Vanadium redox flow battery stack assembly leak detection method and system thereof

Abstract

The invention relates to the field of vanadium redox flow batteries, and discloses a vanadium redox flow battery galvanic pile assembly leak hunting method and system, and the method comprises the following steps: obtaining a reference fluid and chemical parameter distribution in a galvanic pile through constructing a multi-level fluid mechanics model and a chemical reaction kinetic model; collecting and monitoring chemical parameters such as pressure, flow, vanadium ion concentration, oxidation-reduction potential and pH value in the galvanic pile in real time; and performing multi-dimensional feature fusion on the acquired data to generate a data matrix, performing abnormal mode recognition on the data matrix by using a deep learning model of a convolutional neural network and a long-short term memory network, and accurately positioning a leakage position. According to the invention, comprehensive monitoring of fluid and chemical parameters in the galvanic pile is realized, detection requirements of tiny leakage can be rapidly responded, high precision and real-time performance are realized, and effective technical support is provided for safe operation of the vanadium redox flow battery.

Description

Technical Field

[0001] The invention relates to the technical field of vanadium liquid flow batteries, and in particular to a vanadium liquid flow battery stack assembly leak detection method and a system thereof. Background Art

[0002] In existing vanadium flow battery technology, the sealing performance of the stack assembly is crucial, because tiny leaks may lead to electrolyte loss, degradation of stack performance, and even cause safety hazards. Existing leak detection methods usually rely on single monitoring of fluid parameters or simple sensor detection, and lack in-depth analysis of the flow characteristics, pressure changes, and abnormal fluctuations in chemical composition of the fluid in different areas of the stack, resulting in low accuracy of leak detection. At the same time, existing methods are difficult to effectively capture the changes in chemical parameters in the electrolyte of vanadium flow batteries, making it difficult to identify changes in electrolyte composition in a timely manner, affecting the sensitivity of leak monitoring.

[0003] In addition, the existing technologies have limited intelligence in data analysis. Most of them rely only on basic data comparison and lack in-depth mining of data time series and multi-dimensional features, making it difficult to identify small leaks in complex working conditions in real time. Since vanadium flow battery stacks are affected by temperature, pressure and other factors in actual operation, traditional methods are often difficult to adapt to the leakage detection needs under different operating conditions, resulting in poor reliability of leakage detection. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a vanadium liquid flow battery stack assembly leak detection method and system, which solves the problems in the prior art of low accuracy, untimely response and inability to effectively monitor changes in the chemical composition of the electrolyte in vanadium liquid flow battery stack leakage detection.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A vanadium liquid flow battery stack assembly leak detection method, comprising the following steps:

[0006] A multi-level fluid mechanics model is established based on the overall fluid flow state of the vanadium liquid flow battery stack, wherein the fluid mechanics model includes fluid models of macroscopic layer, mesoscopic layer and microscopic layer;

[0007] Obtain the baseline pressure, velocity and flow rate data of the electrolyte at different locations through numerical simulation;

[0008] Collect pressure and flow data at key locations of the fuel cell stack in real time, and compare the real-time data with the baseline data to detect whether there are abnormal pressure or flow fluctuations;

[0009] When abnormal fluctuations are detected, the chemical parameters in the electrolyte are monitored in real time, and the impact of leakage is analyzed through chemical reaction kinetic models;

[0010] Perform multi-dimensional feature fusion on the collected fluid and chemical parameters to form a multi-dimensional data matrix;

[0011] A deep learning model is used to perform abnormal pattern recognition on the multi-dimensional data matrix to determine possible leakage locations.

[0012] Preferably, the macroscopic layer fluid model describes the relationship between the pressure, velocity and height of the fluid at different positions based on the energy conservation principle of the fluid, and determines the pressure and velocity distribution at different positions by Bernoulli's theorem.

[0013] Preferably, the fluid model of the mesoscopic layer is based on the permeability of the fluid in the porous medium, and Darcy's law is used to model the permeation behavior at the seals and interfaces of the stack to estimate the local permeation flow.

[0014] Preferably, the fluid model of the microscopic layer assumes that the leakage path is a tiny capillary, and describes the flow rate and permeation flow of the electrolyte in the capillary or microcrack based on the Poiseuille flow formula.

[0015] Preferably, the step of obtaining the reference pressure, velocity and flow data of the electrolyte at different positions by numerical simulation comprises:

[0016] The fluid mechanics model is solved by numerical simulation software. First, a computational grid is divided in the vanadium liquid flow battery stack, and the entire fluid region is discretized based on the density, viscosity, inlet and outlet boundary conditions of the fluid.

[0017] At the macroscopic level of the battery stack, the energy conservation equation based on fluid density, velocity and height is solved, and the distribution relationship between pressure and flow velocity is established using the Bernoulli theorem satisfied by electrolyte density, flow velocity and pressure at different positions of the battery stack, and the overall flow velocity and pressure distribution of the electrolyte at different positions are calculated;

[0018] At the mesoscopic level, the Darcy's law model is used to calculate the local permeation flow at the seal or interface, and the pressure and flow distribution in the seal or interface area are obtained based on the calculation of the permeability coefficient, fluid viscosity and flow path length;

[0019] At the microscopic level, assuming that the leakage path has the characteristics of a capillary structure, the flow velocity and permeation flow in the microcracks are calculated using the Poiseuille flow formula, and the benchmark flow velocity and flow distribution in the microscopic flow area are established;

[0020] The flow results at each level are integrated to obtain the benchmark fluid pressure, velocity and flow data distribution of the vanadium liquid flow battery stack under normal operating conditions, and the benchmark data is stored in a database for subsequent comparison and analysis with real-time monitoring data.

[0021] Preferably, the step of monitoring the chemical parameters in the electrolyte in real time and analyzing the impact of leakage through a chemical reaction kinetic model when abnormal fluctuations are detected includes:

[0022] The concentration of divalent vanadium ions and trivalent ions in the electrolyte is measured using a vanadium ion selective electrode, and the initial baseline concentration is recorded;

[0023] Establishing a differential equation for concentration change according to a kinetic model of vanadium ion redox reaction, wherein the kinetic model includes oxidation and reduction reaction rate constants;

[0024] The rate constants in the redox kinetic model were used to calculate the rate of change of vanadium ion concentration at different temperatures, where the temperature dependence of the rate constants was described by the Arrhenius equation.

[0025] By real-time monitoring of the changing trend of vanadium ion concentration, abnormal concentration fluctuations can be identified and the impact of leakage can be determined;

[0026] The detected vanadium ion concentration fluctuation data is input into the computing system to determine the significance of the concentration change, so as to identify the concentration change trend caused by the leakage and determine the type of leakage component.

[0027] Preferably, the step of monitoring the chemical parameters in the electrolyte in real time and analyzing the impact of leakage through a chemical reaction kinetic model when abnormal fluctuations are detected further comprises:

[0028] Acquiring redox potential data of the electrolyte in real time, the data being collected by a redox potential sensor;

[0029] Based on the concentration ratio of divalent vanadium ions and trivalent vanadium ions, the redox potential of the electrolyte is calculated by the Nernst equation, wherein the potential calculation in the Nernst equation includes the logarithmic relationship between the standard electrode potential at room temperature and the vanadium ion concentration;

[0030] When it is detected that the redox potential fluctuation exceeds a preset threshold, the fluctuation data is recorded as an abnormal fluctuation, and it is determined whether the fluctuation indicates leakage of the vanadium electrolyte;

[0031] The abnormally fluctuating redox potential data is collected and trend analyzed multiple times to confirm whether the fluctuation of the redox potential continues to be stable within the abnormal range, so as to further determine the changes in the chemical composition of the leakage location.

[0032] Preferably, the step of performing multi-dimensional feature fusion on the collected fluid and chemical parameters to form a multi-dimensional data matrix includes:

[0033] The collected real-time pressure, flow rate data and electrolyte chemical parameter data are segmented according to time series to form a basic data set with time series characteristics, and the data set is normalized;

[0034] Apply convolution operation to the normalized data to extract the spatial feature matrix to obtain the feature distribution at different positions and chemical states;

[0035] A long short-term memory network is used to process the temporal changes of the feature matrix to form a multidimensional data matrix integrating temporal features.

[0036] Preferably, the step of using a deep learning model to perform abnormal pattern recognition on the multidimensional data matrix to determine possible leakage locations includes:

[0037] The multidimensional data matrix is ​​input into the convolution layer of the deep learning model, and the input data is convolved through multiple convolution kernels to generate the initial feature map and perform pooling operation;

[0038] Use the long short-term memory network to perform time series processing on the pooled feature map to form a time series feature vector;

[0039] The time series feature vector is processed using the classification layer to generate the leakage probability distribution of each location;

[0040] The most likely leakage location is determined based on the leakage probability distribution of each location, and a leakage warning signal is issued when the leakage probability exceeds a preset threshold.

[0041] The present invention also provides a vanadium liquid flow battery stack assembly leak detection system, comprising:

[0042] Fluid mechanics modeling module, used to establish a multi-level fluid model of the vanadium liquid flow battery stack and describe the flow characteristics of the electrolyte;

[0043] A numerical simulation module, used to simulate the normal state of the fluid inside the fuel cell stack based on the fluid mechanics model and generate reference fluid data;

[0044] The sensor module includes multiple pressure, flow and chemical sensors for real-time collection of stack fluid, vanadium ion concentration, redox potential and pH data;

[0045] A data processing module is used to compare the real-time monitoring data with the baseline data and input the chemical data of the abnormal fluctuation location into the deep learning model;

[0046] A deep learning module is used to analyze multi-dimensional feature data matrices and perform leak detection and location through convolutional neural networks and long short-term memory networks.

[0047] The present invention provides a vanadium liquid flow battery stack assembly leak detection method and system thereof, which has the following beneficial effects:

[0048] 1. By constructing a multi-level fluid mechanics model and a chemical reaction kinetics model, the present invention can comprehensively describe the fluid and chemical behavior inside the battery stack at the macro, meso and micro levels, ensure sensitive capture of tiny leaks in the vanadium liquid flow battery stack, and significantly improve detection accuracy.

[0049] 2. The method of the present invention utilizes multi-point sensors to collect pressure, flow and chemical parameters in real time, and combines them with numerical simulation data for comparative analysis, thereby realizing real-time monitoring of leakage conditions, ensuring timely identification and response at the early stage of leakage, and helping to prevent further expansion of leakage risks.

[0050] 3. The present invention generates a comprehensive multidimensional data matrix by fusing the multidimensional features of fluid and chemical parameters, which helps the deep learning model to extract spatial and temporal features from complex data and achieve highly robust detection without being disturbed by fluctuations or anomalies in a single data.

[0051] 4. The present invention adopts the deep learning model of convolutional neural network and long short-term memory network to jointly analyze the spatial and temporal characteristics of the data matrix, which can accurately identify the possible location of the leak, thereby improving the accuracy of positioning and providing a direct reference for leak repair.

[0052] 5. The present invention can effectively reduce the safety risks caused by leakage, extend the service life of the battery stack, improve the operating stability of the battery system, and provide technical guarantee for the safe application of vanadium liquid flow batteries through early detection and accurate positioning of the leakage. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0054] Figure 2 It is a schematic diagram of the system structure of the present invention.

[0055] Among them, 100, fluid mechanics modeling module; 200, numerical simulation module; 300, sensor module; 400, data processing module; 500, deep learning module. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Please see attached Figure 1 The embodiment of the present invention provides a vanadium liquid flow battery stack assembly leak detection method, comprising the following steps:

[0058] S1. Establish a multi-level fluid mechanics model based on the overall fluid flow state of the vanadium liquid flow battery stack

[0059] Firstly, the vanadium liquid flow battery stack assembly leak detection method of the present invention is based on the flow state of the electrolyte inside the vanadium liquid flow battery stack, and establishes a multi-level fluid mechanics model so as to systematically describe the flow characteristics of the electrolyte inside the stack from multiple scales such as macroscopic layer, mesoscopic layer and microscopic layer, thereby providing accurate benchmark data for subsequent abnormality detection.

[0060] In this embodiment, a macroscopic fluid model is first constructed to describe the overall flow state of the electrolyte inside the vanadium liquid flow battery stack. The macroscopic model is based on the principle of conservation of energy, regards the flow of the electrolyte as a steady-state incompressible fluid flow, and assumes that the electrolyte in the stack follows Bernoulli's theorem. Therefore, in this macroscopic model, there is a constant energy conservation relationship between the fluid pressure, velocity, and height at different positions. Specifically, according to Bernoulli's theorem, the total energy of the fluid remains unchanged, which can be described as:

[0061]

[0062] Among them, g is the fluid pressure, ρ is the electrolyte density, v is the flow rate, p is the gravitational acceleration, and h is the height. In the specific calculation, the fluid pressure, velocity and height at different positions inside the vanadium liquid flow battery stack are substituted into the Bernoulli equation to obtain the fluid parameter distribution at each position. As an option, the macroscopic layer model is solved using numerical simulation software (such as ANSYS Fluent or COMSOL Multiphysics) to obtain the overall flow rate and pressure distribution of the electrolyte in the stack.

[0063] In this embodiment, a fluid model of the mesoscopic layer is further established to describe the local permeation behavior of the stack seals and interfaces. In general, the flow at the seals and interfaces is closely related to the permeability of the fluid, so Darcy's law is used at the mesoscopic level to model the permeation behavior of the seals and interfaces. Specifically, assuming that there may be small-scale leakage near the seal, Darcy's law is used to evaluate the fluid distribution in the seal area by calculating the permeation flow rate. Darcy's law is as follows:

[0064]

[0065] Among them, Q is the permeation flow rate, j is the permeability coefficient, A is the flow cross-sectional area, μ is the viscosity of the fluid, Δp is the pressure difference between the two ends, and L is the flow path length. Through Darcy's law, the fluid model of the seal and interface of the vanadium liquid flow battery stack is established to obtain the fluid pressure and flow distribution in these areas. As an option, appropriate boundary conditions are applied in the numerical simulation to simulate the possible leakage of the seal and accurately obtain the flow and pressure data in these areas.

[0066] At the microscopic level of this embodiment, it is assumed that there are small cracks or capillary structures in the battery stack to describe the flow in the capillaries or microcracks. The fluid model at the microscopic level is based on the Poiseuille flow equation, assuming that the leakage path is a long and thin capillary structure, and the electrolyte behaves as a laminar flow in the structure, which is suitable for the description of Poiseuille's law. Specifically, the Poiseuille flow formula is as follows:

[0067]

[0068] Among them, Q capillary is the flow rate in the capillary, r is the capillary radius, Δp is the pressure difference at both ends, μ is the viscosity of the fluid, and L is the length of the capillary. The Poiseuille flow formula is used to calculate the flow distribution in microcracks or capillary structures and obtain the permeation flow rate in the microscopic layer. In general, the flow rate of the capillary structure is subject to the channel radius and the viscosity of the fluid. After determining the geometric parameters of the capillary channel, the permeation flow rate can be obtained through the Poiseuille equation.

[0069] As an option, in some embodiments, the three-layer model is integrated to perform a systematic analysis of the fluid flow conditions inside the entire stack. Specifically, the overall results of the three-layer fluid model can be summarized through numerical simulation software to obtain the overall flow velocity, pressure and permeation flow distribution data of the vanadium liquid flow battery stack as the benchmark data in the leak detection method.

[0070] S2. Obtain the reference pressure, velocity and flow rate data of the electrolyte at different locations through numerical simulation

[0071] In order to accurately obtain the fluid state of the vanadium flow battery stack under normal working conditions, the present invention calculates the reference pressure, velocity and flow data of each position through numerical simulation method after constructing a multi-level fluid mechanics model. These reference data will be used as the flow characteristics of the electrolyte during normal operation, providing a reference for the comparison of subsequent real-time data, so as to identify abnormal pressure or flow fluctuations.

[0072] In this embodiment, meshing and parameter initialization are first performed in the fluid region inside the battery stack, with the purpose of discretizing the fluid space of the electrolyte and facilitating subsequent numerical solutions. Specifically, multiple computational grids are divided in the flow region of the electrolyte, where the grid density varies according to the regional characteristics. In general, the flow characteristics of the electrolyte are relatively complex around the seal, in the interface region, and in the capillary structure, so a denser mesh is used in these locations to ensure the calculation accuracy. As an option, a moderately dense meshing scheme can be selected for the overall flow region of the battery stack to strike a balance between accuracy and calculation efficiency.

[0073] After the grid division and parameter initialization are completed, specific boundary conditions are set according to the fluid properties of the electrolyte and the operating conditions of the battery stack. In general, the inlet is set to a known pressure and flow rate boundary condition, and the outlet is a known pressure boundary. Therefore, in this embodiment, the inlet pressure and flow rate are defined as constant values, and normal pressure is set as a boundary condition at the outlet. For the seals and interface areas of the battery stack, the permeability and viscosity of the fluid must also be considered when setting the boundary conditions to ensure accurate simulation of the permeation flow.

[0074] In this embodiment, the multi-level fluid mechanics model is numerically solved by numerical simulation software to calculate the pressure, velocity and flow distribution of the electrolyte at different positions in the battery stack. At the macro level, the fluid model is based on the energy conservation principle of Bernoulli's theorem, assuming that the electrolyte is a steady-state incompressible fluid and satisfies the energy conservation relationship under normal operating conditions. Specifically, in the numerical simulation, according to the Bernoulli equation, the pressure, velocity and height of the electrolyte at each position are substituted into the solution to obtain the overall flow rate and pressure distribution. As an option, the finite volume method or finite element method is used to discretize the model to improve the convergence and accuracy of the simulation results.

[0075] At the mesoscopic level, the fluid model is based on Darcy's law and is applicable to the behavior of the permeating fluid at the seals and interfaces. In general, the permeation behavior in the seal area can be described by the permeation flow rate, which in turn depends on the permeability coefficient, the flow cross-sectional area, and the viscosity of the fluid. Therefore, in the numerical simulation, the permeation parameters of the seal are set according to Darcy's law in order to accurately solve the permeation flow rate and pressure distribution. Specifically, the permeability coefficient in Darcy's law is a material characteristic parameter, which can be set according to the flow characteristics of the electrolyte near the seal to ensure that the calculation accurately reflects the fluid permeation state in the sealing area.

[0076] At the microscopic level, the fluid model assumes that the flow of electrolyte in microcracks or capillaries is laminar, which meets the Poiseuille flow condition. Specifically, the flow in microcracks or capillaries is calculated according to the Poiseuille flow equation, and the radius of the capillary channel and the viscosity parameters of the fluid are set in the simulation to ensure that the calculated flow distribution is consistent with the actual situation. The Poiseuille flow formula is suitable for describing the laminar flow behavior of fluids in capillaries or microcracks, and can accurately simulate the flow characteristics at this level.

[0077] In some embodiments, in order to improve the accuracy of simulation data, boundary constraints of different fluid characteristic regions are considered in numerical simulation. Specifically, a no-slip boundary condition is used for the macroscopic layer model to ensure that the electrolyte flow rate at the boundary is zero; for the permeation model of the mesoscopic layer, an appropriate pressure gradient is applied to reflect the fluid permeation behavior at the seal and interface; for the capillary flow model of the microscopic layer, a capillary pressure difference boundary condition is set to ensure that the fluid generates laminar permeation flow along the direction of the microcracks.

[0078] As an option, the pressure, flow and velocity data obtained from the numerical simulation can be stored in a database to establish a benchmark fluid parameter data set for the vanadium liquid flow battery stack, so as to facilitate real-time data comparison and anomaly detection in the subsequent leak detection process.

[0079] S3. Collect pressure and flow data at key locations of the battery stack in real time, and compare the real-time data with the baseline data to detect abnormal pressure or flow fluctuations.

[0080] In order to achieve real-time monitoring of the operating status of the vanadium liquid flow battery stack, the method of the present invention collects the pressure and flow data of each key position of the stack through sensors in real time on the basis of the benchmark fluid parameter data set obtained by numerical simulation, and compares the real-time data with the benchmark data. This step is intended to quickly identify whether there is abnormal pressure or flow fluctuation inside the stack, so as to determine whether there is a leak. By real-time monitoring and comparing the benchmark data, abnormal conditions of the fluid state can be detected in time, providing a preliminary judgment basis for subsequent chemical parameter monitoring.

[0081] In this embodiment, high-precision pressure and flow sensors are arranged at key positions of the battery stack to collect the flow parameters of the electrolyte in real time. Specifically, the sensors are arranged at the inlet, outlet, near the seals and other nodes where there may be leakage risks. In general, arranging sensors in key areas of the fluid flow path helps to more accurately collect real-time change data of the fluid. Therefore, in this embodiment, nodes with significant flow changes and susceptible to leakage are selected for key arrangement to ensure the integrity and representativeness of real-time data.

[0082] In this embodiment, the real-time data collected by the sensor includes the pressure and flow information of each position. Generally, the sensor needs to have sufficient sampling accuracy and response speed to ensure rapid response to changes in fluid parameters, so high-precision and high-sensitivity pressure and flow sensors are used in sensor selection. As an option, the collected real-time data is compared with the benchmark data stored in the database after data preprocessing to identify whether there are abnormal fluctuations.

[0083] When performing data comparison, the real-time collected fluid parameters are compared one by one with the upper and lower limits of the benchmark data. If the pressure or flow at a certain location exceeds the preset tolerance range, it is determined that there is an abnormal fluctuation at that location. Specifically, the tolerance range of pressure fluctuations is set based on the pressure distribution of the normal flow of the electrolyte and the standard deviation of the flow rate to ensure tolerance for slight fluctuations under normal conditions, but sensitivity to significant fluctuations that exceed the benchmark state. As an option, the real-time pressure signal is spectrally analyzed using data analysis methods such as fast Fourier transform (FFT) or wavelet transform to extract fluctuation characteristics and further verify the abnormal pressure fluctuation at that location.

[0084] In this embodiment, when a significant fluctuation in pressure or flow is detected, the location is preliminarily determined to be a possible leakage area, and the corresponding monitoring data is recorded. In general, the comparison result of real-time data and baseline data will indicate the possible location of the leakage, so every time an abnormal fluctuation is detected, the real-time data of the location is stored for subsequent chemical parameter monitoring and detailed analysis.

[0085] S4. When abnormal fluctuations are detected, the chemical parameters in the electrolyte are monitored in real time, and the impact of leakage is analyzed through the chemical reaction kinetic model.

[0086] In order to further confirm the leakage in the vanadium flow battery stack, the present invention uses a chemical sensor to monitor the chemical parameters in the electrolyte in real time after detecting abnormal pressure or flow fluctuations, and analyzes the impact of leakage on the electrolyte composition in combination with a chemical reaction kinetic model. Through this step, the dynamic changes of key chemical parameters such as vanadium ion concentration, redox potential and pH value in the electrolyte can be obtained, providing an important basis for subsequent leakage component analysis and location determination.

[0087] In this embodiment, firstly, the real-time monitoring of the key chemical parameters of the electrolyte is realized by arranging chemical sensors. Specifically, the vanadium ion concentration, oxidation-reduction potential (ORP) and pH value in the electrolyte are collected by vanadium ion selective electrodes, ORP sensors and pH sensors respectively. Generally, these sensors are arranged near the inlet and outlet of the electrolyte and the seals to monitor the changes in chemical parameters in the area where abnormal fluid fluctuations are detected. Therefore, in this embodiment, chemical sensors are arranged at key nodes on the electrolyte flow path to ensure comprehensive monitoring of the impact of leakage.

[0088] In this embodiment, the monitored chemical parameter changes are analyzed by a chemical reaction kinetic model, especially the concentration change trend of vanadium ions is calculated in real time. Specifically, the redox reaction between vanadium ions in different valence states is regarded as a first-order reaction, and its concentration change rate satisfies the chemical kinetic equation. It is assumed that the concentration change between divalent and trivalent vanadium ions conforms to the following reaction rate equation:

[0089]

[0090] Among them, k ox is the oxidation reaction rate constant, k red is the reduction reaction rate constant. In some embodiments, the reaction rate constant also needs to be corrected according to temperature to ensure the accuracy of concentration changes under different temperature conditions. In general, the relationship between the rate constant and temperature conforms to the Arrhenius equation, so in this model, the rate constant is corrected according to the temperature change to obtain more accurate kinetic model parameters.

[0091] As an option, this embodiment also calculates the redox potential of the electrolyte based on the Nernst equation, and then determines whether the vanadium ion concentration ratio fluctuates significantly. Specifically, the Nernst equation describes the relationship between the ratio of the concentration of divalent and trivalent vanadium ions and the redox potential:

[0092]

[0093] Where E is the actual redox potential, E 0 is the standard electrode potential, R is the gas constant, T is the temperature, n is the number of electron transfers, and F is the Faraday constant. In general, when the concentration ratio of vanadium ions fluctuates due to leakage, the redox potential will change significantly. Therefore, this embodiment monitors the change trend of the redox potential in real time to determine whether the ratio of vanadium ions in the electrolyte deviates from the normal range.

[0094] In some embodiments, in order to improve the detection accuracy, the present embodiment also monitors the pH value as a supplementary chemical parameter. The change in the pH value of the vanadium flow battery can reflect the acid-base balance state of the electrolyte. When the leakage affects the composition of the electrolyte, the pH value may fluctuate abnormally. Specifically, the change in pH value can be associated with the redox reaction of vanadium ions. By comprehensively analyzing the changing trends of vanadium ion concentration, redox potential and pH value, the influence of the chemical composition of the leakage can be further confirmed.

[0095] In summary, this embodiment can effectively track the changes in leakage components through real-time monitoring of chemical parameters and analysis of kinetic models, and provide detailed chemical characteristic data for subsequent leakage location determination.

[0096] S5. Perform multi-dimensional feature fusion on the collected fluid and chemical parameters to form a multi-dimensional data matrix

[0097] In order to improve the accuracy of vanadium flow battery stack leakage detection, the method of the present invention generates a multidimensional data matrix through feature data fusion based on the acquisition of fluid and chemical parameters to provide complete data input for the deep learning model. This step provides high-quality data support for subsequent leakage anomaly pattern recognition by multi-dimensionally fusing the spatial and temporal characteristics of fluid and chemical parameters.

[0098] In this embodiment, the parameter data such as pressure, flow, vanadium ion concentration, redox potential and pH value collected in real time are first preprocessed to ensure the numerical consistency between various dimensions. In general, each parameter data is standardized to eliminate the influence of different dimensions, so this embodiment adopts a standardization method of subtracting the mean and dividing by the standard deviation to ensure that each parameter data is analyzed on the same scale.

[0099] In this embodiment, spatial feature extraction is performed on the standardized multidimensional data matrix. Specifically, the preprocessed data is input into the convolution layer of the convolutional neural network (CNN) model, and the data is convolved using different convolution kernels to capture the spatial features of each node position. The size of the convolution kernel is set according to the distribution characteristics of different parameters to ensure the accuracy of feature extraction. For example, for flow and pressure data with large variations, a larger convolution kernel can be used to extract significant variation features in these data. The initial feature map after convolution contains the spatial feature information of each key position.

[0100] After obtaining the initial feature map, the pooling operation is used in this embodiment to reduce the dimension of the feature map. Specifically, the representative data points in the convolution region are extracted by the pooling operation to retain key features and reduce data redundancy. In general, the pooling operation can enhance the robustness of the feature map, so the maximum pooling method is used in this embodiment to improve the recognition ability of significant features.

[0101] After completing the time series processing, the time series feature vectors are fused with the spatial features to form a multidimensional data matrix. Specifically, the fused multidimensional data matrix contains the feature information of each position in space and time, and the matrix form is X = {p, Q, [V 2+ ],[V 3+ ], pH, E}, where p represents pressure, Q represents flow rate, [V 2+ ] and [V 3+ ] represent the concentrations of divalent and trivalent vanadium ions, pH represents the acidity and alkalinity, and E represents the redox potential. In general, through this multi-dimensional feature fusion, the model can fully grasp the fluid and chemical states at different locations.

[0102] In some embodiments, the generated multi-dimensional data matrix can be further stored as reference data for long-term monitoring and trend analysis of different working conditions of the fuel cell stack to ensure the real-time monitoring effect of the system.

[0103] S6. Use deep learning models to identify abnormal patterns in multidimensional data matrices to determine possible leak locations

[0104] In order to accurately determine the potential leakage location of the vanadium flow battery stack, the method of the present invention uses a deep learning model to perform abnormal pattern recognition based on the generated multidimensional data matrix to determine the possible leakage location. Through this step, the system can comprehensively analyze the fluid and chemical parameter data at each position in the stack, and accurately locate the leakage location based on the abnormal pattern recognition results. The design of the deep learning model is based on a combination of convolutional neural networks (CNN) and long short-term memory networks (LSTM) to extract spatial and temporal features in the data matrix.

[0105] In this embodiment, the multidimensional data matrix is ​​input into the convolution layer of the deep learning model, and the matrix data is convolved by the convolution kernel to extract the spatial features between different positions. In general, the size and step size of the convolution kernel are set according to different data features to ensure the accuracy of feature extraction. Therefore, in this embodiment, for chemical parameters with dense spatial feature distribution (such as vanadium ion concentration and redox potential), a smaller convolution kernel size is used to accurately capture tiny feature differences. The convolution operation generates an initial feature map, which contains the feature distribution information of each position of the battery stack in space.

[0106] In this embodiment, the convolutional feature map is pooled to reduce the size of the feature map and reduce the amount of calculation. Specifically, the pooling operation retains the main features and reduces redundant information by extracting the extreme value or average value of the local area. As an option, the maximum pooling method is used to extract the maximum value from each convolution area to form a pooled feature map with significant features. In general, the pooling operation helps to enhance the robustness of the features, thereby improving the accuracy of the model in extracting spatial features.

[0107] Based on the pooled feature map, a long short-term memory network (LSTM) is used to perform time series processing on the feature map to capture the time variation pattern of fluid and chemical parameter data. Specifically, the pooled feature map is input into the LSTM layer, and the temporal dependency of each position is analyzed through the LSTM unit. The structure of the LSTM unit can process the long and short time variations of historical data, so that the model can adapt to the time dependency of fluid and chemical parameters at different positions in the stack. As an option, the memory step of the LSTM layer is set to the periodic variation time of the stack fluid to ensure that the time characteristics of periodic fluctuations are accurately captured.

[0108] As an option, the hyperparameters of the deep learning model are dynamically adjusted through the Bayesian optimization algorithm to further improve the accuracy of abnormal pattern recognition. Bayesian optimization iteratively updates the hyperparameters according to the detection effect of the model, so that parameters such as convolution kernel size, step size, pooling window, and LSTM memory step size are continuously adjusted during the optimization process to ensure that the model has the best adaptability under various data characteristics.

[0109] In this embodiment, the leakage probability of each position is calculated by a deep learning model. When the leakage probability of a certain position exceeds a preset threshold, the system will determine that there is a leakage at that position and send out an alarm signal. In general, the threshold of the leakage probability is set according to experimental data to ensure sufficient sensitivity to the actual leakage location. In summary, this embodiment achieves accurate positioning of the leakage position of the vanadium liquid flow battery stack through multi-layer convolution, pooling and time series analysis, and can effectively judge the potential leakage area based on the distribution of the leakage probability.

[0110] The present invention constructs a multi-level fluid mechanics model and a chemical reaction kinetics model to collect and monitor the fluid and chemical parameters inside the battery stack in real time, establishes benchmark data through numerical simulation, and uses a deep learning model to identify abnormal patterns in the multidimensional data matrix to accurately locate potential leakage locations. This method combines multi-dimensional data fusion and intelligent analysis technology of fluid and chemical characteristics, and can detect tiny leaks in vanadium flow batteries with high precision, providing effective technical support for the safe operation of the battery stack.

[0111] The vanadium liquid flow battery stack assembly leak detection system described below and the vanadium liquid flow battery stack assembly leak detection method described above can be referenced to each other.

[0112] Please see attached Figure 2 The present invention also provides a vanadium liquid flow battery stack assembly leak detection system, comprising:

[0113] A fluid mechanics modeling module 100 is used to establish a multi-level fluid model of a vanadium liquid flow battery stack to describe the flow characteristics of the electrolyte;

[0114] A numerical simulation module 200, for simulating the normal state of the fluid inside the fuel cell stack based on the fluid mechanics model and generating reference fluid data;

[0115] The sensor module 300 includes a plurality of pressure, flow and chemical sensors for collecting the fluid, vanadium ion concentration, redox potential and pH data of the stack in real time;

[0116] Data processing module 400, for comparing real-time monitoring data with baseline data and inputting chemical data at abnormal fluctuation locations into a deep learning model;

[0117] The deep learning module 500 is used to analyze the multi-dimensional feature data matrix and perform leak detection and positioning through convolutional neural networks and long short-term memory networks.

[0118] The system of this embodiment can be used to execute the above method embodiments, and its principles and technical effects are similar, which will not be repeated here.

[0119] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for leak detection in vanadium flow battery stack assembly, characterized in that: The following steps are involved: A multi-level fluid mechanics model is established based on the overall fluid flow state of the vanadium liquid flow battery stack, wherein the fluid mechanics model includes fluid models of macroscopic layer, mesoscopic layer and microscopic layer; Obtain the baseline pressure, velocity and flow rate data of the electrolyte at different locations through numerical simulation; Collect pressure and flow data at key locations of the fuel cell stack in real time, and compare the real-time data with the baseline data to detect whether there are abnormal pressure or flow fluctuations; When abnormal fluctuations are detected, the chemical parameters in the electrolyte are monitored in real time, and the impact of leakage is analyzed through chemical reaction kinetic models; Perform multi-dimensional feature fusion on the collected fluid and chemical parameters to form a multi-dimensional data matrix; A deep learning model is used to perform abnormal pattern recognition on the multi-dimensional data matrix to determine possible leakage locations.

2. A vanadium liquid flow battery stack assembly leak detection method according to claim 1, characterized in that: The macroscopic layer fluid model describes the relationship between the pressure, velocity and height of the fluid at different positions based on the energy conservation principle of the fluid, and determines the pressure and velocity distribution at different positions through Bernoulli's theorem.

3. A vanadium liquid flow battery stack assembly leak detection method according to claim 1, characterized in that: The fluid model of the mesoscopic layer is based on the permeability of the fluid in the porous medium. The permeation behavior at the seals and interfaces of the stack is modeled through Darcy's law to estimate the local permeation flow.

4. A vanadium liquid flow battery stack assembly leak detection method according to claim 1, characterized in that: The fluid model of the microscopic layer assumes that the leakage path is a fine capillary, and describes the flow rate and permeation flow of the electrolyte in the capillary or microcrack based on the Poiseuille flow formula.

5. A vanadium redox flow battery stack assembly leak detection method according to claim 1, characterized in that: The step of obtaining the reference pressure, velocity and flow rate data of the electrolyte at different positions by numerical simulation comprises: The fluid mechanics model is solved by numerical simulation software. First, a computational grid is divided in the vanadium liquid flow battery stack, and the entire fluid region is discretized based on the density, viscosity, inlet and outlet boundary conditions of the fluid. At the macroscopic level of the battery stack, the energy conservation equation based on fluid density, velocity and height is solved, and the distribution relationship between pressure and flow velocity is established using the Bernoulli theorem satisfied by electrolyte density, flow velocity and pressure at different positions of the battery stack, and the overall flow velocity and pressure distribution of the electrolyte at different positions are calculated; At the mesoscopic level, the Darcy's law model is used to calculate the local permeation flow at the seal or interface, and the pressure and flow distribution in the seal or interface area are obtained based on the calculation of the permeability coefficient, fluid viscosity and flow path length; At the microscopic level, assuming that the leakage path has the characteristics of a capillary structure, the flow velocity and permeation flow in the microcracks are calculated using the Poiseuille flow formula, and the benchmark flow velocity and flow distribution in the microscopic flow area are established; The flow results at each level are integrated to obtain the benchmark fluid pressure, velocity and flow data distribution of the vanadium liquid flow battery stack under normal operating conditions, and the benchmark data is stored in a database for subsequent comparison and analysis with real-time monitoring data.

6. A vanadium liquid flow battery stack assembly leak detection method according to claim 1, characterized in that: The step of monitoring the chemical parameters in the electrolyte in real time and analyzing the impact of leakage through a chemical reaction kinetic model when abnormal fluctuations are detected includes: The concentration of divalent vanadium ions and trivalent ions in the electrolyte is measured using a vanadium ion selective electrode, and the initial baseline concentration is recorded; Establishing a differential equation for concentration change according to a kinetic model of vanadium ion redox reaction, wherein the kinetic model includes oxidation and reduction reaction rate constants; The rate constants in the redox kinetic model were used to calculate the rate of change of vanadium ion concentration at different temperatures, where the temperature dependence of the rate constants was described by the Arrhenius equation. By real-time monitoring of the changing trend of vanadium ion concentration, abnormal concentration fluctuations can be identified and the impact of leakage can be determined; The detected vanadium ion concentration fluctuation data is input into the computing system to determine the significance of the concentration change, so as to identify the concentration change trend caused by the leakage and determine the type of leakage component.

7. A vanadium liquid flow battery stack assembly leak detection method according to claim 6, characterized in that: The step of monitoring the chemical parameters in the electrolyte in real time and analyzing the impact of leakage through a chemical reaction kinetic model when abnormal fluctuations are detected also includes: Acquiring redox potential data of the electrolyte in real time, the data being collected by a redox potential sensor; Based on the concentration ratio of divalent vanadium ions and trivalent vanadium ions, the redox potential of the electrolyte is calculated by the Nernst equation, wherein the potential calculation in the Nernst equation includes the logarithmic relationship between the standard electrode potential at room temperature and the vanadium ion concentration; When it is detected that the redox potential fluctuation exceeds a preset threshold, the fluctuation data is recorded as an abnormal fluctuation, and it is determined whether the fluctuation indicates leakage of the vanadium electrolyte; The abnormally fluctuating redox potential data is collected and trend analyzed multiple times to confirm whether the fluctuation of the redox potential continues to be stable within the abnormal range, so as to further determine the changes in the chemical composition of the leakage location.

8. A vanadium redox flow battery stack assembly leak detection method according to claim 1, characterized in that: The step of fusing the collected fluid and chemical parameters in multiple dimensions to form a multidimensional data matrix includes: The collected real-time pressure, flow rate data and electrolyte chemical parameter data are segmented according to time series to form a basic data set with time series characteristics, and the data set is normalized; Apply convolution operation to the normalized data to extract the spatial feature matrix to obtain the feature distribution at different positions and chemical states; A long short-term memory network is used to process the temporal changes of the feature matrix to form a multidimensional data matrix integrating temporal features.

9. A vanadium redox flow battery stack assembly leak detection method according to claim 1, characterized in that: It is characterized in that The step of using a deep learning model to perform abnormal pattern recognition on the multidimensional data matrix to determine a possible leakage location includes: The multidimensional data matrix is ​​input into the convolution layer of the deep learning model, and the input data is convolved through multiple convolution kernels to generate the initial feature map and perform pooling operation; Use the long short-term memory network to perform time series processing on the pooled feature map to form a time series feature vector; The time series feature vector is processed using the classification layer to generate the leakage probability distribution of each location; The most likely leakage location is determined based on the leakage probability distribution of each location, and a leakage warning signal is issued when the leakage probability exceeds a preset threshold.

10. A vanadium liquid flow battery stack assembly leak detection system, used to perform the method according to any one of claims 1 to 9, characterized in that: include: Fluid mechanics modeling module, used to establish a multi-level fluid model of the vanadium liquid flow battery stack and describe the flow characteristics of the electrolyte; A numerical simulation module, used to simulate the normal state of the fluid inside the fuel cell stack based on the fluid mechanics model and generate reference fluid data; The sensor module includes multiple pressure, flow and chemical sensors for real-time collection of stack fluid, vanadium ion concentration, redox potential and pH data; A data processing module is used to compare the real-time monitoring data with the baseline data and input the chemical data of the abnormal fluctuation location into the deep learning model; A deep learning module is used to analyze multi-dimensional feature data matrices and perform leak detection and location through convolutional neural networks and long short-term memory networks.

Images