Calibration Method and System for Heat Pump Air Conditioning Systems in Commercial Vehicles under Multiple Operating Conditions

Abstract

This invention discloses a calibration method and system for a multi-condition commercial vehicle heat pump air conditioning system. The method includes the following steps: Step 1: Building a characteristic database of operating conditions to obtain a typical vehicle operating condition database; Step 2: Analyzing the influence weights of various parameters of the heat pump air conditioning system on the system's coefficient of performance (COP) and outlet air temperature; Step 3: Operating condition parameter calibration: calibrating the parameters of the heat pump air conditioning system under each operating condition mode to obtain the optimal parameter combination, with the goal of meeting the temperature setting and maximizing the COP. This solution achieves adaptive matching of system parameters through multi-condition hierarchical calibration and dynamic compensation algorithms, combined with iterative optimization using cloud data.

Description

Technical Field

[0001] This invention relates to the field of automotive air conditioning, and in particular to a calibration method and system for a multi-condition commercial vehicle heat pump air conditioning system. Background Technology

[0002] Commercial vehicle heat pump air conditioning systems require more complex calibration processes than passenger vehicle systems due to the need to meet the demands of large spaces for cooling / heating, complex operating conditions (such as high and low temperatures and humidity variations), and high energy efficiency. Traditional calibration methods rely on adjusting parameters under a single operating condition, which presents the following problems:

[0003] 1. It is difficult to balance energy efficiency and comfort: Commercial vehicles have large differences in load and road conditions, and traditional calibration is difficult to dynamically adapt to different load requirements.

[0004] 2. Low heating efficiency at low temperatures: Existing heat pump systems have a drastic drop in heating capacity in environments below -10℃, requiring reliance on electric auxiliary heating, which leads to increased energy consumption.

[0005] 3. Long calibration cycle: It requires a lot of actual vehicle tests to adjust parameters (such as compressor speed, electronic expansion valve opening, heat exchanger air volume), resulting in low efficiency.

[0006] Therefore, an intelligent, multi-objective collaborative calibration method is needed to improve the overall performance of commercial vehicle heat pump air conditioning systems. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a calibration method and system for a multi-condition commercial vehicle heat pump air conditioning system. Through multi-condition hierarchical calibration and dynamic compensation algorithm, combined with cloud data iterative optimization, adaptive matching of system parameters is achieved.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] The calibration method for commercial vehicle heat pump air conditioning systems under multiple operating conditions includes the following steps:

[0010] Step 1: Build a working condition feature library to obtain a typical automotive working condition library;

[0011] Step 2: Analyze the weighting of the influence of each parameter of the heat pump air conditioning system on the system energy efficiency ratio (COP) and outlet air temperature;

[0012] Step 3: Operating Parameter Calibration: With the goal of meeting the temperature setting and maximizing the COP, the parameters of the heat pump air conditioning system under each operating mode are calibrated to obtain the optimal parameter combination.

[0013] The method further includes step 4: the cloud platform performs multi-vehicle data optimization on the optimal parameter combination calibrated in step 3 to generate the final global optimization parameter set.

[0014] In step 1, a typical operating condition library for commercial vehicles is established based on historical data. The classic operating condition library includes various operating conditions for commercial vehicles and their corresponding parameters.

[0015] In step 2, the influence weights of the heat pump air conditioning system parameters on the system energy efficiency ratio (COP) and outlet air temperature are determined by orthogonal experimental design.

[0016] Step 3 includes a basic calibration layer and a dynamic compensation layer. Based on the calibration layer, the parameters of the heat pump air conditioning system under various operating conditions are optimized statically. The dynamic compensation layer compensates and adjusts the optimized static parameters in real time to obtain the calibrated parameters for each operating condition.

[0017] The basic calibration layer solves for the optimal parameter combination under various operating conditions based on the software thermal management control model, maximizing COP and satisfying the temperature setpoint.

[0018] The dynamic compensation layer collects data in real time through on-board sensors and uses fuzzy PID control to dynamically correct parameters and compensate for sudden environmental changes or load fluctuations.

[0019] The calibrated operating condition data are uploaded to the cloud platform, which then distributes them to the test vehicles for operation and analysis to generate a global optimization parameter set as the final global optimization parameters.

[0020] The cloud platform assigns weights to multi-vehicle data and analyzes the collected multi-vehicle data based on these weights to generate the final global optimization parameters.

[0021] The calibration system for commercial vehicle heat pump air conditioning systems under multiple operating conditions includes an operating condition feature library construction module, a parameter sensitivity analysis module, and a hierarchical calibration model. The operating condition feature library construction module is used to build an operating condition feature library to obtain a typical vehicle operating condition library. The parameter sensitivity analysis module is used to analyze the influence weight of each parameter of the heat pump air conditioning system on the system energy efficiency ratio (COP) and outlet air temperature. The operating condition parameter calibration module calibrates the parameters of the heat pump air conditioning system under each operating condition mode with the goal of satisfying the temperature setting and maximizing the energy efficiency ratio (COP) to obtain the optimal parameter combination.

[0022] The advantages of this invention are: 1. By using a virtual calibration environment based on a working condition database, the number of real vehicle tests is reduced, and the calibration cycle is shortened to 1 / 3 of the traditional method;

[0023] 2. Applicable to various commercial vehicle types such as logistics vehicles and intercity buses, with high compatibility.

[0024] 3. Multi-objective collaborative optimization: Simultaneously optimize energy efficiency (COP), comfort (temperature fluctuation ≤ ±1℃) and system reliability (compressor start-stop count).

[0025] 4. Enhanced low-temperature heating: By introducing waste heat recovery and dynamic decoupling control of electronic expansion valve, the heating capacity in the -20℃ environment is increased by more than 30%. Attached Figure Description

[0026] The following is a brief explanation of the contents of each of the accompanying drawings and the markings in the drawings:

[0027] Figure 1 This is a schematic diagram of the heat pump air conditioning system of the present invention.

[0028] Figure 2 This is a flowchart of the calibration method of the present invention. Detailed Implementation

[0029] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and the description of the preferred embodiments.

[0030] The calibration method for a multi-condition commercial vehicle heat pump air conditioning system in this embodiment includes the following steps:

[0031] Step 1: Build a working condition feature library to obtain a typical automotive working condition library;

[0032] Step 2: Analyze the weighting of the influence of each parameter of the heat pump air conditioning system on the system energy efficiency ratio (COP) and outlet air temperature;

[0033] Step 3: Operating Parameter Calibration: With the goal of meeting the temperature setting and maximizing the COP, the parameters of the heat pump air conditioning system under each operating mode are calibrated to obtain the optimal parameter combination.

[0034] Step 4: The cloud platform performs multi-vehicle data optimization on the optimal parameter combination calibrated in Step 3 to generate the final global optimization parameter set.

[0035] In step 1, a typical operating condition library for commercial vehicles is established based on historical data. This classic operating condition library includes various commercial vehicle operating conditions and their corresponding parameters. The typical operating condition library includes parameters such as ambient temperature, humidity, vehicle speed, and battery thermal load (e.g., high-temperature cooling, low-temperature heating, high-humidity defrosting). The core of step 1 is to establish a reusable typical operating condition library for commercial vehicles based on massive historical data. Its technical principle is data-driven, integrating vehicle engineering, heat transfer, and machine learning to condense real road and environmental information into limited but highly representative classic operating conditions, providing standardized input boundaries for subsequent system simulation, control strategy calibration, and component matching. The construction of the operating condition library relies primarily on large-scale, high-quality historical operating data, mainly from: real-vehicle driving data remotely uploaded via the vehicle-mounted T-Box, bench test data precisely controlled within the environmental test chamber, and whole-vehicle road test records. Data fields must cover ambient temperature, relative humidity, vehicle speed, voltage and temperature of each battery cell, inlet and outlet temperatures and flow rates of the battery pack coolant, passenger compartment temperature, air conditioning compressor and PTC power, and water pump and fan duty cycles. To ensure time synchronization, interpolation and resampling will be performed at fixed intervals of 1 to 10 seconds. Median filtering and the Laida criterion will be used to remove outliers for communication packet loss and sensor jumps, ultimately forming a smooth and continuous multidimensional time series. The continuous historical data stream needs to be segmented into independent operating condition segments with complete thermodynamic meaning. The boundary division is based on the transition points of vehicle operating states, such as the transition from "driving cooling" to "parking fast charging cooling" and from "low-temperature cold start" to "stable heating". The start and end of the segments are automatically identified by a change-point detection algorithm using step changes in vehicle speed, high-voltage power-on signal, and air conditioning power. Each segment typically contains a complete thermal management transient process and a subsequent quasi-steady-state phase, ranging in length from several minutes to tens of minutes, ensuring that the segment can fully reflect the dynamic response characteristics of the system.

[0036] Multidimensional feature parameter extraction involves extracting a set of multidimensional parameters characterizing the thermodynamic and driving characteristics of each operating condition segment. These parameters directly form the basis for subsequent clustering and operating condition definition. Core parameters include:

[0037] (1) Average, extreme, and rate of change of ambient temperature;

[0038] (2) The mean and peak values ​​of ambient relative humidity, especially in defrosting conditions, high humidity is the core cause of fogging;

[0039] (3) Vehicle speed distribution characteristics, such as average vehicle speed, maximum vehicle speed, idling time percentage, time proportion of different speed ranges, and vehicle speed standard deviation;

[0040] (4) Battery thermal load is quantitatively described by the battery's heat generation power and heat dissipation requirements. The real-time heat generation can be calculated based on the battery current, open-circuit voltage, and internal resistance, combined with the Bernardi heat generation model. At the same time, the heat dissipation power can be estimated based on the battery temperature, coolant flow rate, and temperature difference. In historical data, this load is often expressed as the heat flow (kW) of the battery pack assembly or a weighted value of the equivalent charge-discharge rate to distinguish between the strong cooling demand at high temperatures and the heat input required at low temperatures.

[0041] After feature extraction, unsupervised machine learning algorithms are used to perform cluster analysis on massive amounts of data. Due to the high dimensionality and varying dimensions of the features, Z-score standardization is first performed, followed by principal component analysis to reduce dimensionality and retain over 95% of the variance. The commonly used K-means algorithm uses silhouette coefficients and the Calinski-Harabasz exponent to determine the optimal number of clusters, ensuring that each cluster corresponds to a typical thermal management condition with clear physical meaning. Domain experts then interpret the high-dimensional feature vectors at the cluster centers and, combined with the time series of representative segments, ultimately define classic operating conditions covering all climates and scenarios. For example, the "low-temperature heating condition" involves an ambient temperature ≤ -10 ℃, medium to high vehicle speeds, external heating of the battery to restore power, and a passenger compartment heat load of 8~10 kW, corresponding to a high-load heating scenario for heat pumps or PTC systems. For each typical cluster, dynamic time warping (DTW) and centroid averaging techniques are used to select complete segments or synthesize the most representative vehicle speed, environmental parameters, and load curves within the cluster, forming formal entries in the classic operating condition library. In addition to the time series, each operating condition entry also stores a feature parameter vector, operating condition label, and applicable vehicle type (such as pure electric light truck, fuel cell heavy truck, cold chain transport vehicle, etc.).

[0042] To ensure that the classic operating condition library can accurately replace massive amounts of real-world test data, the operating conditions within the library need to be replayed and verified through vehicle thermal management system simulation. Typical operating conditions are input as boundary conditions into a one-dimensional or three-dimensional thermal management simulation model. The consistency of key outputs such as coolant temperature, passenger compartment temperature, and compressor power with the original historical data is compared, requiring temperature deviations within ±2℃ and energy consumption deviations not exceeding 5%. For uncovered extreme operating conditions or special scenarios, the library capacity can be expanded by injecting artificially synthesized parameterized short operating conditions, enabling dynamic iteration of the operating condition library. Through the above technical approach, the typical operating condition library for commercial vehicles established in step 1 not only condenses millions of kilometers of complex driving data into a few representative standard operating conditions, but also completely preserves the dynamic coupling relationships of key parameters such as ambient temperature, humidity, vehicle speed, and battery thermal load. This provides a unified and traceable benchmark for subsequent multi-objective thermal management system optimization, hardware-in-the-loop testing of control strategies, and vehicle energy efficiency calibration, significantly improving development efficiency and reducing real-vehicle testing costs.

[0043] Step 2 uses orthogonal experimental design to determine the influence weights of heat pump air conditioning system parameters on the system's coefficient of performance (COP) and outlet air temperature. Based on a typical operating condition database, Step 2 quantifies the influence weights of key heat pump air conditioning parameters on the system's COP and outlet air temperature using orthogonal experimental design. This aims to quickly identify the controlling factors from multiple adjustable parameters, providing a sensitivity ranking basis for subsequent multi-objective optimization. Representative boundaries, such as high-temperature cooling and low-temperature heating, are selected from the operating condition database established in Step 1, and external conditions such as ambient temperature and humidity, vehicle speed, and battery heat load are fixed. Five independent parameters are selected: compressor speed, electronic expansion valve opening, outdoor fan airflow, indoor fan airflow, and superheat setpoint. Three levels of each parameter are selected to cover commonly used calibration ranges. An L18 orthogonal array can be used to arrange the experiments, requiring only 18 simulations or bench tests to achieve balanced distribution of factor levels. Experiments are performed one by one on the heat pump system simulation model or enthalpy difference bench. After the system reaches quasi-steady state, compressor power, heating capacity, and outlet air temperature are collected, and the COP is calculated. Range analysis was performed on the results: the mean response of each factor at different levels was calculated, and the range R was the difference between the maximum and minimum mean of the factor, the magnitude of which directly reflected the magnitude of the influence. Compressor speed typically has the highest weight on COP, as it dominates refrigerant flow and power consumption; the electronic expansion valve opening has the greatest impact on outlet air temperature, directly controlling the evaporation temperature. Compressor speed determines pressure ratio and power; outdoor fan airflow affects heat exchange and low-temperature frosting characteristics; indoor airflow is related to condensing temperature and air supply comfort; and superheat mainly affects system stability. To make the weights more universal, orthogonal experiments were repeated under multiple typical operating conditions such as high temperature, low temperature, and high humidity defrosting, and the weighted average was calculated according to the proportion of actual driving time to obtain the comprehensive weight. High-weight parameters were listed as optimization variables and needed to be adaptively adjusted according to operating conditions; low-weight parameters could use fixed values ​​or simple curves to effectively reduce control complexity. This method systematically quantifies the primary and secondary influences of parameters, directly supporting subsequent response surface or genetic algorithm optimization, significantly improving the matching efficiency and energy consumption performance of heat pump air conditioners in complex scenarios in commercial vehicles.

[0044] Step 3 includes a basic calibration layer and a dynamic compensation layer. The calibration layer performs static parameter optimization on the heat pump air conditioning system parameters under various operating conditions. The dynamic compensation layer compensates and adjusts the optimized static parameters in real time to obtain the calibrated parameters for each operating condition. The basic calibration layer solves for the optimal parameter combination under each operating condition based on a software thermal management control model, maximizing COP and meeting the temperature setpoint. The dynamic compensation layer collects data in real time through onboard sensors and uses fuzzy PID control to dynamically correct parameters, compensating for sudden environmental changes or load fluctuations. The basic calibration layer relies on a high-fidelity thermal management system simulation model (such as a one-dimensional heat pump cycle model built using Modelica or GT-SUITE) to solve for the optimal control parameter vector for each typical operating condition defined in Step 1 (high-temperature cooling, low-temperature heating, high-humidity defrosting, etc.), aiming to maximize COP and meet the outlet air temperature setpoint. Optimization variables focus on high-weight parameters identified in Step 2, including compressor speed, electronic expansion valve opening, outdoor fan airflow, and indoor fan airflow. Low-weight parameters such as superheat setting are fixed empirically to reduce the optimization dimensionality. The constraints include: the deviation of the outlet air temperature from the target value does not exceed ±1.5 ℃, the compressor exhaust temperature is below the limit, the condensing and evaporating pressures are within the safe range, and the temperature constraint of the passenger compartment or battery coolant outlet. A multi-objective optimization algorithm is used to search the Pareto front within the design space. For each typical operating condition, a compromise optimal solution is selected from the Pareto solution set: prioritizing temperature compliance while maximizing COP. After optimization, operating condition characteristic parameters such as ambient temperature, vehicle speed, and battery thermal load are used as indices, and optimal compressor speed, expansion valve opening, and fan duty cycle are used to form a multi-dimensional calibration MAP. This MAP serves as the basic lookup table value for feedforward control, ensuring that the system operates at its efficient point under stable conditions. For atypical operating conditions, initial parameters are obtained through high-dimensional interpolation of adjacent operating condition points to achieve full coverage. The static calibration MAP cannot cope with real-time disturbances such as sudden changes in solar radiation, sudden increases in humidity due to rain showers, and drastic fluctuations in vehicle speed and load; therefore, a dynamic compensation layer is introduced. This layer receives real-time data from multiple onboard sensors, including ambient temperature and humidity, solar radiation intensity, vehicle speed, outlet air temperature, passenger compartment temperature, evaporator outlet superheat, and compressor high and low pressure. The core controller employs a two-dimensional fuzzy PID structure, using the outlet air temperature deviation e (target value - actual value) and the deviation change rate ec as inputs, and outputting the compensation increment Δu of the control parameters, which is directly superimposed on the feedforward value provided by the basic calibration layer. The fuzzy PID self-tunes the proportional, integral, and derivative gains of the PID controller online, or directly outputs the compensation amount. Here, the output compensation method is used, with fuzzy compensators designed separately for the two high-sensitivity actuators: compressor speed and electronic expansion valve opening. Defuzzification uses the centroid method to obtain the accurate compensation amount. To prevent actuator saturation, the compensation amount needs to be limited, and integral separation logic is added: when the deviation is too large, the integral action is canceled to avoid overshoot.Furthermore, to address the potential dangers of the system entering high-pressure protection or excessively low-pressure zones during dynamic processes, the fuzzy compensator incorporates protective rules. For example, when high pressure approaches the limit, the compressor speed is preferentially reduced; this rule has the highest priority to ensure safety. The dual-layer architecture places the time-consuming multi-objective optimization work offline, leaving only lightweight fuzzy inference to run online. This ensures both real-time computational efficiency and achieves a balance between optimal energy efficiency under steady-state conditions and accurate temperature tracking under transient conditions. Compared to single feedback control, feedforward MAP significantly shortens the settling time and avoids large oscillations caused by poor initial PID parameters. Compared to pure feedforward calibration, fuzzy PID provides the system with strong robustness to environmental changes. Thus, the adaptive parameter calibration method formed in step 3 ensures that the heat pump air conditioning system consistently balances high energy efficiency and high comfort throughout its entire lifecycle and in complex and ever-changing usage scenarios. It balances the optimal energy efficiency of the heat pump air conditioning system under steady-state conditions with temperature robustness under transient disturbances. The basic calibration layer provides the optimal combination of static parameters for each typical working condition using a feedforward method, while the dynamic compensation layer uses real-time sensor feedback to correct the feedforward parameters online, achieving full-condition self-adaptation.

[0045] Step 4: Upload the calibrated operating condition data to the cloud platform. The cloud platform then distributes the data to the test vehicles for operation and collects and analyzes multi-vehicle data to generate a global optimization parameter set as the final global optimization parameters. Step 4 establishes a closed-loop iterative mechanism of single-vehicle calibration—cloud platform aggregation—multi-vehicle global optimization, elevating the calibration parameters from single-vehicle static optimality to global optimality driven by big data from multiple vehicles. This step relies on a vehicle-cloud integrated architecture and is completed collaboratively by the vehicle terminal, cloud data platform, and OTA (Over-The-Air) service.

[0046] First, the typical operating condition control parameter sets (including MAP parameters such as compressor speed, expansion valve opening, and fan duty cycle, as well as the membership function parameters and rule weights of fuzzy PID controllers) generated in step 3 under the combined action of the basic calibration layer and the dynamic compensation layer are structurally encapsulated and uploaded to the cloud parameter library via an encrypted tunnel using Protobuf or JSON format through an onboard T-Box. The cloud platform establishes a parameter version management tree, associating operating condition type, applicable vehicle model, software version, and calibration date, and provides traceable differential update capabilities. After the platform performs compliance verification and boundary checks on the uploaded parameters, it marks them as versions to be released.

[0047] Subsequently, the cloud-based OTA engine distributed the parameter set to multiple test vehicles deployed in different climate zones and operating scenarios. The distribution strategy employed a canary release mechanism: first, a small number of vehicles were selected for silent updates and operational monitoring, comparing key indicators such as the average COP (Coefficient of Performance), number of outlet air temperature overshoots, and compressor start-stop frequency within a preset window; if no anomalies were found, the distribution was then gradually expanded to the entire fleet. After updating the parameters, the vehicles reset the corresponding control variables and used checksums to ensure the consistency between the firmware and calibration data.

[0048] During actual road driving, the test vehicle continuously recorded a complete time series of data, including ambient temperature, humidity, solar radiation intensity, vehicle speed, battery thermal load, compressor power, high and low pressure, outlet air temperature, and passenger compartment and battery temperatures, using a high-frequency acquisition channel. After time alignment and outlier filtering at the edge, the data was transmitted in real-time to the cloud data lake via the vehicle-to-everything (V2X) network, and stored in partitions according to vehicle VIN and operating condition labels. The cloud utilized a streaming processing framework to aggregate the data in real-time, constructing a multi-dimensional performance cube with "vehicle-operating condition-time" dimensions. This cube continuously calculated key performance indicators such as actual COP, root mean square error of temperature control, and compressor operating margin for each operating condition segment.

[0049] Building upon this foundation, a global parameter optimization engine runs in the cloud. First, it identifies performance deviations and parameter sensitivities across different vehicles and geographical environments. It employs federated averaging or a distributed optimization strategy based on swarm intelligence: each vehicle can calculate gradients or performance feedback locally, uploading only anonymized feature statistics (such as the local gradient of the loss function). These are then aggregated in the cloud and updated with global parameters through Bayesian optimization or an adaptive evolution strategy based on the covariance matrix. The optimization objective is set as a multi-objective function, maximizing the weighted average COP of all vehicles under comprehensive operating conditions while reducing parameter sensitivity to manufacturing tolerances and aging variances, provided that the outlet air temperature deviation is ≤±1.5 ℃ and system protection constraints are met. Through adversarial verification using large-scale real-vehicle data, it can identify extreme combinations or transient paths not covered in the original calibration MAP, automatically generating compensation corrections and integrating them into the MAP surface.

[0050] After multiple rounds of iterative convergence, the final global optimization parameter set is generated in the cloud. This parameter set not only includes the updated multi-dimensional control MAP and fuzzy rule base, but also embeds the confidence threshold for operating condition identification and fault degradation strategies. After refeedback verification in a simulated digital twin environment, it is again fixed to all applicable vehicles via OTA with one click, and the version is locked as the mass production baseline. Thus, step 4 transforms isolated bench or single-vehicle calibration into a collective intelligent evolution process, enabling the heat pump air conditioning parameters to have continuous self-optimization capabilities, achieving high energy efficiency and robust adaptation to complex and ever-changing usage scenarios throughout the entire life cycle.

[0051] In step S4, the cloud platform assigns weights to the multi-vehicle data and analyzes the collected data based on these weights to generate the final global optimization parameters. In the vehicle-cloud closed loop constructed in step 4, the cloud platform aggregates massive amounts of operational data from multiple test vehicles under different regions, climates, and usage scenarios. If all vehicle data is treated equally, it is difficult to eliminate deviations introduced by individual vehicles due to manufacturing tolerances, sensor offsets, or atypical driving behaviors. The generated global optimization parameters are prone to getting stuck in local optima or lacking generalization ability. Therefore, it is necessary to scientifically assign weights to the multi-vehicle data and then perform weighted analysis accordingly to ensure that the global parameters truly reflect the overall and robust optimal characteristics of the fleet.

[0052] The weighting is based on multi-dimensional vehicle profiles and data quality assessment. First, there's the operating condition coverage weight: the cumulative runtime of each vehicle under various typical operating conditions (high-temperature cooling, low-temperature heating, high-humidity defrosting, etc.) is calculated, assigning higher weights to vehicles with broader coverage and closer to the operating condition boundaries, ensuring the universality of global parameters across all operating conditions. Second, there's the data quality weight: based on data integrity, signal-to-noise ratio, and sensor drift, a sliding window anomaly detection method is used to calculate data reliability indicators; vehicles with higher data quality have higher weights. Third, there's the energy efficiency benchmark weight: a vehicle thermal management system health factor is introduced, such as the degree of degradation in the ratio of long-term compressor power consumption to cooling capacity, prioritizing data from new vehicles or well-maintained vehicles with system conditions close to the design baseline, suppressing interference from aging or faulty vehicles. Fourth, there's the scenario scarcity weight: a higher weight is assigned to a small number of vehicles that provide data on rare but critical scenarios (such as extremely cold starts or heavy rain and high humidity), preventing these edge-condition data from being overwhelmed. Finally, a comprehensive weight coefficient is generated for each vehicle, which is normalized to form a weight vector.

[0053] When generating global optimization parameters, the weight vector is directly embedded in the optimization objective. If a federated averaging architecture is used, when aggregating local gradients or updating parameters in the cloud, a weighted average is performed using vehicle weights as coefficients, replacing the simple arithmetic average. If centralized Bayesian optimization is used, the weighted global objective function (e.g., maximizing the weighted average COP of all vehicles, while simultaneously summing the weighted overshoot penalty for outlet air temperature) is used as the optimization objective, iteratively searching the MAP surface. Weights can also be introduced into the constraints, prioritizing ensuring that the outlet air temperature deviation of high-weight vehicles is ≤±1.5 ℃, while allowing slightly looser boundaries for low-weight vehicles, but not tolerating protective overshoots. After multiple rounds of weighted iteration convergence, the resulting parameter set performs excellently under typical, high-quality operating conditions represented by high-weight data, while retaining sufficient robustness to edge scenarios. The final generated global optimization parameters are essentially the solution for minimizing weighted empirical risk in the multi-vehicle data space, thus effectively suppressing the negative impacts of individual differences and abnormal scenarios while ensuring the optimal average energy efficiency of the entire fleet.

[0054] This embodiment also provides a calibration system for a commercial vehicle heat pump air conditioning system based on multiple operating conditions, including an operating condition feature library construction module, a parameter sensitivity analysis module, and a hierarchical calibration model; wherein the operating condition feature library construction module is used to build an operating condition feature library to obtain a typical vehicle operating condition library; the parameter sensitivity analysis module is used to analyze the influence weight of each parameter of the heat pump air conditioning system on the system energy efficiency ratio (COP) and outlet air temperature; the operating condition parameter calibration module calibrates the parameters of the heat pump air conditioning system under each operating condition mode with the goal of satisfying the temperature setting and maximizing the energy efficiency ratio (COP) to obtain the optimal parameter combination.

[0055] like Figure 2 As shown in the figure, this embodiment presents a multi-condition graded calibration and dynamic compensation algorithm, which combines cloud data iterative optimization to achieve adaptive matching of system parameters.

[0056] The steps include:

[0057] Step 1: Construction of the Operating Condition Feature Library

[0058] A database of typical operating conditions for commercial vehicles has been established based on historical data (such as high-temperature cooling, low-temperature heating, and high-humidity defrosting), covering parameters such as ambient temperature, humidity, vehicle speed, and battery heating and cooling load.

[0059] Step 2: Parameter Sensitivity Analysis

[0060] The influence weights of key parameters (compressor speed, electronic expansion valve opening, and fan speed) on the system COP (coefficient of performance) and outlet air temperature were determined by orthogonal experimental design.

[0061] Step 3: Model calibration

[0062] The calibration process is divided into a basic calibration layer (static parameter optimization) and a dynamic compensation layer (real-time operating condition adaptation):

[0063] Basic calibration layer: Based on the software thermal management control model, the optimal parameter combination under each operating condition is solved to maximize COP and meet the temperature setpoint.

[0064] Dynamic compensation layer: Data is collected in real time by on-board sensors, and fuzzy PID control is used to dynamically correct parameters to compensate for sudden environmental changes or load fluctuations.

[0065] Step 4: Database Co-optimization

[0066] The calibration data is uploaded to the database platform, multi-vehicle data is analyzed, a global optimization parameter set is generated, and implemented on the vehicle terminal.

[0067] The specific working principle is as follows:

[0068] 1. Based on the calibration process and system principle diagrams, and in conjunction with the description of the example calibration, further explanations and descriptions are provided for the specific implementation of the present invention.

[0069] 2. For the operating conditions, locations and configurations of newly developed models, the historical operating condition feature library is used to construct the corresponding usage scenarios and theoretical matching parameters to obtain basic theoretical data accumulation for subsequent vehicle component selection and software function matching.

[0070] 3. Determine the weights of key parameters (compressor speed, electronic expansion valve opening, fan speed, vehicle speed, system pressure, subcooling, superheating expansion valve opening) on ​​system COP (energy efficiency ratio), battery cooling, and outlet air temperature based on orthogonal experimental design.

[0071] 4. Based on the hierarchical calibration model, perform calibration work for the basic calibration layer and the dynamic compensation layer respectively;

[0072] 4.1 Basic Calibration Layer - Static Debugging

[0073] The system was calibrated using basic data to verify the stable operation and reliability of each working mode; the expected parameters of the system were consistent; and the parameters collected by each sensor location point were normal.

[0074] Heating and dehumidification mode: With vehicle doors and windows open, immerse the vehicle in ambient temperature for >1 hour; correctly connect the equipment, have personnel board the vehicle to check that all data collection points and all components are communicating normally, start the vehicle, enable data recording, and begin the heating and dehumidification mode debugging:

[0075] ①: Turn on the condenser fan and observe the battery and motor water temperature to determine whether the battery and motor water pump need to be turned on.

[0076] ②: Blower air volume: High;

[0077] Fresh air damper: External circulation;

[0078] Mode damper: Defrosting;

[0079] ③: Cooling mode;

[0080] ④: Turn on the compressor;

[0081] ⑤: Enable APTC;

[0082] ⑥: After 1 hour of steady-state operation, turn off the compressor and APTC, while keeping other components running.

[0083] Air source heat pump mode adjustment:

[0084] ①: Heating mode;

[0085] ②: Blower air volume: High;

[0086] Fresh air damper: External circulation;

[0087] Mode damper: Blows to feet;

[0088] ③: Adjust the compressor according to its maximum capacity;

[0089] ④: After 1 hour of steady-state operation, turn off the compressor, while keeping other components running;

[0090] Water source heat pump mode debugging:

[0091] ①: Heating mode;

[0092] ②: Blower air volume: High;

[0093] Fresh air damper: External circulation;

[0094] Mode damper: Blows to feet;

[0095] ③: Adjust the compressor according to its maximum capacity (high pressure < 20 bar);

[0096] ④: After 1 hour of steady-state operation, turn off the compressor, while keeping other components running;

[0097] Dual-source heat pump mode adjustment:

[0098] ①: Heating mode;

[0099] ②: Blower air volume: High;

[0100] Fresh air damper: External circulation;

[0101] Mode damper: Blows to feet;

[0102] ③: Adjust the compressor according to its maximum capacity;

[0103] ④: After 1 hour of steady-state operation, turn off the compressor, while keeping other components running;

[0104] Dual-source heat pump + battery WPTC dual-mode debugging:

[0105] ①: Heating mode;

[0106] ②: Blower air volume: High;

[0107] Fresh air damper: External circulation;

[0108] Mode damper: Blows to feet;

[0109] ③: Adjust the compressor according to its maximum capacity (high pressure < 20 bar); WPTC is activated;

[0110] ④: After 1 hour of steady-state operation, turn off the compressor and WPTC, while keeping other components running;

[0111] Water source defrosting mode adjustment:

[0112] ①: Heating mode;

[0113] ②: Blower air volume: Medium;

[0114] Fresh air damper: External circulation;

[0115] Mode damper: Blows to feet;

[0116] ③: Set the compressor outlet air temperature to 30℃; observe the low pressure changes and OHX frosting.

[0117] ④: Once the OHX frosting area exceeds 60%, turn off the compressor and turn on APTC (to maintain the outlet air temperature); cooling mode;

[0118] ⑤: After running the above cycle twice, shut down the compressor, APTC, blower, condenser fan, and power off the entire vehicle.

[0119] 4.2 Dynamic Compensation Layer - Dynamic Debugging:

[0120] Heating and dehumidification (temperature gradient): Data was collected on the heating gradient at medium to high fan speeds under urban road conditions; the windshield remained fog-free throughout the process.

[0121] Heating and dehumidification (maximum heating performance) verifies the maximum heating capacity under highway road conditions; the windshield remains fog-free throughout the test.

[0122] Air source heat pump (temperature gradient) collects the heating gradient of medium to high air volume under urban driving conditions of vehicles; observes the high and low pressure and temperature stability of the system.

[0123] Air source heat pump (maximum heating performance): Verifying the maximum heating capacity under high-speed road conditions;

[0124] Single APTC (Maximum Heating Performance): Verifies the maximum heating capacity under highway road conditions; monitors energy consumption and compares it with the energy consumption of air source heat pumps; supplements mixed-air conditions when performance is insufficient.

[0125] Water source heat pump: Verifying the impact of different ambient temperatures and WPTC water source power on heat pump air conditioning performance;

[0126] Single-cell battery charging liquid heating: Verify the heating efficiency of a single-cell WPTC battery; verify that after fully charging, the water source heat pump is turned on until the water temperature drops to 10℃.

[0127] Dual-source heat pump heating: Verifying the impact of air-source heat pump and WPTC water-source power on heat pump air conditioning performance under different ambient temperatures.

[0128] Dual-mode heating using air source and battery: Verifying the heating performance of the air source heat pump and the thermal performance of the battery under different ambient temperatures.

[0129] Water source heat pump defrosting: During the air source heat pump process, when there are significant changes in air temperature and low pressure, turn on APTC for supplemental heating; confirm system stability; after defrosting for less than 5 minutes, switch back to air source heat pump mode, and keep the outlet air temperature fluctuation <2℃ during the defrosting process.

[0130] The system was calibrated by fine-tuning the control targets, feedforward and feedback quantities of each actuator in the system mode; checking for heat leakage / cross-flow in the system loop; assessing the performance, response speed, and accuracy of battery and coolant temperature control; detecting air conditioning outlet temperature fluctuations during dual-mode switching; and evaluating the performance and priority of battery and coolant temperature control during dual-mode operation. A database of typical operating conditions for commercial vehicles (such as high-temperature cooling, low-temperature heating, and high-humidity defrosting) was also established based on historical data, covering parameters such as ambient temperature, humidity, vehicle speed, and battery thermal load.

[0131] 5. Database collaborative optimization

[0132] Static and dynamic calibration data are uploaded to the database platform, multi-condition data are analyzed, and a global optimization parameter set is generated and implemented on the vehicle terminal.

[0133] The calibration and optimization using the above scheme has the following technical advantages:

[0134] 1. By using a virtual calibration environment based on a working condition database, the number of real vehicle tests is reduced, and the calibration cycle is shortened to 1 / 3 of that of traditional methods;

[0135] 2. Applicable to various commercial vehicle types such as logistics vehicles and intercity buses, with high compatibility.

[0136] 3. Multi-objective collaborative optimization: Simultaneously optimize energy efficiency (COP), comfort (temperature fluctuation ≤ ±1℃) and system reliability (compressor start-stop count).

[0137] 4. Enhanced low-temperature heating: By introducing waste heat recovery and dynamic decoupling control of electronic expansion valve, the heating capacity in the -20℃ environment is increased by more than 30%.

[0138] Obviously, the specific implementation of this invention is not limited to the above-described methods. Any non-substantial improvements made using the inventive concept and technical solution of this invention are within the protection scope of this invention.

Claims

1. A calibration method for a multi-condition commercial vehicle heat pump air conditioning system, characterized in that: Includes the following steps: Step 1: Build a working condition feature library to obtain a typical automotive working condition library; Step 2: Analyze the weighting of the influence of each parameter of the heat pump air conditioning system on the system energy efficiency ratio (COP) and outlet air temperature; Step 3: Operating Parameter Calibration: With the goal of meeting the temperature setting and maximizing the COP, the parameters of the heat pump air conditioning system under each operating mode are calibrated to obtain the optimal parameter combination.

2. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 1, characterized in that: The method further includes step 4: the cloud platform performs multi-vehicle data optimization on the optimal parameter combination calibrated in step 3 to generate the final global optimization parameter set.

3. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 1 or 2, characterized in that: In step 1, a typical operating condition library for commercial vehicles is established based on historical data. The classic operating condition library includes various operating conditions for commercial vehicles and their corresponding parameters.

4. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 1 or 2, characterized in that: In step 2, the influence weights of the heat pump air conditioning system parameters on the system energy efficiency ratio (COP) and outlet air temperature are determined by orthogonal experimental design.

5. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 1 or 2, characterized in that: Step 3 includes a basic calibration layer and a dynamic compensation layer. Based on the calibration layer, the parameters of the heat pump air conditioning system under various operating conditions are optimized statically. The dynamic compensation layer compensates and adjusts the optimized static parameters in real time to obtain the calibrated parameters for each operating condition.

6. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 5, characterized in that: The basic calibration layer solves for the optimal parameter combination under various operating conditions based on the software thermal management control model, maximizing COP and satisfying the temperature setpoint.

7. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 5, characterized in that: The dynamic compensation layer collects data in real time through on-board sensors and uses fuzzy PID control to dynamically correct parameters and compensate for sudden environmental changes or load fluctuations.

8. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 2, characterized in that: The calibrated operating condition data are uploaded to the cloud platform, which then distributes them to the test vehicles for operation and analysis to generate a global optimization parameter set as the final global optimization parameters.

9. The calibration method for a multi-condition commercial vehicle heat pump air conditioning system as described in claim 8, characterized in that: The cloud platform assigns weights to multi-vehicle data and analyzes the collected multi-vehicle data based on these weights to generate the final global optimization parameters.

10. A calibration system for a multi-condition commercial vehicle heat pump air conditioning system, characterized in that: Includes a working condition feature library construction module, a parameter sensitivity analysis module, and a graded calibration model; The operating condition feature library construction module is used to build an operating condition feature library to obtain a typical automotive operating condition library; the parameter sensitivity analysis module is used to analyze the influence weight of each parameter of the heat pump air conditioning system on the system energy efficiency ratio COP and outlet air temperature. The operating condition parameter calibration module calibrates the parameters of the heat pump air conditioning system under various operating conditions with the goal of satisfying the temperature setting and maximizing the energy efficiency ratio (COP) to obtain the optimal parameter combination.

Images