Control method of double-stage charge-refrigerant heat exchange compressor based on carbon dioxide transcritical cycle

Abstract

The application relates to the field of compressor control, in particular to a two-stage air-supply enthalpy-increasing compressor control method based on a carbon dioxide transcritical cycle. The method comprises the following steps: acquiring a set of real-time running state information of a vehicle, based on which, multi-dimensional system state representation and optimal target fusion analysis are carried out to generate a dynamic optimal exhaust pressure target value; based on the dynamic optimal exhaust pressure target value, dynamic coupling and decoupling of a main control channel and an auxiliary control channel and collaborative instruction distribution are carried out to generate a set of collaborative control instruction information of the compressor and an electronic expansion valve; based on the set of collaborative control instruction information, feedforward prediction compensation and timing inertia smoothing management are carried out in combination with measurable disturbance information to generate a set of final actuator driving instruction information after disturbance smoothing; and based on the set of final actuator driving instruction information, safety driving instructions and parameter online fine adjustment learning are carried out to generate a set of two-pole compressor control optimization information. In the compressor control process, the application can maintain the optimal efficiency and pressure stability of the system under transient conditions.

Description

Technical Field

[0001] This application relates to the field of compressor control, and in particular to a two-stage enthalpy-increasing compressor control method based on a transcritical carbon dioxide cycle. Background Technology

[0002] In the field of high-efficiency thermal management of new energy vehicles, CO2 transcritical cycle technology has become a key path to solve the problem of high energy consumption in the thermal management of electric vehicle air conditioning and batteries due to its environmental friendliness, high heat exchange efficiency and excellent performance in a wide temperature range. Its system energy efficiency and operational stability directly determine the vehicle's range and driving comfort, and it is one of the core technologies to enhance the overall competitiveness of electric vehicles.

[0003] However, existing control methods based on CO2 transcritical cycles generally lack a closed-loop optimization mechanism for real-time fusion characterization of multi-dimensional system states, dynamic decoupling and coordination of strong coupling between control channels, and look-ahead compensation and command smoothing for measurable disturbances when facing complex and variable vehicle driving conditions and environmental disturbances. This not only makes it difficult to maintain optimal system efficiency and pressure stability under transient conditions, but may also cause system oscillations due to frequent actuator actions or response lags, thus restricting the full realization of the potential of this technology. Summary of the Invention

[0004] This application provides a two-stage enthalpy-increasing compressor control method based on a transcritical carbon dioxide cycle to solve the aforementioned technical problems. The method includes: acquiring a real-time vehicle operating status information set; performing multi-dimensional system state characterization and optimal target fusion analysis based on the real-time vehicle operating status information set to generate a dynamic optimal exhaust pressure target value; performing dynamic coupling decoupling and collaborative command allocation of the dominant and auxiliary control channels based on the dynamic optimal exhaust pressure target value to generate a collaborative control command information set for the compressor and electronic expansion valve; performing feedforward prediction compensation and temporal inertia smoothing management based on the collaborative control command information set and measurable disturbance information to generate a disturbance-resistant smoothed final actuator drive command information set; and performing online fine-tuning learning of safety drive commands and parameters based on the final actuator drive command information set to generate a two-stage compressor control optimization information set.

[0005] Through the above technical solutions, the CO2 heat pump system can always operate dynamically in the optimal energy efficiency range under complex vehicle operating conditions, significantly improving the vehicle's range; through dynamic decoupling and collaborative control, the inherent oscillation of the system is eliminated, resulting in a faster and more stable response; combined with feedforward disturbance rejection and safety drive, the system's operational safety and component lifespan are guaranteed under extreme conditions; the online learning mechanism ensures that the control performance does not decay throughout the entire life cycle, achieving efficient, stable, reliable and adaptive advanced thermal management.

[0006] Optionally, generating the dynamic optimal exhaust pressure target information set includes: the vehicle real-time operating status information set includes environmental load characteristics, core heat exchange efficiency characteristics, and vehicle load demand characteristics; inputting the environmental load characteristics and the core heat exchange efficiency characteristics into the optimal efficiency point pressure mapping table to obtain a basic optimal exhaust pressure reference value that matches the current operating conditions; and performing target fusion analysis on the basic optimal exhaust pressure reference value based on the load demand characteristics to output the dynamic optimal exhaust pressure target value.

[0007] Optionally, the optimal efficiency point pressure mapping table includes: based on bench experiments of a two-stage gas-injection enthalpy-enhancing compressor under CO2 transcritical cycles, in an environmental simulation chamber covering the full operating spectrum of the vehicle, collecting exhaust pressure data points under different combinations of the environmental load characteristics and the core heat exchange efficiency characteristics, so that the overall energy efficiency ratio of the system reaches a local optimum; performing continuous processing on the exhaust pressure data point set using a surface fitting algorithm based on physical model constraints to generate an initial two-dimensional optimal efficiency pressure reference surface; introducing the system transient response data and long-term operating performance decay coefficient under actual road cycles, performing dynamic robustness correction and durability compensation on the two-dimensional optimal efficiency pressure reference surface, and generating the final multi-dimensional, time-varying coupled optimal efficiency point pressure mapping table.

[0008] Optionally, the target fusion analysis of the basic optimal exhaust pressure reference value includes: constructing a weighted adaptive fusion function containing a system energy efficiency priority mode, a dynamic response priority mode, and a user comfort priority mode based on the load demand characteristics; identifying the specific operating conditions of the vehicle in real time, dynamically adjusting the weight coefficients corresponding to each mode in the weighted adaptive fusion function, and generating a target pressure correction coefficient for the current scenario; analyzing the rate of change of the load demand characteristics in real time, and generating a transient pressure compensation term based on this rate of change, wherein the rate of change characterizes the dynamic trend and urgency of load demand; multiplying the target pressure correction coefficient by the basic optimal exhaust pressure reference value, and superimposing the transient pressure compensation term, to output the dynamic optimal exhaust pressure target value.

[0009] Optionally, generating the collaborative control command information set for the compressor and the electronic expansion valve includes: decomposing the dynamic optimal exhaust pressure target value into a compressor-dominant control target value and an electronic expansion valve auxiliary control target value; identifying the dynamic coupling relationship between the compressor-dominant control target value and the electronic expansion valve auxiliary control target value; decoupling and compensating for the influence of the compressor-dominant control target value based on the dynamic coupling relationship; and merging the compensated auxiliary control target value with the compressor-dominant control target value to generate the collaborative control command information set.

[0010] Optionally, identifying the dynamic coupling relationship between the compressor's primary control target value and the electronic expansion valve's auxiliary control target value includes: constructing a dynamic coupling transfer function based on the first principle of thermodynamics and historical operating data, characterizing the change between the compressor's primary control target value and the electronic expansion valve's auxiliary control target value; updating the parameters of the dynamic coupling transfer function online during real-time operation using a recursive least squares algorithm with a forgetting factor to track changes in the system's dynamic characteristics with operating conditions; analyzing the dynamic coupling gain matrix at the current operating point based on the updated dynamic coupling transfer function, and extracting the gain value and phase lag information of its main coupling channel as a quantitative coupling strength characteristic index; evaluating the coupling strength level between the compressor and the electronic expansion valve based on the quantitative coupling strength characteristic index, and generating a dynamic decoupling trigger command and compensation priority signal according to a preset strength threshold to guide subsequent compensation calculations.

[0011] Optionally, the decoupling compensation calculation based on the dynamic coupling relationship includes: establishing a dynamic coupling model characterizing the impact of compressor speed command changes on the required opening degree of the electronic expansion valve; when the compressor's dominant control target value changes, analyzing in real time based on the dynamic coupling model the electronic expansion valve opening degree compensation amount required to offset the impact of the target value change on the evaporator-side refrigerant flow and system pressure stability; superimposing the electronic expansion valve opening degree compensation amount onto the corresponding opening degree of the reference opening degree command obtained based on the analysis of the electronic expansion valve's auxiliary control target value, thereby generating an anti-coupling interference electronic expansion valve final opening degree command.

[0012] Optionally, generating the final actuator drive command information set after disturbance rejection and smoothing includes: acquiring measurable disturbance information of ambient temperature and vehicle speed, and identifying the real-time changing trend of the measurable disturbance information; based on the amplitude and direction of the real-time changing trend, generating predictive correction values ​​for the compressor speed command and electronic expansion valve opening command in the cooperative control command information set through a feedforward compensator; superimposing the predictive correction values ​​onto the control values ​​of the corresponding control commands to form preliminary disturbance rejection control commands; performing time-series inertial smoothing filtering on the compressor speed command in the preliminary disturbance rejection control commands, wherein the filtering intensity is dynamically adjusted according to the system operating state to generate the smoothed final compressor speed drive command; and combining and encapsulating the final compressor speed drive command and the final electronic expansion valve opening drive command to generate the final actuator drive command information set.

[0013] Optionally, the filtering intensity is dynamically adjusted according to the system operating state, including: monitoring the real-time deviation between the system exhaust pressure and the target value of the dynamic optimal exhaust pressure; when the real-time deviation is less than the steady-state pressure tolerance threshold, determining that the system is in steady-state operation, and increasing the filtering intensity of the time-series inertial smoothing filter to suppress high-frequency fluctuations of the command; when the real-time deviation is greater than or equal to the dynamic adjustment trigger threshold, or when the vehicle load demand characteristic is identified to be in a rapid change phase, determining that the system is in transient operation, and reducing the filtering intensity of the time-series inertial smoothing filter to improve the response speed of the control command.

[0014] Optionally, generating the bipolar compressor control optimization information set includes: based on the final compressor speed drive command and the final electronic expansion valve opening drive command, performing drive range verification and dynamic limiting based on multi-constraint safety boundaries to generate a safe drive command; based on the real-time performance data generated by the system operation driven by the safe drive command, performing closed-loop performance evaluation and online gradient fine-tuning on the benchmark data points in the optimal efficiency point pressure mapping table, the weight coefficients in the weighted adaptive fusion function, and the filter intensity parameters of the time-series inertial smoothing filter; and encapsulating the mapping table, function, and parameter combination after the closed-loop performance evaluation and online gradient fine-tuning into the bipolar compressor control optimization information set. Attached Figure Description

[0015] 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.

[0016] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application;

[0017] Figure 2 A flowchart of a two-stage enthalpy-increasing compressor control method based on a transcritical carbon dioxide cycle provided in an embodiment of this application. Detailed Implementation

[0018] 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.

[0019] 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.

[0020] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0021] Existing control methods based on CO2 transcritical cycles generally lack a closed-loop optimization mechanism for real-time fusion characterization of multi-dimensional system states, dynamic decoupling and coordination of strong coupling between control channels, and look-ahead compensation and command smoothing for measurable disturbances when facing complex and variable vehicle driving conditions and environmental disturbances. This not only makes it difficult to maintain optimal system efficiency and pressure stability under transient conditions, but may also cause system oscillations due to frequent actuator actions or response lags, thus restricting the full realization of the potential of this technology.

[0022] Based on this, this application provides a two-stage enthalpy-increasing compressor control method based on a transcritical CO2 cycle. First, it integrates multi-source vehicle information, analyzes it in real time, and dynamically generates the globally optimal exhaust pressure target value. Second, it establishes a dynamic coupling model between the compressor and the electronic expansion valve, and generates collaborative control commands through a feedforward decoupling algorithm. Next, it introduces feedforward compensation for measurable disturbances such as vehicle speed, and performs adaptive smoothing filtering on the compressor speed command to enhance disturbance rejection and smoothness. Finally, it performs hard safety checks before all commands are executed, and fine-tunes the internal model parameters online based on operating data to generate a two-stage compressor control optimization information set and output it to the user, achieving self-learning and self-adaptation of the controller. This solution enables the CO2 heat pump system to operate dynamically within its optimal energy efficiency range under complex vehicle operating conditions, significantly improving the vehicle's range. Through dynamic decoupling and collaborative control, it eliminates inherent system oscillations, resulting in faster and more stable responses. Combined with feedforward disturbance rejection and safety-driven operation, it ensures the system's operational safety and component lifespan under extreme conditions. The online learning mechanism ensures that control performance does not degrade throughout the entire lifespan, achieving efficient, stable, reliable, and adaptive advanced thermal management.

[0023] Figure 1 This is a schematic diagram illustrating an application scenario provided by this application. In the compressor control process, the method provided in this application can maintain optimal system efficiency and pressure stability under transient operating conditions, significantly improving the vehicle's range.

[0024] Specifically, the method of this application is applied to any server that communicates with the sensor network of the vehicle control system. The server acquires the real-time vehicle operating status information set provided by the sensor network. First, it fuses multi-source vehicle information, analyzes it in real time, and dynamically generates the globally optimal exhaust pressure target value. Second, it establishes a dynamic coupling model between the compressor and the electronic expansion valve, and generates cooperative control commands through a feedforward decoupling algorithm. Next, it introduces feedforward compensation for measurable disturbances such as vehicle speed, and performs adaptive smoothing filtering on the compressor speed command to enhance disturbance rejection and smoothness. Finally, it performs hard safety checks before all commands are executed, and fine-tunes the internal model parameters online based on operating data to generate a bipolar compressor control optimization information set and output it to the user, achieving self-learning and self-adaptation of the controller.

[0025] For specific implementation details, please refer to the following examples.

[0026] Figure 2 This is a flowchart illustrating a two-stage enthalpy-increasing compressor control method based on a transcritical carbon dioxide cycle, as provided in an embodiment of this application. The method of this embodiment can be applied to servers in the above-described scenario. For example... Figure 2 As shown, the method includes:

[0027] S201. Obtain the real-time operating status information set of the vehicle, and based on the real-time operating status information set of the vehicle, perform multi-dimensional system status characterization and optimal target fusion analysis to generate a dynamic optimal exhaust pressure target value.

[0028] The real-time vehicle operating status information set refers to the data collection from the vehicle bus (such as the CAN bus), the air conditioning system sensor network, and the vehicle controller, used to comprehensively describe the operating environment and requirements of the air conditioning system. The data originates from the vehicle control system sensor network. Multi-dimensional system state characterization refers to processing the real-time vehicle operating status information set and transforming it into characteristic quantities that can directly guide control decisions. Optimal target fusion analysis refers to the fact that in the CO2 transcritical cycle, the safe operating range of system performance (especially the coefficient of performance, COP) and key components (such as the compressor and gas cooler) strongly depends on the exhaust pressure (i.e., the high-pressure side pressure). The dynamically optimal exhaust pressure target value can be a pressure target value that is dynamically adjusted according to vehicle conditions (such as high heat dissipation requirements during high-speed driving or urgent cooling requirements during battery fast charging).

[0029] Specifically, CO2 transcritical cycle heat pump systems have become a cutting-edge technology in the field of thermal management for new energy vehicles due to their high efficiency and environmental friendliness. However, their operating characteristics differ significantly from those of traditional refrigerants. System performance (especially the coefficient of performance, COP) and safe operating range are highly dependent on the core parameter of high-pressure side exhaust pressure. Exhaust pressure is not a fixed value that can be freely set; it directly affects cooling / heating capacity, compressor energy consumption, and the safe operating pressure boundaries of key components (such as gas coolers and compressors). In the dynamic and ever-changing vehicle operating environment (such as high-speed driving, rapid acceleration, different ambient temperatures, and battery fast-charging cooling requirements), the comfort requirements of the passenger compartment, the thermal management requirements of the battery pack, and the energy economy requirements of the entire vehicle constitute a dynamic and even conflicting set of multi-dimensional optimization objectives. For example, pursuing the fastest cooling speed of the passenger compartment may require a higher exhaust pressure, but this will sacrifice energy efficiency and increase the system load; prioritizing battery cooling may require adjusting the refrigerant flow path and pressure settings, thereby affecting the comfort of the passenger compartment. Most existing control schemes are based on a single, fixed exhaust pressure setpoint, or rely solely on PID control for a single objective (such as vent temperature). These schemes fail to achieve system-wide dynamic optimization, resulting in poor energy efficiency and reduced driving range, slow response and poor comfort, or operation near safety boundaries with high reliability risks under most complex operating conditions. To address this fundamental challenge, a comprehensive data set of real-time vehicle operating status information, incorporating the vehicle's overall thermal management requirements, is collected and integrated to provide a panoramic data perspective for decision-making. Then, through multi-dimensional system status characterization, the raw data is refined into quantifiable features directly applicable to decision-making, such as "heat load demand level," "disturbance trend intensity," and "multi-objective priority." Finally, through optimal objective fusion analysis, and applying preset, configurable multi-objective optimization rules, a dynamically optimal exhaust pressure target value is calculated in real time that simultaneously satisfies efficiency, speed, safety, and prioritizes the highest current thermal management requirement.

[0030] S202. Based on the dynamic optimal exhaust pressure target value, perform dynamic coupling decoupling and collaborative command allocation of the main and auxiliary control channels to generate a collaborative control command information set for the compressor and electronic expansion valve.

[0031] The primary and secondary control channels can be defined as follows: the primary control channel is the loop that controls the compressor speed to directly regulate the system refrigerant circulation and discharge pressure; the secondary control channel is the loop that controls the electronic expansion valve opening to regulate the evaporator flow and maintain superheat stability. Dynamic coupling decoupling refers to identifying and compensating for the mutual interference between the two control channels. For example, when the compressor speed rapidly increases to meet a new discharge pressure target, it instantly increases the pressure displacement of the low-pressure side from the high-pressure side, causing a sharp decrease in the evaporator inlet flow, which may lead to a surge in superheat, and vice versa. The coordinated control command information set can be defined as a set of commands generated after decoupling and coordination, which simultaneously includes compressor speed commands and electronic expansion valve opening commands.

[0032] Specifically, after determining the system-level dynamic optimal discharge pressure target value, another key challenge is how to coordinate the two core actuators—the compressor (which regulates refrigerant flow and pressure) and the electronic expansion valve (which regulates evaporator flow and superheat)—to accurately and stably achieve this target. In a CO2 transcritical system, the control loops of these two actuators exhibit strong, bidirectional dynamic coupling. For example, significantly increasing the compressor speed to rapidly raise the discharge pressure will instantly increase the compression and pushing of the working fluid on the high-pressure side, causing the refrigerant flow to the evaporator to be "drawn" in the short term. This may cause a sharp increase in the superheat of the evaporator controlled by the electronic expansion valve, or even cause the evaporator to "evacuate," not only reducing efficiency but also potentially leading to the risk of liquid slugging during compressor suction. Conversely, rapid opening of the electronic expansion valve will alter the flow balance between the high and low pressure sides of the system, affecting the stability of the discharge pressure. To address this core coupling problem at the control level, a dynamic coupling decoupling and collaborative command allocation mechanism for the dominant and auxiliary control channels is introduced. The core of this approach lies in establishing a dynamic coupling model in advance, through mechanistic analysis or system identification, to describe the impact of compressor action on superheat (and vice versa). In actual control, when the baseline command for the main channel (compressor) is calculated based on the pressure target, this model simultaneously calculates the estimated disturbance that this action will cause to the auxiliary channel (electronic expansion valve), and immediately generates a feedforward decoupling compensation command with the opposite direction and matching magnitude, which is pre-added to the control calculation of the electronic expansion valve.

[0033] S203. Based on the cooperative control instruction information set, combined with measurable disturbance information, feedforward prediction compensation and timing inertial smoothing management are performed to generate the final actuator drive instruction information set after disturbance rejection and smoothing.

[0034] Measurable disturbance information refers to external factors that can be measured or predicted and have a significant impact on the system, but are not directly controlled variables. Feedforward predictive compensation refers to calculating a correction amount in advance based on the measured disturbance signal and its changing trend before the disturbance actually affects the controlled variable (such as exhaust pressure or superheat), and adding it to the control command. Timing inertia smoothing management specifically refers to the processing of compressor speed commands. Compressors (especially two-stage injection enthalpy-increasing compressors) have large mechanical inertia, and sudden changes in speed can lead to torque shocks, increased vibration and noise, and shortened lifespan. The final actuator drive command information set can be control commands that have undergone two "refined" processes: feedforward disturbance rejection and timing smoothing. It is more "intelligent" and "gentle," and is directly used to drive the actuator.

[0035] Specifically, even with decoupling and coordination between control loops, heat pump systems still face continuous and predictable external disturbances in real-world vehicle environments, as well as constraints imposed by the physical response characteristics of the actuators themselves. The most significant measurable disturbances include vehicle speed and ambient temperature / wind speed. Vehicle speed directly affects the oncoming wind speed of the condenser (gas cooler), thus altering its heat dissipation capacity and causing rapid pressure surges on the high-pressure side; changes in ambient wind also affect the vehicle's heat exchange. If relying solely on error-based feedback control, the system only begins to correct after the disturbance has already caused an impact and deviation, resulting in inherent lag in control response and making pressure and temperature fluctuations during dynamic adjustment unavoidable. Furthermore, as a precision rotating machine, the two-stage enthalpy-increasing compressor experiences frequent and drastic speed changes, leading to torque surges, increased vibration and noise, and accelerated wear, severely impacting its lifespan and NVH performance. Existing technologies typically employ fixed-parameter filters for simple smoothing, but this cannot simultaneously address the requirements of dynamic response speed and smoothness. To address these two types of problems, a dual-enhancement process is designed, which not only significantly improves the system's robustness to changes in the external environment but also ensures the long-term reliable, quiet, and stable operation of actuators (especially expensive and critical compressors). This process can be designed as a robust optimization and actuator management service.

[0036] S204. Based on the final actuator drive instruction information set, perform online fine-tuning learning of safety drive instructions and parameters to generate a bipolar compressor control optimization information set.

[0037] Safety-driven instructions can refer to a final round of safety checks and constraints before sending the final instructions to the actuator. Online parameter fine-tuning learning refers to the system automatically recording the actual control performance under different operating conditions (such as time to reach target pressure, overshoot, and steady-state error) and comparing it with model predictions during long-term operation. The bipolar compressor control optimization information set refers to the set of information generated within the current control cycle by the safety-driven and online learning processes, used to optimize the control performance for the next cycle or long-term operation.

[0038] Specifically, delivering optimized control commands safely and reliably to physical actuators, and endowing the control system with the ability to continuously evolve, is the final and crucial step in ensuring long-term performance and reliability. Hardware actuators have insurmountable safety operating boundaries, such as the compressor's maximum / minimum permissible speed, the electronic expansion valve's limit opening, and the system's maximum permissible exhaust pressure and temperature. Any commands generated by optimization algorithms must ultimately be strictly limited within these physical boundaries; otherwise, equipment damage or even safety accidents may occur. Furthermore, as a complex system, the dynamic characteristics of a vehicle's heat pump system will slowly drift with age, refrigerant leaks, component performance degradation, or external maintenance and replacement. A controller with fixed parameters may not achieve the same control performance as at the time of manufacture later in the vehicle's lifespan. Existing solutions often lack effective online adaptive mechanisms or only possess simple safety protections, failing to achieve "self-optimization" of the controller. This solution constructs a "safety guardian" and "learning brain" for the control closed loop by executing safety-driven commands and online fine-tuning of parameters. The resulting bipolar compressor control optimization information set not only records the safety status and learning results of this control cycle, but also provides valuable data assets for predictive maintenance, energy efficiency analysis, and offline optimization of next-generation control strategies.

[0039] The approach provided in this embodiment first integrates multi-source information from the entire vehicle, analyzes it in real time, and dynamically generates the globally optimal exhaust pressure target value. Second, a dynamic coupling model between the compressor and the electronic expansion valve is established, and collaborative control commands are generated through a feedforward decoupling algorithm. Next, feedforward compensation for measurable disturbances such as vehicle speed is introduced, and adaptive smoothing filtering is applied to the compressor speed commands to enhance disturbance rejection and smoothness. Finally, hard safety checks are performed before all commands are executed, and internal model parameters are fine-tuned online based on operating data to generate a bipolar compressor control optimization information set and output it to the user, achieving self-learning and self-adaptation of the controller. This solution enables the CO2 heat pump system to always operate dynamically within the optimal energy efficiency range under complex vehicle operating conditions, significantly improving the vehicle's range. Through dynamic decoupling and collaborative control, the inherent oscillations of the system are eliminated, resulting in faster and more stable responses. Combined with feedforward disturbance rejection and safety drive, the system's operational safety and component lifespan are guaranteed under extreme conditions. The online learning mechanism ensures that control performance does not decay throughout the entire life cycle, achieving efficient, stable, reliable, and adaptive advanced thermal management.

[0040] In some embodiments, the real-time vehicle operating status information set includes environmental load characteristics, core heat exchange efficiency characteristics, and vehicle load demand characteristics. The environmental load characteristics and core heat exchange efficiency characteristics are input into the optimal efficiency point pressure mapping table to obtain a basic optimal exhaust pressure reference value that matches the current operating conditions. Based on the load demand characteristics, a target fusion analysis is performed on the basic optimal exhaust pressure reference value to output a dynamic optimal exhaust pressure target value.

[0041] Environmental load characteristics are key parameters characterizing the impact of the external environment on the operation of a transcritical CO2 cycle system, and are core indicators reflecting the impact of the external environment on the system load. Core heat exchange efficiency characteristics are key parameters reflecting the heat exchange capacity and operating status of core heat exchange components (such as condensers, evaporators, and make-up gas heat exchangers) in the transcritical CO2 cycle system, directly determining the system's energy transfer efficiency and operational stability. Vehicle load demand characteristics are a set of parameters characterizing the vehicle's current energy output demand on the transcritical CO2 cycle system, directly reflecting the actual functional objectives the system needs to meet, and serving as an important basis for dynamically adjusting the exhaust pressure target value. The optimal efficiency point pressure mapping table is a multi-dimensional data mapping table constructed based on full-condition experimental data of a two-stage make-up gas enthalpy-increasing compressor under transcritical CO2 cycles, reflecting the corresponding relationship between exhaust pressure and the system's overall energy efficiency ratio when it reaches a local optimum under different combinations of environmental load characteristics and core heat exchange efficiency characteristics. The basic optimal exhaust pressure reference value can be the initial exhaust pressure reference value obtained from the optimal efficiency point pressure mapping table when only considering the optimal energy efficiency target under the combination of current environmental load characteristics and core heat exchange efficiency characteristics.

[0042] Specifically, traditional two-stage enthalpy-increasing compressors rely heavily on static targets derived from fixed values ​​or single parameters for exhaust pressure control, resulting in significant limitations in adapting to different operating conditions. For example, in high summer temperatures (e.g., 38°C), a fixed low exhaust pressure can lead to reduced heat exchange efficiency due to high environmental load and dust accumulation on heat exchange components, resulting in insufficient cooling capacity and preventing the vehicle interior temperature from quickly dropping to the set 24°C. In low winter temperatures (e.g., -5°C), a fixed high exhaust pressure can cause energy waste, reducing the system's energy efficiency ratio by more than 30%. Furthermore, traditional methods ignore dynamic changes in vehicle load. For instance, if the number of passengers increases from 2 to 5, the cooling demand surges, and the static pressure target cannot respond in time, significantly reducing comfort. To address the above issues, this step first collects environmental load characteristics using external vehicle temperature sensors (38℃), humidity sensors (60%RH), and sunlight sensors (800W / ㎡). It then collects core heat exchange efficiency characteristics using inlet and outlet pressure sensors (e.g., 10MPa on the high-pressure side), flow sensors (e.g., 5L / min), and wind speed sensors (2m / s) for the heat exchange components. Finally, it collects load demand characteristics using in-vehicle temperature sensors (28℃), occupant detection sensors (5 people), and the set temperature (24℃) obtained from the air conditioning control panel. All data is filtered, noise-reduced, and standardized before being packaged into a real-time operating status information set. (Sampling frequency 1 time / second); then, the environmental and heat exchange characteristic quantities are input into the pre-generated optimal efficiency point pressure mapping table. If there is no perfectly matching operating condition, a linear interpolation algorithm is used to obtain the basic optimal exhaust pressure reference value (e.g., if the pressures of adjacent operating conditions are 9MPa and 11MPa, and the current operating condition matches both operating conditions by 50%, then 10MPa is calculated by (9×0.5+11×0.5)). Finally, combined with the load demand (temperature difference 4℃, high occupancy load), a transient compensation of 0.5MPa is added to the basic reference value to generate a dynamic optimal exhaust pressure target value of 10.5MPa to ensure that it is adapted to the current high load demand.

[0043] The method provided in this embodiment, by comprehensively collecting environmental load characteristics, core heat exchange efficiency characteristics, and vehicle load demand characteristics, constructs a comprehensive set of real-time vehicle operating status information, breaking the limitations of single or fixed parameters in traditional control.

[0044] In some embodiments, based on bench tests of a two-stage enthalpy-increasing compressor under CO2 transcritical cycles, exhaust pressure data points are collected in an environmental simulation chamber covering the full operating spectrum of the vehicle. Under different combinations of environmental load characteristics and core heat exchange efficiency characteristics, the exhaust pressure data points are used to achieve the local optimum of the system's overall energy efficiency ratio. The exhaust pressure data points are then processed into a continuous curve fitting algorithm based on physical model constraints to generate an initial two-dimensional optimal efficiency pressure reference surface. The transient response data of the system under actual road cycles and the long-term operating performance decay coefficient are introduced to perform dynamic robustness correction and durability compensation on the two-dimensional optimal efficiency pressure reference surface, generating the final multi-dimensional, time-varying coupled optimal efficiency point pressure mapping table.

[0045] A two-stage enthalpy-increasing compressor with intermediate gas injection can be a core power component of a transcritical CO2 cycle system. By employing a two-stage compression design with intermediate gas injection, it can enhance the enthalpy change of the refrigerant during the cycle, thereby improving the system's cooling / heating capacity and energy efficiency under a wide range of operating conditions. A transcritical CO2 cycle refers to a thermodynamic cycle mode in which carbon dioxide is used as the refrigerant, and the refrigerant undergoes subcritical, critical, and supercritical state changes during the cycle. Bench testing refers to an experimental method that uses a test platform identical to the actual transcritical CO2 cycle system of a vehicle, built in a laboratory environment, to perform performance tests on the two-stage enthalpy-increasing compressor with intermediate gas injection and the entire system. An environmental simulation chamber is a specialized device used in bench testing to simulate the actual external environmental conditions of a vehicle during operation. It can precisely control environmental parameters such as temperature, humidity, solar radiation intensity, and wind speed to simulate the external environmental conditions of the vehicle under different climates and driving scenarios. The exhaust pressure data set can be a discrete dataset generated during bench testing. This dataset is generated for each combination of environmental load characteristics and core heat transfer efficiency characteristics under the full operating spectrum of the vehicle. The goal is to find the exhaust pressure data that achieves the local optimum of the system's overall energy efficiency ratio through testing. The surface fitting algorithm refers to an algorithm that, when processing the discrete exhaust pressure data set continuously, introduces a thermodynamic physical model of the CO2 transcritical cycle system (such as the refrigerant equation of state, heat transfer equation, compressor working principle, etc.) as constraints, and constructs a continuous pressure reference surface through mathematical fitting methods. The two-dimensional optimal efficiency pressure reference surface can be a two-dimensional surface model obtained by a surface fitting algorithm constrained by a physical model, after processing the discrete exhaust pressure data set continuously. This model reflects the continuous correspondence between environmental load characteristics, core heat transfer efficiency characteristics, and optimal exhaust pressure. System transient response data refers to the dynamic operating data of the CO2 transcritical cycle system recorded in real time by an onboard data acquisition system when the vehicle is driving on actual roads according to typical driving cycles (such as urban roads, highways, suburban roads, etc.). The long-term performance degradation coefficient refers to a coefficient used to quantify the degree of performance decline of core components such as two-stage enthalpy-increasing compressors and heat exchangers during long-term use due to factors such as wear, aging, dust accumulation, and scaling. Dynamic robustness correction involves adjusting the two-dimensional optimal efficiency pressure reference surface based on system transient response data under actual road cycles, enabling the surface model to adapt to the system response characteristics under actual dynamic operating conditions, and improving the adaptability and stability of the mapping table under transient conditions. Durability compensation involves correcting the two-dimensional optimal efficiency pressure reference surface based on the long-term performance degradation coefficient, enabling the surface model to adapt to the performance degradation of core components after long-term operation, ensuring that even with component aging, the mapping table can still provide the exhaust pressure reference value that optimizes the system's overall energy efficiency ratio.

[0046] Specifically, the traditional scheme for constructing the optimal efficiency point pressure mapping table for CO2 transcritical cycle systems has many fatal flaws: data is collected only under a few steady-state conditions such as 25°C and 50% humidity, ignoring extreme environments such as -30°C and 45°C, as well as dynamic driving conditions such as urban congestion, resulting in significant deviations in reference values ​​for vehicles under extreme conditions; the use of purely mathematical fitting without physical constraints may lead to outliers where the 18MPa pressure exceeds the safe range of 8-15MPa for compressors; and the transient changes in actual roads and component degradation are not considered, resulting in a 20% decrease in the heat exchanger's heat transfer coefficient after 5 years of operation, while the mapping table still uses data from the new components, leading to a reduction in system energy efficiency of more than 30%. To address the above issues, this step involves constructing a transcritical CO2 cycle test bench with a two-stage enthalpy-increasing compressor, equipped with high-precision sensors, and placed in an environmental simulation chamber to simulate temperatures from -30°C to 45°C, humidity from 30% to 80%, and solar radiation from 0 to 800 W / m² (environmental load characteristics). The heat exchanger's heat transfer coefficient is adjusted to 800-1200 W / (m²·K), and the refrigerant flow rate is adjusted to 3-7 L / min (core heat exchange efficiency characteristics). Data points on the exhaust pressure that optimize system energy efficiency under all operating conditions are collected (e.g., 25°C, 50% humidity, heat transfer coefficient...). (Optimal pressure at 1000 is 10MPa); after preprocessing to remove outlier data, a least-squares surface fitting algorithm with physical constraints (exhaust pressure 8-15MPa) is used to construct a two-dimensional optimal efficiency pressure reference surface; the surface is corrected using actual road transient response data of sudden wind speed increase during vehicle acceleration, and then durability compensation is performed by introducing the compressor efficiency decay coefficient of 0.85 and the heat exchanger heat transfer coefficient decay coefficient of 0.8 after 5 years of operation, finally forming a multi-dimensional, time-varying coupled mapping table that can quickly output accurate optimal exhaust pressure reference values ​​under any operating conditions.

[0047] The multidimensional, time-varying coupled optimal efficiency point pressure mapping table constructed by the method provided in this embodiment can comprehensively characterize the complex relationship between multidimensional time-varying parameters and optimal exhaust pressure, realizing the transformation from discrete experimental data to a continuous, dynamic, and durable accurate mapping model.

[0048] In some embodiments, based on load demand characteristics, a weighted adaptive fusion function is constructed that includes a system energy efficiency priority mode, a dynamic response priority mode, and a user comfort priority mode; the specific operating conditions of the vehicle are identified in real time, and the weight coefficients corresponding to each mode in the weighted adaptive fusion function are dynamically adjusted to generate the target pressure correction coefficient for the current scenario; the rate of change of load demand characteristics is analyzed in real time, and a transient pressure compensation term is generated based on this rate of change, whereby the rate of change characterizes the dynamic trend and urgency of load demand; the target pressure correction coefficient is multiplied by the basic optimal exhaust pressure reference value, and the transient pressure compensation term is superimposed to output the dynamic optimal exhaust pressure target value.

[0049] The weighted adaptive fusion function is a mathematical function model that dynamically adjusts the weight ratio of different control objectives based on the vehicle's real-time operating conditions. Its core function is to balance three major objectives: system energy efficiency, dynamic response, and user comfort, ensuring that the output target pressure correction coefficient adapts to the current operating condition priority. Specific operating conditions refer to particular operational scenarios formed by the vehicle based on its real-time operating status (such as environmental conditions, driving status, and load demand), serving as the basis for dynamically adjusting the weight coefficients. Typical scenarios include urban congestion, high-temperature exposure, full-load operation, and low-temperature heating. The target pressure correction coefficient can be the output of the weighted adaptive fusion function, representing a proportional coefficient that corrects the baseline optimal exhaust pressure reference value based on the weight allocation of the current operating condition scenario. The transient pressure compensation term can be an additional pressure adjustment value generated based on the rate of change of load demand characteristics. Its core function is to compensate for the lag in the response of the baseline optimal exhaust pressure reference value to transient load changes, ensuring that the dynamically optimal exhaust pressure target value can quickly adapt to sudden increases or decreases in load.

[0050] Specifically, traditional solutions have significant drawbacks in correcting the baseline optimal exhaust pressure reference value: They employ a fixed-weight fusion logic, which fails to adapt to the core needs of different operating conditions. For example, in an emergency cooling scenario with a high temperature of 40°C, an interior temperature of 35°C, and five passengers, a fixed energy efficiency-priority weight leads to a lag in exhaust pressure adjustment, taking 30 minutes to lower the interior temperature to 24°C, severely impacting comfort. Conversely, in a high-speed, constant-speed, energy efficiency-priority scenario with temperatures close to the set value, a fixed comfort weight maintains high exhaust pressure, resulting in over 25% energy waste. Furthermore, traditional solutions lack a transient compensation mechanism; when the load suddenly increases (e.g., from two to five passengers) or decreases, the reference value cannot respond promptly, leading to insufficient cooling or temperature overshoot. To address the above issues, this step first constructs a weighted adaptive fusion function based on load demand characteristics, incorporating three priority modes: system energy efficiency (baseline coefficient 0.9), dynamic response (baseline coefficient 1.1), and user comfort (baseline coefficient 1.0). The function achieves multi-mode fusion through weighted summation. Subsequently, real-time operating data (ambient temperature 40℃, vehicle speed 5km / h, in-vehicle temperature 35℃, 5 occupants) is analyzed and identified as "high-temperature congestion and full-load condition." The weighting coefficients are dynamically adjusted to 0.5 for response, 0.3 for comfort, and 0.2 for energy efficiency, and then calculated using (0.9 × 0.2 + 1.1 × 0.0). The target pressure correction coefficient of 1.04 is calculated by adding 0.5 + 1.0 × 0.3. Then, temperature difference data (10℃ and 12℃) for two consecutive 1-second cycles are collected, and the rate of change (12-10)℃ / 1s = 2℃ / s is calculated. A transient pressure compensation term of 0.4MPa is generated according to the preset correspondence (1℃ / s corresponds to 0.2MPa). Finally, the basic optimal exhaust pressure reference value of 10MPa is obtained, which is corrected by 10MPa × 1.04 to obtain 10.4MPa. The 0.4MPa compensation term is added, and the final output is a dynamic optimal exhaust pressure target value of 10.8MPa.

[0051] The method provided in this embodiment constructs a weighted adaptive fusion function to dynamically adjust the weights and combines the rate of change to generate a transient compensation term, so that the pressure target value accurately matches the working condition priority and transient changes, thereby making up for the shortcomings of the traditional "fixed logic and lack of response" from the root and achieving multi-objective balance.

[0052] In some embodiments, the dynamic optimal exhaust pressure target value is decomposed into a compressor-dominant control target value and an electronic expansion valve auxiliary control target value; the dynamic coupling relationship between the compressor-dominant control target value and the electronic expansion valve auxiliary control target value is identified; based on the dynamic coupling relationship, the influence of the compressor-dominant control target value is decoupled and compensated, and the compensated auxiliary control target value is fused with the compressor-dominant control target value to generate a set of coordinated control command information.

[0053] The compressor-dominant control target value can be the control target parameter achieved primarily by the compressor as the core power component after decomposing the dynamically optimal discharge pressure target value according to the system control division of labor; specifically, it is expressed as a compressor speed command. The electronic expansion valve auxiliary control target value can be the control target parameter achieved primarily by the electronic expansion valve as a flow regulation component after decomposing the dynamically optimal discharge pressure target value; specifically, it is expressed as an electronic expansion valve opening command. Dynamic coupling relationship refers to the dynamic correlation characteristics of mutual influence and constraint between the compressor-dominant control target value and the electronic expansion valve auxiliary control target value. Decoupling compensation refers to the process of calculating and applying targeted adjustments based on the identified dynamic coupling relationship to offset the interference of changes in one control target on the control effect of the other, making the control processes of the two relatively independent and non-interfering.

[0054] Specifically, traditional solutions lack a "dominant-auxiliary" control division of labor, neglect the dynamic coupling relationship between the compressor and the electronic expansion valve, and mostly adopt independent control or static compensation, which has serious defects: for example, when the dynamic optimal discharge pressure target is 10MPa, the compressor increases its speed according to the pressure increase command, but the electronic expansion valve does not adjust its opening at the same time, resulting in a mismatch between refrigerant flow and compression intensity, and the pressure fluctuates between 9.2-10.8MPa for a long time; when it is necessary to increase the pressure from 9MPa to 11MPa and the compressor speed from 1800rpm to 2200rpm, the traditional solution without coupling compensation causes the pressure to overshoot to 11.5MPa and then quickly drop back to 10.3MPa, which aggravates component wear. To address the above issues, this step first obtains the dynamic optimal exhaust pressure target value of 10.8 MPa. Based on the functional positioning of "compressor power-driven and electronic expansion valve flow-assisted," the target is decomposed as follows: the compressor-driven control target is to increase the speed from the current 1900 rpm to 2300 rpm (verified to achieve a 10.5 MPa pressure increase), and the electronic expansion valve-assisted control target is to adjust the opening from 42% to 46% (with the auxiliary matching flow remaining at 0.3 MPa). Based on the thermodynamic principle of CO2 transcritical cycle and 100 sets of historical operating data, the dynamic coupling relationship of "for every 100 rpm increase in compressor speed, the electronic expansion valve opening needs to increase by 1% synchronously" is identified. Since the speed needs to increase by 400 rpm, a 4% decoupling compensation is calculated, and after superposition, the final opening of the electronic expansion valve is 50%. Finally, the decoupled compressor speed command of 2300 rpm and electronic expansion valve opening command of 50% are merged and encapsulated to generate a collaborative control command information set, ensuring that the two operate synchronously and avoiding coupling interference.

[0055] The method provided in this embodiment, based on target decomposition, clear division of labor, identification of coupling relationships, dynamic decoupling compensation, and generation of collaborative instructions, can avoid action conflicts and pressure fluctuations, ensure that the exhaust pressure accurately meets the standards, and fundamentally make up for the shortcomings of traditional technologies that are "undivided, uncoupled, and asynchronous".

[0056] In some embodiments, based on the first principles of thermodynamics and historical operating data, a dynamic coupling transfer function is constructed to characterize the change in the target value of the compressor's dominant control and the change in the target value of the electronic expansion valve's auxiliary control. During real-time operation, the parameters of the dynamic coupling transfer function are updated online using a recursive least squares algorithm with a forgetting factor to track the changes in the system's dynamic characteristics with operating conditions. Based on the updated dynamic coupling transfer function, the dynamic coupling gain matrix at the current operating point is analyzed, and the gain value and phase lag information of its main coupling channel are extracted as quantitative coupling strength characteristic indicators. Based on the quantitative coupling strength characteristic indicators, the coupling strength level between the compressor and the electronic expansion valve is evaluated, and a dynamic decoupling trigger command and compensation priority signal are generated according to a preset strength threshold to guide subsequent compensation calculations.

[0057] The first principle of thermodynamics can be the law of conservation of energy, the core of which is that energy cannot be created or destroyed, but only transformed from one form to another, or transferred from one object to another, with the total energy remaining constant during the transformation and transfer. Historical operating data refers to various related operating parameters recorded by the system data acquisition module during the past operation of the two-stage gas-injection enthalpy-increasing compressor and the electronic expansion valve. The recursive least squares algorithm with a forgetting factor can be an optimization algorithm for online parameter estimation. It introduces a forgetting factor into the traditional recursive least squares algorithm, assigning higher weights to new data and lower weights to old data to achieve dynamic updates of model parameters, adapting to the time-varying changes in system characteristics. The dynamic coupling gain matrix can be a matrix-form parameter extracted at the current operating point based on the updated dynamic coupling transfer function, where each element represents the influence intensity (gain) and phase relationship between the change in one control objective and the change in another control objective. The coupling strength level can be a classification of the strength of the coupling between the compressor and the electronic expansion valve based on quantified coupling strength characteristic indicators and preset strength thresholds, typically divided into three levels: weak coupling, moderate coupling, and strong coupling. The preset strength threshold can be a reference value used to classify the strength levels of coupling effects, including weak-to-moderate coupling thresholds and moderate-to-strong coupling thresholds. Its core function is to establish the correspondence between quantitative indicators and strength levels. The dynamic decoupling trigger command can be a control signal generated based on the coupling effect strength level. It is triggered when the coupling strength level reaches or exceeds the preset threshold, initiating the corresponding decoupling compensation calculation process. The compensation priority signal can be an auxiliary signal generated alongside the dynamic decoupling trigger command, used to clarify the execution priority of decoupling compensation (e.g., level one, level two, level three). Higher priority indicates that decoupling compensation should be executed faster. Its core function is to ensure that coupling interference is handled preferentially when multiple tasks are running in parallel.

[0058] Specifically, traditional solutions often use offline fixed models to identify coupling relationships, lack online parameter update mechanisms, and rely on experience to judge the strength of coupling, which has serious defects. For example, a fixed coupling model built based on a normal operating condition of 25°C may still output a 1% compensation basis when the vehicle enters a high-temperature and strong sunlight condition of 40°C. In reality, "for every 100 rpm increase in compressor speed, the opening of the electronic expansion valve should increase by 1.5%", but the model will not update the compensation basis, resulting in insufficient decoupling and pressure fluctuations exceeding 0.8MPa. After 3 years of operation, component aging causes the coupling gain to drop from 0.9 to 0.7. The fixed model is not updated, and compensation based on the original parameters leads to excess flow and a decrease in energy efficiency of more than 20%. To address the above issues, this step, based on the first principles of thermodynamics and incorporating 1000 sets of historical operating data (including compressor speed changes and electronic expansion valve opening changes under different environments and loads), constructs a dynamic coupling transfer function (e.g., ΔK = 0.01Δn + 0.005Δn', where Δn' is the rate of change of compressor speed) with the input being the compressor speed change (Δn) and the output being the electronic expansion valve opening change (ΔK). A recursive least squares algorithm with a forgetting factor of 0.98 is then employed (98% weight for new data and 2% for old data). Each time a new set of data is collected... (For example, if the speed increases from 1900rpm to 2000rpm and the opening increases from 42% to 43.2%), the model parameters are recursively corrected (e.g., 0.01 is updated to 0.012); the main channel gain of 0.012 and phase lag of 0.2 seconds are extracted from the coupling gain matrix of the current operating point and compared with the preset thresholds (weak-medium 0.008, medium-strong 0.015), and it is determined to be medium coupling; combined with the system's "transient load growth" condition, a secondary decoupling trigger command and corresponding priority signal are generated to provide quantitative basis and execution guidance for subsequent compensation calculations.

[0059] The method provided in this embodiment, based on physical principles and online algorithms, constructs a dynamic model, quantifies coupling strength, and clarifies triggers and priorities, enabling accurate identification of coupling relationships. This provides a scientific basis for decoupling compensation and fundamentally addresses the shortcomings of traditional "offline fixed, unquantified, and unadaptive" methods.

[0060] In some embodiments, a dynamic coupling model is established to characterize the impact of compressor speed command changes on the required opening degree of the electronic expansion valve. When the compressor's dominant control target value changes, the electronic expansion valve opening degree compensation amount required to offset the impact of the target value change on the refrigerant flow rate on the evaporator side and the stability of the system pressure is analyzed in real time based on the dynamic coupling model. The electronic expansion valve opening degree compensation amount is superimposed on the corresponding opening degree of the reference opening degree command obtained based on the analysis of the electronic expansion valve's auxiliary control target value to generate the final opening degree command of the electronic expansion valve that is resistant to coupling interference.

[0061] The dynamic coupling model is a mathematical model built based on dynamic coupling relationships. It specifically characterizes the degree of influence and response law of compressor speed command changes on the required opening degree of the electronic expansion valve, and can quantify the dynamic correspondence between speed changes and opening degree adaptation. The compressor-dominant control target value can be the core control parameter obtained from decomposition, specifically the compressor speed command. Its core function is to adjust the speed to guide the system discharge pressure towards the dynamic optimal target value. The evaporator-side refrigerant flow rate refers to the volume or mass flow rate of refrigerant flowing through the evaporator in a CO2 transcritical cycle system. It directly affects the heat exchange efficiency of the evaporator and the system's cooling / heating capacity, and is one of the core parameters for maintaining system thermal balance. System pressure stability shock refers to the phenomenon where, when the compressor-dominant control target value changes, the coupling relationship causes a mismatch between the evaporator-side refrigerant flow rate and compression intensity, leading to a sudden rise, fall, or fluctuation in system discharge pressure and evaporation pressure, which can disrupt system operational stability. The electronic expansion valve opening compensation amount can be an additional opening adjustment value calculated based on the dynamic coupling model. Its core function is to offset the impact of compressor speed changes on evaporator-side flow rate and system pressure, ensuring that the electronic expansion valve opening is adapted to the new compressor speed. The reference opening command can be an initial opening command derived from the auxiliary control target value of the electronic expansion valve, combined with the current environmental conditions and load requirements. It reflects the ideal opening state of the electronic expansion valve when there is no coupling interference.

[0062] Specifically, traditional decoupling compensation uses static fixed rules or empirical estimations without dynamic model support, which has serious defects: for example, the preset "for every 200 rpm increase in speed, the opening increases by 5%", under the high temperature condition of 38℃, the actual increase is 7% to match the flow rate. Fixed compensation leads to insufficient flow rate and pressure fluctuations exceeding 0.6MPa; while under the low temperature condition of -5℃, only 3% increase is needed. Excessive compensation leads to excess flow rate, and the evaporation pressure is too low, causing frost formation and a decrease in energy efficiency of more than 25%. To address the above issues, this step utilizes the dynamic coupling relationship (the opening needs to increase by 1% for every 100 rpm increase in speed at high temperatures and by 0.8% at low temperatures). Combining this with the thermodynamic characteristics of the CO2 transcritical cycle, a dynamic coupling model is constructed, with the compressor speed change as the input and the electronic expansion valve opening adaptation as the output (e.g., ΔK=0.01Δn for high-temperature conditions). Real-time monitoring shows the compressor's primary control target value adjusting from 2000 rpm to 2400 rpm (Δn=400 rpm). The current environmental sensor determines it to be a 38℃ high-temperature condition, triggering decoupling compensation calculations. The model calculates the opening compensation as 4% (0.01×400 rpm). The decomposed electronic expansion valve auxiliary control target value is obtained, and combined with the 38℃ high temperature and the full-load cooling demand of 5 people, a 45% baseline opening command is derived. The 4% compensation is superimposed on the 45% baseline opening command to generate a 49% anti-coupling interference electronic expansion valve final opening command, ensuring precise adaptation to the compressor's 2400 rpm speed and avoiding insufficient flow and pressure surges.

[0063] By using the method provided in this embodiment, based on the establishment of a dynamic coupling model, the accurate compensation amount is calculated in real time and superimposed with the reference opening degree, which can offset the impact of compressor speed changes on flow and pressure. This fundamentally makes up for the shortcomings of traditional "fixed compensation, empirical estimation, and poor adaptability", ensuring the stable and efficient operation of the system.

[0064] In some embodiments, measurable disturbance information of ambient temperature and vehicle speed is acquired, and the real-time changing trend of the measurable disturbance information is identified. Based on the magnitude and direction of the real-time changing trend, predictive correction quantities are generated for the compressor speed command and electronic expansion valve opening command in the cooperative control command information set through a feedforward compensator. The predictive correction quantities are superimposed on the control quantities of the corresponding control commands to form preliminary disturbance rejection control commands. The compressor speed command in the preliminary disturbance rejection control commands is subjected to time-series inertial smoothing filtering, wherein the filtering intensity is dynamically adjusted according to the system operating state to generate a smoothed final compressor speed drive command. The final compressor speed drive command and the final electronic expansion valve opening drive command are combined and encapsulated to generate a final actuator drive command information set.

[0065] Measurable disturbance information refers to external or internal parameters that can be directly acquired by sensors, are quantifiable, and can interfere with the operating state of a CO2 transcritical cycle system. Its core function is to provide clear disturbance input for feedforward predictive compensation. Real-time change trends refer to the direction and magnitude of changes in measurable disturbance information (ambient temperature, vehicle speed) within a continuous acquisition period, such as rapid temperature increases, constant speed increases, or sudden speed decreases. Its core function is to determine the intensity and urgency of the disturbance's impact. Predictive corrections are additional control adjustments generated by the feedforward compensator based on the real-time change trends of the measurable disturbance. Their core function is to preemptively correct the coordinated control commands, adapting the control actions to the system state changes caused by the disturbance. Preliminary disturbance rejection control commands are transitional commands formed by superimposing the predictive corrections onto the control quantities corresponding to the coordinated control commands. Their key characteristic is that they already include preemptive compensation for the measurable disturbance, initially offsetting the disturbance interference. Timing-inertial smoothing filtering is a signal processing algorithm used to suppress high-frequency fluctuations in control commands. Its core function is to reduce the system impact caused by sudden changes in commands by weighting and smoothing continuous cycle control commands, while dynamically adjusting the filtering strength according to the system state. The final compressor speed drive command can be the compressor control command after being processed by timing-inertial smoothing filtering. Its core feature is that it cancels measurable disturbance interference and suppresses high-frequency fluctuations, enabling stable compressor operation. The final electronic expansion valve opening drive command can be the final execution command after feedforward compensation of the electronic expansion valve opening command in the preliminary disturbance rejection control command (no additional filtering required, or only light filtering). Its core feature is that it adapts to disturbance changes and compressor speed commands, ensuring stable refrigerant flow.

[0066] Specifically, traditional solutions lack a feedforward disturbance rejection mechanism for measurable disturbances, relying solely on feedback control for post-event adjustments. Furthermore, the use of a fixed filter strength has serious drawbacks: for example, when a vehicle moves from a 28°C shaded area into a 38°C exposed road surface, the temperature rises by 10°C within 10 seconds. Feedback control requires more than 20 seconds to offset the decrease in heat exchange efficiency, leading to an increase in the vehicle's interior temperature. When the vehicle speed suddenly increases from 10km / h to 100km / h, traditional solutions fail to adapt to the change in wind speed in advance, causing pressure fluctuations. Simultaneously, a fixed filter strength cannot balance steady-state and transient conditions—in steady-state conditions, weak filtering causes compressor speed fluctuations within 2000rpm±50rpm, increasing wear; in transient conditions, strong filtering prevents the compressor speed from increasing in time, resulting in a slow cooling response. To address the above issues, this step utilizes an ambient temperature sensor (collecting data from 25℃ to 32℃ to 38℃) and the vehicle's CAN bus (acquiring data from 30km / h to 45km / h to 60km / h) to collect measurable disturbance information at a frequency of 1 time per second. It analyzes the trend of rapid temperature increase (13℃ in 3 seconds) and constant vehicle speed increase. The feedforward compensator is then activated, generating predictive corrections of +130rpm for engine speed and +2% for vehicle speed increase (based on a preset correlation of +50rpm for every 5℃ temperature increase and +2% for every 30km / h speed increase). These corrections are then overlaid with a coordinated index. The command (speed 2000 rpm, opening 45%) is used to obtain the initial anti-disturbance command (2130 rpm, 47%). The speed command is then subjected to a time-series inertial smoothing filter with a preset initial strength of 0.5. When the exhaust pressure deviation is detected to be 0.08 MPa (less than the steady-state threshold of 0.1 MPa), the filter strength is adjusted to 0.8. The final speed command of 2130 rpm is obtained by smoothing three consecutive cycles of commands (2130 rpm, 2128 rpm, 2132 rpm) by weighted average. This final speed command of 2130 rpm is then encapsulated with the 47% opening command to form the final actuator drive command information set.

[0067] The method provided in this embodiment generates predictive correction quantities based on the real-time changing trend of measurable disturbances by using a feedforward compensator, thereby achieving "early intervention" for disturbances such as ambient temperature and vehicle speed, avoiding the lag of traditional feedback control. Disturbances such as sudden temperature rises and changes in vehicle speed can be promptly offset, ensuring that the system's operating state is not affected by external interference, and significantly improving the system's stability in complex disturbance environments.

[0068] In some embodiments, the system monitors the real-time deviation between the exhaust pressure and the target value of the dynamic optimal exhaust pressure. When the real-time deviation is less than the steady-state pressure tolerance threshold, the system is determined to be in steady-state operation, and the filtering strength of the time-series inertial smoothing filter is enhanced to suppress high-frequency fluctuations in the command. When the real-time deviation is greater than or equal to the dynamic adjustment trigger threshold, or when the vehicle load demand characteristic is identified to be in a rapid change phase, the system is determined to be in transient operation, and the filtering strength of the time-series inertial smoothing filter is weakened to improve the response speed of the control command.

[0069] The steady-state pressure tolerance threshold can be a preset critical deviation value for determining when the system enters steady-state operation. Its core function is to define the deviation range within which "system operation is stable and pressure fluctuations are acceptable." Steady-state operation can be defined as an operating state where the real-time deviation of the system exhaust pressure is less than the steady-state pressure tolerance threshold, and there are no rapid changes in the vehicle load demand characteristics. Its core characteristics are small fluctuations in system parameters (pressure, flow rate, speed, etc.) and smooth operation. For example, when a vehicle is traveling at a constant speed of 90 km / h on a highway, the ambient temperature is stable at 25℃, and the interior temperature is maintained at a set 24℃, the pressure deviation is stable at 0.05 MPa. The dynamic adjustment trigger threshold can be a preset critical deviation value for determining when the system enters transient operation. Its core function is to define the deviation range within which "system operation deviates from the target and requires rapid response," and its value is greater than the steady-state pressure tolerance threshold. Transient operation can be a state in which the real-time deviation of the system exhaust pressure is greater than or equal to the dynamic adjustment trigger threshold, or when the vehicle load demand characteristics are in a rapidly changing phase. The core characteristic is that the system needs to quickly adjust the control commands to adapt to the changes in demand. For example, if a vehicle suddenly drives from an underground garage at -10°C into an open road at 35°C, the temperature inside the vehicle rises rapidly, and the pressure deviation increases instantly to 0.18MPa.

[0070] Specifically, the traditional solution uses a fixed filter strength to process control commands, which has serious adaptation defects: During steady-state operation (e.g., a vehicle traveling at a constant speed of 90 km / h in a stable environment of 25°C), a fixed filter strength of 0.5 cannot suppress high-frequency fluctuations in compressor speed (±30 rpm), exacerbating bearing wear and wasting energy; during transient operation (e.g., a vehicle moving from a -10°C garage to a 35°C open-air environment, or a sudden increase in load), a fixed filter strength of 0.5 will inhibit command adjustment, preventing the speed from quickly increasing from 2000 rpm to 2300 rpm, resulting in the interior temperature taking 15 minutes to drop to the set value. To address these issues, this step activates the real-time monitoring module, collecting the system exhaust pressure of 10.75 MPa from the compressor exhaust port pressure sensor, calling the dynamic optimal exhaust pressure target value of 10.8 MPa from claim 4, and calculating the real-time deviation absolute value of 0.05 MPa; simultaneously, it collects the vehicle load demand characteristics (set temperature 24°C, actual temperature 25°C), calculating the load change rate of 0.1°C / second. The system invokes preset thresholds (steady-state pressure tolerance threshold 0.1 MPa, dynamic adjustment trigger threshold 0.12 MPa, and rapid change threshold 0.5 °C / sec). Since both the deviation and the rate of change are less than the thresholds, it is determined to be steady-state operation. The initial strength of the timing inertial smoothing filter is increased from 0.5 to 0.8, and the speed commands for three consecutive cycles (2000 rpm, 2002 rpm, and 1998 rpm) are smoothed using a weighted average algorithm to suppress high-frequency fluctuations. If the subsequent deviation increases to 0.15 MPa or the rate of change reaches 0.6 °C / sec, it is determined to be transient, and the filter strength is reduced to 0.3 to ensure that the speed can be quickly increased from 2000 rpm to 2300 rpm, thereby improving the response speed.

[0071] The method provided in this embodiment effectively suppresses high-frequency oscillations in compressor speed commands during steady-state operation, avoids frequent compressor start-stop or sudden speed changes, reduces mechanical wear on core components such as bearings and motors, lowers operating noise and energy waste, and extends the service life of the actuator.

[0072] In some embodiments, based on the final compressor speed drive command and the final electronic expansion valve opening drive command, a drive range verification and dynamic limiting based on multi-constraint safety boundaries are performed to generate a safety drive command. Based on the real-time performance data generated by the system operation driven by the safety drive command, closed-loop performance evaluation and online gradient fine-tuning are performed on the benchmark data points in the optimal efficiency point pressure mapping table, the weight coefficients in the weighted adaptive fusion function, and the filter intensity parameters of the time-series inertial smoothing filter. The mapping table, function, and parameter combination after closed-loop performance evaluation and online gradient fine-tuning are encapsulated into a bipolar compressor control optimization information set.

[0073] Safety-driven instructions can be the final executable instructions generated after drive range verification and dynamic limiting. Their core characteristic is full compliance with multi-constraint safety boundary requirements, preserving the original instruction's control intent while eliminating safety risks. Real-time performance data refers to the system operating parameters and performance indicators collected in real-time by sensors and data acquisition modules during the operation of the safety-driven system. Closed-loop performance evaluation refers to the process of quantifying the suitability of current control parameters (mapping tables, weighting coefficients, filter strength) by comparing them with real-time performance data, using preset system performance indicators (such as energy efficiency ratio thresholds, response time thresholds, and pressure fluctuation thresholds) as a benchmark. Online gradient fine-tuning refers to the process of optimizing control parameters in real-time using a small-step, gradual adjustment method based on the closed-loop performance evaluation results.

[0074] Specifically, traditional solutions lack comprehensive safety verification and online parameter optimization mechanisms, resulting in serious flaws: directly executing drive commands can easily lead to safety risks. For example, sending a 3200rpm speed command directly to the compressor exceeds its 3000rpm safety limit, causing the motor to overheat and burn out; a 98% over-opening command for the electronic expansion valve may cause jamming or frosting. Furthermore, fixed offline core parameters cannot adapt to the system's time-varying characteristics. For instance, after three years of component degradation, continuing to use the original pressure mapping table leads to an energy efficiency decrease of over 30%; fixed weighting coefficients cannot adapt to heating demands in low-temperature regions, resulting in delayed response. To address the above issues, this step obtains the final drive command: compressor speed 3200 rpm, electronic expansion valve opening 98%. The drive range is checked using preset multi-constraint safety boundaries (speed 1000-3000 rpm, opening 5%-95%). If the command exceeds the limits, it is dynamically adjusted to 3000 rpm and 95% to generate a safe drive command. After the drive system is running, real-time performance data (energy efficiency 2.8, pressure fluctuation 0.3 MPa, response time 8 seconds) is collected and compared with preset benchmarks (≥3.0, ≤±0.2 MPa, ≤6 seconds) in a closed-loop evaluation to identify optimization space. Online gradient fine-tuning is used, adjusting the pressure mapping table benchmark value from 10 MPa to 10.1 MPa in 0.05 MPa steps, the dynamic response weight coefficient from 0.5 to 0.55, and the steady-state filter strength from 0.8 to 0.75. Finally, the fine-tuned mapping table, functions, and parameters are encapsulated to generate a bipolar compressor control optimization information set.

[0075] By using the method provided in this embodiment, through the drive range verification and dynamic limiting of multiple constraint safety boundaries, drive commands that exceed the safety limits of the actuator and system are effectively filtered out, avoiding safety risks such as compressor overspeed, electronic expansion valve over-opening, and system overpressure. Furthermore, the actuator always operates within the design safety range, reducing mechanical wear, electrical overload and other faults, extending the service life of core components such as the compressor and electronic expansion valve, and improving the reliability and safety of system operation.

Claims

1. A control method for a two-stage enthalpy-increasing compressor based on a transcritical carbon dioxide cycle, characterized in that, include: Acquire a set of real-time vehicle operating status information, and based on the set of real-time vehicle operating status information, perform multi-dimensional system status characterization and optimal target fusion analysis to generate a dynamic optimal exhaust pressure target value; Based on the target value of the dynamic optimal exhaust pressure, the dynamic coupling and decoupling of the main and auxiliary control channels and the allocation of collaborative commands are performed to generate a set of collaborative control command information for the compressor and the electronic expansion valve. Based on the aforementioned collaborative control instruction information set, feedforward prediction compensation and timing inertial smoothing management are performed in conjunction with measurable disturbance information to generate the final actuator drive instruction information set after disturbance rejection and smoothing. Based on the final actuator drive instruction information set, online fine-tuning learning of safety drive instructions and parameters is performed to generate a bipolar compressor control optimization information set.

2. The method according to claim 1, characterized in that, The generation of the dynamic optimal exhaust pressure target information set includes: The real-time vehicle operating status information set includes environmental load characteristics, core heat exchange efficiency characteristics, and vehicle load demand characteristics. The environmental load characteristics and the core heat exchange efficiency characteristics are input into the optimal efficiency point pressure mapping table to obtain the basic optimal exhaust pressure reference value that matches the current operating conditions. Based on the load demand characteristics, target fusion analysis is performed on the basic optimal exhaust pressure reference value to output the dynamic optimal exhaust pressure target value.

3. The method according to claim 2, characterized in that, The optimal efficiency point pressure mapping table includes: Based on bench tests of a two-stage gas-injection enthalpy-enhancing compressor under CO2 transcritical cycles, exhaust pressure data points were collected in an environmental simulation chamber covering the full operating spectrum of the vehicle, under different combinations of the environmental load characteristics and the core heat exchange efficiency characteristics, to achieve a local optimum in the overall energy efficiency ratio of the system. The exhaust pressure data point set is processed into a continuous surface using a surface fitting algorithm based on physical model constraints to generate an initial two-dimensional optimal efficiency pressure reference surface. By introducing the system transient response data and long-term operating performance degradation coefficient under actual road cycles, dynamic robustness correction and durability compensation are performed on the two-dimensional optimal efficiency pressure reference surface to generate the final multi-dimensional, time-varying coupled optimal efficiency point pressure mapping table.

4. The method according to claim 3, characterized in that, The target fusion analysis of the baseline optimal exhaust pressure reference value includes: Based on the load demand characteristics, a weighted adaptive fusion function is constructed that includes a system energy efficiency priority mode, a dynamic response priority mode, and a user comfort priority mode. The system identifies the specific working conditions of the vehicle in real time, dynamically adjusts the weight coefficients of each mode in the weight adaptive fusion function, and generates the target pressure correction coefficient for the current scenario. The rate of change of the load demand characteristics is analyzed in real time, and a transient pressure compensation term is generated based on this rate of change. The rate of change represents the dynamic trend and urgency of load demand. The target pressure correction coefficient is multiplied by the basic optimal exhaust pressure reference value, and the transient pressure compensation term is added to output the dynamic optimal exhaust pressure target value.

5. The method according to claim 4, characterized in that, The generated set of coordinated control command information for the compressor and electronic expansion valve includes: The dynamic optimal exhaust pressure target value is decomposed into the compressor-dominant control target value and the electronic expansion valve auxiliary control target value; Identify the dynamic coupling relationship between the compressor's primary control target value and the electronic expansion valve's auxiliary control target value; Based on the dynamic coupling relationship, the influence of the compressor's dominant control target value is decoupled and compensated, and the compensated auxiliary control target value is fused with the compressor's dominant control target value to generate the collaborative control instruction information set.

6. The method according to claim 5, characterized in that, The identification of the dynamic coupling relationship between the compressor's primary control target value and the electronic expansion valve's auxiliary control target value includes: Based on the first principles of thermodynamics and historical operating data, a dynamic coupling transfer function is constructed to characterize the change in the target value of the compressor's dominant control and the change in the target value of the electronic expansion valve's auxiliary control. During real-time operation, the parameters of the dynamic coupling transfer function are updated online using a recursive least squares algorithm with a forgetting factor to track the changes in the system's dynamic characteristics with operating conditions. Based on the updated dynamic coupling transfer function, the dynamic coupling gain matrix at the current operating point is analyzed, and the gain value and phase lag information of its main coupling channel are extracted as quantitative coupling strength characteristic indicators. Based on the quantitative coupling strength characteristic index, the coupling strength level between the compressor and the electronic expansion valve is evaluated, and a dynamic decoupling trigger command and compensation priority signal are generated according to the preset strength threshold to guide subsequent compensation calculations.

7. The method according to claim 5, characterized in that, The decoupling compensation calculation based on the dynamic coupling relationship includes: A dynamic coupling model is established to characterize the effect of compressor speed command changes on the required opening degree of the electronic expansion valve; When the target value of the compressor's main control changes, the electronic expansion valve opening compensation amount required to offset the impact of the change in the target value on the refrigerant flow on the evaporator side and the stability of the system pressure is analyzed in real time based on the dynamic coupling model. The electronic expansion valve opening compensation amount is superimposed on the corresponding opening of the reference opening command obtained based on the analysis of the electronic expansion valve auxiliary control target value to generate the final opening command of the electronic expansion valve that is resistant to coupling interference.

8. The method according to claim 7, characterized in that, The generated anti-disturbance smoothed final actuator drive instruction information set includes: Acquire measurable disturbance information of ambient temperature and vehicle speed, and identify the real-time changing trend of the measurable disturbance information; Based on the magnitude and direction of the real-time change trend, predictive corrections are generated for the compressor speed command and electronic expansion valve opening command in the coordinated control command information set through a feedforward compensator. The predictive correction values ​​are respectively superimposed on the control values ​​of the corresponding control commands to form preliminary anti-disturbance control commands; The compressor speed command in the preliminary disturbance rejection control command is subjected to time-series inertial smoothing filtering, wherein the filtering intensity is dynamically adjusted according to the system operating state to generate the smoothed final compressor speed drive command; The final compressor speed drive command and the final electronic expansion valve opening drive command are combined and encapsulated to generate the final actuator drive command information set.

9. The method according to claim 8, characterized in that, The filtering intensity is dynamically adjusted according to the system operating status, including: The real-time deviation between the exhaust pressure of the monitoring system and the target value of the dynamic optimal exhaust pressure; When the real-time deviation is less than the steady-state pressure tolerance threshold, the system is determined to be in steady-state operation, and the filtering strength of the timing inertial smoothing filter is enhanced to suppress high-frequency fluctuations in the command. When the real-time deviation is greater than or equal to the dynamic adjustment trigger threshold, or when the vehicle load demand characteristic is detected to be in a rapid change phase, the system is determined to be in transient operation, and the filtering intensity of the time-series inertial smoothing filter is reduced to improve the response speed of the control command.

10. The method according to claim 9, characterized in that, The generation of the bipolar compressor control optimization information set includes: Based on the final compressor speed drive command and the final electronic expansion valve opening drive command, the drive range verification and dynamic limiting based on multi-constraint safety boundaries are performed to generate a safety drive command; Based on the real-time performance data generated by the system driven by the safety driving command, closed-loop performance evaluation and online gradient fine-tuning are performed on the benchmark data points in the optimal efficiency point pressure mapping table, the weight coefficients in the weight adaptive fusion function, and the filter intensity parameters of the time-series inertial smoothing filter. The mapping table, functions, and parameter combinations after the closed-loop performance evaluation and online gradient fine-tuning are encapsulated into the bipolar compressor control optimization information set.

Images