degradation evaluation for the statistical period. Specifically, the system monitors the changing trend of the driving condition degradation evaluation value for multiple consecutive statistical periods. If the energy efficiency deviation shows a continuously increasing (i.e., continuously decreasing energy efficiency) unidirectional change trend over N consecutive periods (e.g., N=3 or 5), and the change exceeds a set threshold, it is determined that the compressor may have entered an accelerated degradation stage. At this point, the reliability coefficient of the degradation evaluation for the current statistical period is significantly increased. For example, if the degradation evaluation values ​​for three consecutive cycles are 0.05, 0.08, and 0.12, showing a clear monotonically increasing trend, the system will mark the confidence level of this sequence as high and assign a higher confidence weight to this driving-side data during subsequent fusion calibration with parking condition data. Conversely, if the energy efficiency deviation fluctuates violently and without a clear direction across different cycles, the confidence level will be reduced, indicating that more data accumulation may be needed or that parking self-diagnosis may be triggered for verification. This mechanism allows the evaluation system to not only reflect the current energy efficiency level but also to keenly capture the dynamic trend of degradation, providing more input for subsequent fusion calibration.

[0061] Example 7 In one optional implementation, a flowchart for determining the parking self-diagnosis triggering conditions and triggering the parking self-diagnosis is also included, comprising the following steps: Step 1: Parking self-diagnosis trigger conditions include periodic trigger conditions and abnormal trigger conditions; Periodic triggering conditions can refer to a routine diagnostic startup mechanism set based on time or operating load dimensions, used to ensure the basic monitoring frequency of compressor energy efficiency status. Abnormal triggering conditions can refer to a risk-driven diagnostic startup mechanism set based on real-time monitoring results of driving condition degradation evaluation values, used to respond immediately when potential degradation risks are detected. The function of periodic triggering conditions is to perform health checks on the compressor according to a preset rhythm when there are no significant abnormal signs, preventing the slow performance degradation from being missed due to long-term lack of testing. The function of abnormal triggering conditions is to break the routine periodic limit and immediately initiate a high-precision parking standard operating condition test to confirm the degradation status when driving data reflects abnormal fluctuations or a sharp drop in compressor energy efficiency.

[0062] Specifically, the criteria for setting periodic triggering conditions include the time interval since the last successful completion of parking self-diagnosis, or the cumulative operating time of the compressor. For example, the system can be set to trigger periodic self-diagnosis every 7 days or when the compressor has accumulated 50 hours of operation, to ensure that at least one standardized energy efficiency benchmark data is obtained within the normal vehicle usage cycle. The criteria for setting abnormal triggering conditions include an abnormal increase in the driving condition degradation evaluation value, excessive fluctuation of the driving condition degradation evaluation value, or insufficient number of driving condition segments. For example, if the driving condition degradation evaluation value exceeds the preset threshold for three consecutive statistical periods (e.g., energy efficiency deviation greater than 15%), it is judged as an abnormal increase; or if the driving condition degradation evaluation value fluctuates violently in a short period of time (e.g., standard deviation exceeds the set limit), it is judged as excessive fluctuation; or if a sufficient number (e.g., less than 5) of effective steady-state driving condition segments cannot be selected within a long period of time (e.g., two weeks), it is judged as insufficient segment quantity. This may indicate sensor failure or a special vehicle operating mode, requiring parking self-diagnosis to verify the reliability of the data acquisition link or obtain an independent benchmark.

[0063] This step aims to establish a dual-track diagnostic triggering logic, upgrading the originally passive, timed tasks into a proactive risk response mechanism. Periodic triggering ensures minimum diagnostic coverage, mitigating the blind spot risks caused by prolonged periods without diagnostics; abnormal triggering enables dynamic allocation of diagnostic resources, allowing high-risk vehicles to receive priority in-depth testing, thus significantly improving the timeliness and relevance of diagnostics. Simultaneously, insufficient segment quantity, as one of the triggering conditions, also serves as a system self-check function, enabling the timely activation of backup diagnostic solutions when a data acquisition link failure prevents driving evaluation.

[0064] Step 2: Periodic triggering conditions are determined based on the time interval since the last parking self-diagnosis or the cumulative running time of the compressor. Abnormal triggering conditions are determined based on the abnormal increase of the driving condition degradation evaluation value, the excessive fluctuation of the driving condition degradation evaluation value, or the insufficient number of driving condition segments. The time interval since the last parking self-diagnosis is calculated by the difference between the time of the last diagnosis completion recorded by the vehicle clock and the current time. The cumulative compressor runtime is obtained by accumulating the time segments in which the compressor is in operation, reflecting the actual wear and tear of the compressor. An abnormal increase in the driving condition degradation evaluation value is determined by comparing the degradation evaluation value of the current statistical period with the historical baseline or preset warning line. The fluctuation of the driving condition degradation evaluation value exceeding the limit is determined by calculating the variance or range of the degradation evaluation values ​​in multiple recent statistical periods and comparing it with the fluctuation threshold. The insufficient number of driving condition segments is determined by counting the total number of driving condition segments that meet the steady-state or quasi-steady-state conditions within a preset time window and comparing it with the minimum sample size threshold.

[0065] For example, if the time interval threshold is set to 15 days and the cumulative running time threshold is set to 100 hours, then meeting either condition is considered to meet the periodic triggering condition. If the threshold for abnormal increase in degradation evaluation value is set to 20%, the threshold for excessive fluctuation is a standard deviation of 0.05, and the minimum sample size threshold is 3 segments, then when the real-time calculated degradation evaluation value of driving condition suddenly increases to 25%, or the value fluctuates significantly between 5% and 25% in the past 5 evaluations, or only 1 valid segment is collected in the past month, it is considered to meet the abnormal triggering condition.

[0066] This step refines the quantitative criteria for triggering conditions, enabling the system to accurately identify when parking self-diagnostics need to be intervened. By combining rigid constraints in the time / mileage dimension with flexible constraints in the data quality / degradation trend dimension, it avoids unnecessary burdens on vehicle energy and component lifespan caused by overly frequent diagnostics, while ensuring that the true energy efficiency status of the compressor can be captured in a timely manner at critical degradation nodes or in cases of data loss, providing high-quality input data for subsequent degradation index fusion calibration.

[0067] Step 3: When the periodic triggering condition or abnormal triggering condition is met, and the vehicle is in a parking state that allows the compressor to run for a short time, the parking self-diagnosis is triggered.

[0068] The parking state that allows the compressor to operate briefly can be defined as a state where the vehicle is stationary (zero speed), in P or N gear, and the vehicle's thermal management system has the safety authorization to operate. Determining this state requires comprehensive confirmation based on battery SOC (State of Charge), ambient temperature, and the presence of higher-priority thermal management requests (such as emergency battery cooling). Only when the triggering conditions (periodic or abnormal) are met and the vehicle's environmental safety verification is passed will the system officially issue a command to control the compressor and thermal management actuators to enter the parking standard operating condition test combination.

[0069] For example, even if the periodic condition of more than 7 days since the last diagnosis is met, if the vehicle is currently driving at high speed, or the battery SOC is below 20% and insufficient to support the complete test process, or the ambient temperature is below -10℃ and exceeds the test allowable range, the system will temporarily suspend the trigger request until the vehicle enters charging mode or the environmental conditions are suitable. Conversely, if an emergency signal indicating an abnormally high driving condition degradation evaluation value is detected, and the vehicle happens to be parked at a charging station and in a connected state, the system will immediately prioritize resource allocation to initiate parking self-diagnosis to quickly identify the cause of the sudden drop in energy efficiency.

[0070] This step achieves a final closed loop between the triggering logic and the safe execution conditions. By introducing a pre-verification that the vehicle is in a parking state that allows the compressor to operate briefly, it ensures that the self-diagnostic process will not interfere with normal driving operations or endanger the safety of the vehicle's high-pressure system. This dual confirmation mechanism of condition fulfillment + state permission ensures that parking self-diagnostics are executed only at the optimal time, improving the accuracy and repeatability of test data while minimizing the impact on the user's driving experience. This reliably supports the subsequent generation of parking condition degradation evaluation values ​​and the fusion calibration of the compressor's energy efficiency degradation index.

[0071] Example 8 In an optional embodiment, the method further includes performing a standard parking test combination and generating a stage degradation evaluation value, comprising the following steps: Step 1: The parking standard operating condition test combination is executed in sequence from low load to high load, performing multiple standard test stages in sequence; The sequence from low load to high load refers to the pre-set loading logic in the control strategy, designed to avoid interference from thermal inertia or pressure residue under high load conditions on low load test points, ensuring the independence of each test stage and the purity of the data. Specifically, multiple standard test stages typically include a low-speed, low-load stage, a medium-speed, medium-load stage, and a high-speed, high-load stage. For example, the first standard test stage can be set with the compressor speed between 1500 rpm and 2500 rpm, the electronic expansion valve opening controlled within a small range, and the condenser fan and water pump in a low-flow state; the second standard test stage can be set with the compressor speed between 3000 rpm and 4500 rpm, and the actuators entering a medium-flow state; the third standard test stage increases the compressor speed to between 5000 rpm and 7000 rpm, and the actuators entering a high-flow state. This stepped increase in load setting allows the compressor's energy efficiency performance under different stress levels to be captured independently. Through this orderly loading method, the performance degradation of the compressor in specific load ranges can be accurately identified, especially for degradation modes such as increased mechanical friction or seal leakage that are only significantly exposed under high loads.

[0072] Step 2: In each standard test phase, first control the compressor and thermal management actuators to enter the corresponding standard control state, and then determine whether the compressor speed, suction and discharge pressure and discharge temperature have reached the stable conditions; The standard control state refers to a pre-set set of target operating parameters during the testing phase, including target compressor speed, target electronic expansion valve opening, target fan speed, and target water pump flow rate. The system drives each actuator to quickly converge to this state through a closed-loop control algorithm. The stability condition is based on a comprehensive criterion of multi-dimensional physical quantities, rather than the constancy of a single parameter. Specifically, the stability condition must simultaneously meet the following conditions: the deviation between the actual compressor speed and the target speed remains within a set threshold (e.g., ±2%) for a first time period; the rate of change of suction pressure and discharge pressure is lower than a set pressure fluctuation threshold (e.g., a change of less than 0.05 bar per second) for a second time period; and the rate of change of discharge temperature is lower than a set temperature fluctuation threshold (e.g., a change of less than 0.1°C per second) for a third time period. These three dimensions respectively characterize the response stability of the power system, the integrity of the refrigerant fluid system, and the thermal inertial balance state of the thermal management system. For example, in a medium-load test, although the compressor speed had stabilized at 3500 rpm, if the exhaust temperature continued to rise at a rate of 0.5℃ / s, it indicated that the heat exchange between the heat exchanger and the refrigerant had not yet reached dynamic equilibrium. In this case, the system was determined not to have reached stability, and the process continued until the temperature curve flattened. By introducing a triple steady-state criterion of speed, pressure, and temperature, compared to traditional methods that rely solely on speed or pressure, the system's true steady-state characteristics can be reflected more comprehensively. This avoids inappropriate data acquisition timing due to thermal hysteresis or pressure oscillations, thus ensuring the reliability of the measured data.

[0073] Step 3: Once the stable conditions are reached, collect the standard operating condition data for the standard test phase, and generate the corresponding stage degradation evaluation value based on the measured standard energy efficiency value, pressure establishment response, and exhaust temperature response for the standard test phase. Then, integrate the data to obtain the parking condition degradation evaluation value.

[0074] The standard operating condition data is a set of snapshot data collected immediately after confirming that the system has reached the aforementioned three-dimensional steady state. It mainly includes compressor input power, suction pressure, discharge pressure, suction temperature, discharge temperature, electronic expansion valve opening, and fan / water pump status. The measured standard energy efficiency value is usually calculated based on the equivalent COP (coefficient of performance) obtained from the collected data, which is the ratio of equivalent cooling / heating capacity to input power. Pressure build-up response can refer to the time or pressure rise slope required for the system pressure to rise from its initial value to the target steady-state value when starting from zero or switching from the previous stage to the current stage during the test phase. This indicator directly reflects the compressor's volumetric efficiency and whether there is leakage in the system. Discharge temperature response can refer to the trajectory characteristics of the discharge temperature change over operating time or the residual with the theoretical isentropic compression temperature. This indicator sensitively reflects irreversible losses in the compression process, lubrication status, and motor efficiency. When generating the stage degradation evaluation value, not only is the deviation of the measured standard energy efficiency value from the healthy baseline value considered, but also the extended pressure build-up time and the abnormal increase in discharge temperature are weighted and incorporated. For example, if the measured COP decreases by 5% in a certain stage, but the pressure build-up time increases by 20%, the system will assign a higher weight to the pressure build-up response, generating a higher stage degradation evaluation value to warn of potential internal leakage risks. Subsequently, the stage degradation evaluation values ​​generated from low, medium, and high standard test stages are weighted and fused to obtain the final parking condition degradation evaluation value. During the fusion process, weights can be assigned according to the sensitivity of different load stages to degradation; typically, high-load stages have higher weights than low-load stages because compressor degradation is more pronounced under high-stress conditions. Through this multi-dimensional evaluation fusion mechanism, it is possible to effectively distinguish between energy efficiency fluctuations caused by changes in external heat exchange conditions and true degradation caused by a decline in the mechanical or electrical performance of the compressor itself, significantly improving the robustness and diagnostic depth of the parking self-diagnosis results.

[0075] Example 9 In another alternative embodiment, the method further includes fusing calibration and outputting energy efficiency degradation trends, levels, and maintenance windows, including the following steps: Step 1: Determine the reliability of the driving condition degradation evaluation value based on the data coverage, segment stability, and trend consistency; determine the reliability of the parking condition degradation evaluation value based on the test completion time, test stability, and phase consistency. Among them, driving-side reliability is a key indicator for quantifying the quality of driving data, which is calculated through comprehensive multi-dimensional features. Specifically, data coverage refers to the distribution range of steady-state or quasi-steady-state driving condition segments selected within the current statistical period in the equivalent operating condition space. For example, if a segment covers multiple modes such as low-temperature cooling, high-temperature heating, and battery cooling, and the compressor speed covers the entire range of low, medium, and high, then the data coverage is high. Segment stability refers to the variance of the fluctuation of compressor speed, intake and exhaust pressure, and electronic expansion valve opening within each segment. The smaller the fluctuation, the higher the stability score. Trend consistency refers to whether the direction of change of driving condition degradation evaluation values ​​is consistent over multiple consecutive statistical periods. If there are drastic jumps or non-logical deviations from the cumulative running time, the trend consistency is reduced. Parking-side reliability is the basis for evaluating the reliability of parking self-diagnosis results. Test completion time characterizes the timeliness of the data. Test results closer to the current time have a higher reliability weight. For example, a test completed within 24 hours has a reliability of 1.0, while the reliability of tests completed more than 7 days ago decreases over time. Test stability refers to the residual size after the compressor speed, pressure, and temperature reach a stable state in each stage of the parking standard operating condition test combination. The smaller the residual, the less interference the test environment experiences. Stage consistency refers to the degree of dispersion between the degradation evaluation values ​​obtained from multiple standard test stages such as low speed, medium speed, and high speed. If the results of each stage are highly consistent, the stage consistency is high. Through the quantification of the above six dimensions, the system can dynamically distinguish the quality differences between high-frequency but noisy driving data and low-frequency but high-precision parking data.

[0076] Step 2: Dynamically determine the fusion weight based on the driving-side credibility and the parking-side credibility, and reduce the weight of the side with lower credibility or trigger a re-evaluation when the deviation between the driving condition degradation evaluation value and the parking condition degradation evaluation value exceeds the consistency threshold. The determination of the fusion weights is an adaptive decision-making process. The system first calculates the normalized ratio of the driving-side confidence to the parking-side confidence, and then dynamically allocates the fusion weights λ based on this ratio.

[0077] When the parking data has high reliability (e.g., a high-quality self-diagnosis has just been completed) and the driving data coverage is insufficient, the fusion weights are tilted towards the parking condition degradation evaluation value, causing the compressor's energy efficiency degradation index to quickly converge to a high-precision benchmark. Conversely, when the latest parking data is lacking but the driving data shows high stability and consistency, the fusion weights are tilted towards the driving condition degradation evaluation value to maintain the continuity of trend prediction. Furthermore, the system has a preset consistency threshold to monitor for conflicts between the two data sources. Specifically, it calculates the absolute difference or relative deviation rate between the driving condition degradation evaluation value and the parking condition degradation evaluation value. Once this deviation exceeds the set consistency threshold (e.g., a deviation rate greater than 15%), the system determines that there is a one-sided data anomaly, which may stem from sensor drift under driving conditions, model mismatch caused by refrigerant leakage, or environmental interference during parking tests. At this point, a safety circuit breaker mechanism is implemented, automatically and significantly reducing the weight of data from the side with lower credibility, or even temporarily removing that side's data, generating a degradation index only based on the side with higher credibility; alternatively, a re-evaluation command is directly triggered, controlling the vehicle to prioritize performing the parking standard condition test the next time the conditions are met, in order to obtain new calibration data. This dynamic adjustment mechanism effectively avoids misjudgments caused by distortion from a single data source.

[0078] Step 3: Based on the fused and calibrated compressor body energy efficiency degradation index, a time series is generated, and the energy efficiency degradation trend, degradation level and maintenance window are output according to the slope, duration and inflection point characteristics of the time series.

[0079] The compressor's energy efficiency degradation index, after fusion calibration, eliminates operating disturbances and measurement noise, forming highly robust time series data. The system performs deep feature extraction on this time series to predict future states. The slope of change reflects the rate of energy efficiency degradation. By calculating the linear regression slope of the degradation index within several recent time windows, it can be identified whether the compressor is in a slow natural aging stage or an accelerated failure stage. The duration is used to determine the persistence of the degradation state, filtering out short-term, occasional fluctuations. Only when the degradation index remains above a certain level for several consecutive cycles is an increase in the degradation level confirmed. Inflection point features are used to capture abrupt changes in degradation behavior, such as the critical moment when the degradation changes from linear to exponential decline. Based on these characteristics, the system outputs specific energy efficiency degradation trends and predicts the rate of energy efficiency decline within a set mileage or time period. It classifies degradation levels, for example, defining a degradation index below 5% as Level 1 (normal), 5%-10% as Level 2 (minor degradation), 10%-20% as Level 3 (significant degradation), and exceeding 20% ​​or a sharp increase in the slope as Level 4 (rapid degradation / recommendation for maintenance). Based on this, it calculates the maintenance window, that is, the optimal time range or mileage point for recommending users to carry out maintenance or replacement before the degradation index reaches an unacceptable threshold.

[0080] Figure 2This application provides a schematic diagram of the structure of a new energy vehicle compressor energy efficiency degradation trend prediction system according to an embodiment of the present application, as shown below. Figure 2 As shown, the energy efficiency degradation trend prediction system 300 for a new energy vehicle compressor in this embodiment includes: a data acquisition module 301, an operating condition analysis module 302, a benchmark analysis module 303, a degradation analysis module 304, a parking operating condition analysis module 305, a parking degradation analysis module 306, a comprehensive analysis module 307, and a degradation prediction module 308.

[0081] The data acquisition module 301 is used to collect operating data during the operation of the compressor of the new energy vehicle. The operating data includes compressor operating parameters, refrigerant side status parameters, thermal management execution component parameters, and vehicle thermal load parameters. The operating condition analysis module 302 is used to divide the operating data during vehicle driving into segments, and based on compressor speed fluctuations, intake and exhaust pressure fluctuations, electronic expansion valve opening changes, and thermal management mode switching status, to filter out steady-state or quasi-steady-state driving condition segments. The benchmark analysis module 303 is used to generate equivalent operating condition features based on the driving condition segment, and call the benchmark energy efficiency surface under the compressor health state to obtain the benchmark equivalent energy efficiency value corresponding to the driving condition segment. The degradation analysis module 304 is used to calculate the measured equivalent energy efficiency value based on the compressor input power and heat output characterization value in the driving condition segment, compare the measured equivalent energy efficiency value with the benchmark equivalent energy efficiency value, and generate a driving condition degradation evaluation value. The parking condition analysis module 305 is used to control the compressor and thermal management execution components to enter the parking standard condition test combination when the vehicle meets the parking self-diagnosis trigger conditions, and to collect parking standard condition data after the parking standard condition test combination reaches a stable state. The parking degradation analysis module 306 is used to calculate the measured standard energy efficiency value based on the parking standard operating condition data, and compare the measured standard energy efficiency value with the health standard energy efficiency value of the corresponding parking standard operating condition to generate a parking condition degradation evaluation value. The comprehensive analysis module 307 is used to perform fusion calibration based on the driving condition degradation evaluation value and the parking condition degradation evaluation value to generate the compressor body energy efficiency degradation index; The degradation prediction module 308 is used to predict the energy efficiency degradation trend of the compressor based on the time series of the compressor body energy efficiency degradation index and the cumulative operating characteristics of the compressor, and output the energy efficiency degradation prediction result.

Claims

1. A method for predicting the energy efficiency degradation trend of a compressor in a new energy vehicle, characterized in that, include: S1. Collect operating data during the operation of the compressor in the new energy vehicle. The operating data includes compressor operating parameters, refrigerant side status parameters, thermal management execution component parameters, and vehicle thermal load parameters. S2. Divide the operating data during vehicle operation into segments, and based on compressor speed fluctuations, intake and exhaust pressure fluctuations, electronic expansion valve opening changes, and thermal management mode switching status, filter out steady-state or quasi-steady-state driving condition segments. S3. Generate equivalent operating condition features based on the driving condition segment, and call the reference energy efficiency surface under the compressor's healthy state to obtain the reference equivalent energy efficiency value corresponding to the driving condition segment. S4. Calculate the measured equivalent energy efficiency value based on the compressor input power and heat output characterization value in the driving condition segment, compare the measured equivalent energy efficiency value with the benchmark equivalent energy efficiency value, and generate a driving condition degradation evaluation value. S5. When the vehicle meets the parking self-diagnosis triggering conditions, control the compressor and thermal management execution components to enter the parking standard working condition test combination, and collect parking standard working condition data after the parking standard working condition test combination reaches a stable state. S6. Calculate the measured standard energy efficiency value based on the parking standard operating condition data, and compare the measured standard energy efficiency value with the health standard energy efficiency value of the corresponding parking standard operating condition to generate a parking condition degradation evaluation value. S7. Based on the driving condition degradation evaluation value and the parking condition degradation evaluation value, perform fusion calibration to generate the compressor body energy efficiency degradation index; S8. Based on the time series of the compressor body energy efficiency degradation index and the cumulative operating characteristics of the compressor, predict the energy efficiency degradation trend of the compressor and output the energy efficiency degradation prediction result.

2. The method according to claim 1, characterized in that, Before performing the segment division, the collected operating data is synchronized in time and abnormal data is removed. The continuous operating data is divided into several segments to be evaluated, using the compressor start / stop status, thermal management mode status, and data continuity as segmentation boundaries. The segments to be evaluated that exhibit sensor jumps, communication interruptions, or sudden changes in compressor control commands are subject to weight reduction or removal.

3. The method according to claim 1, characterized in that, The steady-state or quasi-steady-state driving condition segment is determined by a steady-state score, which is calculated based on the compressor speed change, intake and exhaust pressure change, electronic expansion valve opening change, and thermal management mode duration. When the steady-state score meets the set scoring conditions, the corresponding segment is determined as a steady-state driving condition segment. When the steady-state score does not meet the set scoring conditions but the change trend of each state quantity in the segment meets the monotonic or slowly changing conditions, the corresponding segment is determined as a quasi-steady-state driving condition segment, and the quasi-steady-state driving condition segment is dynamically corrected.

4. The method according to claim 1, characterized in that, The equivalent operating condition characteristics are generated in the following manner: First, the driving condition segments are divided into cabin cooling, heat pump heating, battery cooling, or combined thermal management conditions according to the thermal management mode. Then, equivalent operating points under the corresponding conditions are generated according to the compressor load status, refrigerant pressure status, and heat exchange side load status, so that the compressor operating segments under different ambient temperatures, different vehicle speeds, and different heat loads are mapped to a comparable equivalent operating condition space.

5. The method according to claim 1, characterized in that, The benchmark energy efficiency surface is constructed using health samples, which are derived from one or more of the following: compressor bench health data, vehicle initial operating data after factory delivery, initial operating data after compressor replacement, and health data of a group of vehicles of the same model. When constructing the benchmark energy efficiency surface, the healthy samples are divided into zones and clustered according to the characteristics of equivalent operating conditions, and the health benchmark energy efficiency value corresponding to each equivalent operating condition zone is determined, so that the benchmark energy efficiency surface can output a benchmark equivalent energy efficiency value that matches the driving condition segment to be evaluated.

6. The method according to claim 1, characterized in that, The driving condition degradation evaluation value is generated in the following manner: The measured equivalent energy efficiency value of each driving condition segment is compared with the corresponding benchmark equivalent energy efficiency value to obtain the energy efficiency deviation of that segment. The segment weight is determined based on the segment's steady-state score, equivalent operating condition coverage, and degree of matching with the benchmark energy efficiency surface. The energy efficiency deviations of multiple segments are weighted and fused, and the reliability of the degradation evaluation of the corresponding statistical period is improved when the energy efficiency deviations change continuously in the same direction.

7. The method according to claim 1, characterized in that, The parking self-diagnostic triggering conditions include periodic triggering conditions and abnormal triggering conditions; The periodic triggering conditions are determined based on the time interval from the last parking self-diagnosis or the cumulative running time of the compressor, and the abnormal triggering conditions are determined based on the abnormal increase of the driving condition degradation evaluation value, the excessive fluctuation of the driving condition degradation evaluation value, or the insufficient number of driving condition segments. When the periodic triggering condition or abnormal triggering condition is met, and the vehicle is in a parking state that allows the compressor to run for a short period of time, parking self-diagnosis is triggered.

8. The method according to claim 1, characterized in that, The parking standard operating condition test combination executes multiple standard test stages sequentially from low load to high load; In each standard test phase, the compressor and thermal management actuators are first controlled to enter the corresponding standard control state, and then it is determined whether the compressor speed, suction and discharge pressure and discharge temperature have reached the stable conditions. Once stable conditions are reached, standard operating condition data from the standard test phase are collected, and corresponding phase degradation evaluation values ​​are generated based on the measured standard energy efficiency values, pressure establishment response, and exhaust temperature response from the standard test phase. These are then integrated to obtain the parking condition degradation evaluation value.

9. The method according to claim 1, characterized in that, The fusion calibration is performed in the following manner: The reliability of the driving condition degradation evaluation value is determined based on the data coverage, segment stability, and trend consistency. The reliability of the parking condition degradation evaluation value is determined based on the test completion time, test stability, and phase consistency. The fusion weight is dynamically determined based on the driving-side credibility and the parking-side credibility. When the deviation between the driving condition degradation evaluation value and the parking condition degradation evaluation value exceeds the consistency threshold, the weight of the side with lower credibility is reduced or a re-evaluation is triggered. A time series is generated based on the compressor body energy efficiency degradation index after fusion calibration, and the energy efficiency degradation trend, degradation level and maintenance window are output according to the change slope, duration and inflection point characteristics of the time series.

10. A system for predicting the energy efficiency degradation trend of a new energy vehicle compressor, characterized in that, The method applied to any one of claims 1-9 includes: The data acquisition module is used to collect operating data during the operation of the compressor in new energy vehicles. The operating data includes compressor operating parameters, refrigerant side status parameters, thermal management execution component parameters, and vehicle thermal load parameters. The operating condition analysis module is used to divide the operating data during vehicle driving into segments, and based on compressor speed fluctuations, intake and exhaust pressure fluctuations, electronic expansion valve opening changes, and thermal management mode switching status, to filter out steady-state or quasi-steady-state driving condition segments. The benchmark analysis module is used to generate equivalent operating condition features based on the driving condition segment, and call the benchmark energy efficiency surface under the compressor's healthy state to obtain the benchmark equivalent energy efficiency value corresponding to the driving condition segment. The degradation analysis module is used to calculate the measured equivalent energy efficiency value based on the compressor input power and heat output characterization value in the driving condition segment, compare the measured equivalent energy efficiency value with the benchmark equivalent energy efficiency value, and generate a driving condition degradation evaluation value. The parking condition analysis module is used to control the compressor and thermal management execution components to enter the parking standard condition test combination when the vehicle meets the parking self-diagnosis trigger conditions, and to collect parking standard condition data after the parking standard condition test combination reaches a stable state. The parking degradation analysis module is used to calculate the measured standard energy efficiency value based on the parking standard operating condition data, and compare the measured standard energy efficiency value with the health standard energy efficiency value of the corresponding parking standard operating condition to generate a parking condition degradation evaluation value. The comprehensive analysis module is used to perform fusion calibration based on the driving condition degradation evaluation value and the parking condition degradation evaluation value to generate the compressor body energy efficiency degradation index; The degradation prediction module is used to predict the energy efficiency degradation trend of the compressor based on the time series of the compressor body energy efficiency degradation index and the cumulative operating characteristics of the compressor, and output the energy efficiency degradation prediction result.

Images