SOFC state monitoring and health diagnosis method and system based on artificial intelligence

Abstract

The invention provides an SOFC state monitoring and health diagnosis method and system based on artificial intelligence, and relates to the technical field of solid oxide battery state monitoring, and the method comprises the steps: collecting the operation parameters of an SOFC in real time; an algorithm based on extended Kalman filtering and information entropy optimization data is adopted, and weighting processing is carried out on the collected operation parameters; on the basis of an improved hybrid neural network model and a reinforcement learning algorithm, the state of the battery is monitored in real time, the expected durability period is predicted, abnormity is recognized, and model parameters are dynamically adjusted through sliding window online learning and a self-correction mechanism; and a feedback control signal is generated by combining a monitoring result, health prediction and an abnormal early warning signal, and the running parameters of the SOFC are dynamically adjusted. The method has the advantages of high precision, strong robustness and adaptivity, and is suitable for intelligent management of the SOFC in energy storage and efficient power generation.

Description

Technical Field

[0001] The present invention relates to the technical field of solid oxide battery state monitoring technology, and in particular to an artificial intelligence-based SOFC state monitoring and health diagnosis method and system. Background Art

[0002] Solid oxide fuel cells (SOFCs) have important applications in energy storage and clean power generation due to their high energy conversion efficiency, environmental friendliness, and long lifespan. Although existing technologies have made significant progress in SOFC monitoring, the following challenges remain in SOFC condition monitoring and health diagnosis: First, data acquisition is limited. Existing technologies are often limited to a single or limited number of parameters (such as voltage and current), lacking comprehensive acquisition and integration of key SOFC operating parameters (such as oxygen partial pressure and ambient humidity), resulting in incomplete information.

[0003] Second, feature extraction is insufficient. Existing technologies mostly use static features (such as mean and variance), which are difficult to capture the time-frequency dynamic characteristics of the SOFC aging process, limiting prediction accuracy.

[0004] Third, the model's adaptability is insufficient. The AI ​​models used in existing technologies are mostly statically trained, lacking online self-correction capabilities and unable to adapt to the nonlinear degradation of SOFCs throughout their life cycle.

[0005] Fourth, there is a lack of closed-loop optimization. Existing technologies mostly remain at the monitoring and diagnosis level, and have not formed a closed-loop system from prediction to operational optimization. Summary of the Invention

[0006] In order to at least partially solve the technical problems in the related art, the present invention provides a SOFC state monitoring and health diagnosis method and system based on artificial intelligence. The present invention aims to solve the technical problems of low state monitoring accuracy and insufficient health prediction ability of solid oxide battery systems under complex operating conditions, and provide an efficient method and system based on artificial intelligence to achieve accurate monitoring of battery operating status and durability prediction. The present invention comprehensively captures the nonlinear degradation process of SOFC through multi-source data fusion and dynamic feature extraction, and achieves high-precision monitoring and durability prediction throughout the life cycle. Compared with traditional methods, the present invention makes full use of the powerful data processing capabilities of artificial intelligence, combined with the operating characteristics of solid oxide batteries, significantly improves the real-time monitoring and prediction accuracy, and provides technical support for intelligent management in the fields of energy storage and efficient power generation.

[0007] In order to achieve the above object, the technical solution adopted by the present invention includes: According to a first aspect of the present invention, a method for SOFC state monitoring and health diagnosis based on artificial intelligence is provided, comprising the following steps: Step S1: real-time collection of SOFC operating parameters, including temperature, voltage, current, internal resistance, oxygen partial pressure, and ambient humidity; Step S2: Using an algorithm based on extended Kalman filtering and information entropy to optimize data, the collected operating parameters are weighted to optimize data consistency and noise resistance; Step S3: Based on the improved hybrid neural network model and reinforcement learning algorithm, the battery status is monitored in real time, the expected endurance cycle is predicted, and anomalies are identified. The model parameters are dynamically adjusted through sliding window online learning and self-correction mechanisms. Step S4: combining the monitoring results, health prediction and abnormal warning signals to generate feedback control signals and dynamically adjust the operating parameters of the SOFC to extend the battery life.

[0008] Optionally, step S4 further includes: outputting the monitoring results, health predictions and abnormal warning signals in a visual form.

[0009] Optionally, in step S1, operating parameters are collected in real time through a sensor network.

[0010] Optionally, step S2 specifically includes: Step S2-1: Process the collected operating parameters according to the following formula: Where, is the predicted state, F is the state transfer matrix, is the updated state at time k-1, is the prediction covariance, is the updated covariance at time k-1, is the transposed matrix of F, Q is the process noise covariance, is the Kalman gain, H is the observation matrix, is the transposed matrix of H, R is the measurement noise covariance, To update the status, is the observed value, is the updated covariance, I is the identity matrix; Step S2-2: Calculate the information entropy and weight of the collected operating parameters according to the following formula: Where E is information entropy, is the normalized value of the i-th sensor data, is the weight of the i-th sensor data, is the information entropy corresponding to the i-th sensor data, is the information entropy corresponding to the j-th sensor data; Step S2-3: Calculate the fusion value according to the following formula: Where, is the fusion value obtained by fusing multiple sensor data, is the data of the i-th sensor.

[0011] Optionally, step S3 specifically includes: Step S3-1: Use discrete wavelet transform to perform time-frequency analysis on the voltage and internal resistance signals, extract transient features, and calculate the capacity decay rate using statistical methods. The calculation formula of discrete wavelet transform is: Where, is the wavelet coefficient of the jth scale and the kth time shift, is the input signal, is the wavelet basis function, is the approximate coefficient of the jth layer, is the detail coefficient of the jth layer, is the scaling function, is the wavelet function; Step S3-2: Based on a hybrid neural network model and reinforcement learning algorithm, the battery status is monitored in real time, the expected endurance cycle is predicted, and anomalies are identified. The hybrid neural network model is a CNN with 4 convolution layers and 3×3 kernels and a LSTM with 3 layers and 128 units. The CNN is used to extract spatial features of multidimensional data, and the LSTM is used to capture long-term dependencies in time series. The performance degradation index PDI and the predicted expected endurance cycle EEC are output. The calculation formulas for the performance degradation index PDI and the predicted expected endurance cycle EEC are: Where, is the capacity-based performance degradation index, C is the current capacity, is the initial capacity; is the performance degradation index based on environmental factors, is the temperature compensation factor, T is the current temperature, is the oxygen partial pressure influencing factor, is the current oxygen partial pressure; is the corresponding weight coefficient; is the performance failure threshold, is the decay rate, β is the cyclic stress factor caused by temperature, and N is the number of cycles.

[0012] Optionally, the Set to 20%.

[0013] Optionally, the step S3-2 further includes: The model weights are updated every several cycles to dynamically adjust the parameters using the latest data, and the prediction strategy is optimized in combination with the reward function to improve long-term stability.

[0014] Optionally, in step S4, the feedback control signal is used to adjust the charging rate and battery control parameters of the SOFC.

[0015] According to a second aspect of the present invention, there is further provided an artificial intelligence-based SOFC state monitoring and health diagnosis system for executing the artificial intelligence-based SOFC state monitoring and health diagnosis method according to any one of the technical solutions in the first aspect of the present invention. The artificial intelligence-based SOFC state monitoring and health diagnosis system comprises: Data acquisition module, used to collect SOFC operating parameters in real time; A data fusion unit is used to perform fusion processing on the collected operating parameters; A feature adaptive screening unit is used to extract multi-scale features related to battery health from the fused data and perform dynamic screening; A self-correcting AI module, which uses artificial intelligence to diagnose battery status and predict durability in real time, achieves model self-correction through online learning, and identifies abnormal conditions; The feedback control unit is used to generate a feedback control signal according to the diagnosis result to optimize the battery operating parameters.

[0016] Optionally, the self-correcting AI module uses an improved hybrid neural network and reinforcement learning algorithm to detect battery status, predict durability and identify anomalies in real time, and the feedback control unit generates a feedback control signal based on the prediction results to dynamically adjust the battery charging rate and battery control parameters.

[0017] Beneficial effects: 1. Through the above technical solution, the present invention can achieve the following technical effects: First, high-precision condition monitoring. This method combines multi-source data fusion (such as temperature, voltage, current, internal resistance, oxygen partial pressure, and ambient humidity) with an extended Kalman filter (EKF) and information entropy weighting algorithm to achieve data consistency exceeding 98% and reduce monitoring error to less than 2%. Compared to traditional single-parameter monitoring methods (e.g., voltage-based monitoring with an error of approximately 5-10%), this significantly improves SOFC monitoring accuracy at high temperatures (600-1000°C) and during dynamic degradation.

[0018] Second, accurate health prediction. This method utilizes a CNN-LSTM hybrid neural network and a reinforcement learning algorithm to predict the expected endurance cycle (EEC) of SOFCs in real time, with an accuracy of up to 95% and an error within ±10 cycles. Through a sliding window online learning mechanism, the model adaptively adjusts parameters to ensure accurate capture of nonlinear aging processes throughout the entire lifecycle, improving long-term stability by 20%.

[0019] Third, it extends battery life. The method of this invention uses closed-loop feedback optimization to dynamically adjust operating parameters based on diagnostic results (such as reducing the charging rate to 70% and controlling the temperature at 600-700°C), effectively slowing the rate of capacity decay. Tests showed that after 1,000 cycles, the capacity decay rate of the feedback-optimized group was reduced to 5% (compared to 8% for the group without feedback), extending battery life by 15-20% (approximately 150-200 additional cycles) and improving operating efficiency by 10%.

[0020] Fourth, strong robustness and real-time performance. The dynamic feature extraction technology of this invention extracts multi-scale features through time-frequency analysis and wavelet transform, improving feature expression by 30%, reducing computational complexity by 70%, and shortening computation time from approximately 500ms to less than 150ms. This ensures the system's real-time monitoring capabilities even with high cycle counts (>1000) and complex environments, and its feature stability is 20% better than traditional static methods.

[0021] Fifth, intelligent management support. This invention provides visual output (internal resistance growth rate graph, temperature impact factor graph, abnormality alarm) and feedback control signals, providing technical support for the intelligent management of SOFC in energy storage and efficient power generation, significantly improving the economic benefits and practicality of the system.

[0022] In summary, the present invention achieves high-precision, high-efficiency, and intelligent SOFC state monitoring and health diagnosis through technological integration. It not only overcomes the limitations of traditional methods under complex conditions, but also provides an innovative solution for the long-term stable operation and life optimization of SOFC, with significant application value.

[0023] 2. Other beneficial effects or advantages of the present invention will be described in detail in the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative labor.

[0025] in: Figure 1 1 is a schematic diagram of a workflow of an artificial intelligence-based SOFC state monitoring and health diagnosis system provided by an exemplary embodiment of the present invention; Figure 2 is a schematic diagram of the change of internal resistance growth rate with cycle number provided by an exemplary embodiment of the present invention. Figure 2 In the figure, the horizontal axis is the "number of cycles" and the vertical axis is the "growth rate (%)", which reflects the dynamic changes of the internal resistance during the battery aging process; Figure 3 is a schematic diagram showing the temperature influence factor changing with time according to an exemplary embodiment of the present invention. Figure 3 In the figure, the horizontal axis is "time (hours)" and the vertical axis is "impact factor", which reflects the real-time impact of temperature on performance. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions 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.

[0027] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are intended to fall within the scope of protection of the present invention.

[0028] In order to facilitate relevant technical personnel to have a clearer and more accurate understanding of the technical solutions of the present invention, the technical problems existing in the prior art are described in more detail below with reference to examples.

[0029] Solid oxide fuel cells (SOFCs) have significant application value in energy storage and clean power generation due to their efficient energy conversion, environmental friendliness, and long lifespan. The development of SOFC technology has undergone a transition from traditional electrochemical monitoring to data-driven intelligence. Early research primarily relied on equivalent circuit models and threshold detection methods to analyze battery status, but these methods exhibit limitations in capturing the complex characteristics of SOFCs at high temperatures (600-1000°C) and during dynamic degradation. With the rise of data-driven technologies and artificial intelligence (AI), the field of battery health monitoring has ushered in new development opportunities.

[0030] Although existing technologies have promoted the development of battery monitoring technology, the following problems and challenges still exist in the field of SOFC condition monitoring and health diagnosis: First, data collection limitations. Existing technologies are often limited to a single or limited set of parameters (such as voltage and current), lacking comprehensive collection and integration of key SOFC operating parameters (such as oxygen partial pressure and ambient humidity), resulting in incomplete information. For example, the Chinese patent document with publication number CN112373352B focuses solely on basic electrical parameters and fails to address the multidimensional impacts of high-temperature SOFC environments.

[0031] Second, feature extraction is inadequate. Existing technologies often rely on static features (such as mean and variance), which struggle to capture the dynamic time-frequency characteristics of SOFC aging, limiting prediction accuracy. While existing technologies have proposed multi-scale analysis, these techniques are not optimized for SOFC material and operating characteristics.

[0032] Third, the model's adaptability is insufficient. Existing AI models are mostly statically trained, lacking online self-correction capabilities and unable to adapt to the nonlinear degradation of SOFCs throughout their lifecycle. For example, Chinese patent application CN113594510B, while incorporated into a cloud platform, fails to implement real-time adaptive adjustments.

[0033] Fourth, there's a lack of closed-loop optimization. Existing technologies largely remain at the monitoring and diagnosis level, failing to form a closed-loop system from prediction to operational optimization. For example, while Chinese patent application CN112373352B improves prediction capabilities, it doesn't feed the results back into system control, resulting in limited lifespan extension.

[0034] To address these challenges, the present invention proposes an artificial intelligence-based SOFC state monitoring and health diagnosis method and system, which has significant technical significance and application value. Through multi-source data acquisition and fusion, dynamic feature extraction, self-correcting AI diagnosis, and closed-loop feedback optimization, the present invention achieves high-precision real-time monitoring and durability prediction of SOFCs under complex conditions. Compared to existing related technologies, the present invention fully utilizes the powerful data processing capabilities of AI and, combined with the high-temperature operating characteristics of SOFCs, significantly improves the robustness of monitoring and the accuracy of predictions. Its technical advantages include: 1. Comprehensiveness: It integrates multi-dimensional parameters such as temperature and oxygen partial pressure to make up for the shortcomings of single parameter monitoring.

[0035] 2. Dynamics: Accurately capture the degradation process through time-frequency analysis and adaptive feature screening.

[0036] 3. Adaptability: AI models learn online and adapt to changes throughout the lifecycle.

[0037] 4. Practicality: Closed-loop feedback optimizes operating strategies and extends SOFC life.

[0038] Therefore, the present invention not only fills the technological gap in the field of SOFC health diagnosis, but also provides an innovative solution for the intelligent management of energy storage and efficient power generation, and has significant scientific value and industrial application prospects.

[0039] The technical solution of the present invention is described in detail below with reference to the accompanying drawings.

[0040] According to a first aspect of the present invention, a method for SOFC state monitoring and health diagnosis based on artificial intelligence is provided, comprising the following steps: Step S1: real-time collection of SOFC operating parameters, including temperature, voltage, current, internal resistance, oxygen partial pressure, and ambient humidity; Step S2: Using an algorithm based on extended Kalman filtering and information entropy to optimize data, the collected operating parameters are weighted to optimize data consistency and noise resistance; Step S3: Based on the improved hybrid neural network model and reinforcement learning algorithm, the battery status is monitored in real time, the expected endurance cycle is predicted, and anomalies are identified. The model parameters are dynamically adjusted through sliding window online learning and self-correction mechanisms. Step S4: combining the monitoring results, health prediction and abnormal warning signals to generate feedback control signals and dynamically adjust the operating parameters of the SOFC to extend the battery life.

[0041] Through the above technical solution, the present invention significantly improves the state monitoring accuracy, health prediction capability, and service life of solid oxide fuel cells (SOFCs) under complex operating conditions by integrating multi-source data acquisition and fusion, dynamic feature extraction, adaptive AI diagnosis, and closed-loop feedback optimization technology. The present invention can achieve the following technical effects: First, high-precision condition monitoring. This method combines multi-source data fusion (such as temperature, voltage, current, internal resistance, oxygen partial pressure, and ambient humidity) with an extended Kalman filter (EKF) and information entropy weighting algorithm to achieve data consistency exceeding 98% and reduce monitoring error to less than 2%. Compared to traditional single-parameter monitoring methods (e.g., voltage-based monitoring with an error of approximately 5-10%), this significantly improves SOFC monitoring accuracy at high temperatures (600-1000°C) and during dynamic degradation.

[0042] Second, accurate health prediction. This method utilizes a CNN-LSTM hybrid neural network and a reinforcement learning algorithm to predict the expected endurance cycle (EEC) of SOFCs in real time, with an accuracy of up to 95% and an error within ±10 cycles. Through a sliding window online learning mechanism, the model adaptively adjusts parameters to ensure accurate capture of nonlinear aging processes throughout the entire lifecycle, improving long-term stability by 20%.

[0043] Third, it extends battery life. The method of this invention uses closed-loop feedback optimization to dynamically adjust operating parameters based on diagnostic results (such as reducing the charging rate to 70% and controlling the temperature at 600-700°C), effectively slowing the rate of capacity decay. Tests showed that after 1,000 cycles, the capacity decay rate of the feedback-optimized group was reduced to 5% (compared to 8% for the group without feedback), extending battery life by 15-20% (approximately 150-200 additional cycles) and improving operating efficiency by 10%.

[0044] Fourth, strong robustness and real-time performance. The dynamic feature extraction technology of this invention extracts multi-scale features (e.g., voltage harmonic components and internal resistance transient response) through time-frequency analysis and wavelet transform. This improves feature expression by 30%, reduces computational complexity by 70%, and shortens computation time from approximately 500ms to less than 150ms. This ensures the system's real-time monitoring capabilities even with high cycle counts (>1000) and complex environments, and its feature stability is 20% better than traditional static methods.

[0045] Fifth, intelligent management support. This invention provides visual output (internal resistance growth rate graph, temperature impact factor graph, abnormality alarm) and feedback control signals, providing technical support for the intelligent management of SOFC in energy storage and efficient power generation, significantly improving the economic benefits and practicality of the system.

[0046] In summary, the present invention achieves high-precision, high-efficiency, and intelligent SOFC state monitoring and health diagnosis through technological integration. It not only overcomes the limitations of traditional methods under complex conditions, but also provides an innovative solution for the long-term stable operation and life optimization of SOFC, with significant application value.

[0047] The present invention is described below with reference to an exemplary embodiment.

[0048] 1. Hardware configuration.

[0049] 1) Temperature sensor: K-type thermocouple (0-1300℃, ±0.2℃), sampling rate 5Hz.

[0050] 2) Voltage sensor: differential amplifier (0-6V, ±0.005V) with anti-interference filter.

[0051] 3) Current sensor: Hall effect sensor (0-150A, ±0.05A), supporting high frequency response.

[0052] 4) Oxygen partial pressure sensor: Zirconia sensor (0-1atm, ±0.01atm).

[0053] 5) Data acquisition system: 32-bit ARM processor, 16-bit ADC, storage capacity 1TB.

[0054] 2. Data fusion and preprocessing.

[0055] 1) Adopt a fusion algorithm based on extended Kalman filter (EKF) and information entropy weighting to reduce sensor noise and redundancy and improve data consistency.

[0056] 2) Mathematical principles: Where, is the predicted state, F is the state transfer matrix, is the updated state at time k-1, is the prediction covariance, is the updated covariance at time k-1, is the transposed matrix of F, Q is the process noise covariance, is the Kalman gain, H is the observation matrix, is the transposed matrix of H, R is the measurement noise covariance, To update the status, is the observed value, is the updated covariance, I is the identity matrix; Among them, E is information entropy, is the normalized value of the i-th sensor data, is the weight of the i-th sensor data, is the information entropy corresponding to the i-th sensor data, is the information entropy corresponding to the j-th sensor data; 3) Fusion value calculation: Where, is the fusion value obtained by fusing multiple sensor data, is the data of the i-th sensor.

[0057] In this process, the state of multi-source data (such as temperature, oxygen partial pressure, etc.) is first estimated through extended Kalman filtering to filter out high-frequency noise; then the entropy value of each sensor data is calculated and weights are assigned; finally, the weighted average is used to obtain the fusion result.

[0058] 3. Feature extraction and time-frequency analysis.

[0059] 1) Use discrete wavelet transform (DWT) to perform time-frequency analysis on voltage and internal resistance signals, extract transient characteristics, and calculate the capacity decay rate using statistical methods.

[0060] 2) Mathematical principles: Mathematical formula: Discrete Wavelet Transform: Where, is the wavelet coefficient of the jth scale and the kth time shift, is the input signal, For the wavelet basis function, "db6" can be selected.

[0061] Signal decomposition: Where, is the approximate coefficient of the jth layer, is the detail coefficient of the jth layer, is the scaling function, is the wavelet function; For example, taking the 6-layer decomposition as an example, the voltage signal is decomposed into a low-frequency approximate component and a high-frequency detail component. The amplitude of the high-frequency component is extracted as a transient feature to reflect the change in the internal resistance of the battery; the low-frequency component is used to calculate the capacity decay rate.

[0062] 4. AI model training and optimization.

[0063] 1) Model structure: CNN (4-layer convolution, 3×3 kernel) + LSTM (3 layers, 128 units), input 5-dimensional features, output PDI and EEC.

[0064] 2) Training parameters: 800 sets of data (1500 cycles), batch size 32, learning rate 0.002.

[0065] 3) Online learning: Update the model every 50 cycles, with a sliding window of 50 and a learning rate decay of 0.95.

[0066] 5. Optimize the display of results and feedback. (e.g. Figure 2 shown) 1) Output: internal resistance growth rate graph, temperature influence factor graph, abnormal alarm.

[0067] 2) Feedback: When PDI < 60%, the charging current is reduced to 70% and the temperature is controlled at 600-700℃.

[0068] In this exemplary embodiment, the advantages of the present invention include: First, high-precision multi-source data fusion. The present invention realizes the efficient fusion of multi-dimensional parameters such as temperature, voltage, current, internal resistance, and oxygen partial pressure by extending the Kalman filter and information entropy weighting algorithm, ensuring the stability of data quality in complex high-temperature environments. The fusion process dynamically adjusts the weight of each parameter to effectively filter out noise interference. For example, under the conditions of temperature fluctuation of ±50°C and oxygen partial pressure change of ±0.05atm, the data consistency remains above 98%. Exemplarily, the monitoring error of the present invention can be reduced to less than 2%, which significantly improves the reliability of the data compared to traditional single parameter monitoring (for example, the error based on voltage alone is about 5-10%), providing a solid foundation for subsequent feature extraction and AI diagnosis.

[0069] Second, the efficiency and real-time performance of dynamic feature extraction. The present invention uses time-frequency analysis and wavelet transform technology to extract multi-scale features, such as voltage harmonic components and internal resistance transient response, and dynamically screens key features through random forest and mutual information methods to eliminate redundant information. The feature dimension is reduced from the initial 15 dimensions to 5 dimensions, and the calculation time is shortened from about 500ms of the traditional method to less than 150ms, ensuring real-time monitoring capabilities. Exemplarily, the feature expression capability of the present invention can be improved by 30%, the computational complexity can be reduced by about 70%, and the feature stability at a high number of cycles (>1000 times) is about 20% better than the static method, which significantly improves the system's ability to capture the aging process.

[0070] Third, the adaptive AI prediction system offers high accuracy and stability. This system combines a CNN-LSTM hybrid neural network with reinforcement learning to analyze battery status in real time and predict the expected endurance cycle (EEC). Through online learning and model self-calibration every 50 cycles, it adapts to the nonlinear changes in the battery from initial operation to deep aging. For example, at a temperature of 650°C and 1500 cycles, the EEC prediction error is controlled within ±10 cycles. As an example, the prediction accuracy reaches 95%, and the long-term stability is improved by 20%, making it more capable of addressing the dynamic degradation of SOFCs throughout their life cycle than traditional static AI models.

[0071] Fourth, closed-loop feedback optimization offers intelligent features and extended battery life. This invention generates feedback signals based on diagnostic results, dynamically adjusting the charging rate (for example, reducing it to 70% of rated current when the PDI is <60%) and temperature control (maintaining 600-700°C), forming a closed-loop optimization system. Tests have shown that after 1,000 cycles, the battery capacity decay rate in the feedback-optimized group was reduced to 5%, compared to 8% in the group without feedback. For example, battery life can be extended by 15-20% (approximately 150-200 additional cycles), and operating efficiency can be improved by 10% (with improved power output stability). Compared to monitoring systems without feedback (such as traditional threshold control), this significantly improves economic benefits and battery life.

[0072] In the above-mentioned embodiments, it should be noted that, in the present invention, the extended Kalman filter estimates the nonlinear system state, filters out sensor noise, and ensures dynamic data consistency. Entropy weighting calculates weights based on the entropy of each parameter, dynamically adjusting the contribution of different data points and enhancing noise immunity. Compared to traditional single Kalman filtering or mean fusion, this method, combining the EKF and entropy, can maintain a fusion accuracy exceeding 98% under complex conditions such as high SOFC temperatures and fluctuating oxygen partial pressures.

[0073] In one embodiment of the present invention, step S4 of the present invention may further include: outputting the monitoring results, health predictions and abnormal warning signals in a visual form.

[0074] In one embodiment of the present invention, in step S1 of the present invention, operating parameters are collected in real time through a sensor network.

[0075] In one embodiment of the present invention, step S2 of the present invention may specifically include: Step S2-1: Process the collected operating parameters according to the following formula: Where, is the predicted state, F is the state transfer matrix, is the updated state at time k-1, is the prediction covariance, is the updated covariance at time k-1, is the transposed matrix of F, Q is the process noise covariance, is the Kalman gain, H is the observation matrix, is the transposed matrix of H, R is the measurement noise covariance, To update the status, is the observed value, is the updated covariance, I is the identity matrix; Step S2-2: Calculate the information entropy and weight of the collected operating parameters according to the following formula: Where E is information entropy, is the normalized value of the i-th sensor data, is the weight of the i-th sensor data, is the information entropy corresponding to the i-th sensor data, is the information entropy corresponding to the j-th sensor data; Step S2-3: Calculate the fusion value according to the following formula: Where, is the fusion value obtained by fusing multiple sensor data, is the data of the i-th sensor.

[0076] In one embodiment of the present invention, step S3 of the present invention may specifically include: Step S3-1: Use discrete wavelet transform to perform time-frequency analysis on the voltage and internal resistance signals, extract transient features, and calculate the capacity decay rate using statistical methods. The calculation formula of discrete wavelet transform is: Where, is the wavelet coefficient of the jth scale and the kth time shift, is the input signal, is the wavelet basis function, is the approximate coefficient of the jth layer, is the detail coefficient of the jth layer, is the scaling function, is the wavelet function; Step S3-2: Based on a hybrid neural network model and reinforcement learning algorithm, the battery status is monitored in real time, the expected endurance cycle is predicted, and anomalies are identified. The hybrid neural network model is a CNN with 4 convolution layers and 3×3 kernels and a LSTM with 3 layers and 128 units. The CNN is used to extract spatial features of multidimensional data, and the LSTM is used to capture long-term dependencies in time series. The performance degradation index PDI and the predicted expected endurance cycle EEC are output. The calculation formulas for the performance degradation index PDI and the predicted expected endurance cycle EEC are: Where, is the capacity-based performance degradation index, C is the current capacity, is the initial capacity; is the performance degradation index based on environmental factors, is the temperature compensation factor, T is the current temperature, is the oxygen partial pressure influencing factor, is the current oxygen partial pressure; is the corresponding weight coefficient; is the performance failure threshold, is the decay rate, β is the cyclic stress factor caused by temperature, and N is the number of cycles.

[0077] In this embodiment, first, for the hybrid neural network model, its model structure is a hybrid network of CNN and LSTM to form a spatiotemporal feature analysis capability, wherein CNN is used to extract the spatial features of multidimensional data (for example, voltage harmonic components, internal resistance transient response, etc.), and LSTM is used to capture long-term dependencies in time series for analyzing SOFC aging trends.

[0078] Second, the self-correction mechanism uses sliding window online learning to achieve model self-correction. Model weights are updated every several cycles (e.g., 50 times), dynamically adjusting parameters using the latest data. At the same time, a reward-based optimization prediction strategy is combined to improve long-term stability.

[0079] Third, in this implementation, the hybrid neural network model outputs real-time monitoring of the performance degradation index (PDI) and prediction of the expected endurance cycle (EEC). Compared to fixed-parameter AI models (such as traditional CNNs), this method, through CNN-LSTM and online learning, adapts to the full lifecycle of SOFCs, achieving a prediction accuracy of 95%.

[0080] In one embodiment of the present invention, the present invention Can be set to 20%.

[0081] In one embodiment of the present invention, step S3-2 of the present invention may further include: The model weights are updated every several cycles to dynamically adjust the parameters using the latest data, and the prediction strategy is optimized in combination with the reward function to improve long-term stability.

[0082] In one embodiment of the present invention, in step S4 of the present invention, the feedback control signal is used to adjust the charging rate of the SOFC and the battery control parameters.

[0083] According to the second aspect of the present invention, see Figure 1 , further provides an artificial intelligence-based SOFC state monitoring and health diagnosis system for executing the artificial intelligence-based SOFC state monitoring and health diagnosis method according to any one of the technical solutions in the first aspect of the present invention, the artificial intelligence-based SOFC state monitoring and health diagnosis system comprising: Data acquisition module, used to collect SOFC operating parameters in real time; A data fusion unit is used to perform fusion processing on the collected operating parameters; A feature adaptive screening unit is used to extract multi-scale features related to battery health from the fused data and perform dynamic screening; A self-correcting AI module, which uses artificial intelligence to diagnose battery status and predict durability in real time, achieves model self-correction through online learning, and identifies abnormal conditions; The feedback control unit is used to generate a feedback control signal according to the diagnosis result to optimize the battery operating parameters.

[0084] In this embodiment, the artificial intelligence-based SOFC state monitoring and health diagnosis system of the present invention is described in an exemplary embodiment.

[0085] 1. Data acquisition module: (1) Function: Collect multi-dimensional parameters such as temperature and oxygen partial pressure through a multi-sensor network (such as K-type thermocouples and zirconia sensors).

[0086] (2) Hardware: 32-bit ARM processor, 16-bit ADC, etc.

[0087] 2. Data fusion unit: (1) Function: Perform weighted fusion of multi-source data to optimize consistency and noise resistance.

[0088] (2) Implementation: The EKF and information entropy algorithms in the above technical solution can be used.

[0089] 3. Feature Adaptive Screening Unit: (1) Function: Combine time-frequency analysis (such as wavelet transform) and statistical methods to extract multi-scale features and dynamically screen them through importance scoring.

[0090] (2) Output: a feature set that is strongly related to SOFC health.

[0091] 4. Self-correcting AI module: (1) Function: Use AI technology (such as CNN-LSTM) to diagnose the status and predict EEC, and make adaptive adjustments through online learning.

[0092] (2) Implementation: Please refer to the above technical solution.

[0093] 5. Feedback control unit: (1) Function: Convert diagnostic results into control signals (e.g., adjusting charging rate, temperature) to form a closed-loop optimization.

[0094] (2) Output: visualization results and operation strategies.

[0095] In this exemplary embodiment, the system of the present invention integrates multi-source data processing, dynamic feature extraction, self-correcting AI and closed-loop feedback, significantly outperforming traditional single module systems.

[0096] In one embodiment of the present invention, the self-correcting AI module of the present invention uses an improved hybrid neural network and reinforcement learning algorithm to detect battery status, predict durability and identify anomalies in real time. The feedback control unit generates a feedback control signal based on the prediction results to dynamically adjust the battery charging rate and battery control parameters.

[0097] The above are only specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for SOFC state monitoring and health diagnosis based on artificial intelligence, characterized in that: The steps include: Step S1: real-time collection of SOFC operating parameters, including temperature, voltage, current, internal resistance, oxygen partial pressure, and ambient humidity; Step S2: Using an algorithm based on extended Kalman filtering and information entropy to optimize data, the collected operating parameters are weighted to optimize data consistency and noise resistance; Step S3: Based on the improved hybrid neural network model and reinforcement learning algorithm, the battery status is monitored in real time, the expected endurance cycle is predicted, and anomalies are identified. The model parameters are dynamically adjusted through sliding window online learning and self-correction mechanisms. Step S4: combining the monitoring results, health prediction and abnormal warning signals to generate feedback control signals and dynamically adjust the operating parameters of the SOFC to extend the battery life.

2. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 1, characterized in that: The step S4 also includes: outputting the monitoring results, health predictions and abnormal warning signals in a visual form.

3. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 1, characterized in that: In step S1, operating parameters are collected in real time through a sensor network.

4. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 1, characterized in that: The step S2 specifically includes: Step S2-1: Process the collected operating parameters according to the following formula: Where, is the predicted state, F is the state transfer matrix, is the updated state at time k-1, is the prediction covariance, is the updated covariance at time k-1, is the transposed matrix of F, Q is the process noise covariance, is the Kalman gain, H is the observation matrix, is the transposed matrix of H, R is the measurement noise covariance, To update the status, is the observed value, is the updated covariance, I is the identity matrix; Step S2-2: Calculate the information entropy and weight of the collected operating parameters according to the following formula: Where E is information entropy, is the normalized value of the i-th sensor data, is the weight of the i-th sensor data, is the information entropy corresponding to the i-th sensor data, is the information entropy corresponding to the j-th sensor data; Step S2-3: Calculate the fusion value according to the following formula: Where, is the fusion value obtained by fusing multiple sensor data, is the data of the i-th sensor.

5. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 1, characterized in that: The step S3 specifically includes: Step S3-1: Use discrete wavelet transform to perform time-frequency analysis on the voltage and internal resistance signals, extract transient features, and calculate the capacity decay rate using statistical methods. The calculation formula of discrete wavelet transform is: Where, is the wavelet coefficient of the jth scale and the kth time shift, is the input signal, is the wavelet basis function, is the approximate coefficient of the jth layer, is the detail coefficient of the jth layer, is the scaling function, is the wavelet function; Step S3-2: Based on a hybrid neural network model and reinforcement learning algorithm, the battery status is monitored in real time, the expected endurance cycle is predicted, and anomalies are identified. The hybrid neural network model is a CNN with 4 convolution layers and 3×3 kernels and a LSTM with 3 layers and 128 units. The CNN is used to extract spatial features of multidimensional data, and the LSTM is used to capture long-term dependencies in time series. The performance degradation index PDI and the predicted expected endurance cycle EEC are output. The calculation formulas for the performance degradation index PDI and the predicted expected endurance cycle EEC are: Where, is the capacity-based performance degradation index, C is the current capacity, is the initial capacity; is the performance degradation index based on environmental factors, is the temperature compensation factor, T is the current temperature, is the oxygen partial pressure influencing factor, is the current oxygen partial pressure; is the corresponding weight coefficient; is the performance failure threshold, is the decay rate, β is the cyclic stress factor caused by temperature, and N is the number of cycles.

6. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 5, characterized in that: described Set to 20%.

7. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 5, characterized in that: The step S3-2 further includes: The model weights are updated every several cycles to dynamically adjust the parameters using the latest data, and the prediction strategy is optimized in combination with the reward function to improve long-term stability.

8. The artificial intelligence-based SOFC state monitoring and health diagnosis method according to claim 1, characterized in that: In step S4 , the feedback control signal is used to adjust the charging rate and battery control parameters of the SOFC.

9. An artificial intelligence-based SOFC state monitoring and health diagnosis system, characterized in that: The method for performing the artificial intelligence-based SOFC state monitoring and health diagnosis method according to any one of claims 1 to 8, wherein the artificial intelligence-based SOFC state monitoring and health diagnosis system comprises: Data acquisition module, used to collect SOFC operating parameters in real time; A data fusion unit is used to perform fusion processing on the collected operating parameters; A feature adaptive screening unit is used to extract multi-scale features related to battery health from the fused data and perform dynamic screening; A self-correcting AI module, which uses artificial intelligence to diagnose battery status and predict durability in real time, achieves model self-correction through online learning, and identifies abnormal conditions; The feedback control unit is used to generate a feedback control signal based on the diagnosis results to optimize the battery operating parameters.

10. The artificial intelligence-based SOFC state monitoring and health diagnosis system according to claim 9, characterized in that: The self-correcting AI module uses an improved hybrid neural network and reinforcement learning algorithm to detect battery status, predict durability and identify anomalies in real time. The feedback control unit generates a feedback control signal based on the prediction results to dynamically adjust the battery charging rate and battery control parameters.

Images