Coordinated control method and system for high-speed motors of multiple cooling fans

Abstract

The present invention relates to the field of motor control technology, and discloses a method and system for cooperative control of high-speed motors of multiple cooling fans. The method comprises: performing multi-dimensional data acquisition and standardized processing on each cooling fan unit in a target cooling system to obtain a system state data matrix, wherein the target cooling system comprises a master cooling fan and multiple slave cooling fans; performing discrete speed response to obtain a continuous speed control strategy; performing multi-dimensional dynamic feature extraction to obtain a system operation feature vector, and performing state feedback equation solution to obtain a mapping relationship matrix; inputting a target control model to perform control parameter analysis to obtain an initial control sequence; inputting a dual Q network model to perform master-slave fan control correction to obtain a control correction amount; performing compensation correction to obtain a target control sequence, and outputting a cooperative control result. The present invention realizes the cooperative control of multiple cooling fans and effectively avoids control conflicts.

Description

Technical Field

[0001] The present invention relates to the field of motor control technology, and in particular to a coordinated control method and system for high-speed motors of multiple cooling fans. Background Art

[0002] Traditional cooling fan control methods mainly use single PWM speed regulation or simple temperature threshold control, which cannot meet the cooling requirements under complex working conditions. In addition, there is a lack of coordination between the fans, which easily leads to system energy waste and low cooling efficiency.

[0003] At present, multi-fan systems generally have noise problems caused by sudden changes in speed, as well as control conflicts caused by the lack of an effective master-slave coordination mechanism. Especially under high temperature conditions, it is difficult to achieve the optimal heat dissipation effect by relying solely on independent control, and a simple temperature threshold control strategy cannot guarantee the smooth operation and energy efficiency optimization of the system. In addition, most existing fan control systems use open-loop control methods, lack the ability to monitor and dynamically adjust the system's operating status in real time, and are unable to cope with complex and changeable heat dissipation environments. At the same time, due to the lack of an effective master-slave fan mapping relationship and collaborative control mechanism, the coordination between the fans is often in a disordered state, affecting the performance of the overall heat dissipation system. Summary of the invention

[0004] The present invention provides a method and system for cooperatively controlling high-speed motors of multiple cooling fans, which are used to realize cooperative control of multiple cooling fans and effectively avoid control conflicts.

[0005] In a first aspect, the present invention provides a method for cooperatively controlling high-speed motors of multiple cooling fans, the method comprising:

[0006] Performing multi-dimensional data collection and standardization processing on each cooling fan unit in a target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and a plurality of slave cooling fans;

[0007] Dividing the temperature data submatrix in the system state data matrix into N control intervals, and performing discrete speed response to obtain a continuous speed control strategy;

[0008] Performing multi-dimensional dynamic feature extraction on the system state data matrix to obtain a system operation feature vector, and solving a state feedback equation to obtain a mapping relationship matrix between the main cooling fan and the plurality of slave cooling fans;

[0009] Inputting the mapping relationship matrix and the continuous speed control strategy into a target control model to perform control parameter analysis to obtain an initial control sequence;

[0010] Inputting the system operation characteristic vector and the initial control sequence into a dual Q network model to perform master-slave fan control correction to obtain a control correction amount;

[0011] The initial control sequence is compensated and corrected based on the control correction amount to obtain a target control sequence, and the target control sequence is sent to the motor drivers of each cooling fan through a CAN bus to output a collaborative control result.

[0012] In a second aspect, the present invention provides a coordinated control system for high-speed motors of multiple cooling fans, the coordinated control system for high-speed motors of multiple cooling fans comprising:

[0013] an acquisition module, used for performing multi-dimensional data acquisition and standardization processing on each cooling fan unit in a target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and a plurality of slave cooling fans;

[0014] A division module, used for dividing the temperature data submatrix in the system state data matrix into N control intervals, and performing discrete speed response to obtain a continuous speed control strategy;

[0015] A solution module, used for performing multi-dimensional dynamic feature extraction on the system state data matrix to obtain a system operation feature vector, and solving a state feedback equation to obtain a mapping relationship matrix between the main cooling fan and the plurality of slave cooling fans;

[0016] An analysis module, used for inputting the mapping relationship matrix and the continuous speed control strategy into a target control model to perform control parameter analysis and obtain an initial control sequence;

[0017] A correction module, used for inputting the system operation characteristic vector and the initial control sequence into a dual Q network model to perform master-slave fan control correction to obtain a control correction amount;

[0018] The output module is used to compensate and correct the initial control sequence based on the control correction amount to obtain a target control sequence, and send the target control sequence to the motor drivers of each cooling fan through a CAN bus to output a collaborative control result.

[0019] A third aspect of the present invention provides a collaborative control device for high-speed motors of multiple cooling fans, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory to enable the collaborative control device for high-speed motors of multiple cooling fans to execute the above-mentioned collaborative control method for high-speed motors of multiple cooling fans.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium, in which instructions are stored, which, when executed on a computer, enable the computer to execute the above-mentioned method for coordinated control of high-speed motors of multiple cooling fans.

[0021] In the technical solution provided by the present invention, by establishing a system state data matrix and multi-dimensional dynamic feature extraction, a comprehensive perception and accurate description of the operating state of the cooling system is achieved, providing a reliable data basis for collaborative control; a master-slave fan mapping relationship matrix and a continuous speed control strategy are combined to solve the speed mutation problem in the multi-fan system and ensure the stability of the system operation; a dual Q network model is introduced to perform master-slave fan control correction, and a scientific master-slave collaborative control mechanism is established to effectively avoid control conflicts; a time-sharing control sequence issuing mechanism based on the CAN bus is designed to ensure the orderly execution of master-slave fan control instructions and improve the real-time and reliability of the system; a complete closed-loop control system is established through multi-level data processing and control strategy optimization, enhancing the system's adaptability to complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0023] Figure 1 A schematic flow chart of a method for cooperatively controlling high-speed motors of multiple cooling fans provided in an embodiment of the present application;

[0024] Figure 2 A schematic block diagram of the structure of a coordinated control system of high-speed motors of multiple cooling fans provided in an embodiment of the present application;

[0025] Figure 3 A schematic block diagram of the structure of a coordinated control device for high-speed motors of multiple cooling fans provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change based on actual conditions.

[0028] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0029] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0030] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0031] See also Figure 1 , Figure 1 A schematic flow chart of a method for cooperatively controlling high-speed motors of multiple cooling fans provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the method for coordinated control of high-speed motors of multiple cooling fans provided in an embodiment of the present application includes steps S100 to S600.

[0032] Step S100, performing multi-dimensional data collection and standardization processing on each cooling fan unit in the target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and multiple slave cooling fans;

[0033] It is understandable that the execution subject of the present invention may be a coordinated control system of high-speed motors of multiple cooling fans, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0034] Specifically, the temperature data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the temperature sensor to obtain the original temperature data, which reflects the cooling effect and current working state of each fan and the entire system. At the same time, the speed sensor is used to collect the speed of the main cooling fan and the slave cooling fan to obtain the original speed data, which provides the control system with the actual running speed information of the fan, which helps to adjust the speed of the fan to meet the cooling demand. In order to ensure the stability of the cooling system and prevent mechanical damage caused by excessive vibration, the vibration conditions of each fan are collected through the vibration sensor to obtain the original vibration data, which reflects the mechanical stability of the fan. The original temperature data, the original speed data and the original vibration data are combined in time series to obtain the original data sequence. In order to eliminate the mean deviation in the original data and unify the dimensions of different data sources, all the original data sequences are subjected to zero mean standardization processing, thereby avoiding the inconsistency of different sensor data in the analysis. The standardized data sequence is subdivided into multiple data subsequences. The standardized data sequence is segmented by the sliding window technology to obtain a set of data subsequences in a series of time windows. Each data subsequence contains temperature, speed and vibration data for a period of time, which can reflect the running state and change trend of the fan during the period. Extract frequency domain features from data subsequences. Through fast Fourier transform, extract the frequency domain features of each data subsequence and form a frequency domain feature matrix. Frequency domain features can not only reveal the frequency characteristics of fan operation, but also help identify frequency patterns related to faults, thereby warning of possible problems in advance. Multi-scale decomposition is performed on the frequency domain feature matrix to extract finer-grained time-frequency features from different time scales. The obtained time-frequency feature matrix can more accurately describe the operating status of the system. In order to reduce the data dimension and extract the most representative features, principal component analysis is performed on the time-frequency feature matrix. By performing linear transformation on high-dimensional data, the data is reduced to a lower dimension while retaining most of the original information to obtain a reduced-dimensional feature matrix. Through the reduced-dimensional feature matrix, the original temperature data, speed data and vibration data are reorganized to form a temperature data submatrix, a speed data submatrix and a vibration data submatrix. The processed and optimized submatrices are concatenated to obtain a complete system status data matrix, which fully reflects the operating status of the target cooling system.

[0035] Step S200, dividing the temperature data submatrix in the system state data matrix into N control intervals, and performing discrete speed response to obtain a continuous speed control strategy;

[0036] Specifically, the temperature data submatrix is ​​subjected to differential calculation of adjacent data in time series to obtain the temperature change rate matrix. By calculating the temperature change between consecutive time points, the temperature change rate is obtained to reflect the temperature dynamic change trend of the cooling system. In order to make the calculation of the temperature change rate more compatible with the actual control task, the sampling time interval normalization operation is performed so that the temperature changes under different time scales have the same calculation basis to ensure the consistency of data processing. The temperature change rate matrix is ​​sorted according to the numerical value, and the maximum inter-class variance method (Otsu method) is used to determine the N-1 segmentation thresholds of the temperature change rate matrix. The maximum inter-class variance method is an image segmentation technology that optimizes the threshold selection by maximizing the inter-class variance, so that the segmented areas have significant differences in the performance of temperature changes. According to the segmentation threshold, the temperature change interval is divided into N control intervals to obtain the temperature control partition matrix, which reflects that the cooling system should adopt different control strategies in different temperature change intervals. A two-dimensional state space is established for the interval corresponding to each temperature control interval. The horizontal axis of the two-dimensional state space represents the temperature value, and the vertical axis represents the temperature change rate. By constructing a two-dimensional state space, the relationship between temperature and temperature change rate is described, and for each temperature state point, a speed response weight is assigned according to its position in the state space. The weight represents the speed response degree that the fan should take under specific temperature and temperature change rate conditions, thereby obtaining a speed response model under each temperature state. Based on the partitioned state space model, the mapping relationship between temperature state and speed response is fitted to obtain an interval discrete speed response function, which describes how the temperature state affects the speed in each temperature interval. The interval discrete speed response function is integrated into the fuzzy rule input. Fuzzy rules can help the system handle the nonlinear relationship between temperature and speed more flexibly, especially when the temperature changes greatly or the uncertainty is high. In order to achieve the adaptability of the fuzzy control system, the triangular membership function is used to fuzzify the temperature state. Through the fuzzification method, the temperature state is converted into a fuzzy set, so that the control system can make a more accurate response in an uncertain environment. The output response curve of the fuzzy controller is the key control information. Through piecewise cubic spline interpolation smoothing and optimization, the control response is made more continuous and stable. Cubic spline interpolation helps to reduce the control instability caused by drastic temperature changes and achieve smooth speed regulation. Based on the control response curve, a gain adaptive PID controller is designed, and the particle swarm optimization (PSO) algorithm is used to optimize the parameters of the PID controller. The particle swarm optimization algorithm simulates the foraging process of bird flocks in nature to find the optimal solution, so that the PID parameters can be automatically adjusted to the most appropriate values, improving the control accuracy and stability of the system. The optimized PID parameter set is weightedly fused with the control rules of the adaptive fuzzy controller to finally generate a continuous speed control strategy.This strategy can handle the control requirements brought about by temperature changes, and can also be dynamically adjusted according to the real-time status of the system to ensure that the system always maintains the optimal heat dissipation effect under different load and environmental conditions.

[0037] Step S300, extracting multi-dimensional dynamic features from the system state data matrix to obtain a system operation feature vector, and solving a state feedback equation to obtain a mapping relationship matrix between a main cooling fan and a plurality of slave cooling fans;

[0038] Specifically, the data matrix of the target cooling system is grouped according to the master and slave fans. The data matrix of the master cooling fan is marked as the master control matrix, and the data matrix of multiple slave cooling fans is marked as the slave control matrix. Grouping helps to clarify the roles of the master cooling fan and the slave cooling fan in the control system, as well as the relationship between them. The master control matrix and the slave control matrix are standardized to eliminate the differences between different fans due to factors such as sampling scale and data unit, making the data more comparable. After standardization, the standardized time series data of the master and slave fans more accurately reflect the state of the system at different time points. The standardized time series data is subjected to spectrum analysis and time domain feature calculation to extract the system operation feature vector. Spectrum analysis helps to identify the frequency components in the system and reveal the periodic behavior of the fan during operation, while time domain feature calculation can reveal the temporal change law of the system. The state space equation of the master and slave fans is established by extracting the system operation feature vector from the standardized time series data. The equation contains state feedback terms and follower feedback terms, which are used to describe the dynamic relationship between the master cooling fan and the slave cooling fan. The state feedback term reflects the influence of the current state of the master cooling fan on its control strategy, while the follower feedback term describes how the slave cooling fan adjusts its behavior according to the state of the master cooling fan. By constructing a control equation containing these two feedbacks, the master-slave coupling state equation is obtained to describe the collaborative working mechanism of the master and slave fans. The master-slave coupling state equation is analyzed for stability to ensure that the system will not be unstable during the dynamic adjustment process. Through stability analysis, it is determined whether the response of the system under different conditions will tend to be stable. After completing the stability analysis, the feedback gain parameter is calculated. The feedback gain parameter determines the sensitivity of the control system to the current state, thereby affecting the collaborative effect between the master and slave fans. According to the feedback gain parameter, the collaborative control parameter of the master and slave fans is obtained. According to the collaborative control parameter, the master-slave mapping function is constructed to determine the control relationship between the master cooling fan and the slave cooling fan. Through the mapping function, it is described how the slave fan responds according to the state of the master fan. The state transfer matrix is ​​calculated to describe the mutual influence between the master fan and the slave fan. The state transfer matrix can reflect the law of state changes of each fan in the system and provide a basis for subsequent control decisions. By performing a symmetric matrix transformation on the control quantity of the slave fan and the state transfer matrix, a more concise and symmetrical representation is obtained. The state transfer matrix is ​​calibrated to improve the control accuracy and robustness of the system. Through calibration, the parameters in the state transfer matrix are adjusted to make it more consistent with the actual operating conditions and ensure that the control relationship between the master and slave fans remains stable under various working environments. Finally, an accurate mapping relationship matrix between the master cooling fan and multiple slave cooling fans is obtained.

[0039] Step S400, inputting the mapping relationship matrix and the continuous speed control strategy into the target control model to perform control parameter analysis to obtain an initial control sequence;

[0040] Specifically, mapping vectors of a master cooling fan and multiple slave cooling fans are extracted based on a mapping relationship matrix. The dynamic relationship between the master cooling fan and the slave cooling fan is identified through the mapping relationship matrix, and a corresponding mapping vector is assigned to each fan, and the mapping vector reflects the control response of each fan under different states. The master-slave mapping vectors are combined to generate a master-slave mapping vector group. The master-slave mapping vector group is input into the target control model, and a state transition cost function of the master-slave fan is established. The state transition cost function is used to describe how the state of the master-slave fan is transferred under a given control input and the control cost required during the transfer process. The cost function can quantify the effectiveness of the control strategy. By combining the state transition cost function with the control parameters in the continuous speed control strategy, the state cost matrix is ​​weighted to construct the cooperative control objective function of the master-slave fan. The control objective function integrates the cost of the fan state transfer and the weight of the control parameters to provide a clear optimization goal to guide the subsequent control strategy solution. Dynamic programming is solved based on the control cost function to obtain the optimal control strategy. As an effective optimization method, dynamic programming can find the optimal solution in multi-stage decision problems and determine the most appropriate control decision for each stage by recursively decomposing the state space. In this process, the optimal control strategy considers all possible control inputs and system state changes to ensure that the system can achieve the target control effect at the lowest cost. The optimal control strategy is time discretized. The continuous control strategy is converted into a discrete control sequence for practical application in the fan control system. The control quantities of the master cooling fan and the slave cooling fan at each discrete time point are extracted respectively to obtain a discrete control sequence. The master-slave synchronization analysis of the discrete control sequence is performed to ensure that the master fan and the slave fan can achieve synchronous control. The speed synchronization check of each fan is performed to ensure that the speed of the master fan and the slave fan is consistent or within the allowable deviation range during their respective control processes. The speed synchronization constraint between the master and slave fans is constructed to ensure that the speed change of the slave fan can keep up with the change of the master fan during the collaborative control process to avoid disconnection or loss of control between the master and slave fans. According to the synchronization control constraint, the discrete control sequence is corrected to ensure that it meets the speed synchronization requirements. The corrected discrete control sequence forms a collaborative control instruction set for the master and slave fans to obtain the initial control sequence.

[0041] Step S500, inputting the system operation characteristic vector and the initial control sequence into the dual Q network model to perform master-slave fan control correction to obtain a control correction amount;

[0042] Specifically, the master-slave component is extracted from the running feature vector of the system. By analyzing the running state of the system, the feature components of the main cooling fan and the feature components of each slave cooling fan are identified, and they are classified and processed respectively to obtain the master-slave feature classification matrix. According to the master-slave feature classification matrix and the initial control sequence, a master-slave collaborative state space is established. This state space regards the main cooling fan as the decision-making subject and the slave cooling fan as the execution subject, forming a master-slave collaborative learning environment. In this environment, the control decision of the main fan affects the execution of the slave fan, and the execution feedback of the slave fan affects the adjustment of the main fan. Through the establishment of the master-slave relationship, a dynamic optimization mechanism based on feedback and learning is formed in the control process. After the master-slave collaborative learning environment is input into the dual Q network model, the hierarchical structure of the master control network and the slave control network is constructed. The dual Q network model adopts the method of deep reinforcement learning and uses neural networks to approximate the Q value function to realize complex control strategy learning. In this model, the master control network is responsible for the decision-making process and generates the optimal control strategy by learning and optimizing the feedback signal from the system. In order to make the master control network more adaptable to the needs of the cooling system, a temperature-driven decision-making unit is embedded in the master control network so that temperature changes can affect the output of the master control strategy. This design helps to improve the accuracy and adaptability of control, especially when the ambient temperature changes greatly. The control strategy of the master cooling fan is evaluated online based on the hierarchical control architecture. The operating status of the system is monitored in real time through the master control network, and the master control correction instruction is generated based on the real-time feedback. The master control correction instruction reflects the adjustment requirements of the current master fan control strategy and is passed down to each slave cooling fan through the master control network to guide the response of the slave fan. At the same time, according to the master control adjustment sequence, the control parameters of multiple slave cooling fans are coordinated and optimized to ensure that the adjustment of the slave fan can be consistent with the adjustment of the master fan. The response deviation of each slave fan is calculated through the slave control network to obtain the slave control deviation sequence. In order to ensure the synergy between the master fan and the slave fan, the master control adjustment sequence and the slave control deviation sequence are time-aligned and phase-corrected to ensure that the response of each fan is consistent on the time axis, avoiding system instability or control imbalance caused by the difference in reaction speed of different fans. Timing alignment and phase correction help to establish a synchronous response relationship between the master and slave fans, ensuring that the fans in the system can work together to achieve the best cooling effect. The control strategy of the master and slave fans is corrected as a whole according to the synchronization control parameters. The correction includes speed correction and phase correction. The speed correction is used to adjust the speed of the fan so that the speed of each fan can be more accurately matched, while the phase correction ensures that the response phase of the fan can be synchronized, thereby avoiding inefficiency or vibration problems caused by timing asynchrony. By controlling the correction process, the coordinated control of multiple cooling fans is achieved to ensure that the system can operate efficiently and stably under different working conditions.

[0043] Step S600: compensate and correct the initial control sequence based on the control correction amount to obtain a target control sequence, and send the target control sequence to the motor drivers of each cooling fan through the CAN bus to output a collaborative control result.

[0044] Specifically, the control correction amount is subjected to master-slave decomposition operation to extract the speed correction parameter of the master cooling fan and the phase following parameter of each slave cooling fan. The correction amount is distributed according to the master-slave relationship through mathematical operation to obtain a matrix containing the correction parameters of the master cooling fan and each slave cooling fan, reflecting the correction requirements of each fan. The master-slave correction parameter matrix is ​​subjected to compensation superposition operation with the initial control sequence. The correction amount is combined with the original control sequence to adjust the control parameters to ensure that each fan can adjust the speed and phase according to the correction requirements during operation, so that the fan system can achieve the best heat dissipation effect. On this basis, the control instructions of the master cooling fan are marked with priority to ensure that the control instructions of the master fan can be executed first in the control sequence to ensure that the master fan, as the core of the system, can lead the adjustment of other slave fans in the collaborative control. Through these operations, the target control sequence is obtained, which contains the corrected fan control instructions and has a reasonable execution priority. The synchronous following instructions of the slave cooling fan are generated according to the control instructions of the master cooling fan in the target control sequence, and the control timing relationship between the master and slave fans is established to obtain the master-slave collaborative control table. The control instructions in the master-slave collaborative control table are converted into CAN bus communication format and converted into data frames suitable for the communication protocol. Through the conversion process, the control data frame of the master cooling fan and the response data frame of the slave cooling fan are generated, and these data frames constitute a communication data frame sequence. Based on the communication data frame sequence, a real-time communication scheduling strategy is constructed. In order to ensure that the data of the master cooling fan can be transmitted first, the data transmission priority of the master fan is set to be higher, while the data of the slave cooling fan is transmitted in sequence according to the phase difference order. This ensures that the control instructions of the master fan can be issued quickly, and ensures that the speed and phase of the slave fan are adjusted synchronously in a predetermined order. Through the communication scheduling strategy, a communication scheduling sequence is obtained. The communication scheduling sequence is packaged according to the address encoding rules of the motor driver, and a synchronization trigger signal is added to ensure that the control signal can accurately reach each motor driver and can be executed synchronously according to the system requirements. The purpose of adding a synchronization trigger signal is to ensure that each motor driver can start executing the control instruction at the same time to avoid inconsistent control effects due to delay differences. By allocating the bus bandwidth of the driver control instructions, it is ensured that the control instructions of the master cooling fan can be issued first, while the control instructions of the slave cooling fan are executed in sequence according to the synchronization order. Through the design of the time-sharing control sequence, it can be ensured that the control instructions of each fan can be transmitted on time and kept synchronized under limited bus bandwidth. Based on the time-sharing control sequence, control instructions are sent to the motor drivers of the main cooling fan and each slave cooling fan respectively through the CAN bus, and the execution feedback of each cooling fan is collected in real time. These feedback data will be returned to the control system to form a closed-loop control, so that the system can monitor and adjust the working status of each fan in real time during operation, thereby outputting the coordinated control results to ensure the stability and efficiency of the cooling system.

[0045] In the embodiment of the present invention, by establishing a system state data matrix and multi-dimensional dynamic feature extraction, a comprehensive perception and accurate description of the operating state of the cooling system is achieved, providing a reliable data basis for collaborative control; a master-slave fan mapping relationship matrix and a continuous speed control strategy are combined to solve the problem of speed mutation in the multi-fan system and ensure the stability of the system operation; a dual Q network model is introduced to perform master-slave fan control correction, and a scientific master-slave collaborative control mechanism is established to effectively avoid control conflicts; a time-sharing control sequence issuing mechanism based on the CAN bus is designed to ensure the orderly execution of master-slave fan control instructions and improve the real-time and reliability of the system; through multi-level data processing and control strategy optimization, a complete closed-loop control system is established to enhance the system's adaptability to complex working conditions.

[0046] In a specific embodiment, the process of executing step S100 may specifically include the following steps:

[0047] The temperature data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the temperature sensor to obtain the original temperature data; the speed data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the speed sensor to obtain the original speed data; the vibration data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the vibration sensor to obtain the original vibration data;

[0048] The original temperature data, the original speed data and the original vibration data are combined in time series to obtain an original data sequence, and the original data sequence is subjected to zero-mean normalization processing to obtain a normalized data sequence;

[0049] Sliding window segmentation is performed on the standardized data sequence to obtain a data subsequence set, and the data subsequence set is subjected to fast Fourier transform to obtain a frequency domain feature matrix;

[0050] Perform multi-scale decomposition on the frequency domain feature matrix to obtain the time-frequency feature matrix, and perform principal component analysis on the time-frequency feature matrix to obtain the reduced-dimensional feature matrix;

[0051] According to the dimension reduction feature matrix, the original temperature data, the original speed data and the original vibration data are reorganized to obtain a temperature data sub-matrix, a speed data sub-matrix and a vibration data sub-matrix;

[0052] The temperature data sub-matrix, the speed data sub-matrix, and the vibration data sub-matrix are concatenated to obtain the system state data matrix.

[0053] Specifically, multi-dimensional data collection is performed on the system. The core of data collection includes three types of data: temperature, speed, and vibration, which are key parameters for evaluating the status of the fan. The temperature sensor is used to collect the temperature data of the main fan and the slave fan, the speed sensor is used to obtain the speed data of the fan, and the vibration sensor is used to monitor the running stability of the fan. These sensors are arranged at each key position of the fan system to ensure that the working status of the entire system is fully and accurately reflected. Through these sensors, the original temperature data, speed data, and vibration data are obtained, which are recorded as , and . Perform time series combination on the original data. Through time series combination, the temperature, speed and vibration data are arranged in time order to form a comprehensive data sequence. The form of the data sequence is expressed as: ;

[0054] The data series contains three types of data at the same time point. The original data series is normalized to zero mean to eliminate the influence of different dimensions or different ranges, making different types of data comparable. Obtained by the following formula:

[0055] ;

[0056] in, and are the mean and standard deviation of the data, expressed as:

[0057] ;

[0058] Through standardization, all data are converted into a standard normal distribution with a mean of zero and a standard deviation of one. Sliding window segmentation is performed on the standardized data sequence. The size of the sliding window needs to be determined according to the actual situation. Suppose the window size is selected as , then the sliding window operation divides the data into multiple subsequences. The length of each subsequence is , and the window moves forward a fixed step size each time it slides. After sliding window processing, multiple data subsequences are obtained ( , each subsequence contains consecutive The data of each time point. The form of each subsequence is:

[0059] ;

[0060] Each data subsequence represents the temperature, speed and vibration data of the master and slave fans in a certain period of time. Perform a fast Fourier transform on each subsequence to convert from the time domain to the frequency domain. The calculation formula for the fast Fourier transform is:

[0061] ;

[0062] in, Indicates subsequences at frequency The Fourier transform result at is an imaginary unit, is the number of points of Fourier transform. The data after Fourier transform It is a frequency domain feature matrix that can reflect the periodic and frequency-related components in the data. Multiscale decomposition is performed on the frequency domain feature matrix. Multiscale decomposition is a technique used to analyze different frequency components of a signal and can extract features at different time scales. Use wavelet transform or other multiscale decomposition methods to decompose the frequency domain feature matrix to obtain the time-frequency feature matrix. , which combines the time and frequency change information to reflect the system's operating status in a more detailed manner. Perform principal component analysis on the time-frequency feature matrix to reduce the data dimension by extracting the principal components of the data while retaining as much original information as possible. Through principal component analysis, the time-frequency feature matrix is ​​reduced in dimension to obtain a reduced dimension matrix containing the most important features. , and its calculation formula is:

[0063] ;

[0064] in, is the principal component matrix, which contains the direction of the largest variance in the data. , reorganize the original temperature data, speed data and vibration data to obtain the temperature data sub-matrix , speed data sub-matrix and the vibration data submatrix These sub-matrices represent the extraction results of temperature, speed and vibration features in the data space after dimensionality reduction. The sub-matrices are concatenated to obtain the final system state data matrix:

[0065] ;

[0066] matrix It is a representation of multi-dimensional system status data, including the dimensionality reduction features of the temperature, speed and vibration data of the master and slave fans in different time windows.

[0067] In a specific embodiment, the process of executing step S200 may specifically include the following steps:

[0068] Perform time series adjacent data difference calculation on the temperature data submatrix, and perform sampling time interval normalization operation to obtain the temperature change rate matrix;

[0069] The temperature change rate matrix is ​​sorted according to the numerical value, and the maximum inter-class variance method is used to determine N-1 segmentation thresholds. The temperature change interval is divided into N control intervals according to the segmentation thresholds to obtain the temperature control partition matrix.

[0070] A two-dimensional state space is constructed for each interval corresponding to the temperature control partition matrix, with the horizontal axis representing the temperature value and the vertical axis representing the temperature change rate. A speed response weight is assigned to each temperature state point to obtain a partition state space model.

[0071] Based on the partitioned state space model, the mapping relationship from temperature state to speed response is fitted to obtain the interval discrete speed response function.

[0072] The interval discrete speed response function is integrated as the fuzzy rule input, and the temperature state is fuzzified using the triangular membership function to obtain an adaptive fuzzy controller. The output response curve of the adaptive fuzzy controller is interpolated by piecewise cubic spline to obtain the control response curve.

[0073] A gain adaptive PID controller is designed based on the control response curve. The particle swarm algorithm is used to optimize the PID parameters to obtain the optimal PID parameter set. The optimal PID parameter set is weightedly fused with the control rules of the adaptive fuzzy controller to generate a continuous speed control strategy.

[0074] Specifically, the temperature data is differentially calculated for adjacent data in time series to obtain the temperature change rate matrix. Assume that the temperature data submatrix , which is represented by temperature data collected at multiple time points, such as:

[0075] ;

[0076] in, Indicated in The temperature value at each time point. Perform differential calculation on adjacent time series data, that is, perform differential calculation on each pair of adjacent temperature values ​​to calculate the temperature change rate. Temperature change rate It is expressed by the following formula:

[0077] ;

[0078] in, Indicated in The temperature change between a time point and the previous time point. By calculating the difference of all adjacent time points, the temperature change rate matrix is ​​obtained: , which is of the form:

[0079] ;

[0080] Normalize the sampling time interval of the temperature change rate matrix to eliminate the influence of different time sampling rates. Assume that the sampling time interval is , the temperature change rate matrix is ​​normalized as follows:

[0081] ;

[0082] Normalized temperature change rate matrix Reflects the speed of temperature change per unit time. In order to facilitate the subsequent control interval division, the normalized temperature change rate matrix is ​​sorted, and the maximum inter-class variance method is used to determine the segmentation threshold of the temperature change interval. The maximum inter-class variance method maximizes the ratio of the inter-class variance to the intra-class variance to find the most appropriate segmentation point and realize the division of the temperature change interval. Suppose the elements of the temperature change rate matrix are The specific steps of the maximum inter-class variance method include: sorting the data according to the temperature change rate; calculating the inter-class variance of different split points, and selecting the split point that maximizes the inter-class variance. Assume that the temperature change interval is divided into subintervals, the split point is ( ). The division of temperature variation range is expressed as:

[0083] ;

[0084] These split points divide the temperature change rate matrix into control intervals and form a temperature control partition matrix , which is of the form:

[0085] ;

[0086] A two-dimensional state space model is established for each control interval. In the state space model, the horizontal axis represents the temperature value. , the vertical axis represents the temperature change rate , and assign a speed response weight to each state point , which reflects the mapping relationship between the temperature change rate and the speed. By fitting the relationship between the temperature state and the speed response, the interval discrete speed response function is obtained: , which is of the form:

[0087] ;

[0088] in, For fitting functions, each temperature state point can be mapped to the corresponding speed response. The interval discrete speed response function is integrated as the fuzzy rule input so that the temperature state can be controlled by the adaptive fuzzy controller. In order to enable the controller to flexibly respond to different temperature changes, the triangular membership function is used to fuzzify the temperature state. The form of the triangular membership function is expressed as:

[0089] ;

[0090] in, are the vertices of the triangle function, is the width of the triangle, and the fuzzy membership function of the temperature control is obtained. Through fuzzy processing, the output response curve of the controller is obtained The response curve represents the temperature change state to the control quantity In order to optimize the controller, a gain adaptive PID controller is designed based on the control response curve, and the PID parameters are optimized by the particle swarm optimization algorithm. The output of the PID controller is expressed as:

[0091] ;

[0092] in, To control the output, is the error, , , The proportional, integral and differential gains are obtained through the particle swarm optimization algorithm. , so that the control system can achieve the best performance while meeting the response speed and steady-state error requirements. The optimal PID parameter set is weightedly integrated with the control rules of the adaptive fuzzy controller to generate the final continuous speed control strategy This strategy can dynamically adjust the fan speed according to the temperature status, and can adaptively adjust the control parameters according to different temperature changes, thereby achieving efficient coordinated control of the fans and ensuring the best cooling effect of the system.

[0093] In a specific embodiment, the process of executing step S300 may specifically include the following steps:

[0094] The system status data matrix is ​​grouped according to the master and slave cooling fans, the data matrix of the master cooling fan is marked as a master control matrix, and the data matrices of multiple slave cooling fans are marked as slave control matrices;

[0095] The master control matrix and the slave control matrix are standardized to obtain the standardized time series data of the master and slave fans, and the standardized time series data are subjected to spectrum analysis and time domain feature calculation to obtain the system operation feature vector;

[0096] The state space equation of the master-slave fan is established according to the system operation characteristic vector, and the control equation including the state feedback term and the follower feedback term is constructed to obtain the master-slave coupling state equation. The stability analysis of the master-slave coupling state equation is performed, and the feedback gain parameters are calculated to obtain the coordinated control parameters of the master-slave fan.

[0097] According to the collaborative control parameters, the master-slave mapping function is constructed, the state transfer matrix is ​​calculated, the follow-up control amount of the slave fan to the master fan is obtained, and the follow-up control amount is transformed into a symmetric matrix to establish the master-slave state correspondence relationship and obtain the state transfer matrix;

[0098] The state transition matrix is ​​calibrated to obtain a mapping relationship matrix between a main cooling fan and a plurality of slave cooling fans.

[0099] Specifically, the system status data matrix is ​​properly processed and grouped. Considering that the target cooling system has both a main cooling fan and multiple slave cooling fans, the system status data matrix is ​​grouped according to the different roles of these fans. Assuming that the temperature, speed, vibration and other data of the target cooling system are collected through various sensors, all these data constitute the system status data matrix For the main cooling fan, its data matrix is ​​expressed as , for the cooling fan, its data matrix is ​​expressed as , and assuming that there are A slave cooling fan, so the slave control matrix The dimension is ,in is the data dimension recorded by each fan. The master control matrix and the slave control matrix are clearly distinguished by data grouping so that targeted operations can be performed in subsequent processing. The grouped data are standardized to eliminate the influence of different data dimensions and dimensions, so that the data between different fans are comparable. For the master control matrix and the slave control matrix, they are respectively subjected to zero mean standardization, that is, each data item Standardize as follows:

[0100] ;

[0101] in, It is The fan data items, and They are In this way, the standardized master control matrix is ​​obtained. and slave control matrix , represents the standardized time series data of each fan. The standardized data matrix and Perform spectrum analysis and time domain feature calculation. Spectrum analysis converts time domain signals into frequency domain signals through fast Fourier transform, revealing the frequency components of the signal. Temperature signal of a fan Perform Fourier transform and get the spectrum:

[0102] ;

[0103] in, represents the frequency domain representation of the signal, is the Fourier transform operator, is a frequency variable. By analyzing the spectrum of the standardized data, the frequency domain feature matrix is ​​obtained and , revealing the energy distribution of the main fan and the slave fan in the frequency domain. The time domain features are calculated, and the time domain characteristics of the signal are characterized by using statistics such as mean, variance, skewness, and kurtosis. The calculated time domain feature matrix and Used to describe the operating status of the fan. These time domain features and frequency domain features are combined to form the system operation feature vector , which is expressed as:

[0104] ;

[0105] The operating status of the fan is reflected through the feature vector. Based on the system's operating feature vector , establish the state space equation of the master and slave fans. The state space equation is a mathematical model that describes the dynamic behavior of the system. Its basic form is:

[0106] ;

[0107] in, is the state vector of the system, is the control input, and are the state matrix and input matrix of the system respectively. In this system, the dynamics of the main cooling fan and the slave cooling fan are coupled, so it is necessary to construct the master-slave coupling state equation. Suppose the state vector of the main control fan is , the state vector of the slave controlled fan is , the control equation is:

[0108] ;

[0109] ;

[0110] In order to describe the coordination behavior between the master and slave fans, state feedback and follower feedback are introduced. These feedback terms are expressed as:

[0111] ;

[0112] ;

[0113] in, and are the feedback gain matrices of the master fan and slave fan, respectively. By solving these state equations and performing stability analysis, the coordinated control parameters of the master and slave fans are obtained, thereby achieving dynamic adjustment and coordination. and , construct the master-slave mapping function to describe the follow-up control amount of the slave fan to the master fan. Assume that the response of the slave fan is , the response of the main control fan is , mapping function Described by the following formula:

[0114] ;

[0115] By calculating the state transfer matrix , obtain the follow-up control quantity of the slave fan, and perform symmetric matrix transformation to construct the master-slave state correspondence. Assume that the response deviation of the slave fan is , and its follow-up control amount is calculated by the following formula:

[0116] ;

[0117] By the state transfer matrix Perform parameter calibration to obtain the mapping relationship matrix of the master and slave cooling fans , reflects the coordinated control relationship between the master and slave fans, and is ultimately used to optimize the collaborative control strategy of the fans.

[0118] In a specific embodiment, the process of executing step S400 may specifically include the following steps:

[0119] Extracting a mapping vector of a master cooling fan and mapping vectors of multiple slave cooling fans based on a mapping relationship matrix, and generating a master-slave mapping vector group;

[0120] The master-slave mapping vector group is input into the target control model, the state transfer cost function of the master-slave fan is established, the state cost matrix is ​​obtained, and the state cost matrix is ​​weighted according to the control parameters in the continuous speed control strategy to construct the coordinated control objective function of the master-slave fan and obtain the control cost function;

[0121] Based on the control cost function, dynamic programming is solved to obtain the optimal control strategy, and the optimal control strategy is time discretized to extract the control amount of the main cooling fan and the slave cooling fan at each time point to obtain a discrete control sequence;

[0122] The master-slave synchronization of the discrete control sequence is analyzed, and the speed synchronization constraints between the master and slave fans are constructed to obtain the synchronization control constraints. The discrete control sequence is modified according to the synchronization control constraints, and the collaborative control instruction set of the master and slave fans is established to obtain the initial control sequence.

[0123] Specifically, the mapping vectors of the main cooling fan and the plurality of slave cooling fans are extracted from the mapping relationship matrix of the system. Contains the relationship between the fans in the system. Assume that the dimension of the mapping relationship matrix is ,in is the number of fans, is the number of control parameters (such as speed, temperature, etc.) corresponding to each fan. By extracting the row vector and column vector in the matrix, the mapping vectors of the main cooling fan and the slave cooling fan are obtained. The mapping vector of the main cooling fan is expressed as , and the mapping vector of each cooling fan is expressed as (in These mapping vectors reflect the control dependency between the master and slave fans. These vectors are organized into master-slave mapping vector groups. Mapping vector group It is expressed as:

[0124] ;

[0125] The mapping vector is input into the target control model to establish the state transition cost function of the master and slave fans. In the control system, the cost function is used to measure the quality of the control strategy, and its purpose is to minimize the energy consumption, error or other related indicators in the control process. State transition cost function Used to describe the transition cost from one state to another, expressed as:

[0126] ;

[0127] in, The system is at time The state vector of is the control input (e.g. speed control quantity), and is the weight matrix, which is used to adjust the weights of state error and control input. is the total time of the control process. By solving the state transition cost function, we can get the state cost matrix of the system: , which represents the cost of system state transition under different control inputs. In order to optimize the control effect, the state cost matrix Weighted processing is performed, taking into account the control parameters in the continuous speed control strategy. Assume that the control parameters in the continuous speed control strategy are and (corresponding to the control gains of the main cooling fan and the slave cooling fan respectively), the weighted state cost matrix is ​​expressed as:

[0128] ;

[0129] in, and Represent the cost matrices of the master cooling fan and the slave cooling fan respectively. The coordinated control objective function of the master and slave fans is constructed through the weighted cost matrix. , which is expressed as:

[0130] ;

[0131] This objective function combines the costs of the main cooling fan and the slave cooling fan and provides a basis for dynamic programming solution. , using the dynamic programming algorithm to solve the optimal control strategy. Dynamic programming solves the optimal control strategy in stages and obtains the optimal control path that minimizes the cost function. At each moment By solving the optimization problem, the control quantities of the main cooling fan and the slave cooling fan are obtained. and The optimal control strategy obtained is expressed as a set of control decisions, namely the optimal control sequence , which is of the form:

[0132] ;

[0133] For optimal control strategy Perform time discretization. In the discretization process, the continuous control quantity is converted into the control value at discrete time points. The obtained discrete control sequence is expressed as:

[0134] ;

[0135] Perform master-slave synchronization analysis to ensure that the speeds of the master cooling fan and the slave cooling fan are consistent to avoid system instability or mutual interference between fans. When analyzing, introduce speed synchronization constraints , which ensures that the speed difference between the master fan and the slave fan does not exceed a certain set threshold ,Right now:

[0136] ;

[0137] By introducing synchronous control constraints, the speed between the master and slave fans is ensured to be always within a reasonable range, thus avoiding instability caused by excessive speed differences. The discrete control sequence is modified according to the synchronous control constraints to obtain the collaborative control instruction set of the master and slave fans, including the control instructions of the master cooling fan and the slave cooling fan at each time point. The initial control sequence is obtained through the optimized control sequence , which represents the optimal control decision of each fan in the whole control process.

[0138] In a specific embodiment, the process of executing step S500 may specifically include the following steps:

[0139] Extract the master and slave components of the system operation feature vector, classify the feature components of the master cooling fan and the feature components of each slave cooling fan, and obtain a master and slave feature classification matrix;

[0140] According to the master-slave feature classification matrix and the initial control sequence, a master-slave collaborative state space is established, and the master cooling fan is set as the decision-making subject and the slave cooling fan is set as the execution subject to obtain a master-slave collaborative learning environment;

[0141] The master-slave collaborative learning environment is input into the dual Q network model to construct a hierarchical structure of the master control network and the slave control network, and a temperature-driven decision-making unit is embedded in the master control network to obtain a hierarchical control architecture.

[0142] Based on the hierarchical control architecture, the control strategy of the master cooling fan is evaluated online, and the master control correction instruction is generated. The control signal is transmitted downward through the master control network to obtain the master control adjustment sequence. The control parameters of multiple slave cooling fans are collaboratively optimized according to the master control adjustment sequence. The response deviation of each slave cooling fan is calculated through the slave control network to obtain the slave control deviation sequence.

[0143] The master control adjustment sequence and the slave control deviation sequence are time-aligned and phase-corrected, a synchronous response relationship between the master and slave fans is established, and the synchronous control parameters are obtained. The control strategy of the master and slave fans is corrected as a whole according to the synchronous control parameters, and a control correction amount including a speed correction amount and a phase correction amount is constructed.

[0144] Specifically, by analyzing the system status data, the characteristic components of the main cooling fan and multiple slave cooling fans are identified. It consists of data collected by multiple sensors (such as temperature, speed, vibration, etc.), which can be expressed as:

[0145] ;

[0146] in, represents the eigenvector of the main cooling fan, and Indicates The feature vectors of the slave cooling fans are divided into the following categories: , and the characteristic components of the cooling fan are classified as , forming a master-slave feature classification matrix , expressed as:

[0147] ;

[0148] Based on master-slave feature classification matrix and initial control sequence Establish a master-slave collaborative state space. In this state space, the master cooling fan is set as the decision-making subject (i.e., the maker of the control strategy), while the slave cooling fan is the execution subject, responsible for responding according to the master control instructions. Define the state space of the system as ,in Represents at the moment According to the master-slave feature matrix and the initial control sequence , establish a master-slave collaborative learning environment , expressed as:

[0149] ;

[0150] In this environment, the master cooling fan generates decisions based on the current state and strategy and passes them to the slave cooling fans. The slave cooling fans perform corresponding control tasks according to the master control signal to achieve the temperature control target of the entire system. The master-slave collaborative learning environment is input into the dual Q network model for further control strategy optimization. DQN is a reinforcement learning method used to optimize control strategies through state-action pairs. The master control network in the model Calculate the control strategy of the main cooling fan, and the slave control network Then the control response of the slave cooling fan is calculated. In order to realize the master-slave collaborative control, a hierarchical structure including the master control and slave control networks is constructed. The master control network generates the control decision of the master cooling fan through the temperature-driven decision unit. The temperature-driven decision unit is modeled in the following way:

[0151] ;

[0152] in, is in state The immediate reward function under is the discount factor, It is the action of the master control network (i.e. the control instruction of the main cooling fan). Through this process, the master control network can generate reasonable control decisions based on the current state and pass the decision to the slave control network. The function of the slave cooling fan is similar to that of the master control network, which calculates the response of the slave cooling fan and adjusts its own control strategy according to the decision of the master control network. The control strategy of the slave control network is described by the following formula:

[0153] ;

[0154] Through calculation, the master control network and the slave control network work together to form a hierarchical control architecture. Based on the hierarchical control architecture, the control strategy of the main cooling fan is evaluated online. Through the evaluation process, the master control correction instruction is generated The control signal is transmitted downward to the slave control network through the master control network. The control signal can adjust the operation of the slave cooling fan so that it can follow the changes of the master cooling fan as accurately as possible, thereby ensuring the coordinated operation between the master and slave fans. The master control adjustment sequence obtained and the control deviation sequence It is the adjustment amount calculated through the network. In order to ensure the control consistency of the master and slave fans, the master control adjustment sequence and the slave control deviation sequence are time-aligned and phase-corrected. Through time alignment, it is ensured that the control instructions of the master cooling fan at each time point can match the response of the slave cooling fan. Phase correction is used to ensure that the response of the slave fan is synchronized with the response of the master fan. Synchronous control parameters It is used to measure the degree of synchronization between the master and slave fans and is calculated using the following formula:

[0155] ;

[0156] when When the speed is too large, the control strategy needs to be adjusted to ensure the synchronization between the fans. The control strategy of the master and slave fans is corrected as a whole according to the synchronous control parameters to obtain the control correction amount including the speed correction amount and the phase correction amount. and phase correction Together they constitute the control correction , which reflects the adjustment required for synchronization between the master and slave fans, and ultimately achieves the temperature control target of the system. The overall corrected control strategy is expressed as:

[0157] ;

[0158] Through this process, an adaptive master-slave fan collaborative control system is built, which can not only handle changes in dynamic environments but also adjust control strategies according to real-time data to ensure that the temperature control effect of the system is always maintained in the optimal state.

[0159] Among them, based on the hierarchical control architecture, the control strategy of the main cooling fan is evaluated online, a main control correction instruction is generated, and the control signal is transmitted downward through the main control network to obtain a main control adjustment sequence, and the control parameters of multiple slave cooling fans are collaboratively optimized according to the master control adjustment sequence, and the response deviation of each slave cooling fan is calculated through the slave control network to obtain a slave control deviation sequence, including: the master control network in the hierarchical control architecture inputs the first type of real-time data and the second type of real-time data to analyze, wherein the first type of real-time data includes the temperature data and speed data of the main cooling fan, and the second type of real-time data includes the ambient temperature coefficient and the load fluctuation coefficient, and a double-layer state reward function is constructed to obtain the state evaluation result of the master control network; according to the state evaluation result, the first feedback channel and the second feedback channel are used for performance calibration, wherein the first feedback channel processes the temperature control target deviation, and the second feedback channel processes the speed response deviation, and the master control performance deviation is obtained through weighted fusion operation; based on the master control performance deviation, a three-level speed correction strategy is established, including a fast response level, a steady-state adjustment level and a fine compensation level, and the optimal speed adjustment amount is calculated according to different intervals of the temperature change trend. , and obtain the master control correction instruction; perform segmented drive parameter conversion on the master control correction instruction, set differentiated control gain coefficients for different temperature gradient intervals, and obtain the master control adjustment sequence through the multi-channel signal transmission mechanism of the master control network; perform signal decomposition of the master control adjustment sequence according to the master-slave mapping relationship, and set dynamic weight coefficients according to the spatial distribution characteristics of the slave cooling fans, generate reference control signals for each slave cooling fan, and obtain the slave control reference sequence; perform three-dimensional feature space comparison calculation on the slave control reference sequence and the real-time operating parameters of each slave cooling fan, including temperature feature dimension, speed feature dimension and vibration feature dimension, establish a multi-layer deviation evaluation matrix, and obtain the slave control state deviation; construct an adaptive compensator according to the slave control state deviation, including a first-level speed compensation unit, a second-level phase compensation unit and a third-level energy consumption compensation unit, and adjust the control parameters of each slave cooling fan in a linked manner to obtain the slave control adjustment amount; input the slave control adjustment amount into the double verification module of the slave control network, wherein the first verification is for control stability and the second verification is for synchronization performance, and calculate the difference between the actual response and the expected response of each slave cooling fan to obtain the slave control deviation sequence.

[0160] In a specific embodiment, the process of executing step S600 may specifically include the following steps:

[0161] Performing master-slave decomposition operation on the control correction amount, extracting the speed correction parameter of the master cooling fan and the phase following parameters of each slave cooling fan, and obtaining the master-slave correction parameter matrix;

[0162] The master-slave correction parameter matrix is ​​compensated and superimposed on the initial control sequence, and the control instructions of the main cooling fan are marked with priority to obtain the target control sequence;

[0163] Generate a synchronous follow-up instruction of the slave cooling fan according to the master cooling fan control instruction in the target control sequence, and establish a control timing relationship between the master and slave fans to obtain a master-slave collaborative control table;

[0164] The control instructions in the master-slave collaborative control table are converted into the CAN bus communication format to generate a control data frame of the master cooling fan and a response data frame of the slave cooling fan, thereby obtaining a communication data frame sequence;

[0165] A real-time communication scheduling strategy is constructed based on the communication data frame sequence, and the data of the main cooling fan is set to be transmitted first, and the slave cooling fans are transmitted in sequence according to the phase difference to obtain a communication scheduling sequence;

[0166] Pack the communication scheduling sequence according to the address encoding rules of the motor driver, add the synchronization trigger signal, and obtain the driver control instruction;

[0167] The bus bandwidth is allocated to the driver control instructions to ensure that the control instructions of the main cooling fan are issued first, and the control instructions of the slave cooling fans are executed in sequence to obtain a time-sharing control sequence;

[0168] Based on the time-sharing control sequence, control instructions are sent to the motor drivers of the main cooling fan and each slave cooling fan through the CAN bus, and the execution feedback of each cooling fan is collected to output the collaborative control results.

[0169] Specifically, specific correction parameters of the main cooling fan and the slave cooling fan are extracted from the overall control correction amount. It is expressed as:

[0170] ;

[0171] in, Indicates the speed correction of the main cooling fan. Indicates the phase following correction of the slave cooling fan. The correction values ​​in the table are classified as the speed correction value of the main cooling fan. and the phase correction from the cooling fan , and organize these parameters into a master-slave correction parameter matrix ,like:

[0172] ;

[0173] This matrix contains the correction parameters of the master cooling fan and all slave cooling fans. With the initial control sequence Perform compensation superposition operation to obtain the target control sequence The compensation superposition process is completed by weighting the master-slave correction parameters and the initial control sequence item by item. The specific operation is expressed as:

[0174] ;

[0175] Initial Control Sequence represents an unmodified control instruction, and represents the corrected control amount. By superimposing the two, we get the new target control sequence In this process, the control instructions of the main cooling fan have a higher priority, and the control instructions of the main cooling fan are marked with a priority to ensure that the instructions of the main cooling fan are executed first in the entire control process. Target control sequence By sorting the priority instruction mark of the main cooling fan and the control instruction of the slave cooling fan, the reasonable operation of the system is ensured. The synchronous follow-up instruction of the slave cooling fan is generated according to the main cooling fan control instruction in the target control sequence. The control instruction of the main cooling fan As a guide signal, it is transmitted to the slave cooling fan. By establishing the control timing relationship between the master and slave fans, the response time of the slave cooling fan is arranged according to the control timing of the master cooling fan. For example, the control instruction of the master cooling fan will precede the response from the cooling fan , and the response of the slave fan needs to be adjusted according to the phase difference with the master fan. The process of establishing this timing relationship is described by the following formula:

[0176] ;

[0177] in, Indicates the control time of the main cooling fan. Indicates the control time of the cooling fan. It is the time delay determined by the phase difference. On this basis, a master-slave collaborative control table is generated, which lists the control relationship and timing arrangement between the master cooling fan and each slave cooling fan. This control table not only contains the control instructions of each fan, but also includes the specific time sequence of instruction execution to ensure that the collaborative operation of all fans can achieve the expected effect. The control instructions in the master-slave collaborative control table are converted to the CAN bus communication format. The control instructions are converted into the format required by the CAN bus, and the control data frames of the master cooling fan and the slave cooling fan are generated. For example, the control data frame of the master cooling fan It is expressed as:

[0178] ;

[0179] in, is the unique identifier of the main cooling fan, It is the control data of the master cooling fan. For the slave cooling fan, the response data frame Similarly expressed as:

[0180] ;

[0181] These data frames are sorted according to the timing to form a complete communication data frame sequence Based on the generated communication data frame sequence, a real-time communication scheduling strategy is constructed. The data of the main cooling fan needs to be transmitted first, while the data of the slave cooling fan should be transmitted in the order of phase difference with the main cooling fan. Communication Scheduling Sequence Build according to the following rules:

[0182] ;

[0183] This scheduling strategy ensures that the data transmission of the main cooling fan is prioritized, and the data of the slave cooling fan is transmitted in phase difference order to maintain the synchronization of the system. After the communication scheduling sequence is built, the communication scheduling sequence is packaged according to the address encoding rules of the motor driver, and a synchronization trigger signal is added to generate the final driver control instruction. Driver control instruction for:

[0184] ;

[0185] In order to ensure communication efficiency, the bus bandwidth is allocated to the driver control instructions to ensure that the control instructions of the main cooling fan are issued first, and the control instructions of the slave cooling fans are executed in sequence. Through bandwidth management, the instructions of the main cooling fan take up a larger bandwidth first, while the instructions of the slave cooling fans are transmitted in sequence according to the phase order, forming a time-sharing control sequence. :

[0186] ;

[0187] Through the time-sharing control sequence, the precise coordinated control of the main cooling fan and the slave cooling fans is finally achieved. The control instructions are sent to the motor drivers of the main cooling fan and each slave cooling fan through the CAN bus, and the execution feedback of each cooling fan is collected to output the coordinated control results. These feedback signals are used to further optimize the control strategy to ensure the stability and efficiency of the system operation.

[0188] See also Figure 2 , Figure 2 A schematic block diagram of the structure of a coordinated control system 200 for high-speed motors of multiple cooling fans provided in an embodiment of the present application, such as Figure 2 As shown, a coordinated control system 200 for high-speed motors of multiple cooling fans includes:

[0189] The acquisition module 210 is used to perform multi-dimensional data acquisition and standardization processing on each cooling fan unit in the target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and multiple slave cooling fans;

[0190] A division module 220 is used to divide the temperature data submatrix in the system state data matrix into N control intervals, and perform discrete speed response to obtain a continuous speed control strategy;

[0191] A solution module 230 is used to extract multi-dimensional dynamic features from the system state data matrix to obtain a system operation feature vector, and solve the state feedback equation to obtain a mapping relationship matrix between the main cooling fan and the plurality of slave cooling fans;

[0192] An analysis module 240 is used to input the mapping relationship matrix and the continuous speed control strategy into the target control model to perform control parameter analysis and obtain an initial control sequence;

[0193] Correction module 250, used for inputting system operation characteristic vector and initial control sequence into dual Q network model to perform master-slave fan control correction to obtain control correction amount;

[0194] The output module 260 is used to compensate and correct the initial control sequence based on the control correction amount to obtain a target control sequence, and send the target control sequence to the motor driver of each cooling fan through the CAN bus to output the coordinated control result.

[0195] Through the collaborative cooperation of the above-mentioned components, by establishing a system status data matrix and multi-dimensional dynamic feature extraction, a comprehensive perception and accurate description of the operating status of the cooling system is achieved, providing a reliable data basis for collaborative control; a combination of the master-slave fan mapping relationship matrix and the continuous speed control strategy is adopted to solve the problem of speed mutation in the multi-fan system and ensure the stability of the system operation; a dual Q network model is introduced to correct the master-slave fan control, and a scientific master-slave collaborative control mechanism is established to effectively avoid control conflicts; a time-sharing control sequence issuing mechanism based on the CAN bus is designed to ensure the orderly execution of the master-slave fan control instructions and improve the real-time and reliability of the system; a complete closed-loop control system is established through multi-level data processing and control strategy optimization, enhancing the system's adaptability to complex working conditions.

[0196] See also Figure 3 , Figure 3A schematic block diagram of the structure of a collaborative control device 300 for high-speed motors of multiple cooling fans provided in an embodiment of the present application, wherein the collaborative control device 300 for high-speed motors of multiple cooling fans includes a processor 301 and a memory 302, wherein the processor 301 and the memory 302 are connected via a system bus 303, wherein the memory 302 may include a non-volatile storage medium and an internal memory.

[0197] The non-volatile storage medium can store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 301, the processor 301 can execute any of the above-mentioned methods for cooperatively controlling high-speed motors of multiple cooling fans.

[0198] The processor 301 is used to provide computing and control capabilities to support the operation of the coordinated control device 300 of the high-speed motors of multiple cooling fans.

[0199] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 301, the processor 301 can execute any of the above-mentioned coordinated control methods for high-speed motors of multiple cooling fans.

[0200] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the collaborative control device 300 for high-speed motors of multiple cooling fans involved in the scheme of the present application. The specific collaborative control device 300 for high-speed motors of multiple cooling fans may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0201] It should be understood that the processor 301 may be a central processing unit (CPU), and the processor 301 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0202] It should be noted that, technical personnel in the relevant field can clearly understand that, for the convenience and conciseness of description, the specific working process of the collaborative control device 300 for high-speed motors of multiple cooling fans described above can refer to the corresponding process of the aforementioned collaborative control method for high-speed motors of multiple cooling fans, and will not be repeated here.

[0203] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement a method for coordinated control of high-speed motors of multiple cooling fans as provided in an embodiment of the present application.

[0204] The computer-readable storage medium may be an internal storage unit of the collaborative control device 300 for high-speed motors of multiple cooling fans in the aforementioned embodiment, such as a hard disk or memory of the collaborative control device 300 for high-speed motors of multiple cooling fans. The computer-readable storage medium may also be an external storage device of the collaborative control device 300 for high-speed motors of multiple cooling fans, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc., equipped with the collaborative control device 300 for high-speed motors of multiple cooling fans.

[0205] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0206] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0207] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for cooperatively controlling high-speed motors of multiple cooling fans, characterized in that: include: Performing multi-dimensional data collection and standardization processing on each cooling fan unit in a target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and a plurality of slave cooling fans; Dividing the temperature data submatrix in the system state data matrix into N control intervals, and performing discrete speed response to obtain a continuous speed control strategy; Performing multi-dimensional dynamic feature extraction on the system state data matrix to obtain a system operation feature vector, and solving a state feedback equation to obtain a mapping relationship matrix between the main cooling fan and the plurality of slave cooling fans; Inputting the mapping relationship matrix and the continuous speed control strategy into a target control model to perform control parameter analysis to obtain an initial control sequence; Inputting the system operation feature vector and the initial control sequence into a dual Q network model to perform master-slave fan control correction to obtain a control correction amount; specifically comprising: extracting master-slave components from the system operation feature vector, classifying the feature components of the master cooling fan and the feature components of each slave cooling fan respectively, and obtaining a master-slave feature classification matrix; A master-slave collaborative state space is established according to the master-slave feature classification matrix and the initial control sequence, and the master cooling fan is set as the decision-making subject and the slave cooling fan is set as the execution subject to obtain a master-slave collaborative learning environment; the master-slave collaborative learning environment is input into the dual Q network model to construct a hierarchical structure of the master control network and the slave control network, and a temperature-driven decision unit is embedded in the master control network to obtain a hierarchical control architecture; based on the hierarchical control architecture, the control strategy of the master cooling fan is evaluated online to generate a master control correction instruction, and the control signal is transmitted downward through the master control network to obtain a master control adjustment sequence, and the control parameters of multiple slave cooling fans are collaboratively optimized according to the master control adjustment sequence, and the response deviation of each slave cooling fan is calculated through the slave control network to obtain a slave control deviation sequence; the master control adjustment sequence and the slave control deviation sequence are time-aligned and phase-corrected to establish a synchronous response relationship between the master and slave fans to obtain synchronous control parameters, and the control strategy of the master and slave fans is overall corrected according to the synchronous control parameters to construct a control correction amount including a speed correction amount and a phase correction amount; The initial control sequence is compensated and corrected based on the control correction amount to obtain a target control sequence, and the target control sequence is sent to the motor drivers of each cooling fan through a CAN bus to output a collaborative control result.

2. The method for cooperatively controlling high-speed motors of multiple cooling fans according to claim 1, characterized in that: The multi-dimensional data collection and standardization processing are performed on each cooling fan unit in the target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and multiple slave cooling fans, including: The temperature data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the temperature sensor to obtain the original temperature data; the speed data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the speed sensor to obtain the original speed data; the vibration data of the main cooling fan and multiple slave cooling fans in the target cooling system are collected through the vibration sensor to obtain the original vibration data; Performing time series combination on the original temperature data, the original rotation speed data and the original vibration data to obtain an original data sequence, and performing zero mean normalization processing on the original data sequence to obtain a normalized data sequence; Performing sliding window segmentation on the standardized data sequence to obtain a data subsequence set, and performing fast Fourier transform on the data subsequence set to obtain a frequency domain feature matrix; Performing multi-scale decomposition on the frequency domain feature matrix to obtain a time-frequency feature matrix, and performing principal component analysis on the time-frequency feature matrix to obtain a reduced-dimensional feature matrix; According to the dimension reduction feature matrix, the original temperature data, the original rotational speed data and the original vibration data are reorganized to obtain a temperature data sub-matrix, a rotational speed data sub-matrix and a vibration data sub-matrix; The temperature data sub-matrix, the rotational speed data sub-matrix, and the vibration data sub-matrix are concatenated to obtain a system state data matrix.

3. The method for coordinated control of high-speed motors of multiple cooling fans according to claim 2, characterized in that: The temperature data submatrix in the system state data matrix is ​​divided into N control intervals, and discrete speed response is performed to obtain a continuous speed control strategy, including: Performing time-series adjacent data difference calculation on the temperature data submatrix, and performing sampling time interval normalization operation to obtain a temperature change rate matrix; The temperature change rate matrix is ​​sorted according to the numerical value, and the maximum inter-class variance method is used to determine N-1 segmentation thresholds, and the temperature change interval is divided into N control intervals according to the segmentation thresholds to obtain a temperature control partition matrix; A two-dimensional state space is constructed for each interval corresponding to the temperature control partition matrix, wherein the horizontal axis represents the temperature value and the vertical axis represents the temperature change rate, and a speed response weight is assigned to each temperature state point to obtain a partition state space model; Based on the partitioned state space model, a mapping relationship from temperature state to speed response is fitted to obtain an interval discrete speed response function; The interval discrete speed response function is integrated as a fuzzy rule input, a triangular membership function is used to perform fuzzy processing on the temperature state to obtain an adaptive fuzzy controller, and a piecewise cubic spline interpolation is performed on the output response curve of the adaptive fuzzy controller to obtain a control response curve; A gain adaptive PID controller is designed based on the control response curve, and the PID parameters are optimized using a particle swarm algorithm to obtain an optimal PID parameter set. The optimal PID parameter set is weightedly fused with the control rules of the adaptive fuzzy controller to generate a continuous speed control strategy.

4. The method for cooperatively controlling high-speed motors of multiple cooling fans according to claim 3, characterized in that: The multi-dimensional dynamic feature extraction is performed on the system state data matrix to obtain the system operation feature vector, and the state feedback equation is solved to obtain the mapping relationship matrix between the main cooling fan and the multiple slave cooling fans, including: Grouping the system status data matrix according to the master and slave cooling fans, marking the data matrix of the master cooling fan as a master control matrix, and marking the data matrices of the plurality of slave cooling fans as slave control matrices; The master control matrix and the slave control matrix are standardized to obtain standardized time series data of the master and slave fans, and the standardized time series data are subjected to spectrum analysis and time domain feature calculation to obtain a system operation feature vector; Establishing state space equations of the master and slave fans according to the system operation characteristic vector, constructing a control equation including a state feedback term and a follower feedback term, obtaining a master-slave coupling state equation, performing stability analysis on the master-slave coupling state equation, calculating feedback gain parameters, and obtaining collaborative control parameters of the master and slave fans; Constructing a master-slave mapping function according to the collaborative control parameters, calculating a state transfer matrix, obtaining a follow-up control amount of the slave fan to the master fan, and performing a symmetric matrix transformation on the follow-up control amount to establish a master-slave state correspondence relationship, and obtaining a state transfer matrix; Parameter calibration is performed on the state transition matrix to obtain a mapping relationship matrix between the main cooling fan and the plurality of slave cooling fans.

5. The method for cooperatively controlling high-speed motors of a plurality of cooling fans according to claim 4, characterized in that: The step of inputting the mapping relationship matrix and the continuous speed control strategy into a target control model to perform control parameter analysis to obtain an initial control sequence includes: Extracting a mapping vector of a master cooling fan and mapping vectors of multiple slave cooling fans based on the mapping relationship matrix, and generating a master-slave mapping vector group; Input the master-slave mapping vector group into the target control model, establish the state transfer cost function of the master-slave fan, obtain the state cost matrix, and perform weighted processing on the state cost matrix according to the control parameters in the continuous speed control strategy, construct the coordinated control objective function of the master-slave fan, and obtain the control cost function; Performing dynamic programming based on the control cost function to obtain an optimal control strategy, and performing time discretization processing on the optimal control strategy to extract the control amount of the main cooling fan and the slave cooling fan at each time point to obtain a discrete control sequence; The master-slave synchronization analysis is performed on the discrete control sequence, and the speed synchronization constraint conditions between the master and slave fans are constructed to obtain the synchronization control constraint. The discrete control sequence is modified according to the synchronization control constraint, and the collaborative control instruction set of the master and slave fans is established to obtain the initial control sequence.

6. The method for cooperatively controlling high-speed motors of a plurality of cooling fans according to claim 1, characterized in that: The initial control sequence is compensated and corrected based on the control correction amount to obtain a target control sequence, and the target control sequence is sent to the motor drivers of each cooling fan through the CAN bus to output a collaborative control result, including: Performing a master-slave decomposition operation on the control correction amount, extracting the speed correction parameter of the master cooling fan and the phase following parameters of each slave cooling fan, and obtaining a master-slave correction parameter matrix; Performing a compensation superposition operation on the master-slave correction parameter matrix and the initial control sequence, and marking the priority of the control instructions of the main cooling fan to obtain a target control sequence; Generate a synchronous follow-up instruction of the slave cooling fan according to the master cooling fan control instruction in the target control sequence, and establish a control timing relationship between the master and slave fans to obtain a master-slave collaborative control table; Performing CAN bus communication format conversion on the control instructions in the master-slave collaborative control table, generating a control data frame of the master cooling fan and a response data frame of the slave cooling fan, and obtaining a communication data frame sequence; Based on the communication data frame sequence, a real-time communication scheduling strategy is constructed, and the data of the main cooling fan is set to be transmitted first, and the data of the slave cooling fans are transmitted in sequence according to the phase difference, so as to obtain a communication scheduling sequence; Packing the communication scheduling sequence according to the address coding rule of the motor driver and adding a synchronization trigger signal to obtain a driver control instruction; The bus bandwidth is allocated to the driver control instructions to ensure that the control instructions of the main cooling fan are issued first and the control instructions of the slave cooling fans are executed in sequence to obtain a time-sharing control sequence; Based on the time-sharing control sequence, control instructions are respectively issued to the motor drivers of the main cooling fan and each slave cooling fan through the CAN bus, and the execution feedback of each cooling fan is collected to output the collaborative control result.

7. A coordinated control system for high-speed motors of multiple cooling fans, characterized in that: A method for cooperatively controlling high-speed motors of a plurality of cooling fans according to any one of claims 1 to 6, comprising: an acquisition module, used for performing multi-dimensional data acquisition and standardization processing on each cooling fan unit in a target cooling system to obtain a system status data matrix, wherein the target cooling system includes a main cooling fan and a plurality of slave cooling fans; A division module, used for dividing the temperature data submatrix in the system state data matrix into N control intervals, and performing discrete speed response to obtain a continuous speed control strategy; A solution module, used for performing multi-dimensional dynamic feature extraction on the system state data matrix to obtain a system operation feature vector, and solving a state feedback equation to obtain a mapping relationship matrix between the main cooling fan and the plurality of slave cooling fans; An analysis module, used for inputting the mapping relationship matrix and the continuous speed control strategy into a target control model to perform control parameter analysis and obtain an initial control sequence; A correction module, used for inputting the system operation characteristic vector and the initial control sequence into a dual Q network model to perform master-slave fan control correction to obtain a control correction amount; The output module is used to compensate and correct the initial control sequence based on the control correction amount to obtain a target control sequence, and send the target control sequence to the motor drivers of each cooling fan through a CAN bus to output a collaborative control result.

8. A coordinated control device for high-speed motors of multiple cooling fans, characterized in that: The cooperative control device of the high-speed motors of the plurality of cooling fans comprises: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the cooperative control device of the high-speed motors of the multiple cooling fans to execute the cooperative control method of the high-speed motors of the multiple cooling fans as described in any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, a coordinated control method for high-speed motors of multiple cooling fans as described in any one of claims 1-6 is implemented.

Images