Alcohol hydrogen fuel cell and super capacitor system of silicon-based electrode and silicon-based chip electroplate

Abstract

The invention belongs to the technical field of self-adaptive control of a super capacitor, and discloses an alcohol hydrogen fuel cell and a super capacitor system of a silicon-based electrode and a silicon-based chip electroplate. Comprising a first processing module for performing thermal state analysis processing based on temperatures of N temperature sampling points of the alcohol hydrogen fuel cell and the super capacitor in unit time, and constructing a four-dimensional thermal sensing model changing along with time; the second processing module is used for performing feature extraction based on the four-dimensional thermal sensing model to obtain four-dimensional thermal sensing feature data; the risk diagnosis module is used for carrying out high-temperature risk area abnormity diagnosis based on the four-dimensional thermal perception feature data and identifying Y high-temperature abnormal areas; the self-adaptive control module is used for carrying out self-adaptive control on Y high-temperature abnormal regions of the alcohol hydrogen fuel cell and the super capacitor; the super capacitor-hydrogen fuel cell coordinated regulation and control structure is constructed, the safety and reliability of the super capacitor and the hydrogen fuel cell at high temperature are improved, and temperature rise control under the fan-free condition is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of supercapacitor adaptive control, and more particularly to an alcohol hydrogen fuel cell and supercapacitor system of a silicon-based electrode and a silicon-based chip electric plate. BACKGROUND

[0002] As a storage element with high power density, long service life and fast charging and discharging characteristics, the stability and reliability of the supercapacitor in a high-temperature and high-impact environment become a key technical bottleneck. Especially when the working environment temperature is in the range of 55℃ to 65℃ for a long time, the device is prone to thermal runaway, capacity decay, and electrode material degradation.

[0003] However, the carbon-based or aluminum electrode supercapacitor structure in the prior art mostly adopts an external air cooling heat dissipation mode, which is difficult to achieve rapid response and precise control of the internal heat source, and is prone to central overheating under thermal accumulation conditions, thereby causing device thermal drift and functional device failure. In addition, the existing technology relies on fixed thresholds or single-point temperature measurement for local high-temperature area identification, which cannot meet the real-time sensing needs of multiple areas and non-uniform temperature fields in dynamic operation, and the existing regulation mechanism is mostly based on fixed logic, lacking closed-loop regulation capability in coordination with temperature evolution trend, and it is difficult to achieve the operation goal of considering both thermal safety and performance under fanless conditions.

[0004] Therefore, it is urgent to propose an adaptive control system for supercapacitors with high integration level, fast thermal response, intelligent sensing and regulation, to solve the problems of thermal safety and reliability in high-temperature operating environment. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purposes, the present application provides the following technical scheme: an alcohol hydrogen fuel cell and supercapacitor system of a silicon-based electrode and a silicon-based chip electric plate, comprising:

[0006] A first processing module performs thermal state analysis and processing based on the temperatures of N temperature sampling points of the alcohol hydrogen fuel cell and supercapacitor of the silicon-based electrode and the silicon-based chip electric plate within a unit time, and constructs a four-dimensional thermal perception model varying with time;

[0007] A second processing module performs feature extraction based on the four-dimensional thermal perception model to obtain four-dimensional thermal perception feature data;

[0008] A risk diagnosis module performs high-temperature risk area anomaly diagnosis based on the four-dimensional thermal perception feature data, and identifies Y high-temperature abnormal areas;

[0009] An adaptive control module performs adaptive control on the Y high-temperature abnormal areas of the alcohol hydrogen fuel cell and supercapacitor of the silicon-based electrode and the silicon-based chip electric plate.

[0010] Further, the method for self-adaptive control of Y high-temperature abnormal areas of the alcohol hydrogen fuel cell and supercapacitor of the silicon-based electrode and silicon-based chip electric plate comprises:

[0011] S600: setting the initial value of y as 1, and the value range of y as 1 to Y;

[0012] S601: obtaining the high-temperature abnormal diagnosis score of the yth high-temperature abnormal area; if the high-temperature abnormal diagnosis score is greater than or equal to the preset high-intensity adjustment score threshold, immediately executing deep power limiting by the power management unit, converting the high-intensity adjustment score threshold into a high-intensity adjustment ratio, proportionally reducing the duty cycle and current amplitude of the pulse current to the value obtained by multiplying the original nominal value by the high-intensity adjustment ratio, and switching the base current borne by the silicon-based electrode and silicon-based chip electric plate supercapacitor to the hydrogen fuel cell stack according to 1-high-intensity adjustment ratio; the cooling circulating liquid of the hydrogen fuel cell stack is introduced into the composite heat conduction jacket outside the copper-molybdenum heat conduction column through the heat exchange conduit, and forms a convection-phase change composite heat dissipation path with the phase change material between the silicon-based electrode and the chip electric plate;

[0013] If the high-temperature abnormal diagnosis score is greater than or equal to the preset medium-intensity adjustment score threshold and less than the preset high-intensity adjustment score threshold, the current power reduction level is maintained unchanged, the medium-intensity adjustment score threshold is converted into a medium-intensity adjustment ratio, and the heat conduction duty cycle of the heat conduction channel is adjusted to the medium-intensity adjustment ratio;

[0014] If the high-temperature abnormal diagnosis score is less than the preset medium-intensity adjustment score threshold, it is recorded as score stable, it is judged whether the score is stable in the last two adjustment cycles, and if the result is yes, all parameters are restored to the original nominal value, the phase change material system is switched to standby state, and the melting-solidification heat cycle is maintained to cope with the next round of temperature rise at any time;

[0015] S602: setting y=y+1, if y is less than or equal to Y, returning to S601 for continuous execution, and if y is greater than Y, ending the current process.

[0016] Further, the identification method of Y high-temperature abnormal areas comprises:

[0017] S500: setting the initial value of the traversal index q as 1, and the value range of q as 1 to Q; setting the initial value of the count variable Y as 0; Q is the number corresponding to the high-temperature risk area;

[0018] S501: Obtain a hotspot label set corresponding to the qth high-temperature risk area; obtain a temperature change rate set in the four-dimensional heat perception feature data; extract the temperature change rate corresponding to each three-dimensional coordinate point in the hotspot label set in the temperature change rate set according to a spatial coordinate matching rule, and construct a regional temperature change rate set of the qth high-temperature risk area; obtain spatial distribution feature data and thermal feature data of the qth high-temperature risk area from the four-dimensional heat perception feature data;

[0019] S502: Input the regional temperature change rate set, the spatial distribution feature data and the thermal feature data of the qth high-temperature risk area into the high-temperature anomaly diagnosis model to obtain a corresponding high-temperature anomaly diagnosis score;

[0020] S503: If the high-temperature anomaly diagnosis score is greater than or equal to a preset high-temperature anomaly diagnosis score threshold, mark the qth high-temperature risk area as a high-temperature anomaly area to obtain a Yth high-temperature anomaly area, and let Y = Y + 1;

[0021] S504: Let q = q + 1, if q is less than or equal to Q, return to S501 to continue execution, if q is greater than Q, obtain Y high-temperature anomaly areas, and end the current process.

[0022] Further, the method for obtaining the four-dimensional heat perception feature data comprises:

[0023] Calculate the temperature change rate of each three-dimensional spatial coordinate in the four-dimensional heat perception model in a unit time, and construct a temperature change rate set to identify potential overheating trends; the three-dimensional spatial coordinates in the four-dimensional heat perception model are not limited to the positions of the sampling points, but are all definable calculation regions in the four-dimensional heat perception model;

[0024] Based on a temperature threshold determination rule, extract three-dimensional spatial coordinates whose temperature exceeds a preset temperature threshold in a unit time, construct a hotspot label, and divide the hotspot label into Q high-temperature risk areas by using a spatial clustering algorithm;

[0025] Based on the spatial position changes of the Q high-temperature risk areas in a time sequence, extract spatial distribution feature data corresponding to the Q high-temperature risk areas; based on the temperature changes of the Q high-temperature risk areas in a time sequence, extract thermal feature data corresponding to the Q high-temperature risk areas;

[0026] Construct the temperature change rate set, the Q spatial distribution feature data and the Q thermal feature data into four-dimensional heat perception feature data.

[0027] Further, the method for obtaining the Q high-temperature risk areas comprises:

[0028] a preset high-temperature identification threshold; in a unit time, traverse each three-dimensional spatial coordinate in the four-dimensional thermal perception model, extract the temperature sequence of the three-dimensional spatial coordinate at M time points, count the proportion of the temperature sequence greater than the high-temperature identification threshold, denoted as the threshold-exceeding proportion, mark the three-dimensional spatial coordinate corresponding to the threshold-exceeding proportion exceeding the preset threshold-exceeding proportion threshold as a hot spot label, and form a hot spot label set by all hot spot labels;

[0029] Based on the hot spot label set, a spatial clustering algorithm is used to group all hot spot labels to obtain Q hot spot label clustering results, each clustering result is defined as an independent high-temperature risk area, and Q spatially continuous high-temperature risk areas are formed.

[0030] Further, the method for obtaining the spatial distribution feature data corresponding to the Q high-temperature risk areas comprises:

[0031] S300: Let the initial value of the traversal index q be 1, and the value range of q be 1 to Q;

[0032] S301: Obtain the hot spot label set of the qth high-temperature risk area corresponding to M time points in a unit time, and the M time points correspond to M hot spot label sets;

[0033] S302: Calculate the spatial distribution diameter, region center point coordinate and average radius distance of the qth high-temperature risk area corresponding to M time points based on the M hot spot label sets of the M time points; calculate the corresponding point distribution density based on the spatial distribution diameter; the spatial distribution diameter includes the distribution diameters of the high-temperature risk area in three directions of horizontal, vertical and vertical directions in three-dimensional space;

[0034] S303: Construct the region center point coordinates of M time points in a unit time into a region center point coordinate set, and the region center point coordinate set is a hot spot migration path; calculate the center point displacement vector average value based on the region center point coordinates of M time points in a unit time, and the center point displacement vector average value is a diffusion direction;

[0035] S304: Construct the hot spot migration path, diffusion direction, spatial distribution diameter corresponding to M time points, region center point coordinate, average radius distance and point distribution density into the spatial distribution feature data of the qth high-temperature risk area;

[0036] S305: Let q=q+1, if q is less than or equal to Q, return to S301 to continue execution, if q is greater than Q, obtain the spatial distribution feature data corresponding to the Q high-temperature risk areas, and end the current process.

[0037] Further, the method for obtaining the thermal feature data corresponding to the Q high-temperature risk areas comprises:

[0038] S400: Let the initial value of the traversal index q be 1, and the value range of q be 1 to Q;

[0039] S401: Obtain the hotspot label set corresponding to the qth high-temperature risk area at M time points in a unit time, and the M time points correspond to M hotspot label sets;

[0040] S402: Based on the temperatures corresponding to the M hotspot label sets at the M time points, construct a hotspot label temperature set corresponding to the M time points;

[0041] S403: Based on the hotspot label temperature set at the M time points, calculate the temperature mean, temperature standard deviation, temperature kurtosis and temperature skewness corresponding to the qth high-temperature risk area at the M time points;

[0042] S404: Construct the temperature mean, temperature standard deviation, temperature kurtosis and temperature skewness corresponding to the M time points as the thermal feature data of the qth high-temperature risk area;

[0043] S405: Let q = q + 1, if q is less than or equal to Q, return to S401 for continuous execution, if q is greater than Q, obtain the thermal feature data corresponding to the Q high-temperature risk areas, and end the current process.

[0044] Further, the method for constructing the four-dimensional thermal perception model changing over time comprises:

[0045] Divide a unit time into M time points to form a time sequence, so as to realize time discrete modeling of the system thermal state;

[0046] At each time point, N spatial temperature sampling points set in the system collect the temperature at the location respectively, forming a temperature sampling data set at the time point; an interpolation algorithm is used to perform spatial fitting on the N temperature sampling values at the time point, to construct a corresponding continuous three-dimensional temperature distribution function, forming a corresponding continuous temperature field;

[0047] Sequence and combine the continuous temperature fields at the M time points to form a four-dimensional thermal perception model changing over time.

[0048] Further, the training method of the high-temperature anomaly diagnosis model comprises:

[0049] Pre-construct a high-temperature anomaly diagnosis data set, the high-temperature anomaly diagnosis data set comprising YC high-temperature anomaly diagnosis data and high-temperature anomaly diagnosis scores corresponding to the YC high-temperature anomaly diagnosis data, YC being a positive integer; divide the high-temperature anomaly diagnosis data set into a training set and a validation set, the training set being used for high-temperature anomaly diagnosis model parameter learning, and the validation set being used for real-time monitoring of the generalization performance and overfitting degree of the high-temperature anomaly diagnosis model;

[0050] A deep neural network with a multilayer perceptron as the core is used as the high temperature anomaly diagnosis model. The high temperature anomaly diagnosis data is standardized and vectorized and then input into the deep neural network. The deep neural network consists of an input layer, a hidden layer, and an output layer. Each hidden layer uses a nonlinear activation function to extract high-order features, and the output layer uses a softmax activation function to obtain the probability distribution corresponding to each high temperature anomaly diagnosis score. Finally, the high temperature anomaly diagnosis score corresponding to the maximum probability is taken as the prediction result of the high temperature anomaly diagnosis model. During the training process, the optimization goal is to minimize the cross entropy loss function, and a gradient descent optimization algorithm is used to update the network weights. An early stopping strategy is set: when the prediction accuracy on the validation set reaches or exceeds the preset threshold, the high temperature anomaly diagnosis model is judged to have converged and the training is terminated.

[0051] Furthermore, the method for constructing the temperature change rate set includes:

[0052] S201: The total number of three-dimensional space coordinate points defined in the four-dimensional thermal sensing model is recorded as R, and the initial value of the traversal index r is set to 1, and the value range of r is 1 to R;

[0053] S202: Obtain the rth three-dimensional spatial coordinate in the four-dimensional thermal perception model, calculate the temperature change rate of the rth three-dimensional spatial coordinate over time based on the temperature sequence formed by the rth three-dimensional spatial coordinate at M time points within a unit time, obtain the temperature change rate corresponding to the rth three-dimensional spatial coordinate within a unit time, and construct a temperature change rate subset of the rth three-dimensional spatial coordinate;

[0054] S203: Increment r by 1. If the current r is less than or equal to R, return to step S202 to continue processing the next three-dimensional space coordinate position; if r is greater than R, the traversal is completed and the current process ends.

[0055] Compared with the existing technology, the technical effects and advantages of the alcohol-hydrogen fuel cell and supercapacitor system of the silicon-based electrode and silicon-based chip plate of the present invention are as follows:

[0056] The alcohol hydrogen fuel cell and super capacitor system of the silicon-based electrode and silicon-based chip electric plate proposed in the application has the system capability of multi-dimensional thermal perception, intelligent risk identification and hierarchical response regulation, which can significantly improve the safety and reliability of super capacitor in high temperature and high power density scenes. By arranging multi-point temperature acquisition units at key positions of the silicon-based electrode, chip electric plate and capacitor core, a four-dimensional thermal perception model is constructed by combining interpolation algorithm to realize dynamic perception of the change of thermal state with time and space. Y high temperature abnormal areas are identified by using spatial clustering and deep neural network technology, and a "deep power limit-adjustable heat conduction-PCM buffer" triple adaptive response control mechanism is started according to the abnormal score grading, so as to accurately adjust the pulse current duty ratio and heat conduction duty ratio, and realize temperature rise control under the condition of no fan.

[0057] By constructing the super capacitor-hydrogen fuel cell collaborative regulation structure and introducing the intelligent power shunt control strategy driven by high temperature diagnosis, the application realizes the adaptive thermal-electric collaborative regulation of super capacitor to hydrogen fuel cell stack, ensures that the stack always operates in the "thermal comfort zone" with structural safety, high reaction efficiency and stable chemical load, and greatly improves the energy supply continuity and thermal control intelligence of the system under high temperature and high load coupling conditions. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 The alcohol hydrogen fuel cell and super capacitor system of the silicon-based electrode and silicon-based chip electric plate proposed in the application has the system capability of multi-dimensional thermal perception, intelligent risk identification and hierarchical response regulation, which can significantly improve the safety and reliability of super capacitor in high temperature and high power density scenes. By arranging multi-point temperature acquisition units at key positions of the silicon-based electrode, chip electric plate and capacitor core, a four-dimensional thermal perception model is constructed by combining interpolation algorithm to realize dynamic perception of the change of thermal state with time and space. Y high temperature abnormal areas are identified by using spatial clustering and deep neural network technology, and a "deep power limit-adjustable heat conduction-PCM buffer" triple adaptive response control mechanism is started according to the abnormal score grading, so as to accurately adjust the pulse current duty ratio and heat conduction duty ratio, and realize temperature rise control under the condition of no fan.

[0059] Figure 2 The alcohol hydrogen fuel cell and super capacitor system of the silicon-based electrode and silicon-based chip electric plate proposed in the application has the system capability of multi-dimensional thermal perception, intelligent risk identification and hierarchical response regulation, which can significantly improve the safety and reliability of super capacitor in high temperature and high power density scenes. By arranging multi-point temperature acquisition units at key positions of the silicon-based electrode, chip electric plate and capacitor core, a four-dimensional thermal perception model is constructed by combining interpolation algorithm to realize dynamic perception of the change of thermal state with time and space. Y high temperature abnormal areas are identified by using spatial clustering and deep neural network technology, and a "deep power limit-adjustable heat conduction-PCM buffer" triple adaptive response control mechanism is started according to the abnormal score grading, so as to accurately adjust the pulse current duty ratio and heat conduction duty ratio, and realize temperature rise control under the condition of no fan.

[0060] Figure 3 The alcohol hydrogen fuel cell and super capacitor system of the silicon-based electrode and silicon-based chip electric plate proposed in the application has the system capability of multi-dimensional thermal perception, intelligent risk identification and hierarchical response regulation, which can significantly improve the safety and reliability of super capacitor in high temperature and high power density scenes. By arranging multi-point temperature acquisition units at key positions of the silicon-based electrode, chip electric plate and capacitor core, a four-dimensional thermal perception model is constructed by combining interpolation algorithm to realize dynamic perception of the change of thermal state with time and space. Y high temperature abnormal areas are identified by using spatial clustering and deep neural network technology, and a "deep power limit-adjustable heat conduction-PCM buffer" triple adaptive response control mechanism is started according to the abnormal score grading, so as to accurately adjust the pulse current duty ratio and heat conduction duty ratio, and realize temperature rise control under the condition of no fan.

[0061] Figure 4 The alcohol hydrogen fuel cell and super capacitor system of the silicon-based electrode and silicon-based chip electric plate proposed in the application has the system capability of multi-dimensional thermal perception, intelligent risk identification and hierarchical response regulation, which can significantly improve the safety and reliability of super capacitor in high temperature and high power density scenes. By arranging multi-point temperature acquisition units at key positions of the silicon-based electrode, chip electric plate and capacitor core, a four-dimensional thermal perception model is constructed by combining interpolation algorithm to realize dynamic perception of the change of thermal state with time and space. Y high temperature abnormal areas are identified by using spatial clustering and deep neural network technology, and a "deep power limit-adjustable heat conduction-PCM buffer" triple adaptive response control mechanism is started according to the abnormal score grading, so as to accurately adjust the pulse current duty ratio and heat conduction duty ratio, and realize temperature rise control under the condition of no fan.

[0062] Figure 5 The alcohol hydrogen fuel cell and super capacitor system of the silicon-based electrode and silicon-based chip electric plate proposed in the application has the system capability of multi-dimensional thermal perception, intelligent risk identification and hierarchical response regulation, which can significantly improve the safety and reliability of super capacitor in high temperature and high power density scenes. By arranging multi-point temperature acquisition units at key positions of the silicon-based electrode, chip electric plate and capacitor core, a four-dimensional thermal perception model is constructed by combining interpolation algorithm to realize dynamic perception of the change of thermal state with time and space. Y high temperature abnormal areas are identified by using spatial clustering and deep neural network technology, and a "deep power limit-adjustable heat conduction-PCM buffer" triple adaptive response control mechanism is started according to the abnormal score grading, so as to accurately adjust the pulse current duty ratio and heat conduction duty ratio, and realize temperature rise control under the condition of no fan. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present application will be described in detail, clearly and completely below with reference to the drawings in the embodiments of the present application. It should be particularly noted that the specific embodiments described below are only used to better illustrate and describe the technical solutions of the present application, and are intended to enable those skilled in the art to better understand and implement the present application, and should not be understood as limiting the protection scope of the present application. Those skilled in the art can modify, adjust or equivalently replace the present application according to the content disclosed in the present application without departing from the spirit and essence of the present application, and these should be regarded as the protection scope of the present application.

[0064] Embodiment 1

[0065] Please refer to Figure 1 As shown in the figure, the embodiment discloses an alcohol hydrogen fuel cell and supercapacitor system of a silicon-based electrode and a silicon-based chip electric plate, which comprises a first processing module, a second processing module, a risk diagnosis module and a self-adaptive control module, each module is connected through wired and / or wireless connection to realize data transmission.

[0066] The first processing module performs thermal state analysis and processing based on the temperatures of N temperature sampling points of the alcohol hydrogen fuel cell and supercapacitor of the silicon-based electrode and silicon-based chip electric plate in a unit time, and constructs a four-dimensional thermal perception model changing with time.

[0067] In the embodiment of the present application, the temperature data of the N temperature sampling points can be obtained by arranging temperature acquisition units at multiple key positions of the silicon-based electrode, supercapacitor core and silicon-based chip electric plate. The temperature acquisition unit can include any type or combination of digital thermal sensor, thermistor sensor or micro-electro-mechanical system (MEMS) temperature sensor, etc., for realizing multi-point perception of the temperature state of the internal and boundary regions of the device. Preferably, the temperature acquisition unit can be arranged at the following regions: including but not limited to the surface region and internal region of the silicon-based electrode (such as the edge position of the nanosilicon sheet), the upper part, bottom and side wall structure of the supercapacitor core, the hydrogen fuel cell stack, and the key functional regions prone to local temperature rise in the silicon-based chip electric plate, such as the positions of the power driving module or control integrated circuit (IC). Through the above multi-point arrangement mode, a representative spatial temperature acquisition matrix can be constructed, providing accurate temperature data basis for subsequent thermal field modeling and construction of spatial thermal perception network.

[0068] The hydrogen fuel cell stack and the supercapacitor construct a functional independent but highly synergistic dual-main-source energy supply structure, and both are in a parallel operation architecture at the same level.

[0069] The hydrogen fuel cell stack is composed of a plurality of silicon-based plates with micro-flow channels to form a membrane electrode assembly, and a plurality of membrane electrode assemblies are stacked in a hot-pressing soldering manner to form a working stack. The stack uses hydrogen as fuel, generates stable direct current power output through electrochemical reaction, is used for maintaining long-term load power supply of the system, and generates controllable heat flow with a temperature in the range of 60-75℃ to provide buffer for system thermal stability.

[0070] The super capacitor has a double-layer electrode material of a highly doped silicon wafer, forms a pulse energy release structure with extremely low equivalent series resistance (ESR), and is used to provide fast energy support during load transient rise or fuel cell stack cold start stage, and at the same time, through a bidirectional DC-DC conversion module, to dynamically compensate the hydrogen fuel cell stack, maintain the hydrogen fuel cell stack membrane hydration, plate temperature and chemical reaction rate in the stable interval.

[0071] In the working mechanism, the super capacitor is mainly responsible for high-frequency pulse load response, and the hydrogen fuel cell stack undertakes stable base current power supply, and the two are complementary and cooperative in the power supply side with frequency division bandwidth logic; in the heat management side, the super capacitor provides a high-density heat source, and the hydrogen fuel cell stack provides stable low-grade heat flow and assists heat extraction. The working temperature of the hydrogen fuel cell stack is controlled at 60-75℃, and the outlet of the cooling water of the hydrogen fuel cell stack is coupled with the copper-molybdenum heat conduction column and the PCM melting layer of the silicon-based capacitor through a heat exchange system, forming a stable and adjustable heat conduction path, greatly reducing the thermal resistance and improving the thermal stability of the system.

[0072] The dual-source coupling structure of the hydrogen fuel cell stack and the super capacitor not only effectively alleviates heat accumulation and improves the stable operation ability of the system in high-temperature environment, but also expands the energy supply flexibility of the system, and exhibits good thermal steady-state control and thermal response dynamic performance in complex working conditions such as new energy vehicles, well logging power supplies and aviation equipment.

[0073] To avoid problems such as membrane dehydration and polarization aggravation of the hydrogen fuel cell stack during low temperature or load fluctuation, the super capacitor module outputs a short-period high-current pulse to the input side of the fuel cell stack in a dynamic feedback control manner, and combines the thermal conduction characteristics of the silicon-based material to perform thermal coupling adjustment through the copper-molybdenum composite heat bridge, to realize bidirectional dynamic compensation of electrical energy and thermal energy, so as to maintain the operation of the stack in the high conversion efficiency interval.

[0074] In summary, the hydrogen fuel cell stack and the ultracapacitor system are coupled in parallel in the embodiment, the hydrogen fuel cell stack and the ultracapacitor are complementary in function, heat control is coordinated and intelligent linkage is achieved, and the mechanism of dynamic power supply of the ultracapacitor to the hydrogen fuel cell is introduced, so that the double-channel coordinated regulation in the energy supply level and the heat control level is achieved, and the technical problems of slow response and temperature control lag of the fuel cell under the traditional one-way energy supply structure are broken through. The application has significant progress in thermal stability, electrochemical energy efficiency and intelligent collaborative control.

[0075] Benefiting from the high temperature resistance of the silicon-based electrode, the straight-through heat dissipation of the vertical heat channel and the high airtightness of the sheet-type vacuum packaging, the ultracapacitor can work continuously at-40°C to 125°C, and the capacity retention rate is still higher than 90% according to the IEC62391-1 high temperature life test (125°C, 1000h). Compared with the rapid decay of the conventional carbon-based or aluminum electrode ultracapacitor at 75°C to 105°C, the high temperature reliability is greatly improved. The silicon-based electrode ultracapacitor has high specific power of the supercapacitor and high specific energy brought by the lithium insertion characteristics of the silicon-based electrode, and realizes the comprehensive performance advantages of high energy density, high power density, low resistance and fast temperature control. The silicon-based electrode and the silicon-based chip electric plate ultracapacitor are suitable for new energy vehicle direct current link energy storage, oil well logging high temperature power supply and aviation power system and other high temperature complex working conditions.

[0076] As shown in Figure 4 , the method for constructing the four-dimensional thermal perception model changing with time comprises the following steps:

[0077] Divide the unit time into M time points to form a time sequence to realize time discrete modeling of the system thermal state;

[0078] At each time point t m , m∈[1,M], the N spatial temperature sampling points set in the system collect the temperature at the location respectively to form a temperature sampling data set at the time point;

[0079] The N temperature sampling values at the time point are spatially fitted by using an interpolation algorithm (such as inverse distance weighting, Kriging interpolation, etc.) to construct a corresponding continuous three-dimensional temperature distribution function to form a corresponding continuous temperature field;

[0080] The continuous temperature fields at the M time points are sequentially combined to form a four-dimensional thermal perception model changing with time.

[0081] It should be noted that the temperature sampling data set is SJJ n，m = {(x n , y n , z n , WD n，m}, wherein (x n , y n , zn ) is the three-dimensional space coordinate of the nth temperature sampling point, WD n，m is the temperature of the nth temperature sampling point at the mth time point, n is the variable index of the number of temperature sampling points. By constructing temperature distribution functions at M time points respectively, the system can obtain a four-dimensional thermal perception model of the thermal state evolution over time, which is used for subsequent temperature rise rate analysis, hot spot migration prediction, local overheating prevention and control, and intelligent control strategy formulation.

[0082] The second processing module performs feature extraction based on the four-dimensional thermal perception model to obtain four-dimensional thermal perception feature data.

[0083] The method for obtaining the four-dimensional thermal perception feature data includes:

[0084] The temperature change rate of each three-dimensional space coordinate in the four-dimensional thermal perception model within a unit time is calculated, and a temperature change rate set is constructed to identify potential overheating trends; the three-dimensional space coordinates in the four-dimensional thermal perception model are not limited to the positions of the sampling points, but are all definable calculation regions within the four-dimensional thermal perception model;

[0085] Based on a temperature threshold judgment rule, three-dimensional space coordinates with temperatures exceeding a preset temperature threshold within a unit time are extracted, a hot spot label is constructed, and the hot spot label is divided into Q high-temperature risk areas using a spatial clustering algorithm;

[0086] Based on the spatial position changes of the Q high-temperature risk areas in the time series, spatial distribution feature data corresponding to the Q high-temperature risk areas are extracted; based on the temperature changes of the Q high-temperature risk areas in the time series, thermal feature data corresponding to the Q high-temperature risk areas are extracted;

[0087] The temperature change rate set, the Q spatial distribution feature data, and the Q thermal feature data are constructed into four-dimensional thermal perception feature data.

[0088] The method for constructing the temperature change rate set includes:

[0089] S201: The total number of three-dimensional space coordinate points defined in the four-dimensional thermal perception model is denoted as R, and the initial value of the traversal index r is set to 1, with the value range of r being 1 to R;

[0090] S202: The rth three-dimensional space coordinate in the four-dimensional thermal perception model is obtained, and based on the temperature sequence of the rth three-dimensional space coordinate at M time points within a unit time, the temperature change rate of the rth three-dimensional space coordinate over time is calculated to obtain the temperature change rate corresponding to the rth three-dimensional space coordinate within a unit time, and a temperature change rate sub-set of the rth three-dimensional space coordinate is constructed;

[0091] The method for calculating the temperature change rate includes:

[0092]

[0093] wherein, BHL r，i is the i-th temperature rate of change of the r-th three-dimensional spatial coordinate, i ranges from 1 to M-1, WD r，i+1 represents the temperature corresponding to the r-th three-dimensional spatial coordinate at the i+1-th time point, WD r，i represents the temperature corresponding to the r-th three-dimensional spatial coordinate at the i-th time point, SJD i+1 represents the i+1-th time point, SJD i represents the i-th time point.

[0094] S203: increase r by 1, if the current r is less than or equal to R, return to step S202 to continue processing the next three-dimensional spatial coordinate position; if r is greater than R, the traversal is completed, and the current process is ended.

[0095] The method for obtaining the Q high-temperature risk areas comprises:

[0096] A high-temperature identification threshold is preset, which can be set according to the temperature resistance characteristics of the silicon-based electrode and the silicon-based chip electric plate supercapacitor, and is preferably set to be between 80% and 95% of the critical working temperature of the silicon-based electrode and the silicon-based chip electric plate supercapacitor;

[0097] In a unit time, each three-dimensional spatial coordinate in the four-dimensional thermal perception model is traversed, the temperature sequence of the three-dimensional spatial coordinate at the M time points is extracted, the proportion of the temperature sequence greater than the high-temperature identification threshold is counted, which is recorded as a threshold-exceeding proportion, the three-dimensional spatial coordinate corresponding to the threshold-exceeding proportion exceeding a preset threshold-exceeding proportion threshold is marked as a hot spot label, and all hot spot labels form a hot spot label set.

[0098] Based on the hot spot label set, a spatial clustering algorithm (preferably a DBSCAN or K-means clustering method) is used to group all hot spot labels, to obtain Q hot spot label clustering results, each clustering result is defined as an independent high-temperature risk area, and Q spatially continuous high-temperature risk areas are formed.

[0099] The method for obtaining the spatial distribution characteristic data corresponding to the Q high-temperature risk areas comprises:

[0100] S300: let the initial value of the traversal index q be 1, and q ranges from 1 to Q;

[0101] S301: obtain the hot spot label set of the q-th high-temperature risk area at the M time points in a unit time, and the M time points correspond to M hot spot label sets;

[0102] S302: Calculate the spatial distribution diameter, the region center point coordinate and the average radius distance of the qth high-temperature risk region at the M time points based on the M hotspot label sets of the M time points; calculate the corresponding point distribution density based on the spatial distribution diameter; the spatial distribution diameter includes the distribution diameters of the high-temperature risk region in three directions of horizontal, vertical and longitudinal directions in three-dimensional space;

[0103] It should be noted that by obtaining the distribution diameters in three axial directions, the three-dimensional spatial structure form and the main diffusion direction of the hotspot region can be accurately represented, thereby providing basic support for subsequent thermal behavior identification and spatial regulation strategy.

[0104] S303: Construct the region center point coordinates of the M time points in a unit time into a region center point coordinate set, and the region center point coordinate set is a hotspot migration path; calculate the average value of the center point displacement vector based on the region center point coordinates of the M time points in a unit time, and the average value of the center point displacement vector is a diffusion direction;

[0105] S304: Construct the hotspot migration path, the diffusion direction, the spatial distribution diameter corresponding to the M time points, the region center point coordinate, the average radius distance and the point distribution density into the spatial distribution feature data of the qth high-temperature risk region;

[0106] S305: Let q=q+1, if q is less than or equal to Q, return to S301 to continue execution, if q is greater than Q, obtain the spatial distribution feature data corresponding to the Q high-temperature risk regions, and end the current process.

[0107] The method for obtaining the spatial distribution diameter comprises:

[0108] XDIA q，m =max(X q，m )-min(X q，m );

[0109] XDIA q，m represents the distribution diameter of the qth high-temperature risk region in the spatial horizontal direction corresponding to the mth time point, max(X q，m ) represents the maximum coordinate of the qth high-temperature risk region in the spatial horizontal direction at the mth time point, and min(X q，m ) represents the minimum coordinate of the qth high-temperature risk region in the spatial horizontal direction at the mth time point.

[0110] YDIA q，m =max(Y q，m )-min(Y q，m );

[0111] YDIA q，mrepresents the distribution diameter of the qth high-temperature risk area in the spatial longitudinal direction corresponding to the mth time point, max(Y q，m ) represents the spatial longitudinal maximum coordinate of the qth high-temperature risk area at the mth time point, min(Y q，m ) represents the spatial longitudinal minimum coordinate of the qth high-temperature risk area at the mth time point.

[0112] ZDIA q，m = max(Z q，m )-min(Z q，m );

[0113] wherein ZDIA q，m represents the distribution diameter of the qth high-temperature risk area in the spatial vertical direction corresponding to the mth time point, max(Z q，m ) represents the spatial vertical maximum coordinate of the qth high-temperature risk area at the mth time point, min(Z q，m ) represents the spatial vertical minimum coordinate of the qth high-temperature risk area at the mth time point.

[0114] The method for calculating the region center point coordinate comprises:

[0115] Center q，m = (Xcen q，m , Ycen q，m , Zcen q，m );

[0116]

[0117] wherein Center q，m is the region center point coordinate of the qth high-temperature risk area at the mth time point, Xcen q，m is the spatial horizontal region center point coordinate of the qth high-temperature risk area at the mth time point, Ycen q，m is the spatial longitudinal region center point coordinate of the qth high-temperature risk area at the mth time point, Zcen q，m is the spatial vertical region center point coordinate of the qth high-temperature risk area at the mth time point, NUM q，m is the number of three-dimensional space coordinates of the qth high-temperature risk area at the mth time point, j is an index variable of the summation formula, X j is the jth spatial horizontal coordinate point, Y j is the jth spatial longitudinal coordinate point, and Z j is the jth spatial vertical coordinate point.

[0118] The method for calculating the average radius distance comprises:

[0119] Calculate the Euclidean distance of each three-dimensional space coordinate in the high-temperature risk area to the center point coordinate of the area:

[0120] Sort all Euclidean distances in ascending order, and select the Euclidean distance corresponding to the last B% to construct an edge point Euclidean distance set, B being a predetermined constant, indicating the selection ratio;

[0121] Calculate the average radius distance by calculating the average of the edge point Euclidean distances in the edge point Euclidean distance set.

[0122] The method for calculating the Euclidean distance comprises:

[0123]

[0124] wherein, JL k represents the kth Euclidean distance in the high-temperature risk area, (X k , Y k , Z k ) is the kth three-dimensional space coordinate in the high-temperature risk area.

[0125] The method for calculating the average of the edge point Euclidean distances in the edge point Euclidean distance set comprises:

[0126]

[0127] wherein, is the average of the edge point Euclidean distances, BYJH num is the number of edge point Euclidean distances in the edge point Euclidean distance set, JL h is the hth edge point Euclidean distance in the edge point Euclidean distance set.

[0128] It should be noted that the average radius distance is used to reflect the concentration degree of the hot spot in the high-temperature risk area in space, and the edge point distance average method can effectively suppress the compression effect of the center gathering point on the overall index, so as to more accurately depict the actual expansion degree of the boundary of the area shape. In this embodiment, by sorting the Euclidean distances of all hot spots to the center point of the area, and selecting the farthest 25% of the points as the edge point set for average processing, the representativeness and statistical stability of the edge shape can be considered, the interference of abnormal points can be avoided, and the divergence trend and expansion rate of the hot area can be effectively represented.

[0129] The method for calculating the point distribution density comprises:

[0130]

[0131] wherein, FBMD q，m is the point distribution density of the qth high-temperature risk area at the mth time point.

[0132] It should be noted that the spatial distribution diameter, the average radius distance and the point distribution density are not directly used as the construction input of the migration path or the diffusion direction, but are used to represent the spatial form characteristics of the high-temperature risk area in a unit time. The spatial distribution diameter, the average radius distance and the point distribution density can assist in determining the type of high-temperature area, and are used to distinguish the heat concentration, the diffusion mode and the dispersion degree, and further support the corresponding response strategy of the thermal control system according to the different thermal region forms. For example, for a central high-temperature type area with a small spatial distribution diameter and a high point density, a fixed-point active heat dissipation scheme can be adopted; for a discrete island type area with a large radius and a low density, an energy limitation and a periodic thermal analysis linkage mechanism can be adopted. By introducing the above parameters, the partition perception ability, the strategy differentiation ability and the accuracy of the overall thermal response of the thermal control system can be enhanced.

[0133] The method for obtaining the thermal feature data corresponding to the Q high-temperature risk areas comprises the following steps:

[0134] S400: setting the initial value of the traversal index q as 1, and the value range of q as 1 to Q;

[0135] S401: obtaining a set of hot spot labels corresponding to the qth high-temperature risk area at M time points in a unit time, and the M time points corresponding to M sets of hot spot labels;

[0136] S402: constructing a set of hot spot label temperatures corresponding to the M time points based on the temperatures corresponding to the M sets of hot spot labels of the M time points;

[0137] S403: calculating the temperature mean value, the temperature standard deviation, the temperature kurtosis and the temperature skewness of the qth high-temperature risk area at the M time points based on the set of hot spot label temperatures of the M time points;

[0138] S404: constructing the temperature mean value, the temperature standard deviation, the temperature kurtosis and the temperature skewness corresponding to the M time points as the thermal feature data of the qth high-temperature risk area;

[0139] S405: setting q = q + 1, if q is less than or equal to Q, returning to S401 for continuous execution, and if q is greater than Q, obtaining the thermal feature data corresponding to the Q high-temperature risk areas, and ending the current process.

[0140] The calculation method of the temperature mean value comprises the following steps:

[0141]

[0142] wherein, is the temperature mean value of the qth high-temperature risk area at the mth time point, WD g is the gth hot spot label temperature in the set of hot spot label temperatures of the qth high-temperature risk area at the mth time point.

[0143] The calculation method of the temperature standard deviation comprises:

[0144]

[0145] WDBC(q, m) = std(Tq(m)) where WDBC q，m is the temperature standard deviation of the qth high-temperature risk area at the mth time point, used to reflect the temperature fluctuation degree.

[0146] The calculation method of the temperature kurtosis comprises:

[0147]

[0148] WDFD(q, m) = kurt(Tq(m)) where WDFD q，m is the temperature kurtosis of the qth high-temperature risk area at the mth time point, used to identify the concentrated peak area.

[0149] The calculation method of the temperature skewness comprises:

[0150]

[0151] WDPD(q, m) = skew(Tq(m)) where WDPD q，m is the temperature skewness of the qth high-temperature risk area at the mth time point, used to judge the distribution bias.

[0152] The risk diagnosis module performs high-temperature risk area anomaly diagnosis based on the four-dimensional heat perception feature data, and identifies Y high-temperature abnormal areas.

[0153] The identification method of the Y high-temperature abnormal areas comprises:

[0154] S500: let the initial value of the traversal index q be 1, the value range of q be 1 to Q; let the initial value of the count variable Y be 0; Q is the number corresponding to the high-temperature risk area;

[0155] S501: obtain the hotspot label set corresponding to the qth high-temperature risk area; obtain the temperature change rate set in the four-dimensional heat perception feature data; extract the corresponding temperature change rate of each three-dimensional coordinate point in the hotspot label set in the temperature change rate set according to the spatial coordinate matching rule, and construct the regional temperature change rate set of the qth high-temperature risk area; obtain the spatial distribution feature data and thermal feature data of the qth high-temperature risk area from the four-dimensional heat perception feature data;

[0156] S502: input the regional temperature change rate set, spatial distribution feature data and thermal feature data of the qth high-temperature risk area into the high-temperature anomaly diagnosis model to obtain the corresponding high-temperature anomaly diagnosis score;

[0157] S503: If the high-temperature anomaly diagnosis score is greater than or equal to a preset high-temperature anomaly diagnosis score threshold, the qth high-temperature risk area is marked as a high-temperature anomaly area, and the Yth high-temperature anomaly area is obtained, and Y is set to Y+1; the high-temperature anomaly diagnosis score threshold is set by combining sample verification, ROC curve analysis and backtest performance evaluation, and its value can be adaptively fine-tuned according to the model sensitivity and error tolerance. In this embodiment, the interval of the high-temperature anomaly diagnosis score is 0 to 1 (including 0 and 1), and the higher the high-temperature anomaly diagnosis score, the higher the degree of anomaly. The high-temperature anomaly diagnosis score threshold can be set to 0.78 to ensure that the high-risk hot spot area is not missed without sacrificing the identification accuracy, and is suitable for the thermal runaway risk warning task of the silicon-based electrode and the silicon-based chip electric plate supercapacitor high-power density running scene.

[0158] S504: q is set to q+1, and if q is less than or equal to Q, S501 is returned to continue execution, and if q is greater than Q, Y high-temperature anomaly areas are obtained, and the current process is ended.

[0159] The training method of the high-temperature anomaly diagnosis model comprises:

[0160] A high-temperature anomaly diagnosis data set is constructed in advance, the high-temperature anomaly diagnosis data set comprises YC groups of high-temperature anomaly diagnosis data and high-temperature anomaly diagnosis scores corresponding to the YC groups of high-temperature anomaly diagnosis data, YC is a positive integer; the high-temperature anomaly diagnosis data set is divided into a training set and a verification set, the training set is used for high-temperature anomaly diagnosis model parameter learning, and the verification set is used for real-time monitoring of the generalization performance and overfitting degree of the high-temperature anomaly diagnosis model;

[0161] A deep neural network with a multilayer perceptron as the core is used as the high-temperature anomaly diagnosis model, the high-temperature anomaly diagnosis data is input into the deep neural network after being standardized and vectorized, and the deep neural network comprises an input layer, a hidden layer and an output layer; each hidden layer uses a nonlinear activation function to extract high-order features, the output layer uses a Softmax activation function to obtain a probability distribution corresponding to each high-temperature anomaly diagnosis score, and finally the high-temperature anomaly diagnosis score corresponding to the maximum probability is taken as the prediction result of the high-temperature anomaly diagnosis model; in the training process, the cross-entropy loss function is minimized as the optimization target, a gradient descent type optimization algorithm is used to update the network weights, and an early stopping strategy is set: when the prediction accuracy on the verification set reaches or exceeds a preset threshold, it is determined that the high-temperature anomaly diagnosis model has converged and the training is terminated.

[0162] It should be noted that by inputting the dynamic behavior of the region temperature, the spatial form characteristics and the thermal statistical characteristics into the intelligent diagnosis model, the potential high temperature runaway area in the silicon-based electrode and the silicon-based chip electric plate supercapacitor is accurately identified and quantitatively evaluated, ensuring the real-time and professional risk determination, and providing reliable decision basis for subsequent adaptive control strategy.

[0163] In the integrated structure of the silicon-based electrode and the silicon-based chip electric plate supercapacitor, high-power pulse current is injected into the silicon-based electrode from the power MOS and the balancing IC in the center of the chip electric plate along the copper-molybdenum composite column, the contact resistance of the electrode-column interface and the ohmic loss of the electrode active layer form a heat source, the high heat flow diffuses outward in a centrifugal manner along the high thermal conductivity path of the silicon substrate, and the packaging shell and the AlN-silicon composite substrate edge are directly coupled with the external heat sink, and the heat dissipation condition is better than the central area, so that the temperature peak in the center area is significantly higher than that in the edge area and shows a typical "center-edge" gradient distribution; if the heat dissipation channel is saturated or the external heat dissipation is limited, the central temperature peak will continue to accumulate in unit time, which is easy to cause thermal runaway at the electrode-chip junction, material fatigue and thermal drift of functional devices, so the "center high temperature type" high temperature risk performance is the most consistent with the thermal behavior characteristics of the silicon-based electrode and the silicon-based chip electric plate supercapacitor, and the "center high temperature type" should be taken as the first monitoring and priority control of thermal runaway type.

[0164] Through the above steps, the spatial area whose temperature exceeds the threshold value can be accurately identified in unit time range from the four-dimensional thermal perception model, and the structured high temperature risk area division is formed, which provides a basis for subsequent heat dissipation device activation, adaptive load reduction control or abnormal warning.

[0165] The adaptive control module performs adaptive control on Y high temperature abnormal areas of the alcohol hydrogen fuel cell and the supercapacitor of the silicon-based electrode and the silicon-based chip electric plate.

[0166] As shown in Figure 5 The method for performing adaptive control on Y high temperature abnormal areas of the alcohol hydrogen fuel cell and the supercapacitor of the silicon-based electrode and the silicon-based chip electric plate includes:

[0167] S600: Let the initial value of y be 1, and the value range of y be 1 to Y;

[0168] S601: Obtain the high-temperature anomaly diagnosis score of the yth high-temperature anomaly region; if the high-temperature anomaly diagnosis score is greater than or equal to a preset high-intensity adjustment score threshold, make the power management unit immediately execute deep power limiting, convert the high-intensity adjustment score threshold into a high-intensity adjustment ratio (for example, the high-intensity adjustment score threshold is 0.9, and the high-intensity adjustment ratio is 90%), proportionally reduce the duty cycle and current amplitude of the pulse current to a value obtained by multiplying the original nominal value (the original nominal value refers to the design operating parameter of the system when no thermal anomaly is triggered) by the high-intensity adjustment ratio, and switch the base current borne by the silicon-based electrode and the silicon-based chip electric plate supercapacitor to the hydrogen fuel cell stack according to 1-high-intensity adjustment ratio; the cooling circulating liquid of the hydrogen fuel cell stack is introduced into the composite heat conduction jacket outside the copper-molybdenum heat conduction column through the heat exchange conduit, and forms a convection-phase change composite heat dissipation path with the phase change material (PCM) between the silicon-based electrode and the chip electric plate; the 1-high-intensity adjustment ratio refers to 1 minus the remaining part of the high-intensity adjustment ratio ratio;

[0169] If the high-temperature anomaly diagnosis score is greater than or equal to a preset medium-intensity adjustment score threshold and less than a preset high-intensity adjustment score threshold, the current power reduction level is maintained unchanged, the medium-intensity adjustment score threshold is converted into a medium-intensity adjustment ratio, and the heat conduction duty cycle of the heat conduction channel is adjusted to the medium-intensity adjustment ratio;

[0170] If the high-temperature anomaly diagnosis score is less than the preset medium-intensity adjustment score threshold, it is recorded as the score remaining stable, it is judged whether the score remains stable in the last two adjustment cycles, and if the result of the judgment is yes, all parameters are restored to the original nominal value, the phase change material system is switched to standby state, and the melting-solidification heat cycle is maintained to cope with the next round of temperature rise at any time;

[0171] S602: Let y=y+1, if y is less than or equal to Y, return to S601 for continuous execution, if y is greater than Y, end the current process.

[0172] It should be noted that the overall architecture of the alcohol hydrogen fuel cell and supercapacitor system of the silicon-based electrode and the silicon-based chip electric plate is shown in Figure 3 . Through the triple passive regulation closed loop of "deep power limiting-adjustable heat conduction bypass-PCM buffer", the central temperature peak can be quickly suppressed in the environment of 85-125℃ without active fan, and the long-term high-temperature reliable operation of the silicon-based electrode-silicon-based chip electric plate supercapacitor is ensured. Relying on the integrated structure of silicon-based electrode-copper-molybdenum column, the equivalent series resistance (ESR) is reduced to 0.20 mΩ (30% aging margin is reserved), and under the condition of peak discharge current (I RMS ) 200A, pulse duty cycle ≤10%, the continuous average power consumption (P avg ) is P avg =I RMS 2XESR≈8W; through the multi-stage conduction-natural convection heat dissipation path composed of a straight-through copper-molybdenum heat conduction column, an AIN-silicon composite substrate and a shell fin, the device-environment thermal resistance can be controlled at 1.8K / W, the theoretical temperature rise obtained by multiplying the continuous average power consumption and the device-environment thermal resistance is only 14K; therefore, the device center temperature of the silicon-based electrode and the silicon-based chip electric plate supercapacitor is not more than 99℃ at a temperature of 85℃, which still has a safety margin of 26K compared with the rated temperature of 125℃, and the device can be stably operated for a long time without a fan.

[0173] In the embodiment, a "supercapacitor-hydrogen fuel cell collaborative regulation structure" is constructed, and the supercapacitor not only serves as a high-frequency pulse power supply unit, but also bears the dynamic regulation support function for the hydrogen fuel cell stack. Specifically, based on the collected overall temperature state information (such as the supercapacitor module shell temperature, the hydrogen fuel cell stack electrode plate temperature, the heat dissipation frame temperature and the environmental reference temperature, etc.), a multi-source temperature feature vector is constructed, and the current thermal operating state of the hydrogen fuel cell stack is evaluated in real time.

[0174] When the system determines that the hydrogen fuel cell stack is in a non-optimal thermal interval, for example, the membrane electrode hydration degree is insufficient, and the electrode plate temperature is lower than the catalytic activity threshold, the controller supplies power to the hydrogen fuel cell stack with a specific dynamic compensation power curve, which comprehensively considers the power response speed, the current density fluctuation tolerance and the stack heat capacity characteristics, so that the hydrogen fuel cell stack can recover to the thermal dynamics suitable interval in a short time, and the compatibility of high conversion efficiency and low polarization loss is realized. Conversely, when the overall temperature state shows that the hydrogen fuel cell stack has entered the high temperature upper limit risk section, the system will appropriately reduce the power supply frequency and amplitude of the supercapacitor to the hydrogen fuel cell stack, and cooperatively regulate the heat conduction channel conduction rate and the PCM phase change material heat absorption rate, to avoid problems such as electrode plate dry cracking and electrolyte degradation due to heat accumulation of the stack.

[0175] Through the above-mentioned dynamic power supply control logic based on the overall temperature state, the present application realizes the adaptive thermal-electric collaborative regulation of the supercapacitor to the hydrogen fuel cell stack, ensures that the stack always operates in the "thermal comfort zone" with structural safety, high reaction efficiency and stable chemical load, and greatly improves the energy supply continuity and thermal control intelligence of the system under high temperature and high load coupling conditions.

[0176] Embodiment 2

[0177] Please refer to Figure 2 As shown in the figure, the present embodiment provides an alcohol hydrogen fuel cell and supercapacitor method of a silicon-based electrode and a silicon-based chip electric plate, which comprises:

[0178] Based on the temperature of N temperature sampling points of the alcohol hydrogen fuel cell and supercapacitor of the silicon-based electrode and the silicon-based chip electric plate per unit time, thermal state analysis and processing are performed, and a four-dimensional thermal perception model changing with time is constructed;

[0179] Feature extraction is performed based on the four-dimensional thermal perception model to obtain four-dimensional thermal perception feature data;

[0180] Based on the four-dimensional thermal perception feature data, high-temperature risk area anomaly diagnosis is performed to identify Y high-temperature abnormal areas;

[0181] The Y high-temperature abnormal areas of the alcohol hydrogen fuel cell and the supercapacitor of the silicon-based electrode and the silicon-based chip electric plate are adaptively controlled.

[0182] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0183] Finally: the above is only a preferred embodiment of the present application and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included within the protection scope of the present application.

Claims

1. Alcohol hydrogen fuel cell and supercapacitor system with silicon-based electrodes and silicon-based chip plates, characterized by: include: The first processing module performs thermal state analysis based on the temperatures of N temperature sampling points of the alcohol-hydrogen fuel cell and supercapacitor of the silicon-based electrode and silicon-based chip board per unit time, and constructs a four-dimensional thermal perception model that changes with time; The second processing module performs feature extraction based on the four-dimensional thermal perception model to obtain four-dimensional thermal perception feature data; The risk diagnosis module diagnoses abnormalities in high-temperature risk areas based on four-dimensional thermal sensing feature data and identifies Y abnormal high-temperature areas; The adaptive control module performs adaptive control on Y high-temperature abnormal areas of the alcohol-hydrogen fuel cell and supercapacitor of the silicon-based electrode and silicon-based chip board.

2. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 1 is characterized in that: The method for adaptively controlling Y abnormal high-temperature areas of an alcohol-hydrogen fuel cell and a supercapacitor of a silicon-based electrode and a silicon-based chip plate includes: S600: Let the initial value of y be 1, and the value range of y be 1 to Y; S601: Obtain a high temperature anomaly diagnostic score for the yth high temperature anomaly area; if the high temperature anomaly diagnostic score is greater than or equal to a preset high intensity regulation score threshold, instruct the power management unit to immediately perform deep power limiting, convert the high intensity regulation score threshold into a high intensity regulation ratio, proportionally reduce the duty cycle and current amplitude of the pulse current to their original nominal values ​​multiplied by the high intensity regulation ratio, and switch the substrate current borne by the silicon-based electrode and silicon-based chip plate supercapacitor to the hydrogen fuel cell stack according to 1-high intensity regulation ratio; the cooling circulating fluid of the hydrogen fuel cell stack is introduced into the composite thermal conductive jacket outside the copper-molybdenum thermal conductive column through a heat exchange conduit, forming a convection-phase change composite heat dissipation path with the phase change material between the silicon-based electrode and the chip plate; If the high temperature abnormality diagnosis score is greater than or equal to the preset medium intensity regulation score threshold and less than the preset high intensity regulation score threshold, the current power reduction level is maintained, the medium intensity regulation score threshold is converted to the medium intensity regulation ratio, and the heat conduction duty cycle of the heat conduction channel is adjusted to the medium intensity regulation ratio; If the high-temperature anomaly diagnosis score is less than the preset medium-intensity regulation score threshold, it is recorded as a stable score. A check is then made to determine whether the score remains stable for two consecutive regulation cycles. If so, all parameters are restored to their original nominal values, and the phase change material system enters a standby state, maintaining the melting-solidification thermal cycle to be ready for the next round of temperature rise. S602: Let y=y+1. If y is less than or equal to Y, return to S601 to continue execution. If y is greater than Y, end the current process.

3. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 1 is characterized in that: The identification methods of Y high temperature abnormal areas include: S500: Let the initial value of the traversal index q be 1, and the value range of q is 1 to Q; let the initial value of the counting variable Y be 0; Q is the number of high temperature risk areas; S501: Obtain a hotspot tag set corresponding to the qth high temperature risk area; obtain a temperature change rate set in the four-dimensional thermal perception feature data; extract the temperature change rate corresponding to each three-dimensional coordinate point in the hotspot tag set in the temperature change rate set according to the spatial coordinate matching rule, and construct a regional temperature change rate set for the qth high temperature risk area; obtain spatial distribution feature data and thermal feature data of the qth high temperature risk area from the four-dimensional thermal perception feature data; S502: Inputting the regional temperature change rate set, spatial distribution characteristic data, and thermal characteristic data of the qth high temperature risk area into a high temperature anomaly diagnosis model to obtain a corresponding high temperature anomaly diagnosis score; S503: If the high temperature anomaly diagnosis score is greater than or equal to the preset high temperature anomaly diagnosis score threshold, mark the qth high temperature risk area as a high temperature anomaly area, and obtain the Yth high temperature anomaly area, and set Y=Y+1; S504: Let q=q+1. If q is less than or equal to Q, return to S501 to continue execution. If q is greater than Q, Y high temperature abnormality areas are obtained and the current process ends.

4. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 1 is characterized in that: The method for acquiring the four-dimensional thermal sensing feature data includes: Calculate the temperature change rate per unit time for each 3D spatial coordinate in the 4D thermal perception model and construct a temperature change rate set to identify potential overheating trends. The 3D spatial coordinates in the 4D thermal perception model are not limited to the locations where sampling points are arranged, but all definable calculation areas within the 4D thermal perception model. Based on the temperature threshold judgment rule, the three-dimensional spatial coordinates of the temperature exceeding the preset temperature threshold within a unit time are extracted, hotspot labels are constructed, and the hotspot labels are divided into Q high temperature risk areas by combining the spatial clustering algorithm; Based on the spatial position changes of Q high-temperature risk areas in the time series, the spatial distribution characteristic data corresponding to the Q high-temperature risk areas are extracted; based on the temperature changes of the Q high-temperature risk areas in the time series, the thermal characteristic data corresponding to the Q high-temperature risk areas are extracted; The temperature change rate set, Q spatial distribution feature data and Q thermal feature data are constructed into four-dimensional thermal perception feature data.

5. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 4 is characterized in that: The methods for obtaining Q high temperature risk areas include: A high temperature recognition threshold is preset; within a unit time, each three-dimensional spatial coordinate in the four-dimensional thermal perception model is traversed, and the temperature sequence of the three-dimensional spatial coordinate at M time points is extracted. The proportion of the temperature sequence greater than the high temperature recognition threshold is counted and recorded as the over-threshold ratio. The three-dimensional spatial coordinate corresponding to the over-threshold ratio exceeding the preset over-threshold ratio threshold is marked as a hotspot label, and all hotspot labels are formed into a hotspot label set; Based on the hotspot tag set, a spatial clustering algorithm is used to group all hotspot tags to obtain Q hotspot tag clustering results. Each clustering result is defined as an independent high temperature risk area, forming Q spatially continuous high temperature risk areas.

6. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 4 is characterized in that: The method for obtaining the spatial distribution characteristic data corresponding to Q high temperature risk areas includes: S300: Let the initial value of the traversal index q be 1, and the value range of q be 1 to Q; S301: Obtain the hotspot tag sets corresponding to M time points in the qth high temperature risk area within a unit time, where M time points correspond to M hotspot tag sets; S302: Based on the M hotspot tag sets at M time points, calculate the spatial distribution diameter, regional center coordinates, and average radius distance corresponding to the qth high temperature risk area at M time points; calculate the corresponding point distribution density based on the spatial distribution diameter; the spatial distribution diameter includes the distribution diameter of the high temperature risk area in the horizontal, vertical, and vertical directions of the three-dimensional space; S303: constructing a set of regional center point coordinates from the coordinates of the regional center points at M time points within a unit time, wherein the set of regional center point coordinates is the hotspot migration path; calculating an average value of the center point displacement vector based on the coordinates of the regional center points at M time points within a unit time, wherein the average value of the center point displacement vector is the diffusion direction; S304: Constructing the spatial distribution characteristic data of the qth high temperature risk area by using the hotspot migration path, diffusion direction, spatial distribution diameter corresponding to M time points, coordinates of the regional center point, average radius distance and point distribution density; S305: Let q=q+1. If q is less than or equal to Q, return to S301 to continue execution. If q is greater than Q, obtain the spatial distribution characteristic data corresponding to Q high temperature risk areas and end the current process.

7. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 4 is characterized in that: The method for obtaining thermal characteristic data corresponding to Q high temperature risk areas includes: S400: Let the initial value of the traversal index q be 1, and the value range of q be 1 to Q; S401: Obtain the hotspot tag sets corresponding to M time points in the qth high temperature risk area within a unit time, where M time points correspond to M hotspot tag sets; S402: constructing a hotspot tag temperature set corresponding to the M time points based on the temperatures corresponding to the M hotspot tag sets at the M time points; S403: Calculate the temperature mean, temperature standard deviation, temperature kurtosis, and temperature skewness corresponding to the qth high temperature risk area at M time points based on the hotspot tag temperature set at M time points; S404: constructing the temperature mean, temperature standard deviation, temperature kurtosis and temperature skewness corresponding to the M time points as thermal characteristic data of the qth high temperature risk area; S405: Let q=q+1. If q is less than or equal to Q, return to S401 to continue execution. If q is greater than Q, obtain the thermal characteristic data corresponding to Q high-temperature risk areas and end the current process.

8. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 1 is characterized in that: The method for constructing the time-varying four-dimensional thermal perception model includes: The unit time is divided into M time points to form a time series to achieve time discrete modeling of the system thermal state; At each time point, the N spatial temperature sampling points set in the system collect the temperature at their respective locations to form a temperature sampling data set at that time point. An interpolation algorithm is used to perform spatial fitting on the N temperature sampling values ​​at that time point to construct a corresponding continuous three-dimensional temperature distribution function and form a corresponding continuous temperature field. The continuous temperature fields at M time points are serialized and combined to form a four-dimensional thermal perception model that changes with time.

9. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 3 is characterized in that: The training method of the high temperature anomaly diagnosis model includes: Pre-constructing a high temperature abnormality diagnosis data set, wherein the high temperature abnormality diagnosis data set includes YC group high temperature abnormality diagnosis data and high temperature abnormality diagnosis scores corresponding to the YC group high temperature abnormality diagnosis data, where YC is a positive integer; dividing the high temperature abnormality diagnosis data set into a training set and a validation set, wherein the training set is used for learning the parameters of the high temperature abnormality diagnosis model, and the validation set is used for real-time monitoring of the generalization performance and overfitting degree of the high temperature abnormality diagnosis model; A deep neural network with a multilayer perceptron as the core is used as the high temperature anomaly diagnosis model. The high temperature anomaly diagnosis data is standardized and vectorized and then input into the deep neural network. The deep neural network consists of an input layer, a hidden layer, and an output layer. Each hidden layer uses a nonlinear activation function to extract high-order features, and the output layer uses a softmax activation function to obtain the probability distribution corresponding to each high temperature anomaly diagnosis score. Finally, the high temperature anomaly diagnosis score corresponding to the maximum probability is taken as the prediction result of the high temperature anomaly diagnosis model. During the training process, the optimization goal is to minimize the cross entropy loss function, and a gradient descent optimization algorithm is used to update the network weights. An early stopping strategy is set: when the prediction accuracy on the validation set reaches or exceeds the preset threshold, the high temperature anomaly diagnosis model is judged to have converged and the training is terminated.

10. The alcohol-hydrogen fuel cell and supercapacitor system of silicon-based electrodes and silicon-based chip panels according to claim 4, characterized in that: The method for constructing the temperature change rate set includes: S201: The total number of three-dimensional space coordinate points defined in the four-dimensional thermal sensing model is recorded as R, and the initial value of the traversal index r is set to 1, and the value range of r is 1 to R; S202: Obtain the rth three-dimensional spatial coordinate in the four-dimensional thermal perception model, calculate the temperature change rate of the rth three-dimensional spatial coordinate over time based on the temperature sequence formed by the rth three-dimensional spatial coordinate at M time points within a unit time, obtain the temperature change rate corresponding to the rth three-dimensional spatial coordinate within a unit time, and construct a temperature change rate subset of the rth three-dimensional spatial coordinate; S203: Increment r by 1. If the current r is less than or equal to R, return to step S202 to continue processing the next three-dimensional space coordinate position; if r is greater than R, the traversal is completed and the current process ends.

Images