Multi-condition-oriented hydraulic cooler pressure self-adaptive regulation optimization method and system

Abstract

This invention relates to the field of hydraulic cooling control technology, and more particularly to a method for adaptive pressure regulation and optimization of hydraulic coolers under multiple operating conditions. This method monitors the operating status in real time and calculates instantaneous operating condition characteristics, automatically identifying the current operating condition category using a condition classifier. For different operating conditions, a corresponding optimization objective function is invoked, aiming to maximize cooling efficiency and minimize pressure fluctuations, to solve for the optimal inlet pressure under constraints. The response speed is determined based on the adjustment deviation and the rate of change of operating conditions. An adaptive feedback control mechanism is used to generate control signals to dynamically compensate for delays and nonlinear effects, driving the actuator to converge the pressure to the optimal value. This invention achieves adaptive optimization and regulation of hydraulic cooler pressure under varying operating conditions, improving cooling efficiency and system stability.

Description

Technical Field

[0001] This invention relates to the field of hydraulic cooling control technology, and in particular to a method and system for adaptive pressure regulation and optimization of hydraulic coolers under multiple operating conditions. Background Technology

[0002] Hydraulic coolers are key thermal management components in hydraulic systems. Their core function is to dissipate the heat generated by the hydraulic oil during operation into the environment through heat exchange, maintaining the hydraulic oil temperature within a reasonable range and ensuring stable and reliable system operation. Hydraulic systems operate under complex and variable conditions. For example, when the actuator is under no-load, light-load, heavy-load, or frequently started and stopped, the system's heat generation power, hydraulic oil flow rate, and temperature will all change significantly. These changes directly affect the state of the oil flowing through the cooler, placing dynamic demands on the cooler's heat dissipation performance and internal pressure distribution.

[0003] In existing technical practices, pressure regulation of hydraulic coolers typically employs on / off control based on a fixed threshold or a simple proportional-integral-derivative (PDI-D) control strategy. A common approach is to compare the cooler outlet oil temperature feedback from a single sensor with a preset temperature threshold. When the oil temperature exceeds the threshold, the system flow rate is altered by opening a bypass valve or adjusting the pump's displacement, thereby indirectly affecting the cooler inlet pressure. Another approach is to use a predetermined pressure-flow curve for feedforward compensation based on the pressure signal from the system's main pump, attempting to maintain relative stability in the cooler piping pressure. These methods treat the cooler as a static or quasi-static component.

[0004] However, the above-mentioned conventional approach has significant drawbacks. Relying on only a single parameter and with fixed control logic, it is difficult to accurately reflect the actual heat load and flow state of the cooler under varying operating conditions. The cooler's heat dissipation efficiency has a strong nonlinear coupling relationship with the internal pressure field, flow rate, and oil temperature. Fixed control parameters cannot maintain optimal control performance under different operating conditions such as sudden load changes and ambient temperature variations. This often leads to lag or over-adjustment in pressure regulation response, resulting in low cooling efficiency or increased pressure fluctuations.

[0005] Specifically, lag and improper pressure regulation can trigger a series of problems. Excessive pressure fluctuations can accelerate fatigue damage to the internal fins and pipes of the cooler, affecting its service life. Simultaneously, suboptimal pressure conditions can lead to a decrease in the heat transfer coefficient, resulting in insufficient heat dissipation capacity of the cooler under some operating conditions, potentially causing the hydraulic oil temperature to exceed the safe range; while under other operating conditions, excessive cooling or excessive flow resistance can cause unnecessary energy loss. Existing methods lack intelligent identification of operating conditions and adaptive optimization capabilities for different conditions, hindering further improvements in the performance and reliability of hydraulic coolers in complex application scenarios. Summary of the Invention

[0006] The present invention provides a method and system for adaptive pressure regulation and optimization of hydraulic coolers for multiple operating conditions, which can solve the problems in the prior art.

[0007] A first aspect of the present invention provides a method for adaptive pressure regulation and optimization of a hydraulic cooler for multiple operating conditions, comprising: The operating status information of the hydraulic cooler is monitored in real time, and the instantaneous operating condition characteristics of the hydraulic cooler are calculated based on the operating status information. The instantaneous operating condition characteristics include pressure drop, flow-pressure ratio, and temperature-pressure coupling coefficient. The instantaneous operating condition features are input into the operating condition classifier. The operating condition classifier automatically identifies the operating condition by establishing a discrimination boundary between the operating condition features and predefined operating condition categories, and outputs the operating condition category to which the current operating condition belongs. For the aforementioned operating condition category, the corresponding pressure optimization objective function is invoked. The pressure optimization objective function aims to maximize cooling efficiency and minimize pressure fluctuations, using inlet pressure as the optimization variable. Under the constraints of the pressure optimization objective function, the optimal inlet pressure value is solved. The constraints include the pressure tolerance limit, flow regulation range, and temperature safety threshold of the hydraulic cooler. The adjustment deviation between the optimal inlet pressure value and the current inlet pressure value is calculated. Based on the adjustment deviation and the rate of change of operating conditions, the response speed of pressure regulation is determined. An adaptive feedback control mechanism is used to generate a pressure regulation control signal. The adaptive feedback control mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The pressure regulation control signal is output to the pressure regulation actuator, which drives the pressure regulation actuator to perform pressure regulation action, so that the inlet pressure converges to the optimal inlet pressure value, thereby realizing the pressure adaptive regulation and optimization of the hydraulic cooler under multiple working conditions.

[0008] The instantaneous operating condition characteristics of the hydraulic cooler are calculated based on the aforementioned operating status information, including: Extract the fluid pressure value measured by the inlet pressure sensor and the fluid pressure value measured by the outlet pressure sensor, and calculate the difference between the two to obtain the pressure drop; Obtain the fluid flow rate value measured by the flow sensor, and calculate the flow rate-pressure ratio by comparing the fluid flow rate value with the pressure drop; The system collects the cooling medium temperature and ambient temperature values ​​measured by a temperature sensor, calculates the temperature difference between the cooling medium temperature and the ambient temperature, couples the temperature difference with the pressure drop, and constructs a temperature-pressure coupling coefficient by establishing the response relationship between temperature change and pressure change. The pressure drop, the flow-pressure ratio, and the temperature-pressure coupling coefficient are output as instantaneous operating condition features to the operating condition classifier.

[0009] The temperature difference and the pressure drop are coupled in a calculation, and a temperature-pressure coupling coefficient is constructed by establishing the response relationship between temperature change and pressure change, including: Within a preset time window, historical change sequences of the cooling medium temperature and pressure drop are collected. These historical change sequences contain paired data of temperature and pressure drop at multiple time sampling points. The temperature value sequence in the historical change sequence is differentiated in the time domain to obtain the temperature change rate sequence, and the pressure drop value sequence is differentiated in the time domain to obtain the pressure change rate sequence; Calculate the correlation metric between the temperature change rate sequence and the pressure change rate sequence, wherein the correlation metric reflects the synchronous response strength of temperature fluctuations and pressure fluctuations; Based on the correlation metric and the current instantaneous values ​​of the temperature difference and the pressure drop, a temperature-pressure coupling coefficient is generated through a weighted fusion operation, in which the correlation metric is used as a weighting factor to adjust the coupling strength. The temperature-pressure coupling coefficient is output as a condition characteristic quantity characterizing the thermodynamic response of the hydraulic cooler.

[0010] The instantaneous operating condition features are input into the operating condition classifier. The classifier automatically identifies the operating condition by establishing a discrimination boundary between the operating condition features and predefined operating condition categories, and outputs the operating condition category to which the current operating condition belongs, including: The pressure drop, the flow-pressure ratio, and the temperature-pressure coupling coefficient are used to construct a multi-dimensional operating condition feature vector. The multidimensional working condition feature vector is mapped to the working condition feature space, which is a multidimensional geometric space composed of each working condition feature quantity as a coordinate axis. In the operating condition feature space, a pre-trained discrimination boundary model is invoked. This discrimination boundary model learns the region division rules for different operating condition categories in the operating condition feature space through historical operating condition data. Calculate the spatial distance between the multidimensional working condition feature vector and the discrimination boundary of each predefined working condition category. The spatial distance reflects the membership degree of the current working condition feature vector to each working condition category. The predefined working condition category corresponding to the discrimination boundary with the smallest spatial distance is selected as the working condition category to which the current working condition belongs.

[0011] In the operating condition feature space, a pre-trained discrimination boundary model is invoked. This discrimination boundary model learns regional division rules for different operating condition categories in the operating condition feature space through historical operating condition data, including: Collect historical operating status information of the hydraulic cooler under various operating conditions, extract historical operating condition feature quantities from the historical operating status information, and label the corresponding operating condition category. The historical working condition features and the working condition category labels are combined to form a training sample set, which contains typical distribution patterns of each working condition category in the working condition feature space. A supervised learning mechanism is used to train the discrimination boundary model. Through iterative optimization, the discrimination boundary model is made to form a hyperplane or nonlinear decision surface in the working condition feature space that can accurately separate different working condition categories. A boundary tolerance mechanism is introduced during training. This mechanism enhances the model's robustness to fluctuations in operating conditions by setting soft-margin regions for the discrimination boundary. Verify the accuracy of the discrimination boundary model in identifying working conditions on the test sample set. Once the accuracy meets the preset performance index, deploy the discrimination boundary model into the working condition classifier for real-time working condition identification.

[0012] For the aforementioned operating condition category, the corresponding pressure optimization objective function is invoked. This objective function aims to maximize cooling efficiency and minimize pressure fluctuations, using inlet pressure as the optimization variable, and includes: According to the operating condition category, the corresponding pressure optimization objective function is retrieved from a pre-established operating condition-objective function mapping library, which stores mathematical expressions for cooling efficiency and pressure fluctuation under different operating condition categories; A cooling efficiency target term is constructed, which establishes a functional relationship between cooling efficiency and inlet pressure based on the heat transfer relationship between the fluid flow rate, the cooling medium temperature, and the inlet pressure. A pressure fluctuation target term is constructed, which quantifies pressure stability by calculating the variance or standard deviation of the inlet pressure over a time series. The cooling efficiency objective and the pressure fluctuation objective are weighted and combined to form a multi-objective optimization function. The weight coefficients of each objective in the multi-objective optimization function are adaptively configured according to the priority requirements of the operating condition category. The inlet pressure is set as the optimization variable of the multi-objective optimization function, and the multi-objective optimization function is output as the pressure optimization objective function to the optimization solution module.

[0013] An adaptive feedback control mechanism is used to generate the pressure regulation control signal. This mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator, including: The actual response data of the pressure regulating actuator during the pressure regulating action is collected in real time. The actual response data includes the position feedback value, velocity feedback value, and real-time measurement value of the inlet pressure of the actuator. The control error is calculated based on the deviation between the actual response data and the optimal inlet pressure value. The trend of the control error over time is analyzed to identify the characteristic patterns of control delay and nonlinear effects. An actuator response characteristic model is established, which describes the delay time constant and nonlinear gain characteristics of the actuator by fitting the dynamic relationship between the actual response data and the control command. An adaptive compensation strategy is designed based on the aforementioned response characteristic model. This adaptive compensation strategy counteracts the control delay and the nonlinear effects by introducing a predictive lead and a nonlinear inverse compensation term into the control signal. The control error, the adjustment deviation, and the response speed are input to the adaptive feedback controller, which, in conjunction with the adaptive compensation strategy, generates a pressure regulation control signal and outputs it to the pressure regulation actuator.

[0014] A second aspect of the present invention provides a multi-condition adaptive pressure regulation and optimization system for hydraulic coolers, comprising: The monitoring and calculation unit is used to monitor the operating status information of the hydraulic cooler in real time, and calculate the instantaneous operating condition characteristics of the hydraulic cooler based on the operating status information. The instantaneous operating condition characteristics include pressure drop, flow-pressure ratio and temperature-pressure coupling coefficient. The working condition identification unit is used to input the instantaneous working condition feature quantity into the working condition classifier. The working condition classifier realizes automatic working condition identification by establishing a discrimination boundary between the working condition feature quantity and the predefined working condition category, and outputs the working condition category to which the current working condition belongs. The target invocation unit is used to invoke the corresponding pressure optimization objective function for the operating condition category. The pressure optimization objective function takes maximizing cooling efficiency and minimizing pressure fluctuation as optimization objectives, and uses inlet pressure as optimization variable. Under the constraints of the pressure optimization objective function, the optimal inlet pressure value is solved. The constraints include the pressure bearing limit, flow regulation range and temperature safety threshold of the hydraulic cooler. The optimization solution unit is used to calculate the adjustment deviation between the optimal inlet pressure value and the current inlet pressure value, determine the pressure regulation response speed based on the adjustment deviation and the rate of change of operating conditions, and generate a pressure regulation control signal using an adaptive feedback control mechanism. The adaptive feedback control mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The deviation calculation unit is used to output the pressure regulation control signal to the pressure regulation actuator, drive the pressure regulation actuator to perform pressure regulation action, so that the inlet pressure converges to the optimal inlet pressure value, and realize the pressure adaptive regulation optimization of the hydraulic cooler under multiple working conditions.

[0015] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0017] This method can monitor the operating status of the hydraulic cooler in real time and calculate key operating condition characteristics, including pressure drop, flow-pressure ratio, and temperature-pressure coupling coefficient. These characteristics comprehensively reflect the dynamic thermo-hydraulic properties of the system, providing a multi-dimensional data foundation for accurate operating condition identification. By automatically identifying the current operating condition category through an operating condition classifier, it effectively overcomes the problem of misjudgment caused by traditional methods relying on a single parameter or empirical threshold, significantly improving the system's adaptability to complex and variable operating conditions and its identification accuracy.

[0018] For each identified specific operating condition category, the corresponding pressure optimization objective function is invoked for solution. This function, with the core objectives of maximizing cooling efficiency and minimizing pressure fluctuations, calculates the optimal inlet pressure value under constraints such as equipment pressure limits, flow rate adjustment ranges, and temperature safety thresholds. This process achieves a precise match between the optimization objective and the operating condition, ensuring that the system can automatically find a pressure setpoint that balances efficient cooling and stable operation under different loads and thermal conditions, thereby improving energy efficiency and equipment reliability globally.

[0019] The pressure regulation response speed is dynamically determined based on the deviation between the optimal pressure value and the current value, as well as the rate of change of operating conditions. The adopted adaptive feedback control mechanism can monitor the actual response of the actuator in real time and dynamically compensate for control delays and nonlinear characteristics. This effectively suppresses overshoot and oscillations caused by system inertia or actuator nonlinearity, making the pressure regulation process smoother, faster, and more stable, and enhancing the system's robustness against disturbances.

[0020] Ultimately, the generated pressure regulation control signal drives the actuator to move, causing the inlet pressure to converge quickly and accurately to the optimal value. The entire process forms a closed-loop optimized control system from condition perception and intelligent decision-making to precise execution, achieving fully automatic and adaptive optimization adjustment of the hydraulic cooler pressure setting under multiple operating conditions. This not only ensures that the cooling system always operates within its high-efficiency range, extending equipment lifespan, but also reduces reliance on manual intervention, improving the overall intelligence level and comprehensive performance of the hydraulic system. Attached Figure Description

[0021] Figure 1 A flowchart illustrating an optimization method for adaptive pressure regulation of hydraulic coolers under multiple operating conditions; Figure 2 This is a flowchart illustrating adaptive feedback control. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0024] Figure 1 This is a flowchart illustrating the adaptive pressure regulation and optimization method for hydraulic coolers under multiple operating conditions according to an embodiment of the present invention. Figure 1 As shown, the adaptive pressure regulation optimization method for hydraulic coolers under multiple operating conditions includes: The operating status information of the hydraulic cooler is monitored in real time, and the instantaneous operating condition characteristics of the hydraulic cooler are calculated based on the operating status information. The instantaneous operating condition characteristics include pressure drop, flow-pressure ratio, and temperature-pressure coupling coefficient. The instantaneous operating condition features are input into the operating condition classifier. The operating condition classifier automatically identifies the operating condition by establishing a discrimination boundary between the operating condition features and predefined operating condition categories, and outputs the operating condition category to which the current operating condition belongs. For the aforementioned operating condition category, the corresponding pressure optimization objective function is invoked. The pressure optimization objective function aims to maximize cooling efficiency and minimize pressure fluctuations, using inlet pressure as the optimization variable. Under the constraints of the pressure optimization objective function, the optimal inlet pressure value is solved. The constraints include the pressure tolerance limit, flow regulation range, and temperature safety threshold of the hydraulic cooler. The adjustment deviation between the optimal inlet pressure value and the current inlet pressure value is calculated. Based on the adjustment deviation and the rate of change of operating conditions, the response speed of pressure regulation is determined. An adaptive feedback control mechanism is used to generate a pressure regulation control signal. The adaptive feedback control mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The pressure regulation control signal is output to the pressure regulation actuator, which drives the pressure regulation actuator to perform pressure regulation action, so that the inlet pressure converges to the optimal inlet pressure value, thereby realizing the pressure adaptive regulation and optimization of the hydraulic cooler under multiple working conditions.

[0025] In one optional implementation, calculating the instantaneous operating characteristic quantities of the hydraulic cooler based on the operating status information includes: Extract the fluid pressure value measured by the inlet pressure sensor and the fluid pressure value measured by the outlet pressure sensor, and calculate the difference between the two to obtain the pressure drop; Obtain the fluid flow rate value measured by the flow sensor, and calculate the flow rate-pressure ratio by comparing the fluid flow rate value with the pressure drop; The system collects the cooling medium temperature and ambient temperature values ​​measured by a temperature sensor, calculates the temperature difference between the cooling medium temperature and the ambient temperature, couples the temperature difference with the pressure drop, and constructs a temperature-pressure coupling coefficient by establishing the response relationship between temperature change and pressure change. The pressure drop, the flow-pressure ratio, and the temperature-pressure coupling coefficient are output as instantaneous operating condition features to the operating condition classifier.

[0026] In practical applications, hydraulic coolers are typically installed in the hydraulic circuits of engineering machinery, marine power systems, or large industrial equipment, and need to adapt to complex operating conditions such as sudden load changes, ambient temperature fluctuations, and frequent flow rate adjustments. To accurately capture the operating characteristics of hydraulic coolers under different working conditions, it is necessary to collect and process their operating status information in real time, and extract quantitative indicators that can reflect the essential characteristics of the working conditions.

[0027] The inlet and outlet pressure sensors are installed at the inlet and outlet of the hydraulic cooler, respectively. Both sensors use the same pressure measurement standard and sampling frequency to ensure time synchronization of the measurement data. to The pressure fluctuates within a certain range, with the specific value depending on the workload of the hydraulic system and the output characteristics of the pump source. The outlet pressure sensor measures the hydraulic oil pressure after passing through the internal flow channels and heat exchange core of the cooler. Due to flow channel resistance, heat exchange tube bundle resistance, and local resistance losses experienced by the fluid inside the cooler, the outlet pressure is inevitably lower than the inlet pressure. The inlet pressure sensor's measurement value is synchronously read through the data acquisition system. Compared with the measured value of the outlet pressure sensor The pressure drop can be obtained by calculating the difference between the two. The calculation formula is: Pressure drop The pressure drop is a key parameter reflecting the internal flow resistance and the degree of blockage in the heat exchange core of a hydraulic cooler. When fouling occurs in the internal flow channels of the cooler, the heat exchange tubes become partially blocked, or the fluid viscosity increases due to a drop in temperature, the pressure drop will increase significantly. Under normal operating conditions, the pressure drop is typically maintained at... to If the pressure drops beyond the range, This may indicate that the cooler needs cleaning and maintenance.

[0028] The flow sensor is installed on the inlet or bypass line of the hydraulic cooler, and uses a turbine, electromagnetic, or ultrasonic flow meter to continuously measure the fluid flow rate. The flow sensor outputs the flow rate value. Measured in liters per minute or cubic meters per hour, this value directly reflects the volume of hydraulic oil passing through the cooler per unit time. Under heavy-load conditions, the hydraulic system requires a larger flow rate to meet the actuator's operational needs; the flow rate may reach [missing value]. to Under light load or standby conditions, the flow rate may decrease to... to To establish the correlation between flow rate and pressure drop, the flow sensor measurement values ​​were... With the calculated pressure drop Perform a ratio calculation to obtain the flow-pressure ratio. The calculation formula is: The physical meaning of this ratio is the flow capacity corresponding to a unit pressure drop, which can be understood as the effective flow coefficient of the cooler channel.

[0029] When the cooler flow channels are unobstructed and the heat exchange core is clean, a relatively small pressure drop can achieve a large flow rate; at this time, the flow rate-pressure ratio is [value missing]. A larger value indicates a lower flow rate; conversely, if there is a blockage inside the cooler or an abnormally high fluid viscosity, a greater pressure drop is required to maintain the same flow rate, resulting in a lower flow rate-pressure ratio. A significant decrease. By monitoring the trend of this ratio, signs of cooling fluid performance degradation can be effectively identified, providing an important basis for operating condition classification.

[0030] The temperature sensor system includes two independent measuring points: a cooling medium temperature sensor and an ambient temperature sensor. The cooling medium temperature sensor is typically installed in the cooling water or cooling air passage on the cooler housing side to measure the actual temperature of the cooling medium responsible for heat exchange. For water-cooled hydraulic coolers, the temperature of the cooling medium is typically within... to Scope: For air-cooled hydraulic coolers, the air temperature passing through the radiator surface must be measured. An ambient temperature sensor is installed at a representative location in the equipment room or work area to measure the surrounding air temperature. This temperature value is affected by seasonal variations, diurnal temperature range, and ventilation conditions in the equipment compartment. The data acquisition system simultaneously reads the measurements from two temperature sensors and calculates the temperature difference between the cooling medium temperature and the ambient temperature. The formula is Temperature difference This reflects the thermal potential difference between the cooling medium and the environment, which directly affects the heat exchange driving force and cooling efficiency of the cooler.

[0031] To establish the response relationship between temperature and pressure changes, it is necessary to analyze how temperature fluctuations affect the internal pressure distribution of the cooler through changes in hydraulic oil viscosity, alterations in the flow state of the cooling medium, and the thermal expansion effect of the heat exchanger tube bundle. Hydraulic oil viscosity decreases with increasing temperature; according to empirical formulas, for every increase in temperature... The viscosity of the hydraulic oil decreases by approximately to Decreased viscosity directly leads to reduced flow resistance, which in turn causes a change in pressure drop. An increase in the temperature of the cooling medium reduces the heat exchange temperature difference, forcing the system to maintain cooling by increasing the flow rate or raising the inlet pressure; this adjustment also alters the pressure drop. This is achieved by collecting temperature difference changes over a time window. With pressure drop change Establish temperature-pressure coupling coefficient The calculation model. The specific calculation method is as follows: in continuous... Within each sampling period, the temperature difference of each period is recorded separately. With pressure drop Calculate the change between adjacent periods: as well as Then, the temperature-pressure coupling coefficient is obtained by fitting using the least squares method. The dimensions of this system are pressure units divided by temperature units, and its value reflects the sensitivity of the pressure system to temperature changes. Under high temperature and high viscosity conditions, even a small change in temperature can cause significant pressure fluctuations. The absolute value is relatively large; however, under normal temperature and low viscosity conditions, the effect of temperature fluctuations on pressure is relatively mild. The absolute value is relatively small.

[0032] After completing the calculation of the above three characteristic quantities, the data processing module will reduce the pressure. Flow-pressure ratio and temperature-pressure coupling coefficient Combined into an instantaneous operating condition feature vector: This eigenvector comprehensively characterizes the current operating state of the hydraulic cooler, where pressure drop reflects the flow resistance level, the flow-pressure ratio reflects the flow channel's guiding capacity, and the temperature-pressure coupling coefficient reflects the interactive response characteristics of thermodynamics and fluid mechanics. Eigenvector The data is then input into a load condition classifier. Based on a pre-trained discrimination boundary model, the classifier categorizes the current load condition into predefined categories such as high load, medium load, low load, cold start, or abnormal load. The execution cycle of the entire feature calculation process is typically set to [time period missing]. to This ensures timely capture of rapid changes in operating conditions while avoiding data redundancy and wasted computing resources due to excessively high sampling frequencies. Sensor measurement data undergoes filtering before entering the calculation process, employing moving average filtering or Kalman filtering algorithms to remove random noise and transient interference, ensuring the stability and reliability of the characteristic quantity calculation results.

[0033] In one optional implementation, the temperature difference and the pressure drop are coupled in a calculation, and a temperature-pressure coupling coefficient is constructed by establishing the response relationship between temperature change and pressure change, including: Within a preset time window, historical change sequences of the cooling medium temperature and pressure drop are collected. These historical change sequences contain paired data of temperature and pressure drop at multiple time sampling points. The temperature value sequence in the historical change sequence is differentiated in the time domain to obtain the temperature change rate sequence, and the pressure drop value sequence is differentiated in the time domain to obtain the pressure change rate sequence; Calculate the correlation metric between the temperature change rate sequence and the pressure change rate sequence, wherein the correlation metric reflects the synchronous response strength of temperature fluctuations and pressure fluctuations; Based on the correlation metric and the current instantaneous values ​​of the temperature difference and the pressure drop, a temperature-pressure coupling coefficient is generated through a weighted fusion operation, in which the correlation metric is used as a weighting factor to adjust the coupling strength. The temperature-pressure coupling coefficient is output as a condition characteristic quantity characterizing the thermodynamic response of the hydraulic cooler.

[0034] During the operation of a hydraulic cooler, a complex coupling relationship exists between temperature and pressure. Changes in the temperature of the cooling medium directly affect the fluid's viscosity characteristics, thereby altering flow resistance and causing corresponding changes in pressure drop. To accurately characterize this thermodynamic coupling, a coupling coefficient is needed to dynamically reflect the response of temperature changes to pressure changes. The calculation process of this coupling coefficient fully considers the dynamic evolution of time-series data, achieving accurate quantification of the instantaneous thermodynamic state through differential analysis and correlation mining of historical data.

[0035] Before starting the calculations, the length of the preset time window is first set. The selection of this time window needs to comprehensively consider the thermal response time constant of the hydraulic cooler and the typical period of pressure fluctuations. For medium-sized hydraulic coolers, the time window is typically set to 30 to 120 seconds to ensure that the complete thermodynamic response process can be captured. Within this time window, the inlet and outlet temperatures of the cooling medium and the pressure drop measured by the pressure sensor are acquired through high-frequency sampling. The sampling frequency is set to 10 to 50 times per second to ensure that the temporal resolution of the data accurately reflects the rapidly changing operating conditions.

[0036] The collected historical change sequences form two sets of time series data, denoted as the temperature value sequence and the pressure drop value sequence, respectively. The temperature value sequence contains the temperature difference of the cooling medium at all sampling points within the time window, with each value corresponding to a precise timestamp. The pressure drop value sequence records the pressure difference between the inlet and outlet of the hydraulic cooler at the same timestamp. These two sets of sequences are perfectly aligned on the time axis, forming a paired dataset, providing a data foundation for subsequent coupled analysis. During data acquisition, the raw signal needs to be preprocessed, including filtering out high-frequency noise and removing anomalous jump points. A moving average filter is used to smooth the raw data, and the filter window length is dynamically adjusted according to the sampling frequency to ensure that effective dynamic information is not lost while suppressing noise.

[0037] After data acquisition, time-domain differentiation is performed on the temperature value sequence. The purpose of differentiation is to extract the rate of temperature change over time, which reflects the dynamic evolution of the thermal state of the cooling medium. Using numerical differentiation, the temperature increment is calculated by differentiating the temperature values ​​between adjacent sampling points. Dividing this increment by the sampling time interval yields the rate of temperature change at that moment. To improve the stability of the differentiation calculation, a central difference scheme is used instead of a simple forward difference scheme. This involves using the temperature values ​​from one sampling point before and after the current moment for the difference calculation, effectively reducing the impact of single-point measurement errors on the differentiation results. By performing point-by-point differentiation on the entire temperature value sequence, a temperature change rate sequence is generated. This sequence is two sampling points shorter than the original sequence because the central difference scheme cannot be used for the first and last points of the sequence.

[0038] The same differential processing method was applied to the pressure drop sequence. The pressure difference between adjacent sampling points in the pressure drop sequence was calculated using differential methods to obtain the rate of change of pressure drop over time. The pressure change rate reflects the dynamic changes in the internal flow resistance of the hydraulic cooler; when the flow rate fluctuates or the internal state of the flow channel changes, the pressure change rate will exhibit significant fluctuations. Using the same central difference scheme as the temperature sequence, the consistency of the numerical calculation methods for the two change rate sequences was ensured. The generated pressure change rate sequence and temperature change rate sequence correspond strictly to each other on the time axis, laying the foundation for subsequent correlation analysis.

[0039] After obtaining the temperature and pressure change rate sequences, the correlation between them is calculated. The Pearson correlation coefficient is used for quantification, effectively reflecting the degree of linear correlation between the two time series. The calculation process involves first calculating the mean for each of the two change rate sequences, then calculating the deviation of each sampling point from the mean. The deviations of the temperature and pressure change rates are multiplied point by point and summed to obtain the numerator of the covariance. The denominator is the square root of the product of the sum of the squares of the deviations of the two sequences. The resulting correlation coefficient ranges from -1 to +1. A positive correlation coefficient indicates that the pressure drop tends to increase with rising temperature, while a negative correlation coefficient indicates an inverse response relationship. The larger the absolute value of the correlation coefficient, the stronger the coupling relationship.

[0040] Under certain operating conditions, the response relationship between temperature and pressure may exhibit a time lag; that is, the pressure change caused by a temperature change is not instantaneous but occurs with a certain time delay. To capture this lag effect, time lag analysis is introduced when calculating the correlation metric. Specifically, the pressure change rate sequence is shifted along the time axis, and the correlation coefficient is calculated for different lag times. The lag time corresponding to the largest absolute value of the correlation coefficient is selected as the optimal lag, and the correlation coefficient at this lag is used as the final correlation metric. This approach more accurately reflects the true response characteristics of the temperature-pressure coupling and avoids underestimating the coupling strength due to ignoring the lag effect.

[0041] After obtaining the correlation metric, a weighted fusion calculation is performed by combining the instantaneous values ​​of temperature difference and pressure drop at the current moment. The instantaneous temperature difference value comes directly from real-time monitoring data, representing the temperature difference between the inlet and outlet of the cooling medium at the current moment. The instantaneous pressure drop value also comes from the measurement value of the real-time pressure sensor, representing the pressure loss of the hydraulic cooler at the current moment. The ratio of the instantaneous temperature difference value to the instantaneous pressure drop value is used as the basic coupling term. The physical meaning of this ratio is the temperature change amplitude corresponding to a unit pressure drop. The larger the value, the greater the temperature reduction can be achieved under the same pressure loss, reflecting the heat exchange efficiency of the cooler.

[0042] In the weighted fusion calculation, the correlation metric is used as a weighting factor to adjust the coupling strength. When the absolute value of the correlation metric is large, it indicates a significant synchronous response relationship between temperature and pressure changes. In this case, the basic coupling term should be given a higher weight so that the temperature-pressure coupling coefficient more sensitively reflects this strong coupling characteristic. Conversely, when the absolute value of the correlation metric is small, it indicates that temperature and pressure changes are relatively independent. The weight of the coupling term should be reduced to avoid misjudging random fluctuations as a strong coupling effect. The specific weighted fusion formula is designed as follows: the basic coupling term is multiplied by the absolute value of the correlation metric, and a normalization coefficient is added to ensure that the numerical range of the coupling coefficient is within a reasonable range. The normalization coefficient is calibrated based on the design parameters of the hydraulic cooler and the statistical characteristics of historical operating data to make the coupling coefficients of different models of coolers comparable.

[0043] In the weighted fusion process, the sign of the correlation metric also needs to be considered. Positive correlation indicates that an increase in temperature is accompanied by an increase in pressure drop. This typically occurs when increased flow rate leads to increased flow velocity, resulting in increased flow resistance and a higher pressure drop. Simultaneously, the higher flow rate brings stronger heat exchange, leading to a greater temperature decrease. Negative correlation may occur when fluid viscosity changes significantly with temperature; increased temperature leads to decreased viscosity, reducing flow resistance and thus lowering the pressure drop. To accurately reflect this physical mechanism, the sign information of the correlation metric is retained when calculating the temperature-pressure coupling coefficient, using the positive or negative sign to distinguish different coupling mechanism types.

[0044] The generated temperature-pressure coupling coefficient, as a comprehensive operating condition feature, can simultaneously contain instantaneous state information of temperature and pressure, as well as their dynamic response relationship, within a single value. The magnitude of this coefficient reflects the strength of thermodynamic coupling under the current operating condition, while the sign of the value reflects the direction of coupling. In subsequent operating condition classification and pressure optimization processes, the temperature-pressure coupling coefficient, along with pressure drop and flow-pressure ratio, constitutes a multi-dimensional operating condition feature vector, providing rich discriminative information for the operating condition classifier. Compared to using temperature or pressure parameters alone, introducing the coupling coefficient can more comprehensively characterize the operating state of the hydraulic cooler. Especially during rapid switching of operating conditions or abnormal fluctuations, changes in the coupling coefficient are often more sensitive than those of a single parameter, helping to identify potential operational risks in advance.

[0045] The calculated temperature-pressure coupling coefficient is output to the set of operating condition characteristics, forming a complete instantaneous operating condition feature vector together with the simultaneously calculated pressure drop and flow-pressure ratio. This feature vector is then input into the operating condition classifier for operating condition category identification, providing an accurate basis for subsequent pressure optimization and adaptive adjustment. The entire calculation process of the temperature-pressure coupling coefficient fully utilizes the statistical characteristics of historical data and the real-time information of instantaneous measurements. It captures dynamic trends through differential analysis, reveals the inherent coupling mechanism through correlation analysis, and achieves the organic integration of multiple information through weighted fusion, ultimately forming key operating condition characteristics that can accurately characterize the thermodynamic response of the hydraulic cooler.

[0046] In one optional implementation, the instantaneous operating condition feature quantity is input into an operating condition classifier. The operating condition classifier automatically identifies the operating condition by establishing a discrimination boundary between the operating condition feature quantity and a predefined operating condition category, and outputs the operating condition category to which the current operating condition belongs, including: The pressure drop, the flow-pressure ratio, and the temperature-pressure coupling coefficient are used to construct a multi-dimensional operating condition feature vector. The multidimensional working condition feature vector is mapped to the working condition feature space, which is a multidimensional geometric space composed of each working condition feature quantity as a coordinate axis. In the operating condition feature space, a pre-trained discrimination boundary model is invoked. This discrimination boundary model learns the region division rules for different operating condition categories in the operating condition feature space through historical operating condition data. Calculate the spatial distance between the multidimensional working condition feature vector and the discrimination boundary of each predefined working condition category. The spatial distance reflects the membership degree of the current working condition feature vector to each working condition category. The predefined working condition category corresponding to the discrimination boundary with the smallest spatial distance is selected as the working condition category to which the current working condition belongs.

[0047] Hydraulic coolers face complex situations with frequent changes in operating conditions during actual operation. To achieve precise adaptive pressure regulation, it is necessary to accurately identify the current operating condition first. After obtaining three core characteristic quantities—pressure drop, flow-pressure ratio, and temperature-pressure coupling coefficient—these quantities are arranged in a preset order to form a multi-dimensional operating condition feature vector. The construction process of this vector must ensure the uniformity of the dimensions of each characteristic quantity to avoid imbalance in feature weights due to differences in numerical scales. To eliminate the influence of dimensions, each characteristic quantity is normalized using a maximum-minimum normalization method, mapping all feature values ​​to the interval between 0 and 1. The three normalized feature values ​​then form the feature vector.

[0048] The constructed multi-dimensional operating condition feature vectors need to be mapped to a specially designed operating condition feature space for analysis. This feature space is a three-dimensional geometric space with normalized pressure drop, flow-pressure ratio, and temperature-pressure coupling coefficient as three orthogonal coordinate axes. In this space, each operating condition corresponds to a unique spatial point location. Statistical analysis of a large amount of historical operating data reveals that different operating condition categories exhibit obvious clustering distribution characteristics in this feature space. For example, high-load operating conditions typically correspond to larger pressure drops and lower flow-pressure ratios, concentrated in specific regions of the feature space; while low-load operating conditions show the opposite characteristic distribution pattern, occupying another independent region; and the feature points of variable load transition operating conditions are scattered in the transition zone between high-load and low-load regions. This spatial distribution pattern provides a physical basis for establishing a discriminant boundary model.

[0049] The training process of the discrimination boundary model relies on the sufficient accumulation of historical working condition data. In the initial stage of system operation, it is necessary to manually label the working condition categories at each time point and record the corresponding feature vectors to gradually build a labeled dataset. Once the labeled data has accumulated to a sufficient scale, the discrimination boundary model is trained using the Support Vector Machine (SVM) algorithm. The SVM identifies the optimal classification hyperplane to delineate the boundaries between different working condition categories in the feature space. For multi-classification problems in three-dimensional feature space, a "one-to-many" strategy is used to construct multiple binary classifiers, each responsible for distinguishing a specific working condition category from all other categories. After training, the discrimination boundary model can divide the feature space into several non-overlapping sub-regions, each corresponding to a predefined working condition category. The boundary determination not only considers classification accuracy but also introduces the concept of boundary width. Narrower boundaries are set in high-confidence classification regions, while the boundaries are appropriately widened in ambiguous areas where working conditions intersect, improving the robustness of classification.

[0050] In the actual working condition identification process, the multi-dimensional working condition feature vectors acquired in real time are substituted into the trained discrimination boundary model for calculation. The core of the calculation is to determine the distance relationship between the position of the feature vector in space and the discrimination boundary of each category. For each predefined working condition category, its discrimination boundary can be represented as a classification hyperplane or a nonlinear boundary surface in the feature space. The geometric distance calculation method from point to hyperplane is adopted. For linear discrimination boundaries, the spatial distance can be directly obtained by the point-to-plane distance formula; for nonlinear boundaries, it is necessary to calculate the Euclidean distance from the feature vector to the nearest point on the boundary surface. The physical meaning of the distance value is to reflect the degree to which the current working condition features deviate from the typical feature center of each category. The smaller the distance, the higher the similarity between the current working condition and the category. In the actual calculation, it is also necessary to consider the scale difference in different directions in the feature space, introduce Mahalanobis distance to replace simple Euclidean distance, and correct the weights of each dimension of features through the covariance matrix to make the distance metric more consistent with the physical characteristics of the actual working condition.

[0051] After calculating the spatial distance between the current feature vector and the discrimination boundaries of each predefined operating condition category, these distance values ​​are sorted and compared. Assuming the system predefines five operating condition categories—high load, medium load, low load, variable load increase, and variable load decrease—the distances from the current feature vector to the discrimination boundaries of these five categories are calculated and denoted as d_1, d_2, d_3, d_4, and d_5, respectively. By comparing these five distance values, the smallest distance is selected; the operating condition category corresponding to this distance is the discrimination result for the current operating condition. To avoid misjudgments in areas with ambiguous boundaries, a distance threshold mechanism is set. When the difference between the smallest and second smallest distances is less than a preset threshold, it is determined that the current operating condition is in a transition phase, and the operating condition category output from the previous moment is temporarily maintained to avoid frequent switching that could cause control strategy oscillations. Simultaneously, the discrimination results for multiple consecutive sampling periods are recorded. The operating condition output is only officially switched after a certain operating condition category appears more than a set number of times, further improving recognition stability.

[0052] In the distance calculation process, the concept of membership degree is introduced to quantify the classification results. Membership degree is defined as the probability that the current feature vector belongs to a certain working condition category, obtained by normalizing the inverse of the distance. Specifically, the sum of the inverses of the distances to the discrimination boundaries of each category is used as the normalization factor, and the membership degree of each category is equal to its inverse distance divided by the normalization factor. The sum of the components of the membership degree vector obtained in this way is 1, and the larger the membership degree, the higher the probability of belonging. When selecting working condition categories, not only the category with the smallest distance is considered, but also its membership degree value is taken into account. When the maximum membership degree exceeds 0.7, the classification confidence is considered high, and the corresponding working condition category can be directly output; when the maximum membership degree is between 0.5 and 0.7, the temporal consistency is checked by combining the historical working condition sequence; when the maximum membership degree is below 0.5, the fuzzy working condition processing mechanism is triggered, and the control parameters are determined by a weighted fusion strategy of multiple adjacent working conditions.

[0053] The output of the working condition classifier includes not only explicit working condition category labels but also classification confidence indices and feature space location information. The confidence indices are calculated from the difference between the maximum and second-largest membership degrees; a larger difference indicates more typical current working condition characteristics and a more reliable classification result. Feature space location information records the coordinates of the current feature vector in three-dimensional space, used for subsequent analysis of working condition evolution trajectories and prediction of working condition change trends. By continuously accumulating working condition identification results, the motion trajectories of working condition feature points in space can be plotted, revealing regular paths of working condition switching and providing data support for further optimization of the discrimination boundary model and improvement of identification accuracy. Upon output of the identification results, the corresponding pressure optimization subroutine for the working condition category is immediately triggered, ensuring that the pressure regulation strategy can quickly respond to changes in working conditions and achieve true multi-working-condition adaptive control. The entire working condition identification process, from feature vector input to category output, has a computation time strictly controlled within 20 milliseconds, meeting the timing requirements of real-time control of hydraulic systems.

[0054] In one optional implementation, a pre-trained discrimination boundary model is invoked in the operating condition feature space. This discrimination boundary model learns region division rules for different operating condition categories in the operating condition feature space through historical operating condition data, including: Collect historical operating status information of the hydraulic cooler under various operating conditions, extract historical operating condition feature quantities from the historical operating status information, and label the corresponding operating condition category. The historical working condition features and the working condition category labels are combined to form a training sample set, which contains typical distribution patterns of each working condition category in the working condition feature space. A supervised learning mechanism is used to train the discrimination boundary model. Through iterative optimization, the discrimination boundary model is made to form a hyperplane or nonlinear decision surface in the working condition feature space that can accurately separate different working condition categories. A boundary tolerance mechanism is introduced during training. This mechanism enhances the model's robustness to fluctuations in operating conditions by setting soft-margin regions for the discrimination boundary. Verify the accuracy of the discrimination boundary model in identifying working conditions on the test sample set. Once the accuracy meets the preset performance index, deploy the discrimination boundary model into the working condition classifier for real-time working condition identification.

[0055] In the multi-condition operation scenarios of hydraulic coolers, the condition classifier needs to rely on a discriminant boundary model to accurately identify the current condition. The construction of the discriminant boundary model essentially involves mining the distribution patterns of different condition categories in the condition feature space through a large amount of historical operating data, thereby forming a mathematical mapping relationship that enables automatic classification. This process involves multiple technical aspects, including data acquisition, feature extraction, model training, boundary optimization, and performance verification.

[0056] The data acquisition phase needs to cover all typical operating conditions that hydraulic coolers may encounter in practical applications. For hydraulic coolers in the construction machinery field, typical operating conditions include heavy-load start-up, high-speed cruising, frequent reversing, low-speed high-torque, and idling standby conditions. During acquisition, the sensor network continuously records operating status information such as pressure, flow rate, and temperature at the inlet and outlet of the hydraulic cooler, while also recording auxiliary information such as ambient temperature, load changes, and system operating frequency. The data acquisition frequency is set to 100Hz to 500Hz to ensure the capture of rapidly changing transient characteristics such as pressure fluctuations and flow pulsations. To ensure the representativeness of the dataset, data is collected for at least 30 minutes of continuous operation for each operating condition category, covering the stable operating phase and the transition phase between operating conditions.

[0057] When extracting operating condition characteristics from the collected historical operating status information, it is necessary to calculate characteristic parameters that reflect the essential characteristics of the operating conditions. For the collected raw pressure signal, the pressure drop is obtained by calculating the inlet and outlet pressure difference, which directly reflects the flow resistance state inside the cooler. The flow rate-pressure ratio is calculated by dividing the measured flow rate by the inlet pressure; this ratio characterizes the cooler's flow capacity under a given pressure condition. The temperature-pressure coupling coefficient is obtained by analyzing the correlation between the inlet and outlet temperature difference and the pressure drop. Specifically, the ratio of the change in temperature difference to the change in pressure drop within a sliding time window is selected; this coefficient reveals the coupling relationship between the heat exchange process and the pressure state. The extracted operating condition characteristics constitute a three-dimensional feature vector, with each set of feature vectors corresponding to the operating condition state at a given moment.

[0058] The labeling of operating condition categories is based on equipment operation logs, operation records, and expert experience. During the labeling process, data segments whose operating status clearly belongs to a certain typical operating condition are labeled with the corresponding operating condition category. For data in the transitional phase of operating condition switching, they are either assigned to the corresponding operating condition or labeled separately as a transitional operating condition category based on the dominant features. After labeling is completed, the historical operating condition feature quantities and operating condition category labels are matched one-to-one to form a training sample set. To enhance the generalization ability of the model, the number of samples in each operating condition category in the training sample set is kept basically balanced to avoid the model being biased towards the classification results of a certain operating condition due to an excessive number of samples in one operating condition.

[0059] The training of the boundary discrimination model employs supervised learning algorithms such as Support Vector Machines (SVMs) or Neural Networks. Taking SVM as an example, the training process aims to find the optimal separating hyperplane in the feature space of work conditions, maximizing the margin between samples from different work condition categories. For linearly separable work condition categories, the hyperplane can be directly obtained by solving a constraint optimization problem. For linearly inseparable categories, a kernel function is introduced to map the feature space to a higher-dimensional space, achieving linear separation in the higher-dimensional space. The radial basis function (RBF) kernel is widely used due to its excellent nonlinear mapping ability. The parameters of the kernel function are determined through cross-validation to ensure that the model accurately fits the training data without overfitting.

[0060] During training, the model parameters are continuously adjusted through iterative optimization algorithms. Taking a neural network model as an example, the input layer receives a 3D working condition feature vector, the hidden layer extracts a deep representation of the features through multiple nonlinear transformations, and the output layer provides the probability distribution of each working condition category. During training, samples are input into the model, and the loss function value between the model output and the true label is calculated. The loss function uses cross-entropy to measure the classification error. The gradient of the loss function with respect to the weight parameters of each layer is calculated using the backpropagation algorithm, and the parameters are updated using gradient descent. The learning rate is set to an adaptive adjustment strategy, using a larger learning rate in the initial stage to accelerate convergence, and gradually decreasing the learning rate in the later stage to fine-tune the parameters. After thousands to tens of thousands of iterations, the classification accuracy of the model on the training set gradually improves, and the loss function value continuously decreases to a stable level.

[0061] The introduction of a boundary tolerance mechanism aims to enhance the model's adaptability to fluctuations in operating condition features. In actual operation, due to factors such as sensor noise and environmental disturbances, the extracted operating condition features fluctuate to some extent, resulting in samples of the same operating condition category not being strictly clustered in the feature space, but exhibiting a certain degree of dispersion. Hard discrimination boundaries require each sample to be strictly located on the correct side of the classification hyperplane, making them sensitive to feature fluctuations and prone to misclassification. Soft margin mechanisms allow some samples to lie within the margin boundary or even cross the hyperplane. By introducing slack variables to quantify the degree to which samples deviate from the correct classification, these deviations are added as penalty terms to the optimization objective function. The penalty coefficient controls the model's tolerance for classification errors. A smaller penalty coefficient makes the model focus more on maximizing the margin, resulting in a smoother classification boundary but potentially some misclassified samples; a larger penalty coefficient forces the model to classify all samples as correctly as possible, but the boundary may become more complex and sensitive to noise. A grid search or Bayesian optimization method is used to select the penalty coefficient value that optimizes the validation set performance, achieving a balance between classification accuracy and robustness.

[0062] After model training, the performance of the discrimination boundary model in identifying work conditions is evaluated using an independent test sample set. The test sample set is separately partitioned from historical data and was not used during training to ensure that the evaluation results reflect the model's true generalization ability. During testing, the work condition features of the test samples are input into the model to obtain the model's output work condition category prediction results, which are then compared with the true labels to calculate the recognition accuracy. Accuracy is defined as the proportion of correctly classified samples out of the total number of samples. In addition to the overall accuracy, recall and precision are also calculated for each work condition category. Recall measures the proportion of samples correctly identified in that category, and precision measures the proportion of samples predicted as belonging to that category that actually do. These metrics are used to evaluate whether the model's recognition performance is balanced across different work condition categories.

[0063] When the overall recognition accuracy on the test set reaches over 95%, and the recall and precision of each working condition category are both not less than 90%, the discrimination boundary model is considered to meet the preset performance indicators. At this point, the trained model parameters are solidified and deployed to the inference module of the working condition classifier. The deployed model receives real-time calculated instantaneous working condition features, calculates the working condition category discrimination result through forward propagation, and the entire inference process takes only milliseconds, meeting the response requirements of real-time working condition recognition. After model deployment, performance is monitored regularly using newly collected running data. When a decrease in recognition accuracy or the emergence of new working condition patterns is detected, the model is updated promptly to ensure that the working condition classifier always maintains high-precision working condition recognition capabilities, providing a reliable basis for subsequent pressure optimization and adjustment.

[0064] In one optional implementation, for the operating condition category, a corresponding pressure optimization objective function is invoked. This pressure optimization objective function aims to maximize cooling efficiency and minimize pressure fluctuations, using inlet pressure as the optimization variable, and includes: According to the operating condition category, the corresponding pressure optimization objective function is retrieved from a pre-established operating condition-objective function mapping library, which stores mathematical expressions for cooling efficiency and pressure fluctuation under different operating condition categories; A cooling efficiency target term is constructed, which establishes a functional relationship between cooling efficiency and inlet pressure based on the heat transfer relationship between the fluid flow rate, the cooling medium temperature, and the inlet pressure. A pressure fluctuation target term is constructed, which quantifies pressure stability by calculating the variance or standard deviation of the inlet pressure over a time series. The cooling efficiency objective and the pressure fluctuation objective are weighted and combined to form a multi-objective optimization function. The weight coefficients of each objective in the multi-objective optimization function are adaptively configured according to the priority requirements of the operating condition category. The inlet pressure is set as the optimization variable of the multi-objective optimization function, and the multi-objective optimization function is output as the pressure optimization objective function to the optimization solution module.

[0065] Based on the operating condition category output by the operating condition classifier, the corresponding pressure optimization objective function needs to be retrieved from a pre-established operating condition-objective function mapping library. This mapping library is established offline during system initialization. It extracts typical characteristics of cooling efficiency and pressure fluctuations under each operating condition category through statistical analysis of historical operating data of the hydraulic cooler under different operating conditions, and stores these characteristics in the mapping library in the form of mathematical expressions. The data structure of the mapping library adopts a hash table format, using the operating condition category identifier as the key and the corresponding pressure optimization objective function parameters as data items, ensuring that the retrieval process can be completed within constant time complexity. For example, when the operating condition classifier outputs the operating condition category as high load, the mapping library returns the specific mathematical relationship parameters between cooling efficiency and pressure fluctuations under that operating condition, including key parameters such as the baseline value of the heat transfer coefficient, the flow-pressure sensitivity coefficient, and the allowable range of pressure fluctuations.

[0066] After obtaining the objective function parameters for the corresponding operating conditions, the cooling efficiency objective term is constructed. The physical essence of cooling efficiency reflects the ability of a hydraulic cooler to remove heat from hydraulic oil per unit time, which is directly affected by the combined effects of fluid flow rate, cooling medium temperature, and inlet pressure. Based on heat transfer principles, cooling efficiency can be expressed as the ratio of heat exchange to input power. Heat exchange is determined by the product of fluid flow rate and temperature difference, while fluid flow rate is directly driven by inlet pressure. To establish the functional relationship between cooling efficiency and inlet pressure, the relationship between flow rate and inlet pressure is first determined according to Bernoulli's equation; the flow rate is positively correlated with the square root of the inlet pressure. Substituting this relationship into the heat transfer calculation formula yields the expression for cooling efficiency with respect to inlet pressure. This expression introduces a heat transfer coefficient parameter, which is affected by the flow state and corrected by the Reynolds number. When the inlet pressure increases, the increased flow rate leads to an increase in the Reynolds number, thereby increasing the heat transfer coefficient and causing the cooling efficiency to exhibit a non-linear growth trend. However, this growth exhibits diminishing marginal returns; that is, when the inlet pressure exceeds a certain threshold, the rate of increase in cooling efficiency gradually slows down.

[0067] To accurately quantify the cooling efficiency target, an evaluation index for cooling efficiency needs to be established. The ratio of actual heat exchange to the theoretical maximum heat exchange is used as the quantitative standard for cooling efficiency, and this ratio varies between zero and one. The actual heat exchange is calculated by monitoring the product of the inlet and outlet temperature difference and the fluid flow rate, while the theoretical maximum heat exchange is determined based on the logarithmic mean temperature difference method. When constructing the target, maximizing cooling efficiency is the optimization objective. The cooling efficiency expression is negatively represented or its reciprocal is taken, transforming the optimization problem into a minimization problem, which is easier to solve using standard optimization algorithms. Considering the differences in flow rate and temperature ranges under different operating conditions, the cooling efficiency target is normalized to eliminate the influence of dimensions and ensure the comparability of the target values ​​under different operating conditions.

[0068] After constructing the cooling efficiency target, the pressure fluctuation target is constructed. Pressure fluctuation reflects the stability of the inlet pressure over time. Frequent pressure fluctuations can cause vibration and noise in the hydraulic system, accelerate the wear of seals and valves, and reduce system reliability. The method for quantifying pressure fluctuation is to perform statistical analysis on the sampled sequence of the inlet pressure within a time window. The length of the time window is determined based on the time scale of the operating conditions; a shorter time window is chosen for rapidly changing conditions, and a longer time window is chosen for steady-state conditions. After obtaining the time series data of the inlet pressure, the variance or standard deviation of the sequence is calculated as a quantitative indicator of pressure fluctuation. Variance reflects the degree to which the data deviates from the mean, while standard deviation has the same dimensions as the original data, facilitating engineering understanding. In practical applications, standard deviation is preferred as the measure of pressure fluctuation.

[0069] The construction of the pressure fluctuation target term needs to consider the dynamic process of pressure change. Simple static variance or standard deviation cannot reflect the rate characteristics of pressure change; therefore, the pressure change rate is introduced as an auxiliary indicator. The pressure change rate is calculated as the ratio of the pressure difference between adjacent sampling points to the sampling time interval, and this indicator can capture the instantaneous trend of pressure change. The weighted sum of the pressure standard deviation and the pressure change rate is used as a comprehensive pressure fluctuation indicator, with the weighting coefficient adjusted according to the dynamic characteristics of the specific operating conditions. For operating conditions requiring rapid response, the weighting coefficient of the pressure change rate is larger, allowing for a certain range of pressure fluctuations to achieve rapid adjustment; for steady-state operating conditions, the weighting coefficient of the pressure standard deviation is larger, emphasizing long-term pressure stability. The pressure fluctuation target term appears in the optimization function in a minimization form, guiding the optimization algorithm to find the inlet pressure setpoint that can maintain pressure stability.

[0070] After constructing the cooling efficiency objective and the pressure fluctuation objective, they are weighted and combined to form a multi-objective optimization function. The general form of the multi-objective optimization function is a linear weighted sum of the individual objective terms, and the balance between different optimization objectives is achieved by adjusting the weight coefficients. The weight coefficients for the cooling efficiency objective are denoted as... The weighting coefficient for the pressure fluctuation objective term is denoted as... The sum of the two equals 1 to ensure the normalization property of the objective function.

[0071] The multi-objective optimization function can be expressed as: ,in This indicates import pressure. This represents the negative or reciprocal value of the normalized cooling efficiency. This represents the normalized pressure fluctuation index. Weighting coefficients. and The value of directly determines the performance characteristics of the optimization result.

[0072] Adaptive configuration of weighting coefficients is a key technical aspect of achieving multi-condition optimization. Different operating conditions have significantly different requirements for cooling efficiency and pressure stability, necessitating dynamic adjustment of weighting coefficients based on the priority requirements of each operating condition. For high-heat-load conditions, where the hydraulic oil temperature approaches its safe upper limit, cooling efficiency becomes the primary optimization objective, and the weighting coefficients for the cooling efficiency objective are crucial. Set to a larger value, such as 0.7 to 0.8, correspondingly increasing the weighting coefficient of the pressure fluctuation target item. Setting it to 0.2 to 0.3 allows for sacrificing pressure stability within a certain range to achieve higher cooling efficiency. For precision control applications, the system is extremely sensitive to pressure fluctuations; even small pressure changes can lead to positioning errors in the actuators. In this case, the weighting coefficient of the pressure fluctuation objective item... Set to a relatively large value, such as 0.6 to 0.7, to ensure that the inlet pressure remains highly stable. For normal operating conditions, both cooling requirements and pressure stability requirements are at a moderate level, with a weighting factor of [missing value]. and Set them to similar values, such as both being 0.5, to achieve a balanced optimization of the two optimization objectives.

[0073] The adaptive weighting mechanism is based on a comprehensive judgment of the operating condition category and the real-time system status. After the operating condition classifier outputs the operating condition category, the initial value of the weighting coefficient for that category is obtained according to a predefined operating condition-weight mapping table. This initial value serves as the baseline weight, which is then fine-tuned based on the real-time system status. The fine-tuning rules consider factors such as the proximity of the current hydraulic oil temperature to the safety threshold and the deviation of the current pressure fluctuation amplitude from the historical average. When the hydraulic oil temperature exceeds 80% of the safety threshold, the weighting coefficient of the cooling efficiency objective is automatically increased, with the increment linearly determined based on the temperature proximity, and a maximum increment not exceeding 0.15. When the pressure fluctuation amplitude exceeds 1.5 times the historical average, the weighting coefficient of the pressure fluctuation objective is automatically increased to ensure the system prioritizes pressure stability recovery. Through this adaptive weighting mechanism, the multi-objective optimization function can flexibly adjust its optimization focus according to the operating condition category and real-time status, meeting the differentiated needs under different operating conditions.

[0074] After constructing the multi-objective optimization function and configuring the weights, the inlet pressure is set as the optimization variable. The feasible region of the inlet pressure as the optimization variable is determined by the physical constraints of the hydraulic cooler, including the pressure limit, the pressure range corresponding to the flow rate adjustment range, and the pressure range limited by the temperature safety threshold. The feasible region of the optimization variable is incorporated into the optimization problem in the form of inequality constraints to ensure that the optimal inlet pressure value obtained by the solution is within the allowable operating range of the system. The initial value of the optimization variable is set to the current inlet pressure value, which helps the optimization algorithm to converge quickly to the vicinity of the optimal solution. The constructed multi-objective optimization function, along with the optimization variables and their constraints, is output to the optimization solution module, which calls the numerical optimization algorithm to solve for the optimal inlet pressure value under the current operating conditions. The entire objective function construction and invocation process is completed within milliseconds, ensuring that the pressure optimization can respond to changes in operating conditions in real time and achieve rapid adaptive adjustment of the hydraulic cooler.

[0075] Figure 2 This is a flowchart illustrating adaptive feedback control. An adaptive feedback control mechanism is used to generate a pressure regulation control signal. This mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator, including: The actual response data of the pressure regulating actuator during the pressure regulating action is collected in real time. The actual response data includes the position feedback value, velocity feedback value, and real-time measurement value of the inlet pressure of the actuator. The control error is calculated based on the deviation between the actual response data and the optimal inlet pressure value. The trend of the control error over time is analyzed to identify the characteristic patterns of control delay and nonlinear effects. An actuator response characteristic model is established, which describes the delay time constant and nonlinear gain characteristics of the actuator by fitting the dynamic relationship between the actual response data and the control command. An adaptive compensation strategy is designed based on the aforementioned response characteristic model. This adaptive compensation strategy counteracts the control delay and the nonlinear effects by introducing a predictive lead and a nonlinear inverse compensation term into the control signal. The control error, the adjustment deviation, and the response speed are input to the adaptive feedback controller, which, in conjunction with the adaptive compensation strategy, generates a pressure regulation control signal and outputs it to the pressure regulation actuator.

[0076] During the pressure regulation process of a hydraulic cooler, the actual response characteristics of the actuator often exhibit time delay and nonlinearity, which can lead to deviations between control commands and actual outputs. To address this issue, a dynamic compensation mechanism targeting the actuator's response characteristics is needed.

[0077] Multi-dimensional response data of the pressure regulating actuator are acquired in real time to construct a complete execution status feedback chain. During the operation of the actuator, a high-precision position sensor continuously acquires the position feedback value of the valve core or piston at a sampling frequency of 50Hz. This position feedback value reflects the actual opening state of the actuator. Simultaneously, a velocity feedback value is acquired through a velocity sensor or by numerical differentiation of the position signal. The velocity feedback value characterizes the movement rate of the actuator. A pressure sensor is installed at the inlet pipeline to measure the inlet pressure value in real time at the same sampling frequency. This measurement value directly reflects the final effect of the regulating action. These actual response data are stored in a time-series dataset with multiple sampling points aligned by timestamps. The data record of the i-th sampling point is in the form of a triplet, containing the measurement values ​​of the three dimensions of position, velocity, and pressure.

[0078] The control error at the current moment is calculated by comparing the collected actual response data with the preset optimal inlet pressure value. The control error is defined as the difference between the optimal inlet pressure value and the real-time measured pressure value; a positive value indicates that the actual pressure is lower than the target value, and a negative value indicates that the actual pressure is higher than the target value. A time series of the control error is constructed, recording the error values ​​at M consecutive time points, typically M being 20 to 50 sampling points. Trend analysis is performed on this error time series, and the error change rate is calculated using the sliding window method. This change rate is obtained by the difference between error values ​​at adjacent time points. When the error change rate remains a stable positive or negative value across multiple consecutive sampling points, it indicates a persistent deviation in the system, a phenomenon usually stemming from control delay. When the error value exhibits periodic oscillations or abrupt changes in the error change rate within a specific pressure range, it indicates that the actuator exhibits nonlinear response characteristics, such as the dead zone effect of the valve core in the small opening range or the saturation effect in the large opening range.

[0079] A mathematical model is established to describe the dynamic response characteristics of the actuator. This model needs to capture two key characteristics: delay and nonlinearity. For the delay characteristic, a first-order inertial element is used to represent the time response process of the actuator. Let the control command signal be... The actual output of the actuator is The dynamic relationship between the two can be expressed as follows: ,in The time constant, For pure time delay, This is the static gain. By systematically identifying historical data and fitting the actual response curve using the least squares method, the delay time constant is obtained. With pure time delay The value. In a typical hydraulic actuator, The range is from 0.1 seconds to 0.5 seconds. The range is from 0.05 seconds to 0.2 seconds. For nonlinear characteristics, a piecewise linearization method is used to establish the input-output mapping relationship. The operating range of the actuator is divided into multiple sub-intervals, and different gain coefficients are used in each sub-interval to describe the relationship between input and output. Specifically, the range of control command values ​​is divided into equal parts. There are several intervals, each corresponding to a local gain value. ,in By analyzing actual response data, the actual output gain of the actuator under different control command amplitudes is identified, and the gain value of each interval is calibrated. When the control command is small, the gain value is usually low, reflecting the influence of dead zone and friction; when the control command is close to the physical limit of the actuator, the gain value also decreases, reflecting the constraints of mechanical limit and flow saturation.

[0080] Based on the established response characteristic model, an adaptive compensation strategy is designed to counteract the adverse effects of delay and nonlinearity. For delay characteristics, a predictive advance compensation mechanism is introduced. This is based on the identified delay time constant. With pure time delay In advance when generating control signals The system outputs control commands in a timely manner, while simultaneously applying phase lead correction. Phase lead correction is achieved by increasing the weighting of the controller output's response to the rate of change of error, thus enabling predictive control actions. Specifically, a weighted term for the rate of change of error is superimposed on the control signal, with the weighting coefficient proportional to the time delay constant. For nonlinear characteristics, a nonlinear inverse compensation module is constructed. An inverse mapping relationship is established based on the local gain values ​​of each interval in the piecewise linearized model. When a specific amplitude actuator output is required, the desired output is divided by the local gain value of the current nonlinear interval to obtain the compensated control command. This inverse compensation operation is essentially a feedforward compensation for the nonlinear characteristics of the actuator, making the relationship between the control command and the actual output approach an ideal linear proportional relationship. As the actuator enters different nonlinear intervals, the compensation module automatically switches the corresponding inverse gain value, achieving nonlinear compensation across the entire operating range.

[0081] The control error, regulation deviation, and response speed are fed as input variables into the adaptive feedback controller. The controller employs an improved proportional-integral-derivative (PI-DE) structure, where the proportional term responds to the current value of the control error, the integral term accumulates historical errors to eliminate steady-state deviation, and the derivative term responds to the error change trend to improve dynamic performance. The controller output is processed by an adaptive compensation strategy to form the final pressure regulation control signal. The adaptive compensation strategy operates on two levels: first, the predictive lead based on the response characteristic model is superimposed on the controller output, causing the control action to be performed earlier. First, the timing is adjusted. Second, the controller output undergoes gain adjustment via a nonlinear inverse compensation module, applying an appropriate inverse gain factor based on the current operating range of the actuator. After these two stages of compensation, the generated pressure regulation control signal effectively counteracts the delay and nonlinear effects of the actuator. The control signal is output to the drive unit of the pressure regulation actuator in digital or analog voltage form. The drive unit converts the control signal into the driving force or current required by the actuator, driving it to complete the pressure regulation action.

[0082] During the control process, the actual response data of the actuator is continuously monitored to form a closed-loop feedback. When a change in the actual response characteristics is detected, such as a change in viscosity due to a change in hydraulic oil temperature affecting the time constant... The drift of the actuator, or the expansion of the dead zone due to changes in frictional characteristics caused by actuator wear, leads to adaptive updates of the response characteristic model through an online parameter identification algorithm. The online identification algorithm employs recursive least squares or a forgetting factor recursive algorithm, re-estimating the model parameters using the most recent Q sampling points, where Q is typically between 100 and 500. The updated model parameters are immediately applied to the calculation of the adaptive compensation strategy, ensuring that the compensation effect always matches the current response characteristics of the actuator. This adaptive mechanism enables the control system to cope with the time-varying characteristics of the actuator and maintain stable control performance during long-term operation. Through the above adaptive feedback control mechanism, the control delay in the pressure regulation process is effectively compensated, overshoot and oscillation caused by actuator response lag are suppressed, control accuracy in the nonlinear region is significantly improved, and the inlet pressure can quickly and accurately converge to the optimal value, meeting the performance requirements of adaptive pressure regulation of the hydraulic cooler under multiple operating conditions.

[0083] A second aspect of the present invention provides a multi-condition adaptive pressure regulation and optimization system for hydraulic coolers, comprising: The monitoring and calculation unit is used to monitor the operating status information of the hydraulic cooler in real time, and calculate the instantaneous operating condition characteristics of the hydraulic cooler based on the operating status information. The instantaneous operating condition characteristics include pressure drop, flow-pressure ratio and temperature-pressure coupling coefficient. The working condition identification unit is used to input the instantaneous working condition feature quantity into the working condition classifier. The working condition classifier realizes automatic working condition identification by establishing a discrimination boundary between the working condition feature quantity and the predefined working condition category, and outputs the working condition category to which the current working condition belongs. The target invocation unit is used to invoke the corresponding pressure optimization objective function for the operating condition category. The pressure optimization objective function takes maximizing cooling efficiency and minimizing pressure fluctuation as optimization objectives, and uses inlet pressure as optimization variable. Under the constraints of the pressure optimization objective function, the optimal inlet pressure value is solved. The constraints include the pressure bearing limit, flow regulation range and temperature safety threshold of the hydraulic cooler. The optimization solution unit is used to calculate the adjustment deviation between the optimal inlet pressure value and the current inlet pressure value, determine the pressure regulation response speed based on the adjustment deviation and the rate of change of operating conditions, and generate a pressure regulation control signal using an adaptive feedback control mechanism. The adaptive feedback control mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The deviation calculation unit is used to output the pressure regulation control signal to the pressure regulation actuator, drive the pressure regulation actuator to perform pressure regulation action, so that the inlet pressure converges to the optimal inlet pressure value, and realize the pressure adaptive regulation optimization of the hydraulic cooler under multiple working conditions.

[0084] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0085] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0086] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-operating condition oriented hydraulic cooler pressure adaptive regulation optimization method, characterized in that, include: The operating status information of the hydraulic cooler is monitored in real time, and the instantaneous operating condition characteristics of the hydraulic cooler are calculated based on the operating status information. The instantaneous operating condition characteristics include pressure drop, flow-pressure ratio, and temperature-pressure coupling coefficient. The instantaneous operating condition features are input into the operating condition classifier. The operating condition classifier automatically identifies the operating condition by establishing a discrimination boundary between the operating condition features and predefined operating condition categories, and outputs the operating condition category to which the current operating condition belongs. For the aforementioned operating condition category, the corresponding pressure optimization objective function is invoked. The pressure optimization objective function aims to maximize cooling efficiency and minimize pressure fluctuations, using inlet pressure as the optimization variable. Under the constraints of the pressure optimization objective function, the optimal inlet pressure value is solved. The constraints include the pressure tolerance limit, flow regulation range, and temperature safety threshold of the hydraulic cooler. The adjustment deviation between the optimal inlet pressure value and the current inlet pressure value is calculated. Based on the adjustment deviation and the rate of change of operating conditions, the response speed of pressure regulation is determined. An adaptive feedback control mechanism is used to generate a pressure regulation control signal. The adaptive feedback control mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The pressure regulation control signal is output to the pressure regulation actuator, which drives the pressure regulation actuator to perform pressure regulation action, so that the inlet pressure converges to the optimal inlet pressure value, thereby realizing the pressure adaptive regulation and optimization of the hydraulic cooler under multiple working conditions.

2. The method of claim 1, wherein The instantaneous operating condition characteristics of the hydraulic cooler are calculated based on the operating status information, including: Extract the fluid pressure value measured by the inlet pressure sensor and the fluid pressure value measured by the outlet pressure sensor, and calculate the difference between the two to obtain the pressure drop; Obtain the fluid flow rate value measured by the flow sensor, and calculate the flow rate-pressure ratio by comparing the fluid flow rate value with the pressure drop; The system collects the cooling medium temperature and ambient temperature values ​​measured by a temperature sensor, calculates the temperature difference between the cooling medium temperature and the ambient temperature, couples the temperature difference with the pressure drop, and constructs a temperature-pressure coupling coefficient by establishing the response relationship between temperature change and pressure change. The pressure drop, the flow-pressure ratio, and the temperature-pressure coupling coefficient are output as instantaneous operating condition features to the operating condition classifier.

3. The method of claim 2, wherein The temperature difference and the pressure drop are coupled and calculated. A temperature-pressure coupling coefficient is constructed by establishing the response relationship between temperature change and pressure change, including: Within a preset time window, historical change sequences of the cooling medium temperature and pressure drop are collected. These historical change sequences contain paired data of temperature and pressure drop values ​​at multiple time sampling points. The temperature value sequence in the historical change sequence is differentiated in the time domain to obtain the temperature change rate sequence, and the pressure drop value sequence is differentiated in the time domain to obtain the pressure change rate sequence; Calculate the correlation metric between the temperature change rate sequence and the pressure change rate sequence, wherein the correlation metric reflects the synchronous response strength of temperature fluctuations and pressure fluctuations; Based on the correlation metric and the current instantaneous values ​​of the temperature difference and the pressure drop, a temperature-pressure coupling coefficient is generated through a weighted fusion operation, in which the correlation metric is used as a weighting factor to adjust the coupling strength. The temperature-pressure coupling coefficient is output as a condition characteristic quantity characterizing the thermodynamic response of the hydraulic cooler.

4. The method of claim 1, wherein The instantaneous operating condition features are input into the operating condition classifier. The classifier automatically identifies the operating condition by establishing a discrimination boundary between the operating condition features and predefined operating condition categories, and outputs the operating condition category to which the current operating condition belongs, including: The pressure drop, the flow-pressure ratio, and the temperature-pressure coupling coefficient are used to construct a multi-dimensional operating condition feature vector. The multidimensional working condition feature vector is mapped to the working condition feature space, which is a multidimensional geometric space composed of each working condition feature quantity as a coordinate axis. In the operating condition feature space, a pre-trained discrimination boundary model is invoked. This discrimination boundary model learns the region division rules for different operating condition categories in the operating condition feature space through historical operating condition data. Calculate the spatial distance between the multidimensional working condition feature vector and the discrimination boundary of each predefined working condition category. The spatial distance reflects the membership degree of the current working condition feature vector to each working condition category. The predefined working condition category corresponding to the discrimination boundary with the smallest spatial distance is selected as the working condition category to which the current working condition belongs.

5. The method of claim 4, wherein In the operating condition feature space, a pre-trained discrimination boundary model is invoked. This discrimination boundary model learns regional division rules for different operating condition categories in the operating condition feature space through historical operating condition data, including: Collect historical operating status information of the hydraulic cooler under various operating conditions, extract historical operating condition feature quantities from the historical operating status information, and label the corresponding operating condition category. The historical working condition features and the working condition category labels are combined to form a training sample set, which contains typical distribution patterns of each working condition category in the working condition feature space. A supervised learning mechanism is used to train the discrimination boundary model. Through iterative optimization, the discrimination boundary model is made to form a hyperplane or nonlinear decision surface in the working condition feature space that can accurately separate different working condition categories. A boundary tolerance mechanism is introduced during training. This mechanism enhances the model's robustness to fluctuations in operating conditions by setting soft-margin regions for the discrimination boundary. Verify the accuracy of the discrimination boundary model in identifying working conditions on the test sample set. Once the accuracy meets the preset performance index, deploy the discrimination boundary model into the working condition classifier for real-time working condition identification.

6. The method of claim 1, wherein For the aforementioned operating condition category, the corresponding pressure optimization objective function is invoked. This objective function aims to maximize cooling efficiency and minimize pressure fluctuations, using inlet pressure as the optimization variable, and includes: Based on the operating condition category, the corresponding pressure optimization objective function is retrieved from a pre-established operating condition-objective function mapping library, which stores mathematical expressions for cooling efficiency and pressure fluctuation under different operating condition categories; A cooling efficiency target term is constructed, which establishes a functional relationship between cooling efficiency and inlet pressure based on the heat transfer relationship between the fluid flow rate, the cooling medium temperature, and the inlet pressure. A pressure fluctuation target term is constructed, which quantifies pressure stability by calculating the variance or standard deviation of the inlet pressure over a time series. The cooling efficiency objective and the pressure fluctuation objective are weighted and combined to form a multi-objective optimization function. The weight coefficients of each objective in the multi-objective optimization function are adaptively configured according to the priority requirements of the operating condition category. The inlet pressure is set as the optimization variable of the multi-objective optimization function, and the multi-objective optimization function is output as the pressure optimization objective function to the optimization solution module.

7. The method of claim 1, wherein An adaptive feedback control mechanism is used to generate a pressure regulation control signal. This mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The mechanism includes: The actual response data of the pressure regulating actuator during the pressure regulating action is collected in real time. The actual response data includes the position feedback value, velocity feedback value, and real-time measurement value of the inlet pressure of the actuator. The control error is calculated based on the deviation between the actual response data and the optimal inlet pressure value. The trend of the control error over time is analyzed to identify the characteristic patterns of control delay and nonlinear effects. An actuator response characteristic model is established, which describes the delay time constant and nonlinear gain characteristics of the actuator by fitting the dynamic relationship between the actual response data and the control command. An adaptive compensation strategy is designed based on the aforementioned response characteristic model. This adaptive compensation strategy counteracts the control delay and the nonlinear effects by introducing a predictive lead and a nonlinear inverse compensation term into the control signal. The control error, the adjustment deviation, and the response speed are input to the adaptive feedback controller, which, in conjunction with the adaptive compensation strategy, generates a pressure regulation control signal and outputs it to the pressure regulation actuator.

8. A multi-condition adaptive pressure regulation and optimization system for hydraulic coolers, used to implement the method as described in any one of claims 1-7, characterized in that, include: The monitoring and calculation unit is used to monitor the operating status information of the hydraulic cooler in real time, and calculate the instantaneous operating condition characteristics of the hydraulic cooler based on the operating status information. The instantaneous operating condition characteristics include pressure drop, flow-pressure ratio and temperature-pressure coupling coefficient. The working condition identification unit is used to input the instantaneous working condition feature quantity into the working condition classifier. The working condition classifier realizes automatic working condition identification by establishing a discrimination boundary between the working condition feature quantity and the predefined working condition category, and outputs the working condition category to which the current working condition belongs. The target invocation unit is used to invoke the corresponding pressure optimization objective function for the operating condition category. The pressure optimization objective function takes maximizing cooling efficiency and minimizing pressure fluctuation as optimization objectives, and uses inlet pressure as optimization variable. Under the constraints of the pressure optimization objective function, the optimal inlet pressure value is solved. The constraints include the pressure bearing limit, flow regulation range and temperature safety threshold of the hydraulic cooler. The optimization solution unit is used to calculate the adjustment deviation between the optimal inlet pressure value and the current inlet pressure value, determine the pressure regulation response speed based on the adjustment deviation and the rate of change of operating conditions, and generate a pressure regulation control signal using an adaptive feedback control mechanism. The adaptive feedback control mechanism dynamically compensates for control delay and nonlinear effects by monitoring the actual response characteristics of the pressure regulation actuator. The deviation calculation unit is used to output the pressure regulation control signal to the pressure regulation actuator, drive the pressure regulation actuator to perform pressure regulation action, so that the inlet pressure converges to the optimal inlet pressure value, and realize the pressure adaptive regulation optimization of the hydraulic cooler under multiple working conditions.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images