A method and system for preventing thermal runaway in lithium batteries based on water-cooled plate temperature control

Abstract

This invention relates to the field of temperature control system technology, specifically to a method and system for preventing thermal runaway of lithium batteries based on water-cooled plate temperature regulation. In this invention, a genetic algorithm is introduced during the multi-condition trajectory generation stage. Candidate screening, mutation perturbation, and optimal solution retention are performed in multiple combinations of cell temperature, water-cooled plate temperature, pump speed, and valve trajectory. This allows the multi-dimensional trajectory to form a solution set with a larger adjustable space while satisfying boundary conditions. Through item-by-item elimination and fitness evaluation processes within the trajectory combinations, the system's control flexibility and coverage stability under complex operating conditions are enhanced. Furthermore, a step-by-step path selection mechanism is constructed based on the tightened trajectory data using a greedy algorithm. Feasibility value screening and local optimal splicing are performed at each time point to ensure that the pump speed and valve control paths are always within the boundaries and continuously feasible, avoiding global computational overhead and ensuring trajectory smoothness and timely feedback.

Description

Technical Field

[0001] This invention relates to the field of temperature control system technology, and in particular to a method and system for preventing thermal runaway of lithium batteries based on water-cooled plate temperature regulation. Background Technology

[0002] The field of temperature control system technology aims to ensure that equipment or systems maintain a stable temperature range in a specific working environment through effective temperature management technology, optimize performance, extend service life and ensure safety, prevent overheating or overcooling, avoid equipment failure, efficiency reduction or safety risks caused by abnormal temperature, and ensure stable and efficient operation of equipment.

[0003] The purpose of a lithium battery thermal runaway prevention method based on water-cooled plate temperature control is to prevent thermal runaway of lithium batteries due to excessive temperature during operation through precise temperature control, thereby improving the safety and reliability of the battery. The method monitors and adjusts the temperature of the lithium battery in real time to ensure that the battery operates within a safe temperature range and avoids thermal runaway caused by excessive temperature.

[0004] In existing technologies, temperature control systems mostly rely on unidirectional control logic triggered by fixed thresholds. During execution, they often adjust a single parameter based on the current temperature state, lacking multi-state interactive modeling capabilities and historical trajectory prediction capabilities. It is difficult to identify the impending over-temperature trend in advance. In variable environments or non-constant load conditions, the control strategy response lag can easily lead the system into a critical state. When the ambient temperature rises sharply or the load current changes abruptly in a short period of time, traditional temperature control strategies cannot establish multi-dimensional linkage paths in real time. Pump and valve regulation often exhibits delay or overcompensation, which can lead to heat accumulation or system oscillation. The temperature control feedback path does not embed dynamic trajectory evaluation and path feasibility verification mechanisms, which can easily cause the control path to become disconnected from the thermal state, affecting the temperature control accuracy and execution reliability. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and propose a method and system for preventing thermal runaway of lithium batteries based on water-cooled plate temperature control.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for preventing thermal runaway of lithium batteries based on water-cooled plate temperature control, comprising the following steps:

[0007] S1: Based on the cell temperature status, water cooling plate temperature status, coolant inlet temperature, pump speed and valve opening, calculate the temperature rate, module temperature difference and inlet / outlet temperature difference, compare the threshold and limit, and obtain thermal boundary constraint data.

[0008] S2: Based on the thermal boundary constraint data, generate the cell temperature trajectory under constant discharge conditions, generate the water-cooled plate temperature trajectory under pulse load conditions, and generate the pump speed trajectory and valve opening trajectory under ambient temperature change conditions. Compare each type of trajectory with the corresponding threshold boundary item by item and retain and filter the suitable combination through a genetic algorithm to obtain the global parameter adaptation solution set.

[0009] S3: Based on the global parameter adaptation solution set, perform constant discharge to generate cell trajectory, perform pulse load to generate water cooling plate trajectory, perform environmental change to generate pump speed and valve trajectory, compare the trajectory with threshold boundary item by item, and obtain working condition trajectory verification record;

[0010] S4: Based on the working condition trajectory verification record, the cell and water-cooled plate sequence is periodically called to calculate and generate the predicted trajectory. The predicted trajectory is compared with the power environment boundary. A greedy algorithm is used to perform a tight solution on the pump speed and valve trajectory. A feasible solution is selected to obtain the predicted execution trajectory set.

[0011] S5: Based on the predicted execution trajectory set, when the cell temperature approaches the upper limit or the prediction is not feasible, the pump speed is switched to a high value and the valve value is limited to a low value, and maintained until the temperature drops, thus establishing an emergency protection operation mode.

[0012] As a further embodiment of the present invention, the thermal boundary constraint data includes cell temperature threshold constraints, pump flow constraints, and valve regulation rate constraints; the global parameter adaptation solution set includes coolant inlet temperature setting, pump speed, valve opening, proportional-integral-derivative parameters, and valve opening feedforward coefficient; the operating condition trajectory verification record includes cell temperature trajectory record, water-cooled plate temperature trajectory record, pump speed trajectory record, and valve opening trajectory record; the predicted execution trajectory set includes pump speed execution trajectory, valve opening execution trajectory, cell temperature prediction trajectory, and water-cooled plate temperature prediction trajectory; and the emergency protection operation mode includes high pump speed operation, low valve speed operation, and cell temperature recovery monitoring.

[0013] As a further aspect of the present invention, the specific steps for generating the thermal boundary constraint data are as follows:

[0014] Based on the cell temperature status, water-cooled plate temperature status, coolant inlet temperature, pump speed, and bypass valve opening, cell temperature points are collected and the temperature difference between adjacent cells in the same module is calculated using the difference calculation rate between adjacent points to generate the module temperature difference, thus obtaining temperature difference data.

[0015] Based on the temperature difference data, the temperature difference between the coolant inlet and outlet is calculated to generate the inlet and outlet temperature difference. The inlet and outlet temperature difference is compared with the cell temperature threshold, pump flow limit, and valve speed limit to obtain thermal boundary constraint data.

[0016] As a further aspect of the present invention, the specific steps for generating the global parameter adaptation solution set are as follows:

[0017] Based on the thermal boundary constraint data, the cell temperature points under constant discharge conditions are collected and connected by continuous points to generate the cell temperature trajectory. The water-cooled plate temperature points under pulse load conditions are collected and spliced ​​by sequence to generate the water-cooled plate temperature trajectory, resulting in a dual-condition temperature trajectory set.

[0018] Based on the dual-condition temperature trajectory set, the pump speed sequence under the ambient temperature change condition is recorded and the pump speed trajectory is generated by sequential arrangement. The valve opening sequence is recorded and the valve trajectory is generated by time-series accumulation, thus obtaining the full-condition operation trajectory set.

[0019] Based on the full-condition operating trajectory set, the cell temperature trajectory is compared point by point with the temperature threshold, the water-cooled plate temperature trajectory is compared point by point with the allowable boundary, the pump speed trajectory is compared point by point with the flow limit, and the valve trajectory is compared point by point with the rate limit. On the candidate solutions, the genetic algorithm is called to perform polynomial mutation and combined with elite retention to screen the suitable combination to obtain the global parameter suitable solution set.

[0020] As a further aspect of the present invention, the genetic algorithm, based on the full-condition operating trajectory set, compares the cell temperature trajectory with the temperature threshold, the water-cooled plate temperature trajectory with the allowable boundary, the pump speed trajectory with the flow limit, and the valve trajectory with the rate limit point by point, filters out trajectory combinations that do not meet the boundary, calls the genetic algorithm in the remaining candidate solutions, performs selection, crossover, and polynomial mutation operations to generate new solutions and combines the solution set with the previous generation, sets the fitness function according to the cell temperature deviation, water-cooled plate response stability, and pump valve control matching degree, sorts them from high to low fitness values, retains the top-ranked solutions through elite retention operation, removes low-fitness solutions, iterates and updates until the convergence condition is met, and obtains the global parameter-fitted solution set.

[0021] As a further aspect of the present invention, the specific steps for generating the working condition trajectory verification record are as follows:

[0022] Based on the global parameter adaptation solution set, the cell temperature is collected and a temperature trajectory is generated under constant discharge conditions, and the water-cooled plate temperature is collected and a temperature trajectory is generated under pulse load conditions, thus obtaining a multi-condition temperature trajectory set.

[0023] Based on the multi-condition temperature trajectory set, a pump speed trajectory and a valve trajectory are established under the condition of ambient temperature change. Multiple trajectories are compared with the set threshold boundaries one by one to obtain the condition trajectory verification record.

[0024] As a further aspect of the present invention, the specific steps for generating the predicted execution trajectory set are as follows:

[0025] Based on the working condition trajectory verification record, the cell temperature sequence is periodically called and the cell prediction trajectory is generated by point-by-point calculation. The water-cooled plate temperature sequence is periodically called and the water-cooled plate prediction trajectory is generated by segment accumulation, thus obtaining the prediction temperature trajectory set.

[0026] Based on the predicted temperature trajectory set, the predicted trajectory of the battery cell is compared point by point with the load power boundary, the predicted trajectory of the water-cooled plate is compared point by point with the ambient temperature boundary, and boundary tightening processing is performed on the pump speed trajectory and valve trajectory to obtain tightened trajectory data.

[0027] Based on the tightening trajectory data, a greedy algorithm is used to retain the pump speed trajectory and valve trajectory that meet the boundary conditions in the comparison results, discard the trajectory that exceeds the boundary, and merge them to form the pump speed and valve execution sequence to obtain the predicted execution trajectory set.

[0028] As a further aspect of the present invention, the greedy algorithm, based on the tightened trajectory data, compares the predicted trajectory of the battery cell with the load power boundary and the predicted trajectory of the water-cooled plate with the ambient temperature boundary point by point, identifies the time segments that meet the boundaries, searches for control adjustment values ​​point by point on the pump speed and valve trajectories and determines whether they are within the feasible range, selects the local optimal combination in chronological order, discards any control value that exceeds the boundary, and continuously splices the pump speed and valve trajectory segments that meet the conditions. By repeatedly executing the selection, judgment and splicing operations, a control path without illegal segments is constructed, and the predicted execution trajectory set is obtained by merging.

[0029] As a further aspect of the present invention, the specific steps for generating the emergency protection operation mode are as follows:

[0030] Based on the predicted execution trajectory set, when the cell temperature approaches the upper limit threshold, the trigger condition is activated, the pump speed command is switched and increased to a high value, and the valve opening command is adjusted and limited to a low value range to obtain the operation switching data.

[0031] Based on the aforementioned operation switching data, the pump is kept running at high speed and the valve is in a low-value state. The battery cell temperature is detected and compared with the threshold. When the detected value falls back to the threshold range, the recovery conditions are confirmed, and an emergency protection operation mode is established.

[0032] A lithium battery thermal runaway prevention system based on water-cooled plate temperature control, wherein the system is used to execute the aforementioned lithium battery thermal runaway prevention method based on water-cooled plate temperature control, and the system includes:

[0033] Boundary acquisition module: Based on the cell temperature status, water-cooled plate temperature status, coolant inlet temperature, pump speed and valve opening, calculate the temperature change rate, inter-module temperature difference and inlet / outlet temperature difference, and acquire thermal boundary constraint data;

[0034] Parameter solution set generation module: Based on the thermal boundary constraint data, it generates temperature and control trajectories under three types of working conditions and compares the boundary conditions. It retains the trajectory combinations that meet the conditions, applies a genetic algorithm to filter and arrange them, and obtains the parameter fitting combination set.

[0035] Trajectory verification module: Reconstructs the trajectory based on the parameter adaptation combination set and compares the boundary records item by item, extracts the out-of-bounds combination information, and obtains the working condition trajectory verification record;

[0036] Predictive tightening module: Based on the working condition trajectory verification record, a temperature trend trajectory is constructed. After comparing it with the power boundary, the control value range between the pump speed and valve trajectory is narrowed. A greedy algorithm is applied to select a feasible path and obtain the predicted execution trajectory set.

[0037] Emergency control module: Based on the predicted execution trajectory set, monitor the cell temperature status. When the high temperature threshold is reached or the trajectory is not feasible, switch the pump speed to a high value and lock the valve at a low value to establish an emergency protection status record.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] In this invention, by introducing a genetic algorithm in the multi-condition trajectory generation stage, candidate screening, mutation perturbation and optimal solution retention are performed in multiple combinations of cell temperature, water-cooled plate temperature, pump speed and valve trajectory, so that the multi-dimensional trajectory can form a solution set with a larger adjustable space on the basis of satisfying the boundary conditions, thus expanding the range of subsequent path options.

[0040] In this invention, the system's control flexibility and coverage stability under complex operating conditions are enhanced through a process of item-by-item elimination and fitness evaluation within trajectory combinations. Furthermore, a step-by-step path selection mechanism is constructed based on the tightened trajectory data using a greedy algorithm. Feasibility value screening and local optimal splicing are performed at each time point to ensure that the pump speed and valve control paths are always within the boundaries and continuously feasible. This avoids global computational overhead and ensures trajectory smoothness and timely feedback, thereby improving the system's ability to generate feasible paths and its response decision efficiency under nonlinear operating conditions. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the workflow of the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0043] Example 1

[0044] Please see Figure 1 This invention provides a technical solution: a method for preventing thermal runaway of lithium batteries based on water-cooled plate temperature control, comprising the following steps:

[0045] S1: Based on the cell temperature status, water cooling plate temperature status, coolant inlet temperature, pump speed and valve opening, calculate the temperature rate, module temperature difference and inlet / outlet temperature difference, compare the threshold and limit, and obtain thermal boundary constraint data.

[0046] S2: Based on thermal boundary constraint data, the cell temperature trajectory is generated under constant discharge conditions, the water-cooled plate temperature trajectory is generated under pulse load conditions, and the pump speed trajectory and valve opening trajectory are generated under ambient temperature change conditions. Each type of trajectory is compared with the corresponding threshold boundary and the suitable combination is retained and screened through a genetic algorithm to obtain the global parameter adaptation solution set.

[0047] S3: Based on the global parameter adaptation solution set, constant discharge is used to generate the cell trajectory, pulse load is used to generate the water cooling plate trajectory, environmental changes are used to generate the pump speed and valve trajectory, and the trajectory is compared with the threshold boundary item by item to obtain the working condition trajectory verification record.

[0048] S4: Based on the working condition trajectory verification record, the cell and water-cooled plate sequence is periodically called to calculate and generate the predicted trajectory. The predicted trajectory is compared with the power environment boundary. A greedy algorithm is used to perform a tight solution on the pump speed and valve trajectory. A feasible solution is selected to obtain the predicted execution trajectory set.

[0049] S5: Based on the predicted execution trajectory set, when the cell temperature approaches the upper limit or the prediction is not feasible, it triggers the switching of the pump speed to a high value and limits the valve to a low value, maintaining this until the temperature drops, thus establishing an emergency protection operation mode.

[0050] The thermal boundary constraint data includes cell temperature threshold constraints, pump flow constraints, and valve regulation rate constraints. The global parameter adaptation solution set includes coolant inlet temperature setting, pump speed, valve opening, proportional-integral-derivative parameters, and valve opening feedforward coefficients. The operating condition trajectory verification records include cell temperature trajectory records, water-cooled plate temperature trajectory records, pump speed trajectory records, and valve opening trajectory records. The predicted execution trajectory set includes pump speed execution trajectory, valve opening execution trajectory, cell temperature prediction trajectory, and water-cooled plate temperature prediction trajectory. The emergency protection operation modes include high pump speed operation, low valve speed operation, and temperature recovery monitoring.

[0051] The specific steps for generating thermal boundary constraint data are as follows:

[0052] Based on the cell temperature status, water-cooled plate temperature status, coolant inlet temperature, pump speed, and bypass valve opening, cell temperature points are collected and the temperature difference between adjacent cells in the same module is calculated using the difference calculation rate between adjacent points to generate the module temperature difference, thus obtaining temperature difference data.

[0053] Based on temperature difference data, the temperature difference between the coolant inlet and outlet is calculated to generate the inlet and outlet temperature difference. The inlet and outlet temperature difference is compared with the cell temperature threshold, pump flow limit, and valve speed limit to obtain thermal boundary constraint data.

[0054] Based on the cell temperature status, water-cooled plate temperature status, coolant inlet temperature, pump speed, and bypass valve opening, a first-order forward difference method is used to extract the difference between adjacent points in each group of cell temperature sampling sequences. The numpy.diff function is called to set the difference order to 1 and fix the sampling interval to 1 second. Each cell temperature data stream is processed sequentially to generate a temperature change sequence. The difference operation is performed on adjacent numbered cell temperature sampling points within the same module, and the np.abs function is called to obtain the temperature difference between each cell. The traversal index range is set to the number of cells in the module minus one. Multi-module cyclic difference processing is completed by sliding the array index window to generate temperature difference data.

[0055] Based on temperature difference data, a vector difference operation method is used to process coolant temperature information. The coolant inlet temperature array and outlet temperature array are read separately. After performing the same-dimensional difference operation using the numpy.subtract function, a temperature difference sequence is formed. The pandas.DataFrame.rolling function is called to set the window width to 10 and perform moving average processing on the temperature difference sequence. The coolant temperature difference sequence is compared item by item with the cell temperature threshold. The numpy.where function is called to set the judgment conditions and extract the index of out-of-bounds segments. At the same time, the pump flow rate data is compared with the maximum flow limit, and the valve opening change rate is compared with the maximum change rate. Boolean expressions and logical AND operators are used to complete the merging and filtering of the three types of judgment conditions to generate thermal boundary constraint data.

[0056] The specific steps for generating a global parameter adaptation solution set are as follows:

[0057] Based on thermal boundary constraint data, the cell temperature points under constant discharge conditions are collected and the cell temperature trajectory is generated by connecting continuous points. The water-cooled plate temperature points under pulse load conditions are collected and the water-cooled plate temperature trajectory is generated by splicing sequences, resulting in a dual-condition temperature trajectory set.

[0058] Based on the dual-condition temperature trajectory set, the pump speed sequence under the ambient temperature change condition is recorded and the pump speed trajectory is generated by sequential arrangement. The valve opening sequence is recorded and the valve trajectory is generated by time-series accumulation, thus obtaining the full-condition operation trajectory set.

[0059] Based on the full-condition operation trajectory set, the cell temperature trajectory is compared with the temperature threshold point by point, the water-cooled plate temperature trajectory is compared with the allowable boundary point by point, the pump speed trajectory is compared with the flow limit point by point, and the valve trajectory is compared with the rate limit point by point. On the candidate solutions, the genetic algorithm is called to perform polynomial mutation and combined with elite retention to screen the suitable combination to obtain the global parameter adaptation solution set.

[0060] Based on thermal boundary constraint data, cell temperature points under constant discharge conditions are collected, and the scipy.interpolate.interp1d method is called to set the kind parameter to linear to perform adjacent temperature point connection processing. The boundary fill value fill_value is set to extrapolate and the sampling interval is set to 1 second to generate a continuous temperature sequence. Water-cooled plate temperature points under pulsed load conditions are collected, and the numpy.concatenate function is called to set the splicing axis to the time axis direction to perform multi-segment temperature sequence sequential splicing operation. The result array is numbered and marked with a unified time index to generate a dual-condition temperature trajectory set.

[0061] Based on the dual-condition temperature trajectory set, the pump speed sequence under the ambient temperature change condition is recorded and the numpy.argsort function is called to set the ascending sort method to reorganize the sampling points according to the time index order. The return index is set to True and the sorting application is completed in the original array to generate the pump speed trajectory. The valve opening sequence is recorded and the numpy.cumsum function is called to set the initial value to 0 and perform cumulative processing on the time series item by item. The cumulative result array is standardized to generate the full-condition operation trajectory set.

[0062] Based on the full-condition operation trajectory set, the cell temperature trajectory is compared point by point with the temperature threshold, the water-cooled plate temperature trajectory is compared point by point with the allowable boundary, the pump speed trajectory is compared point by point with the flow limit, and the valve trajectory is compared point by point with the rate limit. A genetic algorithm is used to adapt and filter the trajectory combinations that meet all boundary conditions. The creator.create method in the deap toolkit is called to create an individual class and set the fitness weight to (1.0,). The toolbox.initRepeat function is used to initialize the individual population and set the individual length to be consistent with the dimension of the trajectory combination. The toolbox.mutate method is executed to set the polynomial mutation probability indpb to 0.2 and the exponent eta to 20 for perturbation operation. After the offspring and parents are combined, the tools.selBest function is called to set the number of individuals to be retained k to the original population size and take the individual with the highest fitness to generate a global parameter adaptation solution set.

[0063] The genetic algorithm, based on the full-condition operating trajectory set, compares the cell temperature trajectory with the temperature threshold, the water-cooled plate temperature trajectory with the allowable boundary, the pump speed trajectory with the flow limit, and the valve trajectory with the rate limit point by point. It eliminates trajectory combinations that do not meet the boundary. In the remaining candidate solutions, the genetic algorithm is called to perform selection, crossover and polynomial mutation operations to generate new solutions and combine the solution set with the previous generation. The fitness function is set according to the cell temperature deviation, water-cooled plate response stability and pump valve control matching degree. The solutions are sorted from high to low fitness value. The top-ranked solutions are retained by elite retention operation and low-fitness solutions are eliminated. The algorithm is iteratively updated until the convergence condition is met to obtain the global parameter-fit solution set.

[0064] Genetic algorithm, according to the formula:

[0065]

[0066] in: Indicates candidate solutions The fitness function value, This represents the candidate parameter solution generated by the genetic algorithm. This represents the discrete time point number in the trajectory. This represents the total number of time points included in the full-condition operation trajectory set. Indicates time Cell temperature trajectory value, Indicates the cell temperature threshold. Indicates time Water-cooled plate temperature trajectory value, This indicates the allowable temperature boundary value for the water-cooled plate. Indicates time Pump speed The corresponding coolant flow rate, Indicates the limit value of coolant flow rate. Indicates time Valve opening rate value, Indicates the valve speed limit value;

[0067] Execution process: First, discrete time points are extracted from the full-condition operation trajectory set. Calculate the cell temperature at each time point. With cell temperature threshold Calculate the temperature of the water-cooled plate by taking the square of the deviation. With Permissible Boundaries The square of the deviation is calculated from the pump speed. The determined flow With flow limit The square of the deviation is calculated, and the valve rate is determined. Valve speed limit The squared deviations are then calculated, and the four deviations are calculated at all time points. to By accumulating the results, candidate solutions are obtained. fitness function value New candidate solutions are generated by polynomial mutation operation, and the elite retention strategy is used to screen solutions with high fitness. After multiple rounds of iteration, a global parameter-fit solution set is obtained, which is used to realize temperature regulation and thermal runaway risk prevention of lithium batteries during operation.

[0068] The specific steps for generating the working condition trajectory verification record are as follows:

[0069] Based on the global parameter adaptation solution set, the cell temperature is collected and a temperature trajectory is generated under constant discharge conditions, and the water-cooled plate temperature is collected and a temperature trajectory is generated under pulse load conditions, thus obtaining a multi-condition temperature trajectory set.

[0070] Based on a multi-condition temperature trajectory set, a pump speed trajectory and a valve trajectory are established under ambient temperature change conditions. Multiple trajectories are compared with set threshold boundaries one by one to obtain the working condition trajectory verification record.

[0071] Based on the global parameter adaptation solution set, the cell temperature is collected under constant discharge conditions, and the `scipy.interpolate.interp1d` method is called with the `kind` parameter set to `linear` to perform linear interpolation between sampling points. The `fill_value` parameter is set to `extrapolate` to fill in boundary point values, and the results are generated into a structured array in time index order. The `numpy.linspace` function is used on the structured array to set the number of interpolations to be consistent with the sampling frequency for resampling, thus completing the construction of the cell temperature trajectory. Under pulse load conditions, the water-cooled plate temperature is collected, and the `pandas.Series.reindex` method is used to set new index values ​​and the `method` parameter is set to `nearest` for time axis alignment. After completing the breakpoint filling of the temperature sequence, the `numpy.polyfit` function is called with the polynomial degree set to 3 for smooth fitting, generating a multi-condition temperature trajectory set.

[0072] Based on a multi-condition temperature trajectory set, under varying ambient temperature conditions, the `numpy.gradient` function is called to set the time step to 1 and perform a first-order derivative operation on the ambient temperature sequence. Pump speed control values ​​are set segment by segment based on the derivative values, and the `numpy.piecewise` function is called to set the conditional expression and pump speed values ​​for each segment to complete the pump speed trajectory construction. Valve opening data under the corresponding time axis is recorded, and the `numpy.convolve` function is used to set the convolution kernel to [0.25, 0.5, 0.25] to perform a three-point moving average. The smoothed result is then processed using the `pandas.Series.cumsum` method with an initial value of 0 to generate the valve trajectory. The constructed pump speed trajectory and valve trajectory are compared point by point with the pump flow limit and valve rate limit set in the temperature control boundary. The cell temperature trajectory and the upper temperature limit, and the water-cooled plate temperature trajectory and the allowable temperature difference boundary are respectively constructed using the `numpy.where` function to construct Boolean index matrices. A logical AND operation is performed on the four comparison results to extract the out-of-bounds point index set, generating a condition trajectory verification record.

[0073] The specific steps for generating the predicted execution trajectory set are as follows:

[0074] Based on the working condition trajectory verification record, the cell temperature sequence is periodically called and the cell prediction trajectory is generated by point-by-point calculation. The water-cooled plate temperature sequence is periodically called and the water-cooled plate prediction trajectory is generated by segment accumulation, thus obtaining the prediction temperature trajectory set.

[0075] Based on the predicted temperature trajectory set, the predicted trajectory of the battery cell is compared point by point with the load power boundary, the predicted trajectory of the water-cooled plate is compared point by point with the ambient temperature boundary, and boundary tightening processing is performed on the pump speed trajectory and valve trajectory to obtain tightened trajectory data.

[0076] Based on the tightened trajectory data, a greedy algorithm is used to retain the pump speed trajectory and valve trajectory that meet the boundary conditions in the comparison results, discard the trajectory that exceeds the boundary, and merge them to form the pump speed and valve execution sequence to obtain the predicted execution trajectory set.

[0077] Based on the working condition trajectory verification records, the cell temperature sequence is periodically called and the numpy.roll function is used to move the temperature sequence forward by one index step along the time axis. The numpy.subtract function is used to perform point-to-point difference operation on the current sequence and the moved sequence respectively, and then added to the current sampling point to synthesize. The future predicted temperature points are generated in a loop according to the time index order. The water-cooled plate temperature sequence is periodically called and the numpy.add.reduceat function is used to set each group of ten consecutive points as an interval. The summation operation is performed in each segment and the numpy.repeat function is used to repeatedly expand the segment result according to the interval length to generate a set of predicted temperature trajectories.

[0078] Based on the predicted temperature trajectory set, the numpy.where function is called to set the judgment condition as the cell temperature exceeds the load power boundary. After judging each point, the index is recorded. The same method is called to perform point-by-point comparison between the water-cooled plate temperature sequence and the ambient temperature boundary and output the boundary outage flag. The numpy.gradient function is called to set the interval step size to 1 for the pump speed trajectory and valve trajectory and to perform the change rate calculation between consecutive points. The numpy.clip function is called to set the maximum change threshold to compress the over-limit control value to an acceptable range and complete the control value reconstruction in time sequence to generate tightened trajectory data.

[0079] Based on the tightening trajectory data, a greedy algorithm is used to extract the feasible pump speed value set and valve opening set at each time index position. The itertools.product function is called to enumerate all combinations of pump speed and valve value. For each combination, a logical judgment function is called to check whether it meets all three conditions of temperature boundary, flow limit and change rate. After retaining the combination that meets all conditions, the sorted function is called to set the sorting key to the sum of the differences between the pump speed and valve value and the value at the previous time. The values ​​are sorted in ascending order and the first combination is extracted as the current step size control output value. The traversal and writing operations are continuously executed until the end of the time index to generate the predicted execution trajectory set.

[0080] The greedy algorithm, based on tightened trajectory data, compares the predicted trajectory of the battery cell with the load power boundary and the predicted trajectory of the water-cooled plate with the ambient temperature boundary point by point, identifies the time segments that meet the boundary, searches for control adjustment values ​​point by point on the pump speed and valve trajectories and determines whether they are within the feasible range, selects the local optimal combination in time sequence, discards any control value that exceeds the boundary, and continuously splices the pump speed and valve trajectory segments that meet the conditions. By repeatedly executing the selection, judgment and splicing operations, a control path without illegal segments is constructed and merged to obtain the predicted execution trajectory set.

[0081] Greedy algorithm, according to the formula:

[0082]

[0083] in: This represents the weighted bias value. Represents a discrete time point index. This represents the total number of discrete points on the trajectory. Indicates time The pump speed trajectory value, Indicates the allowable limit value of pump speed. Indicates time Valve speed trajectory value, Indicates the valve speed limit value. This represents the pump speed deviation weighting coefficient. This represents the valve rate deviation weighting coefficient. This indicates that the trajectory tightening margin coefficient is used for boundary tightening. This indicates that the valve hysteresis compensation parameters are used to correct the actual valve response;

[0084] Execution process: First, establish a discrete time point index in the full-condition operation trajectory set. to Extract the pump speed trajectory value at each moment. And compare it with the tightened pump speed boundary ( 1- Perform the difference calculation, square the difference, and multiply it by the weighting coefficient. The pump speed deviation contribution value is obtained, and the valve speed trajectory value at each moment is extracted. Apply hysteresis compensation bias Obtain the correction value, and then compare the correction value with the tightened valve rate boundary ( 1- Perform the difference calculation, square the difference, and multiply it by the weighting coefficient. Obtain the valve rate deviation contribution value, and then at all time points The weighted deviation contributions of pump speed and valve speed within the range are accumulated point by point to form the weighted deviation cost. To every moment The minimum increment is used as the criterion for trajectory screening. The pump speed trajectory and valve trajectory that meet the conditions are retained point by point and merged in time order to generate a complete set of predicted execution trajectories, which are used to realize thermal management and thermal runaway prevention of lithium batteries during operation.

[0085] The specific steps for generating the emergency protection operation mode are as follows:

[0086] Based on the predicted execution trajectory set, when the cell temperature approaches the upper limit threshold, the trigger condition is activated, the pump speed command is switched and increased to a high value, and the valve opening command is adjusted and limited to a low value range to obtain the operation switching data.

[0087] Based on the operation switching data, the pump is kept running at high speed and the valve is in a low value state. The cell temperature is detected and compared with the threshold. When the detected value falls back to the threshold range, the recovery conditions are confirmed and an emergency protection operation mode is established.

[0088] Based on the predicted execution trajectory set, the numpy.where function is called to set the judgment condition as the cell temperature value being greater than the upper limit threshold minus the set safety margin value, and each time point is judged item by item. After extracting the time index that meets the condition, the pandas.Series.at function is called to locate the corresponding control sequence position and directly assign the pump speed value to the maximum set value, and the valve opening value to the minimum limit value. The data fields corresponding to the two control items are updated synchronously in the original control sequence structure while keeping the time index unchanged. The index segments whose values ​​have been changed are grouped and marked, and the numpy.append function is called to insert the marking information into the original data structure as the running status field, generating running switching data.

[0089] Based on the operation switching data, a state-maintaining judgment method is adopted to call the numpy.logical_and function to perform a Boolean judgment that the pump speed is at its maximum value and the valve is at its minimum value. For the time segment where the current Boolean result is True, a continuous index check operation is performed. For each continuous segment, the cell temperature value is periodically collected and the pandas.Series.rolling function is called to set the window width to five sampling points and calculate the sliding average. In the sliding result, the numpy.where function is called to set the judgment condition that the temperature value is less than or equal to the upper limit threshold and check the time point when the condition is first met in the index order. The status of the five sampling segments after indexing is recorded and written to the confirmation field. After the confirmation field is marked, a fixed status label is inserted into the specified field in the control structure to generate the emergency protection operation mode.

[0090] A lithium battery thermal runaway prevention system based on water-cooled plate temperature control is disclosed. This system is used to execute the aforementioned lithium battery thermal runaway prevention method based on water-cooled plate temperature control. The system includes:

[0091] Boundary acquisition module: Based on the cell temperature status, water-cooled plate temperature status, coolant inlet temperature, pump speed and valve opening, calculate the temperature change rate, inter-module temperature difference and inlet / outlet temperature difference, and acquire thermal boundary constraint data;

[0092] Parameter solution set generation module: Based on thermal boundary constraint data, it generates temperature and control trajectories under three types of working conditions and compares the boundary conditions. It retains the trajectories that meet the conditions, applies a genetic algorithm to filter and arrange them, and obtains the parameter fitting combination set.

[0093] Trajectory verification module: Reconstructs trajectory based on parameter adaptation combination set and compares boundary records item by item, extracts out-of-bounds combination information, and obtains working condition trajectory verification records;

[0094] Predictive tightening module: Based on the working condition trajectory verification record, a temperature trend trajectory is constructed. After comparing it with the power boundary, the control value range is narrowed in the pump speed and valve trajectory. A greedy algorithm is applied to select a feasible path and obtain the predicted execution trajectory set.

[0095] Emergency control module: Based on the predicted execution trajectory set, monitor the cell temperature status. When the high temperature is critical or the trajectory is not feasible, switch the pump speed to a high value and lock the valve at a low value, and establish an emergency protection status record.

[0096] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for preventing thermal runaway of a lithium battery based on temperature regulation of a water-cooled plate, characterized in that, The method comprises the following steps: S1: Based on the temperature state of the battery cell, the temperature state of the water-cooled plate, the inlet temperature of the cooling liquid, the pump speed and the opening degree of the bypass valve, the temperature points of the battery cell are collected and the difference between adjacent points is calculated to obtain the temperature difference of the battery cell in the same module, and the temperature difference data is obtained; Based on the temperature difference data, the temperature difference between the inlet and outlet of the cooling liquid is calculated to obtain the inlet and outlet temperature difference, and the inlet and outlet temperature difference is compared with the battery cell temperature threshold, the pump flow limit and the valve rate limit to obtain the thermal boundary constraint data; S2: Based on the thermal boundary constraint data, the battery cell temperature points under constant discharge conditions are collected and connected to generate a battery cell temperature trajectory, and the water-cooled plate temperature points under pulse load conditions are collected and spliced to generate a water-cooled plate temperature trajectory, and a double-working-condition temperature trajectory set is obtained; Based on the double-working-condition temperature trajectory set, the pump speed sequence under the condition of environmental temperature change is recorded and the pump speed trajectory is generated by sequential arrangement, the valve opening degree sequence is recorded and the valve trajectory is generated by time sequence accumulation, and a full-working-condition running trajectory set is obtained; Based on the full-working-condition running trajectory set, the battery cell temperature trajectory is compared with the temperature threshold point by point, the water-cooled plate temperature trajectory is compared with the allowed boundary point by point, the pump speed trajectory is compared with the flow limit point by point, and the valve trajectory is compared with the rate limit point by point, and the genetic algorithm is called on the candidate solution to perform polynomial mutation and combine with elite reservation to screen the adaptive combination, and a global parameter adaptive solution set is obtained; S3: Based on the global parameter adaptive solution set, the battery cell temperature under constant discharge conditions is collected and a temperature trajectory is generated, and the water-cooled plate temperature under pulse load conditions is collected and a temperature trajectory is generated, and a multi-working-condition temperature trajectory set is obtained; Based on the multi-working-condition temperature trajectory set, the pump speed trajectory and the valve trajectory are established under the condition of environmental temperature change, and the multi-tracks are compared with the set threshold boundary item by item to obtain the working condition trajectory verification record; S4: Based on the working condition trajectory verification record, the battery cell temperature sequence is periodically called and the battery cell prediction trajectory is generated by point-by-point calculation, and the water-cooled plate temperature sequence is periodically called and the water-cooled plate prediction trajectory is generated by section accumulation, and a prediction temperature trajectory set is obtained; Based on the prediction temperature trajectory set, the battery cell prediction trajectory is compared with the load power boundary point by point, the water-cooled plate prediction trajectory is compared with the environmental temperature boundary point by point, and the boundary tightening process is performed on the pump speed trajectory and the valve trajectory, and the tightening trajectory data is obtained; Based on the tightening trajectory data, the greedy algorithm is used to retain the pump speed trajectory and the valve trajectory that meet the boundary conditions and discard the trajectories that exceed the boundary in the comparison results, and the pump speed and valve execution sequence are combined to obtain a prediction execution trajectory set; S5: Based on the prediction execution trajectory set, when the battery cell temperature approaches the upper threshold, the trigger condition is triggered, the pump speed instruction is switched and raised to a high value, and the valve opening degree instruction is adjusted and limited in a low value range, and running switching data is obtained; Based on the running switching data, the pump is kept at high speed and the valve is kept at low value, the battery cell temperature is detected and compared with the threshold, and when the detected value falls within the threshold range, the recovery condition is confirmed, and an emergency protection running mode is established. 2.The lithium battery thermal runaway prevention method based on water cold plate temperature regulation according to claim 1, wherein, The thermal boundary constraint data includes a battery cell temperature threshold constraint, a pump flow constraint, and a valve adjustment rate constraint, the global parameter adaptation solution set includes a cooling liquid inlet temperature setting, a pump rotation speed, a valve opening, a proportional integral derivative parameter, and a valve opening feedforward coefficient, the working condition trajectory verification record includes a battery cell temperature trajectory record, a water-cooled plate temperature trajectory record, a pump rotation speed trajectory record, and a valve opening trajectory record, the predicted execution trajectory set includes a pump speed execution trajectory, a valve opening execution trajectory, a battery cell temperature prediction trajectory, and a water-cooled plate temperature prediction trajectory, and the emergency protection operation mode includes a pump rotation speed high value operation, a valve low value operation, and a battery cell temperature recovery monitoring. 3.The lithium battery thermal runaway prevention method based on water cold plate temperature regulation according to claim 1, wherein, The genetic algorithm, based on the full working condition operation trajectory set, respectively compares the battery cell temperature trajectory with the temperature threshold, the water-cooled plate temperature trajectory with the allowable boundary, the pump speed trajectory with the flow limit, and the valve trajectory with the rate limit point by point, filters out the trajectory combination that does not satisfy the boundary, calls the genetic algorithm in the remaining candidate solutions, performs selection, crossover, and polynomial mutation operations, generates new solutions, and combines the solution set with the previous generation, sets the fitness function according to the battery cell temperature deviation, the water-cooled plate response stability, and the pump valve control matching degree, sorts the fitness values from high to low, retains the top solutions through the elite retention operation, eliminates the low fitness solutions, iteratively updates until the convergence condition is met, and obtains the global parameter adaptation solution set. 4.The lithium battery thermal runaway prevention method based on water cold plate temperature regulation according to claim 1, wherein, The greedy algorithm, based on the tightened trajectory data, respectively compares the battery cell prediction trajectory with the load power boundary and the water-cooled plate prediction trajectory with the ambient temperature boundary point by point, identifies the time sequence fragments that satisfy the boundary, finds the control adjustment amount on the pump speed and valve trajectory point by point and judges whether it is located in the feasible interval, selects the local optimal combination in time sequence in turn, discards any control value that exceeds the boundary, continuously splices the pump speed and valve trajectory paragraphs that satisfy the conditions, constructs a control path that does not contain illegal segments through repeated execution of selection, judgment, and splicing operations, and merges to obtain the predicted execution trajectory set.

5. A lithium battery thermal runaway prevention system based on water-cooled plate temperature regulation, characterized in that, The lithium battery thermal runaway prevention method based on water-cooled plate temperature regulation according to any one of claims 1-4, the system comprises: A boundary acquisition module: based on the battery cell temperature state, the water-cooled plate temperature state, the cooling liquid inlet temperature, the pump rotation speed, and the bypass valve opening, the battery cell temperature points are collected and the adjacent point difference calculation rate is calculated, the temperature difference of adjacent battery cells in the same module is calculated to generate a module temperature difference, and temperature difference data is obtained. Based on the temperature difference data, the temperature difference between the cooling liquid inlet and outlet is calculated to generate an inlet and outlet temperature difference, the inlet and outlet temperature difference is compared with the battery cell temperature threshold, the pump flow limit, and the valve rate limit, and the thermal boundary constraint data is obtained. A parameter solution set generation module: based on the thermal boundary constraint data, the battery cell temperature points under constant discharge conditions are collected and connected by consecutive points to generate a battery cell temperature trajectory, the water-cooled plate temperature points under pulse load conditions are collected and spliced by sequence to generate a water-cooled plate temperature trajectory, and a double working condition temperature trajectory set is obtained. Based on the double-working condition temperature trajectory set, record the pump speed sequence under the condition of environmental temperature change and generate the pump speed trajectory through sequential arrangement, record the valve opening sequence and generate the valve trajectory through time sequence accumulation, and obtain the full-working condition operation trajectory set; Based on the full-working condition operation trajectory set, compare the cell temperature trajectory with the temperature threshold point by point, compare the water-cooled plate temperature trajectory with the allowable boundary point by point, compare the pump speed trajectory with the flow limit point by point, compare the valve trajectory with the rate limit point by point, call the genetic algorithm on the candidate solution to perform polynomial mutation and combine the elite reservation to screen the adaptive combination, and obtain the global parameter adaptive solution set; The trajectory verification module: based on the global parameter adaptive solution set, collect the cell temperature under the condition of constant discharge and generate the temperature trajectory, collect the water-cooled plate temperature under the condition of pulse load and generate the temperature trajectory, and obtain the multi-working condition temperature trajectory set; Based on the multi-working condition temperature trajectory set, establish the pump speed trajectory and establish the valve trajectory under the condition of environmental temperature change, compare the multiple trajectories with the set threshold boundary item by item, and obtain the working condition trajectory verification record; The prediction tightening module: based on the working condition trajectory verification record, periodically call the cell temperature sequence and generate the cell prediction trajectory through point-by-point calculation, periodically call the water-cooled plate temperature sequence and generate the water-cooled plate prediction trajectory through section accumulation, and obtain the prediction temperature trajectory set; Based on the prediction temperature trajectory set, compare the cell prediction trajectory with the load power boundary point by point, compare the water-cooled plate prediction trajectory with the environmental temperature boundary point by point, and perform boundary tightening processing on the pump speed trajectory and the valve trajectory, and obtain the tightening trajectory data; Based on the tightening trajectory data, adopt the greedy algorithm, retain the pump speed trajectory and the valve trajectory that meet the boundary condition in the comparison result, discard the trajectories that exceed the boundary, and combine to form the pump speed and valve execution sequence, and obtain the prediction execution trajectory set; The emergency control module: based on the prediction execution trajectory set, trigger the condition when the cell temperature approaches the upper threshold, switch the pump speed instruction and raise to a high value, adjust the valve opening instruction and limit in a low value interval, and obtain the operation switching data; Based on the operation switching data, keep the pump high-speed operation and the valve low value state, detect the cell temperature and compare the threshold value, confirm the recovery condition when the detected value falls to the threshold range, and establish the emergency protection operation mode.

Images