Fault diagnosis method of fuel cell uninterruptible power supply system and storage medium

Abstract

The invention discloses a fault diagnosis method of a fuel cell uninterruptible power supply system and a storage medium. The method comprises the following steps: acquiring historical fault data and real-time operation data of a fuel cell; determining a correlation coefficient threshold based on a pre-constructed hybrid model according to the historical fault data; calculating an actual correlation coefficient based on a pre-constructed hybrid model according to the real-time operation data; and determining whether the fuel cell uninterruptible power supply system breaks down or not according to the size relationship between the actual correlation coefficient and the correlation coefficient threshold value. Compared with a traditional fault diagnosis method, the technical scheme provided by the embodiment of the invention can more scientifically and efficiently determine the running state of the fuel cell UPS system, and solves the problem that the traditional fault diagnosis method has certain limitation on empirical judgment, so that the running reliability of the fuel cell UPS system is relatively poor; and the reliability and the safety of the operation of the fuel cell UPS system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of new energy technology, and in particular to a fault diagnosis method and storage medium for a fuel cell uninterruptible power supply system. Background Art

[0002] As a high-efficiency and environmentally friendly power supply alternative, the fuel cell uninterruptible power supply (UPS) system needs to be highly reliable and safe while providing power. With the development of fuel cell technology, system failures may involve many aspects, including electrochemical reactions, material fatigue, temperature changes, etc. Therefore, fault diagnosis is crucial to ensure stable system operation and timely repair of potential problems.

[0003] Traditional fault diagnosis methods often rely on experience and judgment, which has certain limitations and leads to poor reliability of fuel cell UPS system operation. Summary of the invention

[0004] The present invention provides a fault diagnosis method and storage medium for a fuel cell uninterruptible power supply system, so as to solve the problem that the conventional fault diagnosis method has certain limitations, resulting in poor reliability of the operation of the fuel cell UPS system.

[0005] According to one aspect of the present invention, a fault diagnosis method for a fuel cell uninterruptible power supply system is provided, comprising:

[0006] Obtain historical fault data and real-time operation data of fuel cells;

[0007] Determining a correlation coefficient threshold value based on the historical fault data and a pre-built hybrid model;

[0008] Calculating an actual correlation coefficient based on a pre-built hybrid model according to the real-time operating data;

[0009] Whether a failure occurs in the fuel cell uninterruptible power supply system is determined according to the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold.

[0010] Optionally, determining the correlation coefficient threshold value based on the historical fault data based on a pre-built hybrid model includes:

[0011] Calculating a training correlation coefficient based on the historical fault data and simulation data obtained based on a pre-built hybrid model under the same fault condition;

[0012] Setting the initial value of the correlation coefficient threshold based on a preset step size;

[0013] A correlation coefficient threshold is determined according to the correlation coefficient threshold initial value and the training correlation coefficient.

[0014] Optionally, determining the correlation coefficient threshold value according to the correlation coefficient threshold initial value and the training correlation coefficient includes:

[0015] Establishing a confusion matrix table according to the correlation coefficient threshold initial value and the training correlation coefficient;

[0016] Calculate the recall rate and the precision rate according to the confusion matrix table;

[0017] Calculate the F1 score according to the recall rate and the precision rate;

[0018] The maximum value of the F1 score is determined as the correlation coefficient threshold.

[0019] Optionally, calculating the actual correlation coefficient based on the real-time operation data based on a pre-built hybrid model includes:

[0020] The actual correlation coefficient is calculated according to a preset number of the real-time operation data and the real-time simulation data obtained based on the pre-built hybrid model according to the real-time operation data.

[0021] Optionally, determining whether a failure occurs in the fuel cell uninterruptible power supply system according to the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold value includes:

[0022] Comparing the actual correlation coefficient with the correlation coefficient threshold, and if the actual correlation coefficient is less than the correlation coefficient threshold, determining that the fuel cell uninterruptible power supply system fails;

[0023] If the actual correlation coefficient is greater than or equal to the correlation coefficient threshold, it is determined that the fuel cell uninterruptible power supply system operates normally.

[0024] Optionally, the calculating the actual correlation coefficient according to the preset number of real-time operation data and the real-time simulation data obtained based on the pre-built hybrid model according to the real-time operation data includes:

[0025] The actual correlation coefficient is calculated using the following formula:

[0026]

[0027] In the formula, is the actual correlation coefficient, is the covariance between the real-time simulation data and the real-time operation data output by the hybrid model, is the variance of the real-time simulation data output by the hybrid model, is the variance of real-time running data, Mmi is the real-time simulation data output by the hybrid model of the i-th sample, is the mean of the real-time simulation data output by the hybrid model, M ti is the real-time running data of the i-th sample, is the mean of the real-time running data of the i-th sample.

[0028] Optionally, after determining that the fuel cell uninterruptible power supply system fails, the method further includes:

[0029] The fuel cell uninterruptible power supply system sends a fault signal and obtains current fault information;

[0030] The correlation coefficient threshold is updated according to the current fault information.

[0031] Optionally, before acquiring the historical fault data and real-time operation data of the fuel cell, the method further includes:

[0032] Acquiring historical data of the fuel cell; the historical data including current, temperature, humidity, oxygen pressure, oxygen flow rate, hydrogen pressure and hydrogen flow rate;

[0033] According to the current based on the RC equivalent circuit model, a Kalman filter algorithm is used to establish a mechanism model of the fuel cell;

[0034] Establishing a neural network model of the fuel cell using a neural network algorithm according to the temperature, the humidity, the oxygen pressure, the oxygen flow rate, the hydrogen pressure and the hydrogen flow rate;

[0035] The mechanism model and the neural network model are mixed to obtain a hybrid model of the fuel cell.

[0036] Optionally, the mixing of the mechanism model and the neural network model to obtain a hybrid model of the fuel cell includes:

[0037] The hybrid model should satisfy the following formula:

[0038] M m =βM e +(1-β)M n 0≤β≤1;

[0039] Where M m is the output matrix of the hybrid model, β is the model selection weight, M e is the output matrix of the mechanism model, M n is the output matrix of the neural network model.

[0040] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the fault diagnosis method for a fuel cell uninterruptible power supply system described in any embodiment of the present invention when executed.

[0041] The technical solution provided by the embodiment of the present invention determines the correlation coefficient threshold value by using the pre-built hybrid model based on the historical fault data. The hybrid model can not only deeply analyze the working state of the system through the electro-chemical reaction and material transfer process inside the fuel cell, but also enable the system to better adapt to the changes in data and the complexity of the system operation during the actual operation process. The correlation coefficient threshold value can characterize the sensitivity and accuracy of the system fault diagnosis. According to the real-time operation data, based on the pre-built hybrid model, the actual correlation coefficient is calculated, and the actual correlation coefficient is used to judge the difference between the prediction result output by the hybrid model and the real-time operation data, so as to further determine whether the fuel cell uninterruptible power supply system fails in a timely and accurate manner according to the size relationship between the actual correlation coefficient and the correlation coefficient threshold value. Compared with the traditional fault diagnosis method, the technical solution provided by the embodiment of the present invention can determine the operating state of the fuel cell UPS system more scientifically and efficiently, solves the problem that the traditional fault diagnosis method relies on experience judgment and has certain limitations, resulting in poor reliability of the operation of the fuel cell UPS system, and effectively improves the reliability and safety of the operation of the fuel cell UPS system.

[0042] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A flowchart of a method for diagnosing a fault in a fuel cell uninterruptible power supply system provided by an embodiment of the present invention;

[0045] Figure 2 A flow chart of another method for diagnosing a fault in a fuel cell uninterruptible power supply system provided by an embodiment of the present invention;

[0046] Figure 3 A flowchart of another method for diagnosing a fault in a fuel cell uninterruptible power supply system provided by an embodiment of the present invention;

[0047] Figure 4 A schematic diagram of the structure of a fuel cell uninterruptible power supply system provided by an embodiment of the present invention;

[0048] Figure 5 A schematic diagram of the structure of a DC / DC boost converter in a fuel cell uninterruptible power supply system provided by an embodiment of the present invention;

[0049] Figure 6 A flowchart of a fault diagnosis method for a fuel cell uninterruptible power supply system provided by an embodiment of the present invention;

[0050] Figure 7 A schematic diagram of a hybrid model of a fault diagnosis method for a fuel cell uninterruptible power supply provided by an embodiment of the present invention;

[0051] Figure 8 The embodiment of the present invention provides Figure 7 A schematic structural diagram of an electronic device for a fuel cell uninterruptible power supply fault diagnosis method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0053] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0054] Figure 1 This is a flow chart of a method for diagnosing a fault in a fuel cell uninterruptible power supply system provided by an embodiment of the present invention. This embodiment can be applied to the evaluation of the operating status of a fuel cell uninterruptible power supply system. Figure 1 As shown, the fault diagnosis method includes:

[0055] S110, acquiring historical fault data and real-time operation data of the fuel cell.

[0056] Among them, historical fault data includes current and voltage under different fault conditions; real-time operation data includes current and voltage under the current operation state. Specifically, the current and voltage of the fuel cell under the current operation state are collected in real time through current sensors and voltage sensors; by calling the fuel cell system database, the pre-stored current and voltage of the fuel cell under different fault conditions can be obtained.

[0057] S120 . Determine a correlation coefficient threshold value based on a pre-built hybrid model according to historical fault data.

[0058] Among them, the hybrid model includes a mechanism model and a neural network model of the fuel cell. The mechanism model uses physical laws to describe the electro-chemical reaction and material transfer process inside the fuel cell. It can deeply reflect the physical process inside the fuel cell. It can also be called a white box model. The neural network model is a model that analyzes the operating data of the fuel cell and uses a neural network algorithm as a data-driven method. Specifically, the pre-built hybrid model is trained using historical fault data to obtain training results, and the correlation coefficient threshold is determined based on the training results. The purpose of determining the correlation coefficient threshold is to ensure the sensitivity and accuracy of system fault diagnosis.

[0059] S130 , calculating an actual correlation coefficient based on the pre-built hybrid model according to the real-time operation data.

[0060] Specifically, the purpose of calculating the actual correlation coefficient is to use the actual correlation coefficient to determine the difference between the prediction results output by the hybrid model and the real-time operation data; the higher the actual correlation coefficient, the closer the real-time operation data is to the prediction results output by the hybrid model; the lower the actual correlation coefficient, the more contrary the real-time operation data is to the prediction results output by the hybrid model.

[0061] S140: Determine whether a fuel cell uninterruptible power supply system fails based on the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold.

[0062] Specifically, the actual correlation coefficient is compared with the correlation coefficient threshold to obtain a comparison result, and whether the fuel cell uninterruptible power supply system fails is determined based on the comparison result. Exemplarily, when the actual correlation coefficient is less than the correlation coefficient threshold, it means that the real-time operating data is inconsistent with the predicted result output by the hybrid model, and it is determined that the fuel cell uninterruptible power supply system fails; when the actual correlation coefficient is greater than or equal to the correlation coefficient threshold, it means that the real-time operating data is consistent with the predicted result output by the hybrid model, and it is determined that the fuel cell uninterruptible power supply system is in normal operation.

[0063] The technical solution provided by the embodiment of the present invention determines the correlation coefficient threshold value by using the pre-built hybrid model based on the historical fault data. The hybrid model can not only deeply analyze the working state of the system through the electro-chemical reaction and material transfer process inside the fuel cell, but also enable the system to better adapt to the changes in data and the complexity of the system operation during the actual operation process. The correlation coefficient threshold value can characterize the sensitivity and accuracy of the system fault diagnosis. According to the real-time operation data, based on the pre-built hybrid model, the actual correlation coefficient is calculated, and the actual correlation coefficient is used to judge the difference between the prediction result output by the hybrid model and the real-time operation data, so as to further determine whether the fuel cell uninterruptible power supply system fails in a timely and accurate manner according to the size relationship between the actual correlation coefficient and the correlation coefficient threshold value. Compared with the traditional fault diagnosis method, the technical solution provided by the embodiment of the present invention can determine the operating state of the fuel cell UPS system more scientifically and efficiently, solves the problem that the traditional fault diagnosis method relies on experience judgment and has certain limitations, resulting in poor reliability of the operation of the fuel cell UPS system, and effectively improves the reliability and safety of the operation of the fuel cell UPS system.

[0064] Figure 2 This is a flow chart of another method for diagnosing a fault of a fuel cell uninterruptible power supply system provided by an embodiment of the present invention. The embodiment of the present invention further refines the above embodiment on the basis of the above embodiment. Figure 2 , the fault diagnosis method comprises:

[0065] S210: Acquire historical fault data and real-time operation data of the fuel cell.

[0066] S220, calculating a training correlation coefficient according to historical fault data and simulation data obtained based on a pre-built hybrid model under the same fault condition.

[0067] Specifically, the following formula is used to calculate the actual correlation coefficient:

[0068]

[0069] In the formula, is the training correlation coefficient, is the covariance of the simulation data and historical fault data output by the hybrid model, is the variance of the simulated data output by the hybrid model, is the variance of historical fault data, M mi is the simulated data output by the mixed model of the i-th sample, is the mean of the simulated data output by the hybrid model, M ti is the historical fault data of the i-th sample, It should be noted that the above formula is a general formula for calculating the training correlation coefficient and the actual correlation coefficient. When the calculation target is different, the meaning of each parameter in the above formula is also different.

[0070] S230: Setting an initial value of a correlation coefficient threshold based on a preset step size.

[0071] The preset step size can be preset according to the system calculation requirements. Specifically, since the correlation coefficient is a statistical indicator used to measure the strength of the linear relationship between two variables, its value range is [-1, 1]. Therefore, the value range of the correlation coefficient threshold is [0, 1]. Within the value range of [0, 1], a value within the value range is selected every preset step size and set as the initial value of the correlation coefficient threshold.

[0072] S240: Determine a correlation coefficient threshold value according to the initial value of the correlation coefficient threshold value and the training correlation coefficient.

[0073] Optionally, determining the correlation coefficient threshold according to the correlation coefficient threshold initial value and the training correlation coefficient includes:

[0074] A confusion matrix table is established based on the initial value of the correlation coefficient threshold and the training correlation coefficient.

[0075] Among them, the confusion matrix table is a specific table that is widely used in machine learning, pattern recognition, statistics and other fields to evaluate the performance of classification models.

[0076] Specifically, multiple correlation coefficient threshold initial values ​​are set within the value range of [0, 1] according to the preset step size, and the training correlation coefficient is all correlation coefficients under different fault conditions. Each training correlation coefficient is compared with the correlation coefficient threshold initial value. When the training correlation coefficient is less than the correlation coefficient threshold initial value, it means that there is a weak correlation between the historical fault data and the simulation data obtained based on the pre-built hybrid model under the same fault condition, and the system is determined to be faulty. When the training correlation coefficient is greater than or equal to the correlation coefficient threshold initial value, it means that there is a strong correlation between the historical fault data and the simulation data obtained based on the pre-built hybrid model under the same fault condition, and the system is determined to be normal. The comparison results are divided into four situations, including: predicted to be normal and actually normal, predicted to be normal and actually faulty, predicted to be faulty and actually normal, and predicted to be faulty and actually faulty. Based on this, a confusion matrix table is established, as shown in Table 1 below.

[0077] Table 1

[0078] The actual fault Actually normal Predicted failure TP FP Predicted to be normal FN TN

[0079] TP represents the number of samples that are actually faulty and predicted to be faulty; FN represents the number of samples that are actually faulty and predicted to be normal; FP represents the number of samples that are actually normal and predicted to be faulty; TN represents the number of samples that are actually normal and predicted to be normal.

[0080] Calculate the recall and precision based on the confusion matrix table.

[0081] Specifically, the recall rate is calculated using the following formula:

[0082]

[0083] In the formula, re is the recall rate, TP is the number of samples that are actually faulty and predicted to be faulty, and FN is the number of samples that are actually faulty and predicted to be normal.

[0084] The accuracy is calculated using the following formula:

[0085] In the formula, pr is the precision, and FP represents the number of samples that are actually normal and predicted to be faulty.

[0086] The F1 score is calculated based on the recall and precision.

[0087] Specifically, the F1 score is calculated using the following formula:

[0088] Where F1 represents the F1 score.

[0089] The maximum value of the F1 score is determined as the correlation coefficient threshold.

[0090] Specifically, based on the above process, multiple F1 scores can be calculated, and the maximum value can be selected as the correlation coefficient threshold. In practical applications, if the correlation coefficient threshold is set too low, the system fault diagnosis will not be sensitive enough, and if the threshold is set too high, it will lead to misdiagnosis. Therefore, the F1 score indicator is used to measure the appropriateness of the correlation coefficient threshold setting, and then determine the final correlation coefficient threshold.

[0091] S250: Calculate the actual correlation coefficient based on the pre-built hybrid model according to the real-time operation data.

[0092] S260: Determine whether a fuel cell uninterruptible power supply system fails based on the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold.

[0093] The technical solution provided in the embodiment of the present invention, in order to take into account the sensitivity and accuracy of fault diagnosis, abstracts the problem of how to obtain the optimal threshold into a two-classification problem, uses the confusion matrix table and the F1 score indicator to screen the set initial value of the correlation coefficient threshold, and then determines the optimal correlation coefficient threshold, thereby further improving the reliability of the fuel cell UPS system operation and the accuracy of fault diagnosis.

[0094] Figure 3 This is a flow chart of another method for diagnosing a fault of a fuel cell uninterruptible power supply system provided by an embodiment of the present invention. The embodiment of the present invention further refines the above embodiment on the basis of the above embodiment. Figure 3 , the fault diagnosis method comprises:

[0095] S310, obtaining historical fault data and real-time operation data of the fuel cell.

[0096] S320: Determine a correlation coefficient threshold value based on the historical fault data and a pre-built hybrid model.

[0097] S330 , calculating an actual correlation coefficient according to a preset number of real-time operation data and real-time simulation data obtained based on a pre-built hybrid model according to the real-time operation data.

[0098] Considering that it is impossible to collect all the data at one time in practical applications, an appropriate amount of data, that is, a preset number of data, needs to be used in the calculation. The preset number can be pre-set according to the system calculation requirements. The more the preset number, the greater the difficulty of the system calculation, and the fewer the preset number, the easier the system calculation.

[0099] Specifically, the following formula is used to calculate the actual correlation coefficient:

[0100]

[0101] In the formula, is the actual correlation coefficient, is the covariance between the real-time simulation data and the real-time operation data output by the hybrid model, is the variance of the real-time simulation data output by the hybrid model, is the variance of real-time running data, M mi is the real-time simulation data output by the hybrid model of the i-th sample, is the mean of the real-time simulation data output by the hybrid model, M ti is the real-time running data of the i-th sample, is the mean of real-time running data.

[0102] S340, determine whether the actual correlation coefficient is less than the correlation coefficient threshold, if so, execute S350; if not, execute S360.

[0103] S350: Determine that a fuel cell uninterruptible power supply system fails.

[0104] S360. Determine that the fuel cell uninterruptible power supply system operates normally.

[0105] S370. The fuel cell uninterruptible power supply system sends a fault signal and obtains current fault information.

[0106] Among them, the fault information includes the fault type, the fault current and voltage corresponding to the fault type. Specifically, when a fault occurs, the fuel cell uninterruptible power supply system sends a fault signal, obtains the fault current and voltage under the current fault condition through the current sensor and voltage sensor, and determines the current fault type based on the hybrid model according to the fault current and voltage data. In addition, after receiving the fault signal, the fault module needs to be isolated in time to prevent it from affecting the overall system. Figure 4 A schematic diagram of the structure of the fuel cell uninterruptible power supply system provided by the embodiment of the present invention is shown in FIG. Figure 4 The fuel cell uninterruptible power supply system includes multiple fuel cell stacks and DC / DC boost converters. One fuel cell stack is connected to one DC / DC boost converter and is connected in parallel with a DC bus. The output of the DC / DC boost converter is connected to a load through an inverter, and the AC power output is connected to a rectifier. Figure 5 A schematic diagram of the structure of a DC / DC boost converter in a fuel cell uninterruptible power supply system according to an embodiment of the present invention is shown in FIG. Figure 5 , switch S r It is in the closed state under normal conditions. When a module fails, the controller sends a signal and the switch S r Disconnect to isolate the faulty module to prevent it from affecting the entire system and minimize the damage to the entire system. Considering that the current and voltage spikes generated by hard switching are large, they may cause high electromagnetic interference and affect system performance. r It is a zero current switch. When the current in the fault state drops to zero, the switch is disconnected to improve electromagnetic compatibility.

[0107] S380. Update the correlation coefficient threshold according to the current fault information.

[0108] Specifically, the current fault information is input into the fault database, and S320 is re-executed according to the current fault information to update the correlation coefficient threshold.

[0109] The technical solution provided by the embodiment of the present invention can update the correlation coefficient threshold in real time after a fault occurs, which can better cope with the actual operation of the fuel cell uninterruptible power supply system and further effectively improve the accuracy of fault diagnosis. In addition, the system adopts a modular system structure, which is convenient for fault isolation after fault diagnosis, and effectively improves the safety and service life of the system.

[0110] Figure 6 This is a flow chart of another method for diagnosing a fault of a fuel cell uninterruptible power supply system provided by an embodiment of the present invention. The embodiment of the present invention further refines the above embodiment on the basis of the above embodiment. Figure 6 , the fault diagnosis method comprises:

[0111] S610: Obtain historical data of the fuel cell.

[0112] The historical data includes current, temperature, humidity, oxygen pressure, oxygen flow rate, hydrogen pressure and hydrogen flow rate. Specifically, by calling the fuel cell system database, the pre-stored historical data of the fuel cell under different working conditions can be obtained.

[0113] S620, establishing a fuel cell mechanism model based on the current-based RC equivalent circuit model and using a Kalman filter algorithm.

[0114] Specifically, the RC equivalent circuit model is adopted, that is, a voltage source V OCV , a fuel cell internal resistance R cell and a double layer capacitor C dl The series connection establishes the following relationship:

[0115]

[0116] The actual historical current I(t) is used as the input of the RC equivalent circuit model, and the Kalman filter algorithm and the model relationship are used to calculate the simulated voltage of the model. The simulated voltage is compared with the actual historical voltage to determine whether the simulated voltage fits the actual historical voltage within the allowable error range. If so, the output parameter V OCV , R cell and C dl If not, modify the parameter V OCV , R cell and C dl Until the simulation voltage is within the allowable error range to match the actual historical voltage, the mechanism model is established and the output is M e Matrix, including voltage and current:

[0117]

[0118] S630, establishing a neural network model of the fuel cell using a neural network algorithm according to temperature, humidity, oxygen pressure, oxygen flow rate, hydrogen pressure and hydrogen flow rate.

[0119] Specifically, the neural network algorithm is used to model the fuel cell, and a multi-layer perceptron structure is adopted, including an input layer, several hidden layers and an output layer. The hidden layer can perform nonlinear transformation on the input data of the input layer, mainly to complete feature learning and representation learning. The more hidden layers there are, the more complex features the model can learn, but this will also increase the computational complexity of the model. In specific application scenarios, the appropriate number of hidden layers can be selected as needed. The input data of the input layer are the temperature T, humidity H, and oxygen pressure of the fuel cell stack. Oxygen flow rate Hydrogen pressure and hydrogen flow rate That is, it has six input nodes, and the output is selected as the fuel cell stack voltage V n and current I n , the neural network model should satisfy the following loss function relationship:

[0120]

[0121] in, is the predicted output voltage and current of the neural network model for the i-th sample, is the actual voltage and current data of history, J(W,b) is the loss function, which is an indicator used to measure the difference between the predicted results of the neural network model and the actual results. The smaller the value of the loss function, the closer the predicted results of the model are to the actual results, and the better the performance of the model. Update the weight W and bias b through back propagation:

[0122]

[0123] Among them, α is the learning rate, is the partial derivative of the loss function with respect to the weight, is the partial derivative of the loss function with respect to the bias. Through continuous iteration, the neural network model is finally established, and the output is M n Matrix, including voltage and current:

[0124]

[0125] S640, mixing the mechanism model and the neural network model to obtain a hybrid model of the fuel cell.

[0126] The mixed model should satisfy the following formula:

[0127] M m =βM e +(1-β)Mn 0≤β≤1;

[0128] Where M m is the output matrix of the hybrid model, β is the model selection weight, M e is the output matrix of the mechanism model, M n is the output matrix of the neural network model. β can be determined according to the amount of data that can be collected by the system. When more operating data can be collected, the neural network model is selected, and the value of β is closer to 0. Otherwise, the mechanism model is more emphasized, and the value of β is closer to 1. At this point, the hybrid model has been constructed, such as Figure 7 As shown, Figure 7 A schematic diagram of a hybrid model of a fault diagnosis method for a fuel cell uninterruptible power supply provided by an embodiment of the present invention.

[0129] S650: Obtain historical fault data and real-time operation data of the fuel cell.

[0130] S660: Determine a correlation coefficient threshold value based on the historical fault data and a pre-built hybrid model.

[0131] S670: Calculate the actual correlation coefficient based on the pre-built hybrid model according to the real-time operation data.

[0132] S680: Determine whether a fuel cell uninterruptible power supply system fails based on the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold.

[0133] The technical solution provided by the embodiment of the present invention forms a hybrid model by establishing a mechanism model and a neural network model of a fuel cell, combining the output results of the two models to adapt to the actual situation. The advantage of the hybrid model is that it gives full play to the advantages of the two modeling methods. The mechanism model provides a deep understanding of the internal operating mechanism of the system, which helps to deeply analyze the working state of the system at the theoretical level. At the same time, the neural network model makes the system more adaptable, able to adapt to the changes in data and the complexity of system behavior in actual work, and effectively improves the robustness and reliability of the system.

[0134] Figure 8 The embodiment of the present invention provides Figure 7A schematic diagram of the structure of an electronic device for a method for diagnosing a fuel cell uninterruptible power supply fault provided by an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0135] like Figure 8 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0136] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0137] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a fault diagnosis method for a fuel cell uninterruptible power supply system.

[0138] In some embodiments, the fault diagnosis method for the fuel cell uninterruptible power supply system can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the fault diagnosis method for the fuel cell uninterruptible power supply system described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the fault diagnosis method for the fuel cell uninterruptible power supply system in any other appropriate manner (for example, by means of firmware).

[0139] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0140] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0141] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0142] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0143] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0144] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0145] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0146] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for diagnosing a fault in a fuel cell uninterruptible power supply system, characterized in that: include: Obtain historical fault data and real-time operation data of fuel cells; Determining a correlation coefficient threshold value based on the historical fault data and a pre-built hybrid model; Calculating an actual correlation coefficient based on a pre-built hybrid model according to the real-time operating data; Whether a failure occurs in the fuel cell uninterruptible power supply system is determined according to the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold.

2. The fault diagnosis method according to claim 1, characterized in that: Determining the correlation coefficient threshold according to the historical fault data based on a pre-built hybrid model includes: Calculating a training correlation coefficient based on the historical fault data and simulation data obtained based on a pre-built hybrid model under the same fault condition; Setting the initial value of the correlation coefficient threshold based on a preset step size; A correlation coefficient threshold is determined according to the correlation coefficient threshold initial value and the training correlation coefficient.

3. The fault diagnosis method according to claim 2, characterized in that: Determining the correlation coefficient threshold value according to the correlation coefficient threshold initial value and the training correlation coefficient includes: Establishing a confusion matrix table according to the correlation coefficient threshold initial value and the training correlation coefficient; Calculate the recall rate and the precision rate according to the confusion matrix table; Calculate the F1 score according to the recall rate and the precision rate; The maximum value of the F1 score is determined as the correlation coefficient threshold.

4. The fault diagnosis method according to claim 1, characterized in that: The calculating the actual correlation coefficient according to the real-time operation data based on the pre-built hybrid model includes: The actual correlation coefficient is calculated according to a preset number of the real-time operation data and the real-time simulation data obtained based on the pre-built hybrid model according to the real-time operation data.

5. The fault diagnosis method according to claim 1, characterized in that: Determining whether a failure occurs in the fuel cell uninterruptible power supply system according to the magnitude relationship between the actual correlation coefficient and the correlation coefficient threshold comprises: Comparing the actual correlation coefficient with the correlation coefficient threshold, and if the actual correlation coefficient is less than the correlation coefficient threshold, determining that the fuel cell uninterruptible power supply system fails; If the actual correlation coefficient is greater than or equal to the correlation coefficient threshold, it is determined that the fuel cell uninterruptible power supply system operates normally.

6. The fault diagnosis method according to claim 4, characterized in that: The calculating of the actual correlation coefficient according to the preset number of real-time operation data and the real-time simulation data obtained based on the pre-built hybrid model according to the real-time operation data comprises: The actual correlation coefficient is calculated using the following formula: In the formula, is the actual correlation coefficient, is the covariance between the real-time simulation data and the real-time operation data output by the hybrid model, is the variance of the real-time simulation data output by the hybrid model, is the variance of real-time running data, M mi is the real-time simulation data output by the hybrid model of the i-th sample, is the mean of the real-time simulation data output by the hybrid model, M ti is the real-time running data of the i-th sample, is the mean of the real-time running data of the i-th sample.

7. The fault diagnosis method according to claim 5, characterized in that: After determining that the fuel cell uninterruptible power supply system fails, the method further includes: The fuel cell uninterruptible power supply system sends a fault signal and obtains current fault information; The correlation coefficient threshold is updated according to the current fault information.

8. The fault diagnosis method according to claim 1, characterized in that: Before acquiring the historical fault data and real-time operation data of the fuel cell, the method further includes: Acquiring historical data of the fuel cell; the historical data including current, temperature, humidity, oxygen pressure, oxygen flow rate, hydrogen pressure and hydrogen flow rate; According to the current based on the RC equivalent circuit model, a Kalman filter algorithm is used to establish a mechanism model of the fuel cell; Establishing a neural network model of the fuel cell using a neural network algorithm according to the temperature, the humidity, the oxygen pressure, the oxygen flow rate, the hydrogen pressure and the hydrogen flow rate; The mechanism model and the neural network model are mixed to obtain a hybrid model of the fuel cell.

9. The fault diagnosis method according to claim 8, characterized in that: The hybrid model of the fuel cell obtained by mixing the mechanism model and the neural network model comprises: The hybrid model should satisfy the following formula: M m =βM e +(1-β)M n 0≤β≤1; Where M m is the output matrix of the hybrid model, β is the model selection weight, M e is the output matrix of the mechanism model, M n is the output matrix of the neural network model.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the fault diagnosis method for a fuel cell uninterruptible power supply system according to any one of claims 1 to 9 when executed.

Images