Load matching optimization method of multi-channel gate driver and related equipment thereof

Abstract

The invention relates to a load matching optimization method of a multi-channel gate driver and related equipment thereof. The method comprises the following steps of: acquiring grid current, grid voltage and switching frequency of the multi-channel parallel grid driver to obtain driver load operation data; performing region division processing on the operation space to obtain a safe operation region and a transition region; inputting a first adaptive prediction model for feature extraction to obtain a load balancing prediction result; parallel optimization calculation is carried out on the control strategy to obtain each driving channel control parameter, and each driving channel control parameter comprises a grid driving voltage regulating variable and a driving time sequence regulating variable; performing real-time adjustment on each driving channel control parameter to obtain a load balance adjustment instruction; and performing performance index evaluation calculation to obtain a performance optimization parameter, and updating the first adaptive prediction model to obtain a second adaptive prediction model. According to the invention, through cooperative control of voltage regulation and time sequence compensation, dynamic balance of each channel load is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of gate drivers, and in particular to a load matching optimization method for a multi-channel gate driver and related equipment. Background Art

[0002] With the rapid development of high-power inverter technology, multi-channel parallel gate drive technology has become an important means to improve power density and switching performance. However, in practical applications, due to the discreteness of device parameters, asymmetric wiring, and uneven temperature distribution, there is often a load imbalance problem between the drive channels, resulting in a decrease in drive capability and an increase in switching losses, which seriously affects the reliability of the system.

[0003] Traditional load balancing methods mainly rely on static compensation and empirical parameter adjustment, which is difficult to cope with load fluctuations under dynamic conditions. At the same time, there are complex electromagnetic coupling and thermal coupling effects between the drive channels. These coupling effects will change dynamically with the changes in switching frequency and load conditions, bringing great challenges to load balancing control. Summary of the invention

[0004] The main purpose of the present invention is to provide a load matching optimization method for a multi-channel gate driver and related equipment. The present invention realizes dynamic balance of the load of each channel through coordinated control of voltage regulation and timing compensation.

[0005] To achieve the above object, the present invention provides a load matching optimization method for a multi-channel gate driver, comprising the following steps: The gate current, gate voltage and switching frequency of the multi-channel parallel gate driver are collected to obtain driver load operation data, wherein the driver load operation data includes impedance parameters, coupling parameters and temperature parameters; Based on the load operation data of the driver, the operation space is divided into regions to obtain a safe operation region and a transition region; Inputting the driver load operation data into a first adaptive prediction model for feature extraction to obtain a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value; Performing parallel optimization calculation on the control strategy according to the load balancing prediction result to obtain control parameters of each drive channel, wherein the control parameters of each drive channel include a gate drive voltage adjustment amount and a drive timing adjustment amount; The control parameters of each drive channel are adjusted in real time to obtain a load balancing adjustment instruction; A performance indicator evaluation calculation is performed based on the load balancing adjustment instruction to obtain a performance optimization parameter, and the first adaptive prediction model is updated according to the performance optimization parameter to obtain a second adaptive prediction model.

[0006] The present invention also provides a load matching optimization system for a multi-channel gate driver, comprising: An acquisition module is used to acquire gate current, gate voltage and switching frequency of a multi-channel parallel gate driver to obtain driver load operation data, wherein the driver load operation data includes impedance parameters, coupling parameters and temperature parameters; A region division module, used for performing region division processing on the operation space based on the load operation data of the driver to obtain a safe operation region and a transition region; A feature extraction module, used for inputting the driver load operation data into a first adaptive prediction model for feature extraction to obtain a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value; A calculation module, used for performing parallel optimization calculation on the control strategy according to the load balancing prediction result to obtain control parameters of each drive channel, wherein the control parameters of each drive channel include a gate drive voltage adjustment amount and a drive timing adjustment amount; An adjustment module, used for adjusting the control parameters of each drive channel in real time to obtain a load balancing adjustment instruction; An updating module is used to perform performance indicator evaluation calculation based on the load balancing adjustment instruction to obtain performance optimization parameters, and to update the first adaptive prediction model according to the performance optimization parameters to obtain a second adaptive prediction model.

[0007] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0008] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0009] In summary, the technical solution provided by the present invention adopts a high-precision data acquisition and parameter extraction method, combined with a comprehensive analysis of impedance parameters, coupling parameters and temperature parameters, to accurately obtain the load operation characteristics of the multi-channel parallel gate driver, and establishes a judgment mechanism for the safe operation area and the transition area based on the regional division of the operation space and the dynamic boundary optimization, so as to effectively prevent the system from entering an unsafe working state; an adaptive prediction model is used to extract load features, and through multi-level feature decoupling and fusion, the load distribution and fluctuation trend between channels are accurately predicted, providing a decision-making basis for the optimization of the control strategy; a parallel optimization algorithm is introduced to perform real-time calculation of the control strategy, and the dynamic balance of the load of each channel is achieved through the coordinated control of voltage regulation and timing compensation; a hierarchical control mechanism and a priority scheduling strategy are designed to ensure the smooth execution of control instructions and avoid system instability due to sudden changes in the adjustment process; a complete performance evaluation and model update mechanism is established, and the performance of the prediction model is continuously optimized through online learning and parameter adaptive adjustment to ensure the long-term stable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 1 is a schematic diagram of the steps of a load matching optimization method for a multi-channel gate driver according to an embodiment of the present invention; Figure 2 is a structural block diagram of a load matching optimization system of a multi-channel gate driver in one embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0011] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0013] Reference Figure 1 , this embodiment provides a load matching optimization method for a multi-channel gate driver, comprising the following steps: S1, collecting gate current, gate voltage and switching frequency of the multi-channel parallel gate driver to obtain driver load operation data, the driver load operation data includes impedance parameters, coupling parameters and temperature parameters; Among them, the operating parameters of the multi-channel parallel gate driver are collected and processed, which mainly include the measurement of gate current, gate voltage and switching frequency to obtain the load operation data of the driver. For the sampling of gate current, high-precision current sensors and high-speed data acquisition modules are used to capture the current changes of the multi-channel gate driver at different time points to form a gate current sampling sequence. This sequence digitally represents the dynamic characteristics of the current over time. At the same time, the gate voltage is sampled by configuring a high-bandwidth voltage probe and a precise sampling circuit to convert the voltage signal of the multi-channel gate driver into a digital gate voltage sampling sequence. This sequence records the voltage waveform changes of each drive channel for subsequent calculation of resistance, capacitance and inductance parameters. Based on the collected gate current sampling sequence and gate voltage sampling sequence, the formula (in is the impedance, and The impedance parameters of the driver are calculated by combining the gate resistance, Miller capacitance and parasitic inductance values ​​(voltage and current, respectively). In order to refine the results, the gate resistance, Miller capacitance and parasitic inductance values ​​are extracted in combination with the circuit model. The gate resistance reflects the power loss characteristics of the driver, the Miller capacitance describes the phase relationship between voltage and current, and the parasitic inductance reveals the dynamic characteristics of the circuit under high frequency conditions. After obtaining the impedance parameters of a single drive channel, the coupling calculation is performed on adjacent drive channels. The coupling calculation includes the analysis of electromagnetic coupling and thermal coupling. The calculation of the electromagnetic coupling coefficient is achieved by measuring the mutual inductance and capacitance values ​​between adjacent channels, and the formula is used. (in is the electromagnetic coupling coefficient, It is mutual induction. and The thermal coupling coefficient is calculated by establishing a heat conduction model, combining the thermal power consumption and temperature rise of each channel, to analyze the transfer rate and influence of heat between adjacent channels. These coupling parameters help identify interference and mismatch problems that occur in the operation of multi-channel drivers. Temperature monitoring is performed on each drive channel of a multi-channel parallel gate driver. By arranging high-precision temperature sensors at each key node of each channel, temperature changes are continuously recorded to form a temperature monitoring data sequence. Based on this sequence, the temperature distribution of each drive channel is calculated by spatial interpolation method and thermal distribution model to obtain specific temperature parameters. The temperature parameters include the temperature values ​​of each channel, and also the temperature gradient value, that is, the rate of change of temperature between channels. This information can reflect the thermal stability and heat dissipation balance of the driver. The impedance parameters, coupling parameters and temperature parameters are subjected to data normalization. The data of different dimensions are converted into dimensionless values ​​through the data normalization method to generate the load operation data of the driver.

[0014] S2, dividing the operating space into regions based on the load operating data of the drive to obtain a safe operating region and a transition region; Specifically, the impedance parameters in the driver load operation data are analyzed, the key characteristics of each drive channel are extracted, and the maximum allowable power value is calculated. By combining the impedance characteristics of each channel such as resistance, capacitance, inductance, and the dynamic current and voltage data during operation, based on the formula (in is the maximum allowable power, and where gate voltage and impedance are gate voltage and impedance respectively) to obtain the power limit of each channel. On this basis, the power deviation relationship between channels is used to construct a load deviation threshold matrix, which is used to describe the degree of load deviation of each channel and provide a basic basis for demarcating the operating area. The boundary threshold of the operating area is obtained by threshold optimization of the load deviation threshold matrix. Based on the optimization algorithm, such as through linear programming or nonlinear optimization technology, the data in the deviation matrix is ​​mapped to a reasonable boundary range to ensure that the demarcated safety area can meet the requirements of operating stability. The operating area boundary threshold is used to mark the safe range of the load at this stage, and to identify the boundary conditions for entering the transition area. Based on the coupling parameters in the driver load operation data, the degree of interference between each drive channel is quantified and calculated to generate a channel coupling strength matrix. Analyze key parameters such as the electromagnetic coupling coefficient and the thermal coupling coefficient, and use the formula (in For Channel and The coupling strength, and The interference intensity is quantified by the electric field and magnetic field strength between the two channels, respectively, and the mutual coupling relationship of all channels is expressed in matrix form. The channel coupling strength matrix can reflect the interactive characteristics of the driver under different operating conditions. The matrix is ​​used to correct the threshold of the operating area boundary to ensure that the influence of the interference between channels is taken into account. After completing the coupling correction, the temperature parameter is introduced to optimize the threshold of the operating area boundary. Based on the temperature information in the load operation data of the driver, the temperature distribution of each drive channel is subjected to thermal stress analysis to calculate the thermal stress distribution matrix. The calculation of thermal stress distribution is based on material mechanics and heat conduction models, combined with actual temperature monitoring data, to analyze the influence of temperature gradient on channel reliability. By converting the temperature gradient value into a stress distribution matrix, temperature compensation is implemented on the threshold of the operating area boundary, so as to more accurately reflect the safety range under actual operating conditions. Based on the operating area boundary threshold and the load deviation threshold matrix, the initial boundary of the operating space is preliminarily divided. Multi-dimensional feature extraction is performed on the initial operating space boundary to comprehensively describe the characteristics of the operating space. The characteristic vector of the operating space is extracted by calculating indicators such as load balance, temperature uniformity and coupling strength. Among them, load balancing indicates the uniformity of load distribution in each channel, temperature uniformity reflects the thermal distribution state of the driver, and coupling strength describes the degree of interference between channels. The regional division results are optimized according to the operating space feature vector, and the optimized operating space division results are obtained by adjusting boundary conditions, re-evaluating feature weights, and other methods. Based on the optimized operating space division results, the load operation data of the driver is monitored in real time, and it is dynamically matched with the optimized regional division results to determine whether the current operating state is in a safe area or a transition area. If the system operating state is in the safe area, the driver can maintain efficient and stable operation; if it enters the transition area, the early warning mechanism is triggered, and further control strategy adjustments are initiated to prevent system instability.

[0015] S3, inputting the driver load operation data into the first adaptive prediction model for feature extraction to obtain a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value; It should be noted that the driver load operation data is input into the load decoupling layer of the first adaptive prediction model. This layer adopts an autoencoder structure to perform nonlinear separation on the impedance parameters, coupling parameters and temperature parameters in the driver operation data. The separation process decomposes the original complex multi-dimensional input data into a set of load feature components with independent physical meanings through the hidden layer embedding operation of the autoencoder, thereby effectively eliminating redundant information in the data and retaining key features. The load feature components are input into the multi-channel complementary feature extraction layer of the first adaptive prediction model. In this layer, the model analyzes the independent characteristics of each drive channel through parallel channel feature subnetworks. Each subnetwork adopts a residual connection structure to avoid the gradient vanishing problem and improve the efficiency of feature extraction. At the same time, in order to capture the correlation information between channels, these subnetworks communicate through lateral information transmission paths, so that the characteristics of a certain channel can be dynamically adjusted to adapt to the status of other channels, thereby extracting more comprehensive channel correlation features. Channel correlation features can reflect the mutual coupling performance of each drive channel during operation. The channel correlation features are input into the dynamic coupling modeling layer of the first adaptive prediction model, which is designed with a graph convolutional network structure. The graph convolutional network can model complex multi-channel associations based on the representation of nodes and edges, and capture the load coupling effect of the driver by dynamically constructing the association map between channels. Through the graph convolution operation, the model can extract global and local load characteristics in different graph structures, obtain coupling dynamic features, and reflect the coupling relationship of the multi-channel system under complex operating conditions. The coupling dynamic features are input into the multi-scale temporal feature fusion layer, which contains multiple parallel long short-term memory unit groups. Each unit group processes the features of a specific time scale and can extract the temporal characteristics of the load from short-term fluctuations and long-term trends. The cross-scale attention mechanism is used to weightedly fuse features of different time scales to ensure that the model can have a comprehensive and accurate description of the temporal dynamics. The fused multi-scale temporal features reflect the changing law of the driver load in the time dimension. In the thermal-electric coupling analysis layer, the model uses a dual-stream network structure to process thermal and electrical features respectively. The thermal feature flow mainly focuses on the distribution and change of temperature parameters and analyzes the operating performance of the driver under thermal conditions; while the electrical feature flow analyzes the dynamic characteristics of the electrical load based on impedance and coupling parameters. Through the adaptive weight fuser, these two feature streams are merged into a unified thermoelectric coupling feature to achieve a comprehensive description of the operating status of the multi-channel driver under thermal and electrical conditions. The thermoelectric coupling features are input into the distribution prediction layer and the fluctuation analysis layer respectively. In the distribution prediction layer, the model adopts a conditional variational autoencoder structure to generate a probability model of the load distribution between channels by learning the implicit distribution of the current load state, and obtains the predicted value of the load distribution between channels. This value describes the distribution of the load between the channels in the form of probability. In the fluctuation analysis layer, the model uses a wavelet neural network structure to capture load fluctuations of different frequencies through multi-resolution analysis to generate load fluctuation prediction values.The advantage of wavelet neural network is that it can handle high-frequency short-term fluctuations and low-frequency long-term trends at the same time. The load distribution prediction value between channels and the load fluctuation prediction value are integrated into the load balancing prediction result, which reflects the current load distribution status and provides early warning capabilities for future fluctuations.

[0016] S4, performing parallel optimization calculation on the control strategy according to the load balancing prediction result to obtain control parameters of each drive channel, wherein the control parameters of each drive channel include a gate drive voltage adjustment amount and a drive timing adjustment amount; Specifically, the load distribution prediction value between channels is extracted from the load balancing prediction result, and the load imbalance of each driving channel is calculated by data analysis method. The load imbalance reflects the degree to which the load of each driving channel deviates from the balanced state, and is an important input variable for optimizing the driving voltage and timing control. Based on the load imbalance, the voltage compensation value required for each channel is calculated to adjust the voltage deviation of the channel load and improve its load balancing performance. At the same time, a dynamic compensation model is established based on the load fluctuation prediction value. The core goal of this model is to design a driving timing compensation strategy to cope with the dynamic changes of the load between different channels. The driving timing compensation strategy specifically includes dead time adjustment and switch delay compensation. The dead time adjustment is used to optimize the safe interval time of the power device during the switching process, while the switch delay compensation reduces the timing deviation caused by load fluctuation by fine-tuning the trigger time of the channel switch. These dynamic compensation strategies ensure the timing stability and load response capability of the driver during operation. The calculated load imbalance is input into the weight distributor to generate the voltage adjustment coefficient matrix of each driving channel. The matrix determines the voltage adjustment range of each channel through weighted calculation, providing a reference for the specific adjustment of the gate drive voltage. At the same time, the drive timing compensation strategy is decomposed in parallel to obtain the channel-level timing control sequence. The channel-level timing control sequence cooperates with the voltage regulation coefficient matrix to optimize the voltage control quantity and timing control quantity of each channel, so that it can achieve consistent dynamic performance under complex multi-channel load conditions. Orthogonal optimization is performed based on the voltage regulation coefficient matrix and the channel-level timing control sequence, and a multi-channel collaborative control matrix is ​​generated through a multi-objective optimization algorithm. This matrix contains both voltage control quantity and timing control quantity, which is used to fully describe the optimized control strategy of the driver. In this process, the role of the orthogonal optimization algorithm is to ensure that the mutual influence between voltage and timing can be effectively processed, so that the generated control strategy has the best coordination in the global scope. The multi-channel collaborative control matrix is ​​input into the constraint processor to perform boundary constraint processing on the control parameters. The boundary constraints include the upper and lower limits of the voltage amplitude and the minimum and maximum allowable ranges of the timing interval, so as to avoid the control parameters exceeding the physical limitations of the actual hardware. After constraint processing, the generated initial values ​​of the control parameters meet the basic hardware operation requirements. The initial values ​​of the control parameters are optimized by multi-objective iteration. Through the multi-objective optimization algorithm, a balance is sought between load balancing and switching loss to ensure that the load between channels is evenly distributed and the power loss during the switching process is minimized. The result of iterative optimization is a set of optimal control parameter combinations, including the voltage adjustment amount and timing adjustment amount of each channel. Based on the optimal control parameter combination, the gate drive voltage and drive timing are allocated to specific drive units according to the channel to obtain the control parameters of each drive channel.

[0017] S5, adjusting the control parameters of each drive channel in real time to obtain a load balancing adjustment instruction; Among them, the control parameters of each drive channel are monitored, and the real-time operation state matrix is ​​constructed by collecting the dynamic operation data of the driver such as current, voltage, temperature and timing in real time. The matrix includes the current load balance and temperature distribution data, which respectively reflect the balance of multi-channel load distribution and the thermal state of each channel. Based on the real-time operation state matrix, the working area of ​​the system is determined. By analyzing the matrix data, the current operation parameters are compared with the pre-defined safe operation area threshold to generate a region identification signal. The region identification signal is used to determine whether the system is in the safe operation area. If the system parameters exceed the boundary of the safe operation area, the identification signal will prompt to enter the transition area and trigger the corresponding control mode adjustment. The region identification signal is input into the control strategy selector to generate a control mode signal. The control mode signal is dynamically adjusted according to the current operation state of the system, including steady-state control mode and fast response mode. When the system is in the safe operation area, the steady-state control mode will maintain the current control strategy and fine-tune the voltage and timing to maintain load balance; when the system enters the transition area, the fast response mode will start immediately and quickly adjust the key parameters to avoid performance degradation or instability. According to the control mode signal, the control parameters of each drive channel are classified and processed to generate a hierarchical control sequence. The hierarchical control sequence includes a voltage adjustment sequence and a timing adjustment sequence, wherein the voltage adjustment sequence calculates the adjustment amount of the gate voltage according to the load deviation of each channel, and the timing adjustment sequence optimizes the switching timing of the drive signal according to the dynamic response requirements of the system. The classification process ensures that different types of control parameters can be optimized separately and solve different load imbalance problems in a targeted manner. The hierarchical control sequence is prioritized to generate a control execution sequence table. The control execution sequence table sorts the adjustment instructions according to the urgency of load balancing, and puts the adjustment instructions with the greatest impact on system performance in the priority execution position. The urgency is determined based on the size of the load imbalance in the real-time operating state matrix and the rate of change of the temperature gradient, thereby ensuring that the most urgent problems can be solved in a timely manner. The control execution sequence table is input into the actuator scheduling module to generate a specific timing execution plan. The timing execution plan determines the specific execution time of the control instructions of each drive channel by comprehensively considering the hardware response time, the dependency of the adjustment instructions, and the synergy between multiple channels. The formulation of this plan needs to balance the needs of fast response and stable transition, avoiding conflicts caused by too much synchronous execution and ensuring that the adjustment can take effect as soon as possible. The timing execution plan is buffered and an instruction cache queue is generated. The instruction cache queue stores and schedules the adjustment instructions to achieve smooth switching of instructions, thereby avoiding system instability caused by frequent parameter adjustments. Based on the instruction cache queue, the adjustment instructions are synthesized, the voltage and timing adjustment information of all channels are integrated, and the load balancing adjustment instructions are generated.

[0018] S6, performing performance indicator evaluation calculation based on the load balancing adjustment instruction to obtain performance optimization parameters, and updating the first adaptive prediction model according to the performance optimization parameters to obtain a second adaptive prediction model.

[0019] Specifically, the performance of the execution effect of the load balancing adjustment instruction is evaluated, and the performance indicator matrix is ​​calculated by collecting the changes in the key parameters of the system after adjustment. The performance indicator matrix contains three core dimensions: load balancing index, temperature consistency index and response time index. The load balancing index reflects the uniformity of the load distribution of each drive channel and is a direct reflection of the evaluation of the load balancing optimization effect; the temperature consistency index describes the difference in thermal distribution between each channel and is a key measure of the thermal stability of the system; the response time index is used to evaluate the time required for the system to achieve load adjustment from receiving the adjustment instruction to verify the real-time and efficiency of the adjustment. Based on the performance indicator matrix, the model error analysis is carried out to quantify the prediction accuracy of the first adaptive prediction model under actual load conditions. By comparing the model prediction output with the actual operation results, the prediction error distribution is generated. The prediction error distribution includes two parts: load distribution prediction error and fluctuation prediction error. The load distribution prediction error reflects the model's prediction deviation of the load balancing situation of each channel, while the fluctuation prediction error reflects the model's insufficient prediction ability of dynamic changes in the load. Through the analysis of these errors, the performance bottlenecks and shortcomings of the model in different scenarios are identified, providing a clear direction for subsequent optimization. The prediction error distribution is input into the adaptive corrector, and the model parameter correction is calculated. The adaptive corrector uses the error feedback mechanism to adjust the network parameters of each layer of the model in a targeted manner to improve the prediction ability of the model. The model parameter correction provides a specific adjustment basis for each layer, and updates the parameters of each layer differently according to different error distribution characteristics. The corrected parameters are used to optimize the overall performance of the model to make it more suitable for the actual operation requirements of the drive. Based on the model parameter correction, the load decoupling layer of the first adaptive prediction model is updated to optimize its load feature extraction structure. The load decoupling layer is the basic part of the entire model, and the nonlinear separation of the drive operation data is achieved through the autoencoder structure. During the update process, the corrected parameters will improve the decomposition accuracy of the decoupling layer for impedance parameters, coupling parameters and temperature parameters, and generate more accurate load feature components. The multi-channel complementary feature extraction layer of the first adaptive prediction model is dynamically weighted to optimize the channel feature network. The multi-channel complementary feature extraction layer extracts the independent characteristics of each drive channel through a parallel channel feature subnetwork, and captures the mutual influence between channels through lateral information transfer. Through dynamic weight adjustment, the information transfer weights between sub-networks are optimized according to the feedback of the performance indicator matrix, thereby improving the complementary ability and robustness of the channel feature network under complex operating conditions. Based on the performance indicator matrix, the graph structure of the dynamic coupling modeling layer of the first adaptive prediction model is optimized. The dynamic coupling modeling layer models the coupling effect between channels through the graph convolutional network structure, and its performance directly determines the model's ability to predict dynamic changes in load.By adjusting the node representation and edge weights in the graph structure, the optimized coupling modeling network can more accurately capture the correlation characteristics between channels, thereby improving the modeling accuracy of the system dynamic behavior. The updated load feature extraction structure, the optimized channel feature network and the updated coupling modeling network are integrated to construct the second adaptive prediction model. With the support of parameter optimization and structural adjustment at each layer, the updated model can more accurately reflect the complex operating state of the multi-channel gate driver.

[0020] In one example, the gate current, gate voltage and switching frequency of a multi-channel parallel gate driver are collected to obtain driver load operation data, where the driver load operation data includes impedance parameters, coupling parameters and temperature parameters, including: Sampling the gate current of the multi-channel parallel gate driver to obtain a gate current sampling sequence, and sampling the gate voltage of the multi-channel parallel gate driver to obtain a gate voltage sampling sequence; Based on the gate current sampling sequence and the gate voltage sampling sequence, impedance calculation is performed to obtain impedance parameters, where the impedance parameters include gate resistance value, Miller capacitance value and parasitic inductance value; The impedance parameters of adjacent driving channels are coupled and calculated to obtain coupling parameters, where the coupling parameters include electromagnetic coupling coefficient and thermal coupling coefficient; Performing temperature monitoring on each driving channel of the multi-channel parallel gate driver to obtain a temperature monitoring data sequence, and performing temperature distribution calculation based on the temperature monitoring data sequence to obtain temperature parameters, wherein the temperature parameters include a temperature value and a temperature gradient value of each driving channel; The impedance parameters, coupling parameters and temperature parameters are processed by data standardization to obtain the driver load operation data.

[0021] In this example, the gate current of multi-channel parallel gate drivers is sampled by an accurate sensor system. Assume that the sampled gate current signal is , indicating time The gate current value at the moment. After analog-to-digital conversion, the current signal forms a digital gate current sampling sequence , where each Represents discrete time points At the same time, a high-precision voltage probe is used to sample the gate voltage to obtain a sampling signal The signal is also converted into a gate voltage sampling sequence through analog-to-digital conversion. ,in Represents discrete time points The voltage value of the driver is obtained by calculating the impedance using the gate current and voltage data obtained by sampling. ,in is the impedance value, is the voltage value, is the current value, calculated point by point Get the impedance value sequence at the corresponding time point In order to extract the specific impedance parameters, the circuit model is used to decompose the impedance into gate resistance 、Miller capacitance and parasitic inductance The gate resistance is obtained by steady-state measurement as ,in and Represent the average values ​​of voltage and current respectively. The Miller capacitance is calculated by frequency response analysis as ,in represents the gate charge change, represents the gate voltage change. The parasitic inductance is derived using the phase relationship between high-frequency voltage and current as ,in is the voltage change across the inductor, is the change in the current change rate. After obtaining the impedance parameters of a single channel, the coupling relationship between adjacent drive channels is analyzed to calculate the electromagnetic coupling coefficient and thermal coupling coefficient. Through the analysis channel and Channel Mutual inductance and self-perception , The ratio is defined as ,in is the mutual inductance between channels, indicating that the channel The current in the channel The magnetic field effect generated on the surface. Thermal coupling coefficient Based on the heat conduction equation, the temperature gradient between adjacent channels and heat transfer power Calculated as ,in It is from the channel To channel The temperature sensors installed in each drive channel monitor the temperature data of the drive in real time and generate a temperature monitoring data sequence ,in Indicates channel The temperature gradient is calculated by analyzing the spatial distribution of temperature data. , thus generating temperature parameters, including the temperature value and temperature gradient value of each channel. These temperature parameters can intuitively reflect the thermal distribution state and thermal imbalance of the multi-channel system. The acquired impedance parameters, coupling parameters and temperature parameters are processed by data standardization to generate driver load operation data. Assuming a certain parameter set is , the normalization process is done by linear normalization formula ,in represents the standardized value, and Represent the minimum and maximum values ​​of the original data respectively. The parameters of different dimensions and ranges are unified into the range of [0,1] through processing.

[0022] In one example, the operating space is divided into regions based on the driver load operating data to obtain a safe operating region and a transition region, including: Analyze the impedance parameters in the driver load operation data to obtain the maximum allowable power value of each drive channel, and calculate the load deviation threshold matrix based on the maximum allowable power value; Based on the load deviation threshold matrix, the threshold is optimized to obtain the operating area boundary threshold, which is used to divide the safe operating area and the transition area; Based on the coupling parameters in the driver load operation data, the interference degree between each drive channel is quantitatively calculated to obtain a channel coupling strength matrix, which is used to correct the boundary threshold of the operation area; Performing thermal stress analysis on the temperature distribution of each drive channel according to the temperature parameters in the drive load operation data to obtain a thermal stress distribution matrix, and performing temperature compensation on the boundary threshold of the operation area based on the thermal stress distribution matrix; Based on the operation area boundary threshold and the load deviation threshold matrix, the region is divided to obtain the initial operation space boundary, and multi-dimensional feature extraction is performed on the initial operation space boundary to obtain the operation space feature vector. The operation space feature vector includes load balance, temperature uniformity and coupling strength. The region is optimized according to the operating space feature vector to obtain an optimized operating space division result, and the operating state is determined in real time based on the optimized operating space division result to obtain a safe operating area and a transition area.

[0023] In this example, the maximum allowable power value for each drive channel is obtained by analyzing the impedance parameters in the driver load operating data. Maximum allowable power value It is calculated based on the impedance parameters of the channel, reflecting the maximum power load that the channel can withstand. The formula is ; in, It is a channel The maximum allowable voltage, It is a channel The gate resistance, is the switching frequency, and is the Miller capacitance. By calculating the , the maximum power carrying capacity of the system is obtained, which provides boundary conditions for load balancing and safe operation. Based on the maximum allowable power value, the load deviation threshold matrix is ​​calculated. Assume that there are The actual load power of the channel is , the deviation threshold matrix Each element of is defined as: ; in, Indicates channel and Channel By constructing the matrix , quantifies the degree of load balance between different channels. If the matrix element exceeds the preset threshold, it means that the system enters the transition region. Perform threshold optimization to obtain the boundary threshold of the operating area. The optimization process of the boundary threshold is based on the optimization algorithm, such as finding the threshold combination that can maximize the safe operating space through linear programming. Assuming that the optimization goal is to minimize the proportion of the transition area, the optimization problem is described as: ; in, It is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0; is the boundary threshold variable, and its optimal value is determined by optimization calculation. At the same time, based on the coupling parameters in the driver load operation data, the interference degree between each drive channel is quantitatively calculated to generate the channel coupling strength matrix The electromagnetic coupling strength between channels is given by the formula: ; Calculated, where It is a channel and The mutual induction, and are the self-inductance of the channel respectively. Meanwhile, the thermal coupling strength is given by the formula: ; Calculated, where It is a channel and The heat flow between is the temperature difference between the two channels. The electromagnetic coupling and thermal coupling strength together constitute the coupling strength matrix , which is used to correct the boundary threshold of the operating area. In order to enhance the accuracy of the operating area division, the temperature parameters in the drive load operation data are subjected to thermal stress analysis to generate a thermal stress distribution matrix The calculation of thermal stress is based on the theory of thermoelasticity, using the formula ; in, is the thermal expansion coefficient of the material, is Young's modulus, is the temperature gradient. Thermal stress distribution matrix It can quantify the thermal load state of each channel and perform temperature compensation on the operating area boundary threshold. Through the above calculations, the region division is completed based on the operating area boundary threshold and the load deviation threshold matrix to obtain the initial operating space boundary. On this basis, multi-dimensional feature extraction is performed on the initial operating space boundary to generate the operating space feature vector ,in is the load balancing degree, is the temperature uniformity, is the coupling strength. These eigenvectors can fully describe the operating state of the system. According to the operating space eigenvectors, the region is optimized to obtain the optimized operating space division result. By determining the operating state in real time, the system operating state is divided into a safe operating area or a transition area.

[0024] In one example, the driver load operation data is input into the first adaptive prediction model for feature extraction to obtain a load balancing prediction result, which includes a load distribution prediction value between channels and a load fluctuation prediction value, including: The load operation data of the driver is input into the load decoupling layer of the first adaptive prediction model. The load decoupling layer adopts an autoencoder structure to perform nonlinear separation on the impedance parameter, coupling parameter and temperature parameter to obtain the load characteristic component. Inputting the load feature component into the multi-channel complementary feature extraction layer of the first adaptive prediction model, the multi-channel complementary feature extraction layer includes parallel channel feature sub-networks, each channel feature sub-network adopts a residual connection structure, and establishes a lateral information transmission path between channels to obtain channel correlation features; The channel correlation features are input into the dynamic coupling modeling layer of the first adaptive prediction model. The dynamic coupling modeling layer adopts a graph convolutional network structure to capture the load coupling effect by dynamically constructing the correlation map between channels to obtain the coupling dynamic features. The coupled dynamic features are input into the multi-scale temporal feature fusion layer of the first adaptive prediction model. The multi-scale temporal feature fusion layer includes a parallel group of long short-term memory units. Each group of units processes features of different time scales and fuses features through a cross-scale attention mechanism to obtain multi-scale temporal features. The multi-scale time series features are input into the thermoelectric coupling analysis layer of the first adaptive prediction model. The thermoelectric coupling analysis layer adopts a dual-stream network structure to process thermal features and electrical features respectively, and merges the two feature streams through an adaptive weight fusion device to obtain thermoelectric coupling features. The thermoelectric coupling feature is input into the distribution prediction layer of the first adaptive prediction model. The distribution prediction layer adopts a conditional variational autoencoder structure to generate a probability model of the load distribution between channels based on the current load state to obtain a predicted value of the load distribution between channels. The thermoelectric coupling characteristics are input into the fluctuation analysis layer of the first adaptive prediction model. The fluctuation analysis layer adopts a wavelet neural network structure and captures the load fluctuation characteristics of different frequencies through multi-resolution analysis to obtain the load fluctuation prediction value. Output the load balancing prediction result based on the load distribution prediction value and load fluctuation prediction value between channels.

[0025] In this example, the load operation data of the driver is input into the load decoupling layer of the first adaptive prediction model. This layer adopts an autoencoder structure, which can reduce the dimension and perform nonlinear separation on complex multi-dimensional input data. The load operation data of the driver consists of impedance parameters, coupling parameters and temperature parameters. Assume that the input data is a matrix ,in , respectively represent impedance related data (such as gate resistance 、Miller capacitance , parasitic inductance ), coupling parameters (electromagnetic coupling coefficient , thermal coupling coefficient ) and temperature parameters (temperature distribution of each channel ). The encoder part of the autoencoder compresses the input data into the latent space through a nonlinear mapping function, and the formula is ,in is the load characteristic component, is the weight matrix of the encoder. The decoder part is implemented by the function Reconstruct the input data to verify the accuracy of feature extraction. It can capture the nonlinear characteristics of the input data. Input the multi-channel complementary feature extraction layer, which consists of multiple parallel channel feature sub-networks, each of which processes the data of a driving channel. In order to avoid the gradient vanishing problem and improve the feature extraction efficiency, the sub-network adopts a residual connection structure, and its formula is ,in It is The feature representation of the layer, is the weight matrix is the activation function. In order to capture the dynamic correlation between channels, the channel feature sub-networks establish interactions through lateral information transmission paths, and the weight of the interaction is determined by the shared parameter Represents, generating channel correlation feature matrix . Each element in the matrix Indicates channel and Channel The feature correlation between them can reflect the complementarity of load distribution. Enter the dynamic coupling modeling layer, which dynamically constructs the correlation map between channels through the graph convolutional network structure to capture the complex load coupling effect. The core formula of the graph convolutional network is ,in It is The node feature matrix of the layer, is the adjacency matrix of the graph, is the layer weight matrix, is the activation function. Through multi-layer graph convolution operations, the model can effectively aggregate the neighborhood information of nodes, extract dynamic correlation features between channels, and output coupled dynamic features. . Couple the dynamic features The multi-scale time series feature fusion layer is input, which processes features of different time scales through a parallel long short-term memory network (LSTM) group. The state update formula of each LSTM unit is ,in It's time The hidden state of and They are state weight and input weight respectively. Through multiple groups of LSTM running in parallel, different units process short-term dynamic and long-term trend features. These features are weighted and fused through the cross-scale attention mechanism to generate multi-scale time series features. , which can comprehensively reflect the time dynamics of the load. Multi-scale time series characteristics It is transferred to the thermoelectric coupling analysis layer. This layer uses a dual-flow network structure to process and fuse thermal and electrical features separately. and current characteristics They are extracted through independent sub-networks, and the formula is ,in is the temperature data, The heat flow and current features are combined into a thermoelectric coupling feature through an adaptive weight fuser. , providing input for subsequent predictions. Input the distribution prediction layer and the fluctuation analysis layer respectively. In the distribution prediction layer, the conditional variational autoencoder generates a probability model of load distribution by learning the potential distribution, and its formula is ,in is a hidden variable. The distribution prediction layer outputs the predicted value of the load distribution between channels to describe the distribution of loads between channels. The fluctuation analysis layer captures load fluctuations of different frequencies based on wavelet neural networks. Its formula is: ,in is the wavelet transform function, is the wavelet kernel parameter, outputs the load fluctuation prediction value, and describes the dynamic behavior of the system. The load distribution prediction value between channels and the load fluctuation prediction value are used as the output of the model to generate the load balancing prediction result.

[0026] In one example, the control strategy is optimized and calculated in parallel according to the load balancing prediction result to obtain the control parameters of each drive channel, and each drive channel control parameter includes a gate drive voltage adjustment amount and a drive timing adjustment amount, including: Performing data analysis on the predicted values ​​of load distribution between channels in the load balancing prediction results to obtain the load imbalance of each drive channel, which is used to calculate the drive voltage compensation value; A dynamic compensation model is established based on the load fluctuation prediction value to obtain a drive timing compensation strategy, which includes a dead time adjustment amount and a switch delay compensation amount; The load imbalance is input into the weight distributor to obtain the voltage regulation coefficient matrix of each driving channel, and the voltage regulation coefficient matrix is ​​used to calculate the regulation range of the gate driving voltage; The driving timing compensation strategy is decomposed in parallel to obtain a channel-level timing control sequence, which cooperates with the voltage adjustment coefficient matrix; Based on the voltage regulation coefficient matrix and the channel-level timing control sequence, orthogonal optimization is performed to obtain a multi-channel collaborative control matrix, which includes a voltage control amount and a timing control amount; The multi-channel cooperative control matrix is ​​input into the constraint processor, and the voltage amplitude and the timing interval are bounded to obtain the initial value of the control parameter; Perform multi-objective iterative optimization on the initial values ​​of the control parameters to obtain the optimal control parameter combination, which satisfies the balance requirements of load balancing and switching loss; Based on the optimal control parameter combination, the gate drive voltage and drive timing of each drive channel are allocated to obtain the control parameters of each drive channel.

[0027] In this example, the predicted values ​​of the load distribution between channels in the load balancing prediction results are analyzed to extract the load information of each drive channel, thereby calculating the load imbalance. Assume that the predicted load distribution is ,in Indicates channel The load imbalance is calculated by calculating the deviation of each channel's load from the global average. Expressed as ; in It is the average value of the load power of all channels. The load imbalance reflects the degree to which the load of each channel deviates from the balanced state, which serves as the basis for calculating the drive voltage compensation value. Based on the load fluctuation prediction value , establish a dynamic compensation model to design the drive timing compensation strategy. Load fluctuation prediction value Indicates channel By analyzing the fluctuation characteristics, the dynamic compensation model includes the dead time adjustment and switch delay compensation , and its calculation formula is: ; in and is the proportionality coefficient, is the load imbalance, is the maximum allowable power of the channel. These adjustments are used to dynamically compensate the switching timing of each channel to reduce load mismatch caused by fluctuations. Input weight distributor to generate voltage adjustment coefficient matrix for each drive channel ,in Indicates channel The voltage regulation coefficient of the channel Dependency of load imbalance. Voltage regulation range Calculated by the following formula: ; in Indicates channel The gate drive voltage adjustment amount. The voltage adjustment coefficient matrix is ​​optimized by weight distribution to ensure that the voltage adjustment of each channel is consistent with the overall load state of the system. The drive timing compensation strategy is decomposed in parallel to obtain the channel-level timing control sequence of each channel. The timing control sequence includes the switch dead time and delay compensation time, and is combined with the voltage adjustment coefficient matrix The voltage regulation coefficient matrix and the channel-level timing control sequence are input into the orthogonal optimization module, and a multi-channel collaborative control matrix is ​​generated through a multi-objective optimization algorithm. ,in It is a channel The driving voltage control quantity, is the dead time adjustment, is the delay compensation amount. Orthogonal optimization achieves a trade-off between load balancing and power consumption minimization through the following objective function: ; in and are weight coefficients, which adjust the priority of load balancing and switching loss respectively. The multi-channel cooperative control matrix is ​​input into the constraint processor to impose boundary constraints on the voltage amplitude and timing interval. The constraints include the upper and lower limits of the voltage and the safety range of the timing interval . The initial values ​​of the control parameters generated after processing meet the operating limitations of the physical hardware. Based on the initial values ​​of the control parameters, the optimal control parameter combination that satisfies the load balancing degree and the switching loss balance is calculated through multi-objective iterative optimization. The optimization algorithm uses a genetic algorithm or a particle swarm optimization algorithm to iteratively update the parameters and gradually approach the global optimal solution. The final optimization result provides the voltage adjustment amount and timing adjustment amount of each channel. According to the optimal control parameter combination, the gate drive voltage and drive timing of each drive channel are allocated to obtain the control parameters of each drive channel.

[0028] In one example, the control parameters of each drive channel are adjusted in real time to obtain a load balancing adjustment instruction, including: Monitor the status of each drive channel control parameter to obtain a real-time operation status matrix, which includes the current load balance and temperature distribution data; Based on the real-time operation status matrix, the working area is determined to obtain an area identification signal, and the area identification signal is used to determine whether the system is in a safe operation area; Inputting the area identification signal into the control strategy selector to obtain a control mode signal, the control mode signal includes a steady-state control mode and a fast response mode; Classify and process the control parameters of each driving channel based on the control mode signal to obtain a hierarchical control sequence, which includes a voltage adjustment sequence and a timing adjustment sequence; Prioritize the hierarchical control sequence to obtain a control execution sequence table, which determines the execution order of each adjustment instruction according to the urgency of load balancing; Input the control execution sequence table into the actuator scheduling module to obtain a timing execution plan, which includes the specific execution time of each drive channel control instruction; The timing execution scheme is buffer-managed to obtain an instruction cache queue, which ensures smooth switching of control instructions. Instructions are synthesized based on the instruction cache queue to obtain load balancing adjustment instructions.

[0029] In this example, the control parameters of each drive channel are monitored to obtain the real-time operation status matrix. drive channels, and the operating status of each channel includes real-time load power and temperature The real-time operation state matrix is ​​expressed as: ; in It is a channel The real-time load power, It is a channel By collecting this data online and dynamically updating the matrix , reflecting the operating status of the entire system in real time. Based on the real-time operating status matrix, calculate the load balancing degree and temperature uniformity , used to determine whether the system is running within a safe range. The calculation formula for load balancing is: ; in is the average load power of all channels, expressed as: ; The calculation of temperature distribution uniformity is similar and is defined as: ; in is the average temperature of all channels, expressed as: ; Through and Threshold judgment to generate area identification signal It is used to indicate whether the current system is in the safe operation area. The rules of the area identification signal are: ; in and It is the safety threshold of load balance and temperature distribution. The area identification signal is input to the control strategy selector to determine the control mode signal of the system. The control strategy includes steady-state control mode and fast response mode, and the signal rules are: ; In steady-state mode, the system optimizes control parameters with a smaller adjustment step; in fast response mode, the system will prioritize emergency parameters to restore balance. According to the control mode signal, the control parameters of each drive channel are classified and processed to generate a hierarchical control sequence. Assume that the control parameters include voltage regulation and timing adjustment , the hierarchical control sequence is expressed as: ; The priority of the hierarchical control sequence is based on the urgency of load balancing. The priority is determined by the load imbalance. Decide, press Arrange in descending order to generate a control execution sequence table , represents the execution order of the adjustment instructions. The control execution order table is input into the executor scheduling module to generate a timing execution plan. Assume that each instruction The execution time is , the timing execution plan is: ; in is the total number of instructions, To ensure smooth switching of control instructions, the sequential execution scheme generates instruction cache queues through buffer management. Instruction cache queues It is used to store unexecuted instructions and optimize the execution efficiency of instructions through queue management. The cache mechanism ensures that the adjustment instructions of multiple channels will not conflict. The instructions are synthesized through the instruction cache queue to generate the final load balancing adjustment instructions. The adjustment instructions include voltage and timing adjustment of all channels, according to the timing scheme Execute step by step.

[0030] In one example, a performance indicator evaluation calculation is performed based on a load balancing adjustment instruction to obtain a performance optimization parameter, and a first adaptive prediction model is updated according to the performance optimization parameter to obtain a second adaptive prediction model, including: Performing performance evaluation on the execution effect of the load balancing adjustment instruction to obtain a performance indicator matrix, which includes a load balancing degree indicator, a temperature consistency indicator, and a response time indicator; Perform model error analysis based on the performance indicator matrix to obtain prediction error distribution, which includes load distribution prediction error and fluctuation prediction error; Inputting the prediction error distribution into the adaptive corrector to obtain the model parameter correction amount, and the model parameter correction amount is used to update the network parameters of each layer of the first adaptive prediction model; The load decoupling layer of the first adaptive prediction model is updated based on the model parameter correction amount to obtain an updated load feature extraction structure; Dynamically adjusting the weights of the multi-channel complementary feature extraction layer of the first adaptive prediction model to obtain an optimized channel feature network; Based on the performance indicator matrix, the graph structure of the dynamic coupling modeling layer of the first adaptive prediction model is optimized to obtain an updated coupling modeling network; The updated load feature extraction structure, the optimized channel feature network and the updated coupling modeling network are integrated to obtain a second adaptive prediction model.

[0031] In this example, the performance of the execution effect of the load balancing adjustment instruction is evaluated to generate a performance indicator matrix for evaluating the actual effect of the adjustment strategy. The performance indicator matrix includes load balancing index, temperature consistency index and response time index. Assume that the system contains drive channels, the actual load power of each channel is , the ideal load power distribution is , then the load balancing index Defined as: ; in Indicates the average load power. This formula calculates the deviation of each channel's load distribution from the ideal distribution and is used to quantify the load balancing effect. Temperature consistency index It is used to evaluate the uniformity of temperature distribution of each drive channel, which is defined as: ; in It is a channel The actual temperature, Indicates the average temperature. Response time indicator It represents the time required for the system to achieve the target state from receiving the adjustment instruction. Assume that the adjustment start time is , the steady state time is ,but: ; Through the above formula, the generated performance indicator matrix is ​​expressed as: ; Based on the performance indicator matrix, the model error analysis is performed to calculate the prediction error distribution. The prediction error distribution includes the load distribution prediction error and the fluctuation prediction error. Defined as actual load With the model prediction value The differences: ; The volatility forecast error Indicates actual load fluctuation and forecast fluctuations The differences: ; The prediction error distribution is represented by a vector ; The prediction error distribution is input into the adaptive corrector to generate the model parameter correction. The adaptive corrector is based on the gradient descent method and uses the prediction error to optimize the loss function of the model. Assume that the loss function of the model is: ; in is the weight parameter of the fluctuation error. By taking the derivative of the loss function, we get the correction value : ; in is the learning rate, is the parameter set of the model. The correction is used to update the network parameters of each layer of the first adaptive prediction model. According to the correction, the load decoupling layer of the model is updated to optimize the load feature extraction structure. The load decoupling layer adopts an autoencoder structure, and its update formula is ; in is the encoder weight matrix, is the input data, is the updated feature representation. At the same time, the multi-channel complementary feature extraction layer is dynamically weighted to optimize the channel feature network. The feature extraction network of each channel is based on the residual connection structure, and its dynamic adjustment formula is: ; in It is Layer features, is the weight matrix, is the correction amount. Based on the performance indicator matrix, the graph structure of the dynamic coupling modeling layer is optimized and the adjacency matrix of the graph convolutional network is updated. The optimized graph structure is achieved by updating the adjacency weights: ; in Indicates the correction amount adjusted according to the performance index. The updated load feature extraction structure, the optimized channel feature network and the updated coupling modeling network are integrated to generate a second adaptive prediction model.

[0032] Reference Figure 2 , this embodiment provides a load matching optimization system for a multi-channel gate driver, including: The acquisition module 1 is used to collect the gate current, gate voltage and switching frequency of the multi-channel parallel gate driver to obtain the driver load operation data, and the driver load operation data includes impedance parameters, coupling parameters and temperature parameters; The area division module 2 is used to divide the operation space into areas based on the load operation data of the driver to obtain a safe operation area and a transition area; Feature extraction module 3, used for inputting driver load operation data into the first adaptive prediction model for feature extraction, and obtaining a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value; The calculation module 4 is used to perform parallel optimization calculation on the control strategy according to the load balancing prediction result to obtain the control parameters of each drive channel, wherein the control parameters of each drive channel include the gate drive voltage adjustment amount and the drive timing adjustment amount; The adjustment module 5 is used to adjust the control parameters of each drive channel in real time to obtain a load balancing adjustment instruction; The updating module 6 is used to perform performance indicator evaluation calculation based on the load balancing adjustment instruction to obtain performance optimization parameters, and update the first adaptive prediction model according to the performance optimization parameters to obtain a second adaptive prediction model.

[0033] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0034] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0035] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0036] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0037] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0038] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0039] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A load matching optimization method for a multi-channel gate driver, characterized in that: The following steps are involved: The gate current, gate voltage and switching frequency of the multi-channel parallel gate driver are collected to obtain driver load operation data, wherein the driver load operation data includes impedance parameters, coupling parameters and temperature parameters; Based on the load operation data of the driver, the operation space is divided into regions to obtain a safe operation region and a transition region; Inputting the driver load operation data into a first adaptive prediction model for feature extraction to obtain a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value; Performing parallel optimization calculation on the control strategy according to the load balancing prediction result to obtain control parameters of each drive channel, wherein the control parameters of each drive channel include a gate drive voltage adjustment amount and a drive timing adjustment amount; The control parameters of each drive channel are adjusted in real time to obtain a load balancing adjustment instruction; A performance indicator evaluation calculation is performed based on the load balancing adjustment instruction to obtain a performance optimization parameter, and the first adaptive prediction model is updated according to the performance optimization parameter to obtain a second adaptive prediction model.

2. The load matching optimization method for a multi-channel gate driver according to claim 1, characterized in that: The gate current, gate voltage and switching frequency of the multi-channel parallel gate driver are collected to obtain driver load operation data, wherein the driver load operation data includes impedance parameters, coupling parameters and temperature parameters, including: Sampling the gate current of the multi-channel parallel gate driver to obtain a gate current sampling sequence, and sampling the gate voltage of the multi-channel parallel gate driver to obtain a gate voltage sampling sequence; Perform impedance calculation based on the gate current sampling sequence and the gate voltage sampling sequence to obtain impedance parameters, wherein the impedance parameters include a gate resistance value, a Miller capacitance value, and a parasitic inductance value; Performing coupling calculation on the impedance parameters of adjacent driving channels to obtain coupling parameters, wherein the coupling parameters include an electromagnetic coupling coefficient and a thermal coupling coefficient; Performing temperature monitoring on each driving channel of the multi-channel parallel gate driver to obtain a temperature monitoring data sequence, and performing temperature distribution calculation based on the temperature monitoring data sequence to obtain temperature parameters, wherein the temperature parameters include a temperature value and a temperature gradient value of each driving channel; The impedance parameter, the coupling parameter and the temperature parameter are subjected to data standardization processing to obtain driver load operation data.

3. The load matching optimization method of a multi-channel gate driver according to claim 2, characterized in that: The process of dividing the operating space into regions based on the driver load operating data to obtain a safe operating region and a transition region includes: Analyzing the impedance parameters in the driver load operation data to obtain the maximum allowable power value of each drive channel, and calculating the load deviation threshold matrix based on the maximum allowable power value; Performing threshold optimization based on the load deviation threshold matrix to obtain an operating area boundary threshold, wherein the operating area boundary threshold is used to divide a safe operating area and a transition area; Quantitatively calculate the interference degree between each driving channel based on the coupling parameters in the driver load operation data to obtain a channel coupling strength matrix, and the channel coupling strength matrix is ​​used to correct the operating area boundary threshold; Performing thermal stress analysis on the temperature distribution of each drive channel according to the temperature parameters in the driver load operation data to obtain a thermal stress distribution matrix, and performing temperature compensation on the operating area boundary threshold based on the thermal stress distribution matrix; Performing region division based on the operating region boundary threshold and the load deviation threshold matrix to obtain an initial operating space boundary, and performing multi-dimensional feature extraction on the initial operating space boundary to obtain an operating space feature vector, wherein the operating space feature vector includes load balance, temperature uniformity, and coupling strength; The region is optimized according to the operating space feature vector to obtain an optimized operating space division result, and the operating state is determined in real time based on the optimized operating space division result to obtain a safe operating region and a transition region.

4. The load matching optimization method for a multi-channel gate driver according to claim 3, characterized in that: The driver load operation data is input into the first adaptive prediction model for feature extraction to obtain a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value, including: Inputting the driver load operation data into the load decoupling layer of the first adaptive prediction model, wherein the load decoupling layer adopts an autoencoder structure to perform nonlinear separation on impedance parameters, coupling parameters and temperature parameters to obtain load characteristic components; Inputting the load feature component into a multi-channel complementary feature extraction layer of the first adaptive prediction model, wherein the multi-channel complementary feature extraction layer includes parallel channel feature sub-networks, each channel feature sub-network adopts a residual connection structure, and establishes a lateral information transmission path between channels to obtain channel correlation features; Inputting the channel association features into the dynamic coupling modeling layer of the first adaptive prediction model, wherein the dynamic coupling modeling layer adopts a graph convolutional network structure, captures the load coupling effect by dynamically constructing an association map between channels, and obtains coupling dynamic features; Inputting the coupled dynamic features into a multi-scale temporal feature fusion layer of the first adaptive prediction model, wherein the multi-scale temporal feature fusion layer comprises a parallel group of long short-term memory units, each group of units processes features of different time scales, and performs feature fusion through a cross-scale attention mechanism to obtain a multi-scale temporal feature; Inputting the multi-scale time series features into the thermoelectric coupling analysis layer of the first adaptive prediction model, wherein the thermoelectric coupling analysis layer adopts a dual-stream network structure to process thermal features and electrical features respectively, and merges the two feature streams through an adaptive weight fuser to obtain thermoelectric coupling features; Inputting the thermoelectric coupling feature into the distribution prediction layer of the first adaptive prediction model, wherein the distribution prediction layer adopts a conditional variational autoencoder structure, generates a probability model of the load distribution between channels based on the current load state, and obtains a predicted value of the load distribution between channels; Inputting the thermoelectric coupling characteristics into the fluctuation analysis layer of the first adaptive prediction model, wherein the fluctuation analysis layer adopts a wavelet neural network structure and captures load fluctuation characteristics of different frequencies through multi-resolution analysis to obtain a load fluctuation prediction value; A load balancing prediction result is output according to the inter-channel load distribution prediction value and the load fluctuation prediction value.

5. The load matching optimization method for a multi-channel gate driver according to claim 4, characterized in that: The control strategy is optimized and calculated in parallel according to the load balancing prediction result to obtain control parameters of each drive channel, wherein each drive channel control parameter includes a gate drive voltage adjustment amount and a drive timing adjustment amount, including: Performing data analysis on the predicted value of the load distribution between channels in the load balancing prediction result to obtain the load imbalance of each driving channel, wherein the load imbalance is used to calculate the driving voltage compensation value; A dynamic compensation model is established based on the load fluctuation prediction value to obtain a driving timing compensation strategy, wherein the driving timing compensation strategy includes a dead time adjustment amount and a switch delay compensation amount; Inputting the load imbalance into a weight distributor to obtain a voltage adjustment coefficient matrix of each driving channel, wherein the voltage adjustment coefficient matrix is ​​used to calculate the adjustment range of the gate driving voltage; Parallel decomposition of the driving timing compensation strategy is performed to obtain a channel-level timing control sequence, wherein the channel-level timing control sequence cooperates with the voltage adjustment coefficient matrix; Performing orthogonal optimization based on the voltage regulation coefficient matrix and the channel-level timing control sequence to obtain a multi-channel collaborative control matrix, wherein the multi-channel collaborative control matrix includes a voltage control amount and a timing control amount; Inputting the multi-channel cooperative control matrix into a constraint processor, performing boundary constraints on voltage amplitude and timing interval, and obtaining initial values ​​of control parameters; Performing multi-objective iterative optimization on the initial values ​​of the control parameters to obtain an optimal control parameter combination, wherein the optimal control parameter combination satisfies the balance requirements of load balancing and switching loss; Based on the optimal control parameter combination, the gate drive voltage and drive timing of each drive channel are allocated to obtain the control parameters of each drive channel.

6. The load matching optimization method of a multi-channel gate driver according to claim 5, characterized in that: The real-time adjustment of the control parameters of each drive channel to obtain a load balancing adjustment instruction includes: Performing status monitoring on the control parameters of each drive channel to obtain a real-time operation status matrix, wherein the real-time operation status matrix includes current load balance and temperature distribution data; Performing a working area determination based on the real-time operating state matrix to obtain an area identification signal, wherein the area identification signal is used to determine whether the system is in the safe operating area; Inputting the area identification signal into a control strategy selector to obtain a control mode signal, wherein the control mode signal includes a steady-state control mode and a fast response mode; Classify and process the control parameters of each driving channel based on the control mode signal to obtain a hierarchical control sequence, wherein the hierarchical control sequence includes a voltage adjustment sequence and a timing adjustment sequence; Prioritizing the hierarchical control sequences to obtain a control execution sequence table, wherein the control execution sequence table determines the execution sequence of each adjustment instruction according to the urgency of load balancing; Input the control execution sequence table into the actuator scheduling module to obtain a timing execution plan, wherein the timing execution plan includes a specific execution time of each drive channel control instruction; The timing execution scheme is buffered and managed to obtain an instruction cache queue, the instruction cache queue ensures smooth switching of control instructions, and instruction synthesis is performed based on the instruction cache queue to obtain a load balancing adjustment instruction.

7. The load matching optimization method for a multi-channel gate driver according to claim 6, characterized in that: The performing of performance indicator evaluation calculation based on the load balancing adjustment instruction to obtain performance optimization parameters, and updating the first adaptive prediction model according to the performance optimization parameters to obtain a second adaptive prediction model, includes: Performing a performance evaluation on the execution effect of the load balancing adjustment instruction to obtain a performance indicator matrix, wherein the performance indicator matrix includes a load balancing degree indicator, a temperature consistency indicator, and a response time indicator; Performing model error analysis based on the performance indicator matrix to obtain a prediction error distribution, wherein the prediction error distribution includes a load distribution prediction error and a fluctuation prediction error; Inputting the prediction error distribution into an adaptive corrector to obtain a model parameter correction amount, wherein the model parameter correction amount is used to update the network parameters of each layer of the first adaptive prediction model; Based on the model parameter correction amount, the load decoupling layer of the first adaptive prediction model is updated to obtain an updated load feature extraction structure; Dynamically adjusting the weights of the multi-channel complementary feature extraction layer of the first adaptive prediction model to obtain an optimized channel feature network; Optimizing the graph structure of the dynamic coupling modeling layer of the first adaptive prediction model based on the performance indicator matrix to obtain an updated coupling modeling network; The updated load feature extraction structure, the optimized channel feature network and the updated coupling modeling network are integrated to obtain a second adaptive prediction model.

8. A load matching optimization system for a multi-channel gate driver, characterized in that: For implementing the steps of the method according to any one of claims 1 to 7, the system comprises: An acquisition module is used to acquire gate current, gate voltage and switching frequency of a multi-channel parallel gate driver to obtain driver load operation data, wherein the driver load operation data includes impedance parameters, coupling parameters and temperature parameters; A region division module, used for performing region division processing on the operation space based on the load operation data of the driver to obtain a safe operation region and a transition region; A feature extraction module, used for inputting the driver load operation data into a first adaptive prediction model for feature extraction to obtain a load balancing prediction result, wherein the load balancing prediction result includes a load distribution prediction value between channels and a load fluctuation prediction value; A calculation module, used for performing parallel optimization calculation on the control strategy according to the load balancing prediction result to obtain control parameters of each drive channel, wherein the control parameters of each drive channel include a gate drive voltage adjustment amount and a drive timing adjustment amount; An adjustment module, used for adjusting the control parameters of each drive channel in real time to obtain a load balancing adjustment instruction; An updating module is used to perform performance indicator evaluation calculation based on the load balancing adjustment instruction to obtain performance optimization parameters, and to update the first adaptive prediction model according to the performance optimization parameters to obtain a second adaptive prediction model.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images