Hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning

Abstract

The invention discloses a hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning, and relates to the technical field of fault diagnosis. The system comprises a data acquisition unit, a data analysis and feature extraction unit and a fault prediction unit. The data acquisition unit is used for acquiring operation data of the hydrogen energy cogeneration unit in real time; the data analysis and feature extraction unit is used for performing electrochemical performance analysis based on a polarization curve according to the operation data and establishing a polarization model of the hydrogen energy cogeneration unit; calculating the electric efficiency and the thermal efficiency of the hydrogen energy cogeneration unit, and constructing a fault feature vector of the hydrogen energy cogeneration unit; and the fault prediction unit is used for judging the probability that various types of faults occur in the next time step. According to the method, dynamic changes in unit operation can be captured in a self-adaptive mode, potential faults can be warned in advance, and therefore the safety, stability and operation efficiency of the system are greatly improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of fault diagnosis, and in particular to a fault diagnosis and prediction system for a hydrogen energy cogeneration unit based on deep learning. Background Art

[0002] As a clean and efficient energy carrier, hydrogen energy has played an important role in the transformation of the global energy structure in recent years. Based on fuel cell technology, hydrogen cogeneration units can provide both electricity and heat energy, improve energy efficiency and reduce carbon emissions, and are widely used in distributed energy systems, industrial waste heat recovery, and household energy supply. As the core component of hydrogen cogeneration units, the operating stability and fault diagnosis capabilities of fuel cells directly affect the overall performance of the unit. However, the operation of fuel cells is affected by many factors, including electrochemical reactions, gas transport, catalyst activity, thermal management systems, and environmental changes. Therefore, how to achieve efficient and intelligent fuel cell fault diagnosis and prediction has become one of the key technologies to ensure the stability and life of hydrogen cogeneration units.

[0003] At present, the fault diagnosis and prediction methods for fuel cells mainly include methods based on physical models, methods based on signal processing, and methods based on machine learning. Among them, the methods based on physical models mainly rely on the electrochemical equations and kinetic characteristics of fuel cells, such as polarization curve models, equivalent circuit models, etc. For example, the Nernst equation and the Tafel equation can be used to describe the open circuit voltage and activation polarization loss of fuel cells, the Ohm's law can be used to characterize the ohmic polarization of fuel cells, and the Fick's law and the Butler-Volmer equation can be used to calculate the transport characteristics of the gas diffusion layer. This type of method can accurately describe the operating characteristics of fuel cells, but its main problem is that the calculation complexity is high and it is highly dependent on the model parameters. In practical applications, due to the aging of fuel cell materials, the accumulation of pollutants, and changes in environmental factors, the parameters in the physical model often need to be continuously adjusted, resulting in poor real-time and adaptability of fault diagnosis. Summary of the invention

[0004] The purpose of the present invention is to provide a hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning. In view of the various fault types, complex fault mechanisms, and difficulties in timely prediction in the operation and maintenance of hydrogen energy cogeneration units, an efficient, accurate and adaptive fault diagnosis and prediction system is proposed by integrating multi-dimensional feature construction, deep learning modeling and time series analysis. It overcomes the shortcomings of traditional technologies in feature selection, fault identification timeliness and model scalability, so that the unit can obtain reliable operation and maintenance guarantees at a lower cost and higher efficiency in practical applications. In addition, the system can also issue risk warnings before major failures occur in the unit, allowing the operation and maintenance strategy to change from passive response to active protection.

[0005] The technical solution of the present invention is achieved in this way: A deep learning-based fault diagnosis and prediction system for hydrogen energy cogeneration units, the system comprising: a data acquisition unit, a data analysis and feature extraction unit and a fault prediction unit; the data acquisition unit is used to collect the operating data of the hydrogen energy cogeneration unit in real time; the data analysis and feature extraction unit is used to perform electrochemical performance analysis based on polarization curves according to the operating data, and establish a polarization model of the hydrogen energy cogeneration unit; the electrical efficiency and thermal efficiency of the hydrogen energy cogeneration unit are calculated according to the polarization model, and then the deep learning model is used to construct a fault feature vector of the hydrogen energy cogeneration unit; the fault prediction unit is used to judge the probability of various types of faults occurring in the next time step according to the fault feature vector and the feature template vectors of each type of fault set.

[0006] Furthermore, the types of faults include: membrane dehydration fault, membrane flooding fault, catalyst poisoning fault, gas diffusion layer blockage fault and thermal management system fault.

[0007] Furthermore, the data acquisition unit collects the operating data of the hydrogen energy cogeneration unit in real time, including: the single cell voltage in V , load current in A , Stack temperature in K , Hydrogen pressure in Pa , oxygen pressure in Pa , Hydrogen flow rate in mol / s , oxygen chemical equivalent ratio , coolant inlet temperature in K , coolant outlet temperature in K , fuel cell operating temperature in K , water vapor partial pressure in Pa , Ohm resistance in Ω and the recovered heat in W ;in, For time.

[0008] Furthermore, the polarization model is expressed using the following formula: ; in, is the standard potential, in V; is the gas constant, which is 8.314 J / (mol·K); is the Faraday constant, which is 96485C / mol; is the transfer coefficient, ranging from 0.3 to 0.7; is the exchange current density, in A / cm²; is the first concentration polarization parameter, the unit is V, and the value range is 0.02 to 0.08; It is the second concentration polarization parameter, with the unit of cm² / A and the value range is 10 to 30.

[0009] Furthermore, the exchange current density is calculated using the following formula: ; in, is the influence index of hydrogen partial pressure on exchange current density, ranging from 0.7 to 1.2; is the influence index of oxygen partial pressure on exchange current density, ranging from 0.3 to 0.5; is the reference pressure, which is 1 atm or Pa; is the activation energy, in J / mol; The exchange current density prefactor reflects the exchange current density under ideal conditions without an activation energy barrier. The unit is A / cm² and the value range is arrive .

[0010] Furthermore, the transfer coefficient Calculated using the following formula: ; in, is the reorganization energy, which represents the energy adjustment of the catalyst and electrolyte interface during charge transfer. The unit is eV and the value range is 0.2 to 0.6, where 1eV=96.5kJ / mol; is the activation free energy, which is defined as the minimum Gibbs free energy required for the reaction to proceed, in eV, ranging from 0.1 to 0.4; is the electron transfer coefficient, which is related to the charge transfer rate and ranges from 0.5 to 0.7; is the surface coverage; surface coverage Calculated using the following formula: ; in, is the hydrogen adsorption equilibrium constant, ,in It is the adsorption free energy of hydrogen on the catalyst surface, measured in eV, and ranges from -0.3 to -0.1.

[0011] Furthermore, the electrical efficiency and thermal efficiency Use the following formulas for calculation: ; ; in, is the calorific value of hydrogen.

[0012] Furthermore, the process of constructing the fault feature vector of the hydrogen energy cogeneration unit using the deep learning model specifically includes: first constructing the following initial fault vector: ; in, is the reference voltage; is the reference electrical efficiency; is the reference thermal efficiency; then a single-layer neural network in deep learning is used to extract Processing is performed to obtain the fault feature vector .

[0013] Furthermore, the probability of each type of failure occurring at each time step is calculated using the following formula: ; in, express Failures occur all the time The probability of represents the time step; Represents the fault feature vector With failure Mahalanobis distance of feature template vector; Represents the fault feature vector With failure The square of the Euclidean distance of the feature template vector; is the set historical time window; is the integration variable; and All are integer subscript indices.

[0014] The hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning of the present invention has the following beneficial effects: At the fault feature construction level, the present invention introduces multiple indicators that are closely related to the electrochemical and thermal management processes of the unit, and normalizes and differentially calculates these indicators, thereby achieving a comprehensive characterization of the full-operation state of the unit. In this process, not only the dynamic characteristics of core processes such as charge transfer and gas diffusion in the fuel cell stack are considered, but also the regulating effect of the cooling system on temperature distribution is fully taken into account. This multi-factor fusion makes the fault features more representative and distinguishable.

[0015] The present invention introduces a feature extraction mechanism in a deep learning model, and nonlinearly maps the constructed multidimensional fault vector through a single-layer neural network, so that the differences in different fault modes are retained in a high-dimensional space. Compared with traditional feature engineering methods, the present invention pays more attention to the integration of data-driven and principle, relatively simplifies the process of setting artificial rules, and makes the system more scalable when facing complex or unknown fault types. Through such a feature extraction strategy, the system can use the multidimensional sensor data continuously acquired in actual operation to perform self-learning and self-updating, avoiding diagnostic errors caused by environmental changes or working condition switching.

[0016] In terms of fault identification and prediction, the present invention can accurately evaluate the similarity between each fault mode and the current operating state in a multi-dimensional feature space by setting feature templates of multiple known fault modes and combining the distance measurement method of feature vectors. Unlike the traditional pure rule comparison method, the present invention not only considers the matching degree between the fault characteristics and the template at the current moment, but also introduces a correction term for time evolution. This approach can effectively deal with the progressive characteristics of fuel cell failures. Whether it is catalyst deactivation, abnormal membrane state, gas diffusion layer blockage, or potential failures in the thermal management link, early signals can be captured through continuous monitoring and trend analysis of time series. While taking into account historical data, the present invention takes into account short-term fluctuations and long-term evolution, providing a more reliable basis for unit fault prediction, thereby helping operation and maintenance personnel to take intervention measures before large-scale failures break out. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of the system structure of a deep learning-based hydrogen energy cogeneration unit fault diagnosis and prediction system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present disclosure clearer and more understandable, the present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present disclosure and are not used to limit the present disclosure.

[0019] Example 1, reference Figure 1 : A hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning, the system comprising: a data acquisition unit, a data analysis and feature extraction unit and a fault prediction unit; the data acquisition unit is used to collect the operating data of the hydrogen energy cogeneration unit in real time; the data analysis and feature extraction unit is used to perform electrochemical performance analysis based on polarization curves according to the operating data, and establish a polarization model of the hydrogen energy cogeneration unit; according to the polarization model, the electrical efficiency and thermal efficiency of the hydrogen energy cogeneration unit are calculated, and then the deep learning model is used to construct the fault feature vector of the hydrogen energy cogeneration unit; the fault prediction unit is used to judge the probability of various types of faults occurring in the next time step according to the fault feature vector and the feature template vectors of each type of fault set.

[0020] Specifically, during the operation of the hydrogen energy cogeneration unit, the electrochemical reaction process of the fuel cell stack is strongly affected by the supply status of hydrogen and oxygen. The temperature distribution inside the battery stack, gas diffusion, catalyst activity, membrane hydration state and load changes will play an important role in the final output voltage, power and the thermoelectric conversion efficiency of the overall system. Therefore, the data acquisition unit must be able to obtain key parameters such as single cell voltage, load current, fuel cell operating temperature, hydrogen and oxygen pressure, hydrogen flow, water vapor partial pressure, ohmic resistance, etc. with high precision to build a polarization model of the fuel cell, and further use it to extract the health status characteristics of the unit. The real-time and accuracy of these parameters are crucial for the subsequent deep learning model, because the fault feature vectors relied on for training the deep learning model are extracted from these basic data, and the acquisition deviation of any variable will lead to misleading model training, thereby reducing the prediction accuracy and even causing misdiagnosis. Therefore, in the design of the data acquisition unit, it is necessary to use high-precision sensors and improve the noise resistance and reliability of the data through multi-sensor data fusion technology. In order to ensure the integrity and consistency of the data, the data acquisition unit adopts a high-frequency data sampling strategy to capture the dynamic changes in the system in a short period of time during operation. Since the failure of the hydrogen energy cogeneration unit is often caused by the accumulation of small abnormalities, such as membrane dehydration may start from the slow increase of the temperature of the battery stack, and catalyst poisoning may be manifested as a slight decrease in oxygen utilization, if the data acquisition frequency is too low, these early abnormal signals may not be captured, causing the system to issue an alarm only when a real failure occurs, missing the best time for intervention. Therefore, the data acquisition unit of this system can synchronously measure all key variables with a sampling period of milliseconds, and at the same time build a time series in combination with historical data, so that the subsequent deep learning model can learn the trend and fault development law in the time dimension. At the same time, in order to ensure the stability of the data acquisition process, the system also introduces an outlier detection mechanism to filter the mutation data to avoid erroneous input caused by accidental sensor noise or external interference, which affects the accuracy of fault diagnosis. In terms of data transmission, the data acquisition unit and the data analysis unit are connected through a high-speed communication interface to ensure that the data can be transmitted to the subsequent modeling and calculation modules without loss after acquisition. Since hydrogen cogeneration units usually have a distributed architecture, multiple fuel cell stacks, gas supply systems, cooling systems and other subsystems need to work together, so the sensors of the data acquisition unit are distributed in multiple different physical locations, and the data is aggregated through wireless or wired networks. During the data synchronization process, in order to avoid data misalignment caused by transmission delays, the system uses timing alignment technology to ensure that the data of all variables corresponds to the same operating state at the same timestamp, thereby ensuring that subsequent steps such as polarization curve calculation, electrical efficiency and thermal efficiency calculation, and fault feature extraction can be performed based on high-quality input.

[0021] In the system architecture of the present invention, the first key task of the data analysis and feature extraction unit is to establish a polarization model of the fuel cell. The polarization model is a mathematical expression that describes the relationship between the voltage and load changes of the fuel cell. It comprehensively considers multiple factors such as electrochemical thermodynamics, ohmic losses inside the battery, kinetic characteristics of charge transfer, and gas diffusion, and can reflect the performance changes of the fuel cell under different operating conditions. By analyzing the unit operation data, extracting variables such as single cell voltage, load current, fuel cell temperature, gas pressure, and combining the key parameters such as commutation potential, activation polarization, ohmic polarization, and concentration polarization calculated by electrochemical principles, the unit can construct a complete polarization curve model to provide data support for subsequent feature extraction. The change of the polarization curve can intuitively reflect the health status of the fuel cell. For example, when the catalyst is poisoned, the activation polarization loss will increase significantly, resulting in a drop in the battery voltage, and when the membrane is dehydrated or flooded, the ohmic polarization or concentration polarization loss will fluctuate abnormally. Therefore, establishing an accurate polarization model can not only quantify the electrochemical performance of the system, but also provide a theoretical basis for the construction of fault characteristics. After building the polarization model, the second task of the data analysis and feature extraction unit is to calculate the electrical efficiency and thermal efficiency of the hydrogen cogeneration unit. Since the energy conversion process of hydrogen fuel cells involves the coupling of electrical energy and thermal energy, it is difficult to fully characterize the operating status of the system by relying solely on a single voltage or power variable. This unit calculates the electrical efficiency and thermal efficiency of the unit by combining real-time measured fuel consumption, recovered heat, output power and other data to obtain the overall energy utilization of the system. The electrical efficiency characterizes the ability to convert hydrogen chemical energy into electrical energy, while the thermal efficiency describes the performance of the waste heat recovery system. These variables are not only used to evaluate the energy utilization level of the unit, but are also important features for fault diagnosis. For example, when the electrical efficiency drops abnormally, it may mean catalyst poisoning or gas diffusion layer blockage, while abnormal fluctuations in thermal efficiency may indicate a fault in the cooling system or unstable fuel supply. Therefore, through the calculation of electrical efficiency and thermal efficiency, this unit can further explore the potential fault factors that affect the performance of the unit and provide fault feature input for subsequent deep learning models. After completing the calculation and modeling of basic data, the core task of the data analysis and feature extraction unit is to construct a feature vector for fault diagnosis. The key to this process is how to extract the most representative features from complex, multivariate time series data so that the deep learning model can accurately identify the fault mode. To this end, this unit constructs a fault feature vector containing multiple dimensions based on the polarization model, electrical efficiency and thermal efficiency calculation results, including standardized voltage, efficiency, temperature difference, gas supply status, and voltage-current curve derivative information. These variables not only cover the basic parameters of unit operation, but also include key indicators reflecting the health status of the system.For example, the first-order derivative of voltage to current can be used to characterize the dynamic response characteristics of the battery, while the change of the second-order derivative can reveal the stability of the system under different load conditions. In addition, considering that the operating state of the hydrogen energy cogeneration unit is a time-related process, the data at a single time point cannot fully describe the health trend of the system. Therefore, this unit uses a sliding window method to extract features from historical data to construct a more predictive fault feature vector. By combining time series analysis, this unit can effectively capture the long-term change trend of the system, thereby providing more timely input data for the deep learning model.

[0022] The working principle of the fault prediction unit is based on time-correlation modeling. Its input is the fault feature vector constructed by the data analysis and feature extraction unit, and the output is the probability of occurrence of different fault types. Since the faults of hydrogen energy cogeneration units have complex nonlinear characteristics, and different types of faults often have cross-influences, such as catalyst poisoning may lead to reduced battery activity, which in turn causes blockage of the gas diffusion layer, and abnormal cooling system may cause imbalance in membrane hydration state, which in turn affects the polarization characteristics of the battery. Therefore, it is difficult to accurately predict faults by relying only on a single variable or data at a single moment. The present invention utilizes the pattern recognition ability of the deep learning model, and by training a large amount of operating data, enables the system to automatically learn the evolution laws of different fault modes, and combines the historical trends of time series to predict possible future faults. In the specific implementation process, the fault prediction unit uses a neural network to perform nonlinear mapping on the fault feature vector at the current moment, calculates its matching degree with various fault templates, and further combines the feature change trend within the historical time window to evaluate the probability of occurrence of different faults. This process not only considers the similarity between the current feature vector and the known fault mode, but also introduces the Mahalanobis distance metric to measure the rate of change of the fault feature vector in the time dimension, so as to achieve more refined trend analysis. For example, in the early stage of membrane dehydration failure, the ohmic resistance of the fuel cell may show a slow upward trend, but the voltage drop is small at this time. If it only relies on static data analysis, the system may not be able to determine that the fault has occurred. However, through the time correlation modeling of the fault prediction unit, it can be determined that the risk of membrane dehydration is increasing by identifying the trend of resistance change and its historical evolution pattern, and give early warning, so that the operation and maintenance personnel can take intervention measures before the failure occurs.

[0023] Embodiment 2: The types of faults include: membrane dehydration fault, membrane flooding fault, catalyst poisoning fault, gas diffusion layer blockage fault and thermal management system fault.

[0024] Specifically, membrane dehydration failure is one of the common problems in the operation of fuel cells, mainly occurring under low humidity, high temperature or high load conditions. The proton exchange membrane of the fuel cell relies on the hydration state to maintain good proton conductivity. If the membrane is dehydrated, the proton transfer rate will drop significantly, resulting in an increase in ohmic polarization loss, which in turn causes a drop in output voltage. In the process of constructing the fault feature vector, the system of the present invention pays special attention to parameters such as the ohmic resistance, operating temperature, hydrogen flow rate and water vapor partial pressure of the fuel cell. The system learns the changing trends of these variables through a deep learning model, so that the system can identify early signs of membrane dehydration and issue an alarm before the failure occurs. In contrast to membrane dehydration, membrane flooding failure is caused by an imbalance in water management inside the fuel cell. Excessive water absorption by the membrane or accumulation of liquid water on the cathode side will lead to obstructed gas diffusion, thereby affecting the oxygen supply efficiency and ultimately reducing the power output of the battery. Membrane flooding is usually manifested as large voltage fluctuations under low load conditions and reduced oxygen excess coefficients. In the feature extraction process, the present invention focuses on analyzing the changes in the oxygen pressure, coolant temperature, and polarization curve of the fuel cell, and combines the time series analysis method to establish the characteristic mode of membrane flooding failure, so that the system can automatically detect the occurrence of membrane flooding and predict its development trend based on historical data. Catalyst poisoning failure is one of the important factors affecting the life of the fuel cell. It is mainly caused by the adsorption of impurities (such as carbon monoxide, sulfide, etc.) in the fuel gas on the catalyst surface, reducing its activity, thereby causing the electrode reaction kinetics to deteriorate. The typical characteristics of catalyst poisoning are reduced exchange current density, increased activation polarization loss, and slow decay of voltage under constant load. The present invention constructs the fault characteristics of catalyst poisoning by monitoring the exchange current density, load current, hydrogen and oxygen pressure, and activation energy related parameters of the fuel cell, and uses a deep learning model to classify and predict it, thereby providing intervention measures before the catalyst completely fails and extending the service life of the system. Gas diffusion layer blockage failure usually occurs when the fuel cell has been running for a long time or the ambient humidity is high. Due to the accumulation of impurity particles, carbides or excessive water in the gas diffusion layer, the supply of reaction gas is limited, which in turn affects the uniformity of the electrode reaction. This fault usually manifests as an abnormally steep drop in the polarization curve of the fuel cell in the high load area, abnormal fluctuations in the oxygen chemical equivalence ratio, and nonlinear changes in the derivative of voltage to current. The present invention constructs a fault feature vector containing polarization curve derivative information, and combines the pattern recognition capability of the deep learning model to enable the system to accurately identify gas diffusion layer blockage and predict its development trend, thereby providing maintenance personnel with accurate diagnostic information. Thermal management system failure is one of the key factors affecting the stable operation of hydrogen cogeneration units, mainly manifested as cooling system failure, reduced heat exchange efficiency or abnormal temperature control. When the thermal management system fails, the operating temperature of the fuel cell may rise rapidly or fluctuate excessively, thereby affecting the membrane hydration state, catalyst activity and electrochemical reaction rate.The fault characteristic vector of the present invention specifically introduces key parameters such as the coolant inlet and outlet temperature difference, the fuel cell temperature gradient, and the recovered heat, and combines it with time series trend analysis, so that the system can detect abnormal conditions of the thermal management system and predict its possible impact in advance, thereby ensuring that the unit operates within the optimal temperature range.

[0025] Embodiment 3: Data acquisition unit, real-time acquisition of operating data of hydrogen energy cogeneration unit includes: single cell voltage in V , load current in A , Stack temperature in K , Hydrogen pressure in Pa , oxygen pressure in Pa , Hydrogen flow rate in mol / s , oxygen chemical equivalent ratio , coolant inlet temperature in K , coolant outlet temperature in K , fuel cell operating temperature in K , water vapor partial pressure in Pa , Ohm resistance in Ω and the recovered heat in W ;in, For time.

[0026] Specifically, the voltage of a single cell is the most direct indicator of the health status of a fuel cell. It reflects the driving force of the electrochemical reaction and the influence of polarization loss. Under normal operating conditions, the voltage of the fuel cell shows a certain regularity as the load changes, but if faults such as decreased catalyst activity, membrane dehydration or diffusion layer blockage occur, the voltage will fluctuate abnormally. Therefore, the system needs to accurately collect the voltage of the single cell and combine it with the time series analysis method to determine whether its change trend conforms to normal operating conditions. At the same time, the load current, as the output parameter of the battery, determines the rate of the electrochemical reaction and directly affects the change of the polarization curve. Therefore, the data acquisition unit must synchronously record the load current to ensure that the calculation of the polarization model can accurately reflect the operating status of the system. The stack temperature is one of the core parameters of the internal thermal management of the fuel cell. It directly affects the electrochemical reaction rate, catalyst performance and the hydration state of the proton exchange membrane. If the temperature is too high, it may cause membrane dehydration and accelerate material aging, while if the temperature is too low, it may cause a decrease in catalyst activity and a decrease in fuel utilization. Therefore, this system collects the stack temperature in real time to ensure that the thermal management system can maintain the best operating conditions. In addition, the supply status of hydrogen and oxygen is also an important factor affecting the performance of the fuel cell. Hydrogen pressure and oxygen pressure determine the gas diffusion rate of the electrode reaction and are closely related to the polarization loss of the fuel cell. When the gas supply is insufficient or the pressure is abnormal, the polarization curve may change significantly, affecting the stability of the system. Therefore, the data acquisition unit must monitor the hydrogen and oxygen pressures in real time and analyze their correlation with other variables in combination with the deep learning model to determine whether there is a potential gas supply failure. Hydrogen flow rate is one of the core input parameters of the fuel cell, which determines the hydrogen supply rate of the electrochemical reaction and affects the overall efficiency of the fuel cell.

[0027] Under normal operating conditions, the hydrogen flow rate should maintain a stable corresponding relationship with the load current, but the flow rate may fluctuate when the gas diffusion layer is blocked or the hydrogen supply is abnormal. Therefore, this system collects the hydrogen flow rate in real time, and calculates the fuel utilization rate in combination with the oxygen chemical equivalent ratio variable to ensure that the fuel cell's gas supply system is in the best state. In addition, the thermal management state of the fuel cell directly affects the stability of the system, and the inlet and outlet temperatures of the coolant can reflect the efficiency of heat transfer and the health of the heat dissipation system. When the coolant temperature is abnormal, it may mean that the thermal management system is blocked, the coolant flow is insufficient, or the performance of the heat exchanger is reduced. Therefore, the data acquisition unit must accurately monitor the inlet and outlet temperatures of the coolant to ensure that the thermal management system of the unit operates normally. Water vapor partial pressure is one of the important parameters affecting the hydration state of the proton exchange membrane, which determines the conductivity of the membrane and the ohmic polarization loss of the fuel cell. When the water vapor partial pressure is too low, membrane dehydration may occur, resulting in an increase in the resistance of the battery, thereby reducing the output voltage; when the water vapor partial pressure is too high, it may cause membrane flooding failure, resulting in gas diffusion obstruction. Therefore, this system collects water vapor partial pressure in real time, and combines the stack temperature and coolant temperature to calculate the hydration state of the membrane and predict its changing trend. Ohmic resistance is a parameter that characterizes the internal resistance of the fuel cell. It is not only affected by the hydration state of the membrane, but also related to the catalyst layer, electrode structure and current collector material. When the ohmic resistance increases abnormally, it usually means membrane damage, material aging or internal contamination of the battery. Therefore, this system monitors the ohmic resistance in real time through the data acquisition unit, and calculates its impact on the output power in combination with the polarization model to determine whether the system is in normal operation. Recovered heat is a key parameter to measure the efficiency of waste heat utilization in fuel cells. It reflects whether the thermal management system of the unit is operating normally. When the recovered heat is lower than the normal level, it may mean that the performance of the heat exchanger is reduced, the cooling system is blocked or the operating parameters are not matched. Therefore, the data acquisition unit needs to measure the recovered heat in real time and optimize the energy management strategy of the unit in combination with the calculation model of electrical efficiency and thermal efficiency.

[0028] Embodiment 4: The polarization model is expressed using the following formula: ; in, is the standard potential, in V; is the gas constant, which is 8.314 J / (mol·K); is the Faraday constant, which is 96485C / mol; is the transfer coefficient, ranging from 0.3 to 0.7; is the exchange current density, in A / cm²; is the first concentration polarization parameter, the unit is V, and the value range is 0.02 to 0.08; It is the second concentration polarization parameter, with the unit of cm² / A and the value range is 10 to 30.

[0029] Specifically, the standard potential Represents the theoretical open circuit voltage of the fuel cell under standard conditions, which determines the maximum voltage output of the fuel cell. When the fuel and oxidant exist in standard conditions and there is no internal loss, the voltage of the fuel cell can reach However, in actual operation, due to the existence of various losses, the operating voltage of the fuel cell is always lower than the theoretical value. Therefore, the subsequent losses are used to correct the actual voltage so that the model can accurately describe the actual working conditions. The first loss of the polarization model comes from the influence of gas pressure. This term is an electrochemical equilibrium potential correction term derived from the Nernst equation, which describes the change in the output potential of the fuel cell under different gas pressure conditions. is the gas constant, represents the temperature of the fuel cell, is the Faraday constant. These parameters together determine the effect of gas pressure on electric potential. and Represent the partial pressures of hydrogen and oxygen respectively, while the molecular part is the partial pressure of water vapor. Based on the mathematical form of this term, it can be concluded that the higher the partial pressures of hydrogen and oxygen, the higher the output voltage of the fuel cell, and an increase in the partial pressure of water vapor will reduce the cell voltage. This is because the partial pressures of hydrogen and oxygen determine the concentration gradient of the electrode reaction, and an increase in water vapor may affect the reaction kinetics. If the fuel supply system fails, such as a drop in hydrogen pressure or restricted oxygen diffusion, the value of this term will change abnormally, causing the open circuit potential of the fuel cell to drop. The second term stands for activation polarization loss, which describes the kinetic loss of the fuel cell in the low current region. The activation polarization is mainly determined by the activity of the catalyst and the electrode reaction rate, where is the transfer coefficient, which reflects the kinetic characteristics of the electrode reaction. is the exchange current density, which represents the equilibrium current density of the forward and reverse reactions on the electrode surface. When the catalyst activity decreases or gas transport is limited, the exchange current density will decrease, resulting in increased activation polarization losses, which in turn reduces the output voltage of the fuel cell. The logarithmic form of this term shows that as the load current As the activation polarization loss increases, it gradually increases, especially in the low current region, where this loss dominates. If the fuel cell catalyst is poisoned, such as CO molecules adsorbed on the catalyst surface, the electrode reaction slows down. will be significantly reduced, causing the value of this item to increase, thereby causing the voltage of the fuel cell to drop.

[0030] Item 3 is the Ohmic polarization loss, which describes the resistive loss inside the fuel cell. It is mainly determined by factors such as the ion conduction resistance of the proton exchange membrane, the resistance of the electrode material, and the contact resistance of the current collector. The mathematical form of this term is a linear relationship, indicating that when the load current As increases, the ohmic losses will increase proportionally. If the hydration state of the proton exchange membrane decreases, for example due to membrane dehydration, resulting in a decrease in the proton transport capacity, then will increase, thereby increasing the ohmic loss and causing the battery voltage to drop. Therefore, the change of this item can be used as a key indicator of the membrane hydration state. The system of the present invention can identify membrane dehydration in advance and provide early warning information for maintenance by continuously monitoring the change of this item and combining it with a deep learning prediction model. is the concentration polarization loss, which mainly describes the voltage drop caused by the limited gas transport. When the fuel cell is running under high load, the gas diffusion layer may not be able to provide enough hydrogen and oxygen in time, resulting in an increase in the gas concentration gradient, which reduces the electrode reaction rate and causes the output voltage of the fuel cell to drop sharply. The exponential function form of this term shows that when the load current When it increases to a certain extent, the concentration polarization loss will show a nonlinear rapid growth, resulting in a sharp decline in fuel cell performance. and The first concentration polarization parameter and the second concentration polarization parameter respectively determine the severity of concentration polarization. If the gas diffusion layer of the fuel cell is clogged, for example, due to long-term operation, pollutants are deposited on the electrode surface, which hinders gas transport, the value of this item will increase significantly, resulting in a decrease in the voltage of the fuel cell.

[0031] Example 5: The exchange current density is calculated using the following formula: ; in, is the influence index of hydrogen partial pressure on exchange current density, ranging from 0.7 to 1.2; is the influence index of oxygen partial pressure on exchange current density, ranging from 0.3 to 0.5; is the reference pressure, which is 1 atm or Pa; is the activation energy, in J / mol; The exchange current density prefactor reflects the exchange current density under ideal conditions without an activation energy barrier. The unit is A / cm² and the value range is arrive .

[0032] Specifically, the first It is the pre-factor of the exchange current density, which reflects the theoretical maximum exchange current density of the fuel cell under ideal conditions, that is, without an activation energy barrier. This parameter mainly depends on the intrinsic activity of the catalyst, the properties of the electrode material, and the catalyst loading. In practical applications, catalyst aging, contamination, or changes in the catalyst layer structure can lead to Therefore, in the diagnostic system of the present invention, the deep learning model is used to monitor the The changing trend of degradation patterns, thus identifying early signs of catalyst deactivation or poisoning. It reflects the effect of temperature on the electrode reaction rate. This term is derived from the Arrhenius equation, which describes the exponential relationship between the reaction rate and temperature. is the activation energy, which represents the energy barrier that the electrode reaction needs to overcome. is the gas constant, is the operating temperature of the fuel cell. When the temperature of the fuel cell increases, the value of this exponential term increases, indicating that the electrode reaction activity increases and the exchange current density increases; on the contrary, when the temperature decreases, the reaction rate decreases and the exchange current density decreases, resulting in a decrease in the output power of the fuel cell. Therefore, in the system of the present invention, the change of temperature-related terms can be used to monitor the thermal management state of the fuel cell. For example, when the cooling system fails, resulting in When the index rises abnormally, the electrochemical reaction rate of the fuel cell will temporarily increase, but long-term high temperature will accelerate the degradation of the catalyst and proton exchange membrane, eventually leading to a decline in system performance. By analyzing the changing trend of this index over time, the deep learning model can identify thermal management failures in advance and predict possible long-term impacts, thereby improving the safety and reliability of the system.

[0033] Item 3 reflects the effect of hydrogen partial pressure on the exchange current density, where is the hydrogen partial pressure on the cathode side of the fuel cell, is the reference pressure, is the influence index of hydrogen on the exchange current density, usually between 0.7 and 1.2. The physical meaning of this term is that the higher the hydrogen concentration, the faster the electrode reaction rate, and therefore the greater the exchange current density. However, when the hydrogen supply is insufficient, for example due to an abnormality in the fuel supply system, The value of this item decreases, thereby reducing the exchange current density, resulting in an increase in activation polarization loss and a decrease in the voltage of the fuel cell. The system of the present invention can monitor the hydrogen partial pressure in real time and analyze the changing trend of this item in combination with deep learning, identify the abnormality of the fuel supply system, and predict the impact of fuel shortage on the performance of the fuel cell. describes the effect of oxygen partial pressure on the exchange current density, where is the oxygen partial pressure on the cathode side of the fuel cell, is the effect index of oxygen on the exchange current density, usually between 0.3 and 0.5. Similar to hydrogen, the higher the oxygen concentration, the faster the electrode reaction rate of the fuel cell, so the exchange current density increases; conversely, if the oxygen supply is insufficient, for example due to a failure in the air supply system or a blockage in the gas diffusion layer, The value of this item decreases, resulting in a decrease in the exchange current density, which ultimately affects the electrochemical performance of the fuel cell. The system of the present invention can monitor the oxygen partial pressure in real time and analyze the dynamic characteristics of the oxygen supply system in combination with a deep learning model to identify gas transport problems and predict the possibility of failure. The calculation formula covers key factors such as catalyst activity, temperature effect, hydrogen partial pressure and oxygen partial pressure, each of which affects the overall performance of the fuel cell. The present invention uses deep learning technology to monitor the various parameters of the formula in real time and analyze their changing trends to accurately identify the health status of the fuel cell. For example, when the catalyst ages, The value of will gradually decrease, resulting in a decrease in the exchange current density; when the fuel supply is insufficient, decreases, causing the activation polarization loss to increase; when the cooling system fails, It may increase abnormally, increasing the exchange current density in the short term, but may accelerate material degradation in the long term. After being processed by the deep learning model, this information can be used for early fault diagnosis and provide long-term performance prediction, thereby improving the operational reliability of the fuel cell system.

[0034] Example 6: Transfer coefficient Calculated using the following formula: ; in, is the reorganization energy, which represents the energy adjustment of the catalyst and electrolyte interface during charge transfer. The unit is eV and the value range is 0.2 to 0.6, where 1eV=96.5kJ / mol; is the activation free energy, which is defined as the minimum Gibbs free energy required for the reaction to proceed, in eV, ranging from 0.1 to 0.4; is the electron transfer coefficient, which is related to the charge transfer rate and ranges from 0.5 to 0.7; is the surface coverage; surface coverage Calculated using the following formula: ; in, is the hydrogen adsorption equilibrium constant, ,in It is the adsorption free energy of hydrogen on the catalyst surface, measured in eV, and ranges from -0.3 to -0.1.

[0035] Specifically, in the calculation formula of the transfer coefficient, is the reorganization energy, which represents the energy barrier that needs to be overcome during the transfer of charge between the catalyst and the electrolyte interface, and is measured in eV. In the fuel cell reaction, electrons are transferred from hydrogen molecules to the catalyst surface and through the electrolyte to the cathode, where they complete the reduction reaction under the action of oxygen. The size of the reorganization energy directly affects the rate of charge transfer. If If the value is too large, it means that the charge transfer process is blocked, resulting in a decrease in the electrode reaction kinetics, which in turn increases the activation polarization loss and reduces the operating voltage of the fuel cell. In the system of the present invention, the change in reorganization energy can be used to identify changes in catalyst activity, such as when the catalyst is aged or poisoned. will gradually increase, thus affecting the transfer coefficient This will eventually lead to a decrease in the performance of the fuel cell. Time-varying trends can be analyzed to identify early signs of catalyst aging and predict its impact on fuel cell lifetime. is the activation free energy, which indicates the minimum Gibbs free energy required for the electrode reaction, that is, the minimum energy required for the electron to transfer from the fuel molecule to the catalyst. The magnitude of the activation free energy determines the kinetic speed of the fuel cell reaction. The larger the value, the higher the energy barrier required for the reaction, resulting in a decrease in the exchange current density, which in turn increases the activation polarization loss of the fuel cell. In the diagnostic system of the present invention, Changes in can reflect changes in catalyst performance. For example, when contamination occurs on the catalyst surface, the activation free energy may increase, causing the electrode reaction to slow down, thus affecting the overall efficiency of the fuel cell. The deep learning model can predict the performance trend of the fuel cell under different load conditions by monitoring the changes of this parameter over time and combining it with other operating data.

[0036] Index term It reflects the effect of temperature on the transfer coefficient. This term comes from the Arrhenius equation, which describes the exponential relationship between chemical reaction rate and temperature. is the gas constant, is the operating temperature of the fuel cell. When the temperature rises, the value of this exponential term increases, indicating that the rate of the electrode reaction increases and the transfer coefficient increases, which reduces the activation polarization loss of the fuel cell; on the contrary, when the temperature decreases, the transfer coefficient However, excessively high temperature may accelerate the aging of the catalyst, so in the system of the present invention, the change of temperature-related items can be used to monitor the thermal management state of the fuel cell. Abnormal increase may increase the transfer coefficient in the short term, but may damage the electrode material in the long term. The deep learning model of the present invention can identify the dynamic change pattern of the index term and combine it with the time series analysis method to evaluate and predict the state of the thermal management system. Multiplication factor The relationship between the electron transfer process and hydrogen adsorption on the catalyst surface is described. is the electron transfer coefficient, which determines the rate of charge transfer between the electrode and electrolyte interface, and is usually between 0.5 and 0.7. If the electron transfer coefficient is small, it means that charge transfer is hindered, the exchange current density of the fuel cell is reduced, and the activation polarization loss of the system increases. On the other hand, is the surface coverage, which indicates the proportion of the catalyst surface adsorbed by hydrogen. The calculation formula for surface coverage is The effect of hydrogen pressure on the reaction activity of the catalyst surface is explained. is the hydrogen adsorption equilibrium constant, given by (the adsorption free energy of hydrogen on the catalyst surface). The value range of adsorption free energy is -0.3 to -0.1 eV, indicating that the adsorption of hydrogen on the catalyst is an exothermic reaction. The increase in temperature will reduce the amount of hydrogen adsorption, while the increase in pressure will increase the amount of hydrogen adsorption. It can be seen from the expression of surface coverage that when the hydrogen partial pressure When increasing, Gradually approaching 1, it means that the catalyst surface is completely covered by hydrogen, which can ensure the efficient operation of the fuel cell. However, if the hydrogen supply is insufficient, decline, will also decrease, thus affecting the reaction activity of the catalyst and causing the transfer coefficient The deep learning system of the present invention monitors the output voltage of the fuel cell in real time. changes, combined with The calculation results can detect anomalies in the fuel supply system and predict possible performance degradation problems.

[0037] Example 7: Electrical Efficiency and thermal efficiency Use the following formulas for calculation: ; ; in, is the calorific value of hydrogen.

[0038] Specifically, electrical efficiency The calculation formula reflects the power conversion capability of the fuel cell. represents the actual electrical power output of the fuel cell, where is the operating voltage of a single cell, is the load current, and their product is the power provided by the battery at a certain moment. represents the total energy of hydrogen input, where represents the hydrogen flow rate, is the calorific value of hydrogen, that is, the total energy released when each mole of hydrogen is completely burned. Since the working principle of the fuel cell is that hydrogen and oxygen undergo an electrochemical reaction to generate electricity, this ratio can be used to measure the ability of the fuel cell to convert the input hydrogen energy into electricity. From the formula structure, the electrical efficiency of the fuel cell is affected by multiple factors. If the operating voltage If the hydrogen flow rate is too low, for example due to catalyst poisoning, membrane dehydration, or gas transport limitations, the electrical power output of the fuel cell will decrease, resulting in a decrease in electrical efficiency. The load current has increased but the electric power has not increased year-on-year, indicating that the fuel utilization rate has decreased, resulting in a decrease in electrical efficiency. It will also affect the electrical efficiency. When the load current increases, the polarization loss may increase, resulting in a decrease in the operating voltage, thereby reducing the electrical efficiency. Therefore, the fault diagnosis system of the present invention can be analyzed through deep learning. The system can identify whether the battery has performance degradation and predict possible failure trends by combining historical change patterns of the polarization curve and gas supply data. For example, if the electrical efficiency continues to decrease under the same operating conditions, it may indicate catalyst aging or abnormal gas supply system. The system trains a large amount of operating data through deep learning models, enabling it to autonomously learn these patterns and provide early warnings before failures occur.

[0039] Thermal efficiency The calculation formula reflects the waste heat utilization capacity of the fuel cell system. The molecular part Represents the heat recovered by the system, the denominator It represents the remaining energy that is not converted into electricity in the fuel cell reaction, that is, the part that can be used for waste heat recovery. The physical meaning of this formula is that the fuel cell can not only be used as an electricity supply device, but also use the waste heat to improve the overall energy efficiency of the system. One of the core advantages of hydrogen cogeneration units is that they can efficiently utilize the chemical energy of the fuel, not only to provide electricity, but also for heating or industrial processes. Therefore, thermal efficiency calculation is crucial for overall system optimization. From the perspective of thermal efficiency calculation, if Decrease, that is, the recovered heat decreases, indicating that the waste heat utilization efficiency decreases, which may be caused by failure of the cooling system, blockage of the heat exchanger or abnormal temperature control. The system of the present invention can analyze the long-term change trend of thermal efficiency through a deep learning model, combined with data such as coolant temperature and fuel cell temperature gradient, to identify possible problems with the thermal management system. For example, if the thermal efficiency of the fuel cell continues to decrease under the same hydrogen input, it may indicate that the heat transfer efficiency of the heat exchange system is reduced, and may even indicate that the coolant circulation is blocked. Based on these data, the intelligent diagnostic system of the present invention can discover the early characteristics of faults through pattern recognition methods, and provide maintenance suggestions in advance to avoid the system being affected before a serious fault occurs.

[0040] Embodiment 8: The process of constructing the fault feature vector of the hydrogen energy cogeneration unit using the deep learning model specifically includes: first constructing the following initial fault vector: ; in, is the reference voltage; is the reference electrical efficiency; is the reference thermal efficiency; then a single-layer neural network in deep learning is used to extract Processing is performed to obtain the fault feature vector .

[0041] Specifically, the initial fault feature vector It consists of seven key variables, each of which has a specific physical meaning and can effectively reflect the health status of the fuel cell system. First, represents the normalized voltage value of the fuel cell, where is the reference voltage, which is used to eliminate the voltage scale differences between different fuel cell systems. The physical meaning of this ratio is to measure the deviation of the current fuel cell operating voltage relative to the normal operating state. When the catalyst ages, the membrane is dehydrated, or the gas transport is blocked, the operating voltage is usually lower than the reference voltage, so this parameter can be used to detect the performance degradation of the fuel cell. The second variable represents the normalized value of the electrical efficiency of the fuel cell, where is the reference electrical efficiency. Electrical efficiency is an important indicator for measuring the energy conversion capacity of fuel cells. Its reduction may indicate catalyst poisoning, insufficient fuel supply or increased polarization loss. This normalized value is used to eliminate the impact of different operating conditions on electrical efficiency, allowing the system to identify abnormal patterns based on relative change trends. For example, under the same hydrogen input, if the electrical efficiency decreases but the load power remains unchanged, it may mean that the catalyst surface activity is reduced or gas transport is limited. The third variable represents the normalized value of thermal efficiency, where is the reference thermal efficiency. Thermal efficiency is used to measure the waste heat utilization of the fuel cell system. Its decline may be caused by cooling system failure, reduced heat exchange efficiency or abnormal membrane hydration state. Since the thermal management of the fuel cell directly affects its long-term stability, the change trend of this parameter can reflect the health of the system. For example, at the same input power, if the thermal efficiency drops significantly, it may mean that the cooling system is blocked or the heat exchanger is damaged.

[0042] The fourth variable It reflects the difference between the stack temperature and the coolant temperature and is normalized to the ratio of the load current. The physical meaning of this ratio is to describe the working efficiency of the thermal management system. The larger the temperature difference, the more likely the heat dissipation capacity of the fuel cell is to be limited. If this parameter continues to rise under the same load conditions, it may mean that the coolant flow is insufficient or the heat exchange efficiency is reduced, causing the operating temperature of the fuel cell to deviate from the optimal range. The fifth variable The ratio of hydrogen pressure to flow is used to measure the health of the hydrogen transport system. When the hydrogen pressure is high but the flow is low, it may mean that the hydrogen supply is blocked, such as a blockage in the hydrogen supply pipeline or a blockage in the gas diffusion layer. On the contrary, if the flow is high but the pressure is low, it may mean a leak or abnormal regulation of the gas supply system. The sixth variable Represents the first derivative of the fuel cell voltage to the load current. This parameter is used to characterize the slope of the polarization curve and reflects the dynamic response of the fuel cell under different load conditions. When the polarization curve of the fuel cell becomes steeper, the value of this parameter decreases, indicating that the working state of the fuel cell is limited, such as the degradation of catalyst performance or the obstruction of gas transport. The seventh variable Represents the second-order derivative of the fuel cell voltage to the load current. This parameter is used to characterize the curvature of the polarization curve and reflects the nonlinear characteristics of the fuel cell under high load conditions. When the absolute value of this parameter increases, it means that the polarization characteristics of the fuel cell have changed dramatically, which may be caused by changes in the membrane hydration state, abnormal gas transport, or increased polarization loss.

[0043] Initial Fault Vector As the input of the neural network, the vector dimension is (Right now The number of variables included). The weight matrix of the input vector passing through the hidden layer and the bias vector For linear transformation, the formula is as follows: ; in: is the initial fault feature vector, with dimension (In the example of the present invention ). is the weight matrix with dimension ,in is the number of neurons in the hidden layer. is the bias vector with dimension In order to enhance the expressive power of the neural network, a nonlinear activation function is used to The present invention adopts the ReLU activation function, which is calculated as follows: ; The characteristic of the ReLU activation function is that it maintains a linear relationship for positive inputs and outputs 0 for negative inputs, which can enhance the learning ability of the neural network and avoid the gradient vanishing problem. It is further transformed linearly in the output layer to generate the final fault feature vector: ; in: is the weight matrix of the output layer, with dimension ,in is the dimension of the final feature vector. In the present invention, the dimension is kept unchanged and is taken as 7. is the bias vector of the output layer, with dimension . is the final fault feature vector with dimension After the above calculations, we finally get is a dimensional vector, which extracts the initial fault feature vector The core information in the vector is obtained, and more complex fault modes are learned through neural network training. This vector can be used for subsequent fault classification, anomaly detection or time series prediction to determine whether the hydrogen energy cogeneration unit is in a healthy state, or to predict the probability of future failures. In order to improve the prediction performance, the present invention can adopt an MLP structure with regularization, for example, in the weight matrix The L2 regularization term is added to the calculation of to prevent overfitting. In addition, the number of neurons in the output layer It can be adjusted according to different application requirements, for example, Corresponding to five typical faults (membrane dehydration, membrane flooding, catalyst poisoning, gas diffusion layer blockage, and thermal management system failure).

[0044] Example 9: Use the following formula to calculate the probability of each type of failure occurring at each time step: ; in, express Failures occur all the time The probability of represents the time step; Represents the fault feature vector With failure Mahalanobis distance of feature template vector; Represents the fault feature vector With failure The square of the Euclidean distance of the feature template vector; is the set historical time window; is the integration variable; and All are integer subscript indices.

[0045] Specifically, the present invention first uses the fault feature vector constructed by deep learning , this feature vector is a high-dimensional representation extracted from multiple operating parameters of the fuel cell, including voltage, current, temperature, pressure and other information. In order to evaluate the similarity between the current state and the known failure mode, the system calculates With each known failure mode Mahalanobis distance The Mahalanobis distance has more advantages than the Euclidean distance. It not only considers the covariance information between variables, but also normalizes the features of different scales, so that the system can still accurately measure the similarity between different fault modes when facing complex data distribution. Its calculation method is essentially based on the transformation of the covariance matrix, so that those fault modes with similar statistical characteristics in multidimensional space can be distinguished more accurately. The smaller the distance, the closer the current state is to the fault. The closer the characteristics are, the higher the probability of the fault. In order to map this distance relationship to the probability distribution, the present invention adopts the exponential normalization method, that is, taking the negative exponent of the Mahalanobis distance and normalizing it so that the sum of the probabilities of all faults is always 1. This method is similar to Softmax normalization, but on this basis, negative exponential operations are added, so that the state closer to the known fault mode has a higher probability. This means that if the feature vector at a certain moment is close to a specific fault, If the fault is very close to the feature template, the probability of the fault will be amplified, while the cases with lower similarity to other faults will be suppressed. This ensures that the system can correctly identify the most likely fault type.

[0046] However, it is not enough to rely on the current data for classification, because the failure of the hydrogen energy cogeneration unit does not occur instantly, but evolves over time. Therefore, the present invention introduces a time evolution correction term, which is calculated by a time window. Inner eigenvector With fault template , in order to measure the temporal stability of the system's operating status. Specifically, this item calculates the rate of change of the Euclidean distance between the fault feature vector and the specific fault template in the past time period, and accumulates it by integration. If the Euclidean distance has changed greatly in the past period of time, it means that the system state fluctuates greatly and has not yet stabilized towards a certain fault mode, then the system will reduce the predicted probability of the fault; and if the distance gradually decreases over time, it means that the system is converging to the fault mode, then the probability of the fault will increase. This mechanism effectively enhances the system's ability to capture the trend of fault occurrence, ensures that the system will not generate false alarms due to short-term abnormal fluctuations, and can accurately judge the development trend of the fault. In practical applications, the physical meaning of the time correction term can be understood as: if in time If the operating state of the fuel cell is constantly approaching a certain fault mode, such as a membrane dehydration failure causing the battery's ohmic resistance to gradually increase, and this trend persists over multiple time steps, then the correction term will increase the probability of membrane dehydration failure, allowing the system to give an early warning. However, if the fault feature vector has experienced large fluctuations over a period of time, such as instantaneous voltage fluctuations due to load changes, but the system then returns to normal operation, then the correction term will reduce the predicted probability of the fault, thereby avoiding misjudgment. This analysis method based on time-varying trends makes the prediction model of the present invention more robust than traditional static diagnostic methods, and can identify potential development trends of faults in advance.

[0047] Membrane dehydration failure is mainly manifested as increased ohmic resistance, decreased fuel cell voltage, and reduced electrical efficiency. Insufficient membrane hydration will lead to decreased proton conductivity, which in turn causes greater ohmic polarization losses. ; (operating voltage drops by 15%); (Electrical efficiency drops by 20%); (Thermal efficiency drops by 10%); (Stack temperature rises); (Hydrogen utilization rate increases); the slope of the polarization curve decreases , the second-order derivative tends to be flat Membrane flooding failure is manifested by obstructed gas diffusion, drastic changes in polarization curves, and increased voltage fluctuations, which usually occurs under low load or low temperature environments. ; The gas diffusion layer is blocked by liquid water, resulting in (Hydrogen transfer is limited); the polarization curve drops sharply in the high load region Catalyst poisoning failure will lead to increased activation polarization loss, decreased exchange current density, and reduced catalyst activity. Typically, it is manifested as a more obvious voltage drop in the low current density area. ; The exchange current density decreases, resulting in an increase in activation polarization losses ; Gas diffusion layer blockage failure leads to oxygen diffusion obstruction, a steep drop in the polarization curve, and severe voltage fluctuations, which is usually caused by long-term operation or accumulation of pollutants. ; The oxygen diffusion capacity decreases, resulting in ; The polarization curve drops steeply ;The thermal management system failure is manifested by reduced cooling efficiency, abnormal stack temperature, and large voltage drop. ; Coolant temperature deviation is large ; Thermal efficiency is significantly reduced .

[0048] The preferred embodiments of the present disclosure are described above with reference to the accompanying drawings, but the scope of the present disclosure is not limited thereby. Any modification, equivalent substitution and improvement made by those skilled in the art without departing from the scope and essence of the present disclosure shall be within the scope of the present disclosure.

Claims

1. A hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning, characterized in that: The system includes: a data acquisition unit, a data analysis and feature extraction unit and a fault prediction unit; the data acquisition unit is used to collect the operation data of the hydrogen energy cogeneration unit in real time; the data analysis and feature extraction unit is used to perform electrochemical performance analysis based on polarization curves according to the operation data, and establish a polarization model of the hydrogen energy cogeneration unit; according to the polarization model, the electrical efficiency and thermal efficiency of the hydrogen energy cogeneration unit are calculated, and then the deep learning model is used to construct the fault feature vector of the hydrogen energy cogeneration unit; the fault prediction unit is used to judge the probability of various types of faults occurring in the next time step according to the fault feature vector and the feature template vectors of each type of fault set.

2. The deep learning-based hydrogen energy cogeneration unit fault diagnosis and prediction system according to claim 1, characterized in that: The types of faults include: membrane dehydration fault, membrane flooding fault, catalyst poisoning fault, gas diffusion layer blockage fault and thermal management system fault.

3. The hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning as claimed in claim 2, characterized in that: The data acquisition unit collects the operating data of the hydrogen energy cogeneration unit in real time, including: the single cell voltage in V , load current in A , Stack temperature in K , Hydrogen pressure in Pa , oxygen pressure in Pa , Hydrogen flow rate in mol / s , oxygen chemical equivalent ratio , coolant inlet temperature in K , coolant outlet temperature in K , fuel cell operating temperature in K , water vapor partial pressure in Pa , Ohm resistance in Ω and the recovered heat in W ;in, For time.

4. The deep learning-based hydrogen energy cogeneration unit fault diagnosis and prediction system according to claim 3, characterized in that: The polarization model is expressed using the following formula: ； in, is the standard potential, in V; is the gas constant, which is 8.314 J / (mol·K); is the Faraday constant, which is 96485C / mol; is the transfer coefficient, ranging from 0.3 to 0.7; is the exchange current density, in A / cm²; is the first concentration polarization parameter, the unit is V, and the value range is 0.02 to 0.08; It is the second concentration polarization parameter, with the unit of cm² / A and the value range is 10 to 30.

5. The hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning according to claim 4, characterized in that: The exchange current density is calculated using the following formula: ； in, is the influence index of hydrogen partial pressure on exchange current density, ranging from 0.7 to 1.2; is the influence index of oxygen partial pressure on exchange current density, ranging from 0.3 to 0.5; is the reference pressure, which is 1 atm or Pa; is the activation energy, in J / mol; The exchange current density prefactor reflects the exchange current density under ideal conditions without an activation energy barrier. The unit is A / cm² and the value range is arrive .

6. The hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning according to claim 5, characterized in that: Transfer coefficient Calculated using the following formula: ； in, is the reorganization energy, which represents the energy adjustment of the catalyst and electrolyte interface during charge transfer. The unit is eV and the value range is 0.2 to 0.6, where 1eV=96.5kJ / mol; is the activation free energy, which is defined as the minimum Gibbs free energy required for the reaction to proceed, in eV, ranging from 0.1 to 0.4; is the electron transfer coefficient, which is related to the charge transfer rate and ranges from 0.5 to 0.7; is the surface coverage; surface coverage Calculated using the following formula: ； in, is the hydrogen adsorption equilibrium constant, ,in It is the adsorption free energy of hydrogen on the catalyst surface, measured in eV, and ranges from -0.3 to -0.

1.

7. The hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning according to claim 6, characterized in that: Electrical efficiency and thermal efficiency Use the following formulas for calculation: ； ； in, is the calorific value of hydrogen.

8. The deep learning-based hydrogen energy cogeneration unit fault diagnosis and prediction system according to claim 7, characterized in that: The process of constructing the fault feature vector of the hydrogen energy cogeneration unit using the deep learning model specifically includes: first constructing the following initial fault vector: ； in, is the reference voltage; is the reference electrical efficiency; is the reference thermal efficiency; then a single-layer neural network in deep learning is used to extract Processing is performed to obtain the fault feature vector .

9. The hydrogen energy cogeneration unit fault diagnosis and prediction system based on deep learning according to claim 8, characterized in that: The probability of each type of failure occurring at each time step is calculated using the following formula: ； in, express Failures occur all the time probability; represents the time step; Represents the fault feature vector With failure Mahalanobis distance of feature template vector; Represents the fault feature vector With failure The square of the Euclidean distance of the feature template vector; is the set historical time window; is the integration variable; and All are integer subscript indices.

Images