Leakage detection and collection method of flow battery energy storage system electric pile

Abstract

The invention relates to the technical field of new energy storage, and discloses a liquid leakage detection and collection method for a galvanic pile of a flow battery energy storage system, which comprises the following steps of: arranging a humidity sensor, a gas sensor, a temperature sensor, a visual sensor and a pressure sensor in a key area of the galvanic pile to form a high-density detection network of multi-source data; sensor data are collected and transmitted to a control center in real time, multi-source data are fused through Kalman filtering, Bayesian estimation and a deep learning model, and feature information of liquid leakage is obtained; and based on feature information of multi-source data fusion, a multi-point interpolation method and three-dimensional space modeling are used to realize positioning of a liquid leakage area. The intelligent grading response and control technical scheme is adopted, and factors such as leakage diffusion and environmental conditions are dynamically analyzed, so that grading self-adaptive response of operations such as isolation, cooling, load shedding and shutdown is realized, and compared with the scheme that only single isolation or shutdown is adopted to cope with a leakage event in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of new energy storage technology, and in particular to a method for detecting and collecting leakage of a stack of a flow battery energy storage system. Background Art

[0002] Liquid flow battery is a type of energy storage battery that uses liquid electrolyte as an energy carrier and stores and releases energy through chemical reactions. Its basic working principle is to dissolve the active substance in the electrolyte, transport it to the battery electrodes through the flow of the electrolyte, and perform redox reactions in the electrode reactions to store and release electrical energy. Liquid flow battery systems are widely used in the field of new energy due to their high energy storage density and fast response speed. However, due to the special properties of the electrolyte, once leakage occurs in the battery stack, it may not only affect the safety of the equipment, but may also lead to reduced energy conversion efficiency, thereby affecting the overall performance of the energy storage system.

[0003] Most existing leakage treatment solutions rely only on simple fault response methods, such as shutdown or basic isolation measures. However, this "one-size-fits-all" response approach is difficult to adapt to the diversity of leakage events and cannot flexibly adjust treatment measures according to the severity of the leakage, the speed of spread and environmental conditions. Especially when the leakage is mild, direct shutdown or extensive isolation will cause waste of system resources and unnecessary interruptions. In the case of severe leakage, simple isolation alone is often difficult to effectively suppress the spread.

[0004] Current leakage detection solutions mostly use a single or a few sensors for monitoring, such as humidity sensors, temperature sensors, etc. This single data source is easily interfered with when facing complex environmental conditions and is difficult to provide comprehensive leakage information, resulting in insufficient monitoring sensitivity and accuracy. A single sensor may produce false alarms or missed alarms when environmental conditions fluctuate, and it is difficult to effectively distinguish normal fluctuations from leakage signals.

[0005] Based on the above analysis, the present invention proposes a leakage detection and collection method for a flow battery energy storage system stack. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a leakage detection and collection method for a flow battery energy storage system stack, which solves the problem that when the leakage is slight, direct shutdown or extensive isolation will cause waste of system resources and unnecessary interruption.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for detecting and collecting leakage of a flow battery energy storage system stack, comprising the following steps:

[0008] Humidity sensors, gas sensors, temperature sensors, visual sensors, and pressure sensors are deployed in key areas of the fuel cell stack to form a high-density detection network with multi-source data.

[0009] Collect sensor data and transmit it to the control center in real time. Use Kalman filtering, Bayesian estimation and deep learning models to fuse multi-source data and obtain characteristic information of leakage.

[0010] Based on the feature information of multi-source data fusion, multi-point interpolation method and three-dimensional spatial modeling are used to locate the leakage area and generate a three-dimensional heat map to intuitively display the leakage diffusion area and its concentration distribution, and provide the spatial location information of the leakage.

[0011] Preferably, according to the real-time update of the three-dimensional heat map, the diffusion speed of the leakage, the concentration change and the environmental factors are analyzed. The system identifies the potential high-risk areas and the possible paths of the leakage diffusion according to the diffusion trend and dynamic change information, and predicts the diffusion range, providing data support for adjusting the detection strategy;

[0012] Adaptively adjust the sensor's detection threshold and sampling frequency according to environmental conditions, divide the battery stack into multiple independent modules, and perform intelligent isolation operations on the detected leakage area, including local isolation, regional isolation and complete isolation modes;

[0013] Activate the bypass circuit of the isolation module to achieve bidirectional power supply, ensure the continuity of data and power after isolation, and perform graded response according to the fault level. Response measures include cooling, exhaust, load reduction and shutdown.

[0014] Preferably, the hierarchical response comprises the following steps:

[0015] According to the leakage volume, diffusion speed and environmental parameters, the risk level of the fault is calculated through the analytic hierarchy process and grey relational analysis;

[0016] When the fault risk level is level I, only a warning signal is issued and data monitoring is started;

[0017] When the fault risk level is II, local isolation operation is initiated to reduce the power of the affected modules;

[0018] When the fault risk level is level III, regional isolation and cooling operations are initiated;

[0019] When the fault risk level is IV, the fault area is completely isolated and the surrounding modules are unloaded;

[0020] When the fault risk level is V, the system will be completely shut down and the exhaust device will be started to reduce the internal pressure.

[0021] Preferably, the characteristic information of the multi-source data fusion includes multi-dimensional data of humidity, gas concentration, temperature, image information and pressure signals. After smoothing and noise filtering through Kalman filtering and Bayesian estimation, the fused data is further subjected to feature extraction using a deep learning model to generate comprehensive leakage features.

[0022] Preferably, the three-dimensional heat map generated by the multi-point interpolation method combined with the three-dimensional space modeling is used to display the diffusion area of ​​the leakage, and the visual information of the dynamic diffusion of the leakage is provided by updating the diffusion model and the changes of the heat map in real time.

[0023] Preferably, the intelligent isolation operation includes isolating the detected leakage area and automatically activating the bypass circuit after the fault is isolated to ensure the normal operation of other modules of the system. The bypass circuit is provided with a bidirectional power supply design to support bidirectional data transmission and continuous power supply during isolation, ensuring that isolation does not affect the overall operation of the system.

[0024] Preferably, the bypass circuit is switched by a fast electronic switch, and the electronic switch adopts a control mechanism for intelligently adjusting the current flow to prevent instantaneous high current from affecting data transmission and power supply of other modules, ensuring seamless power and data transmission between modules.

[0025] Preferably, the collection method further comprises the following steps:

[0026] The leakage data, isolation information and response data are stored in multiple levels, and the historical data are analyzed and predicted based on time series analysis, cluster analysis and deep learning models to optimize the adaptive parameter configuration of the system;

[0027] The data analysis and feedback optimization includes time series analysis and deep learning model prediction of historical data, and automatically updating the system's response threshold and sensor sampling frequency according to the data change trend. The deep learning model is a long short-term memory network LSTM, which is used to identify changes in leakage patterns in complex environments.

[0028] Preferably, the adaptive parameter configuration includes adjusting the detection threshold of the sensor, the sampling frequency, and the response strength of the isolation, cooling and exhaust operations:

[0029] Sensor threshold and sampling frequency adjustment to adapt to changing environmental conditions;

[0030] Isolation mode response intensity adjustment, including automatic selection of local isolation or complete isolation based on the detected leakage diffusion speed to reduce the impact on other modules of the system;

[0031] The adaptive configuration of cooling and exhaust operation adjusts the startup frequency and power of cooling and exhaust equipment in real time according to the leakage location and temperature.

[0032] A liquid leakage detection and collection system, comprising:

[0033] Multi-source data acquisition module, which is used to collect humidity, gas, temperature, visual and pressure signals from a multi-source sensor array in real time and send the collected environmental data to the control center;

[0034] A data processing and fusion module, a control center connected to the multi-source data acquisition module, is used to perform multi-source data fusion and feature extraction operations, specifically including Kalman filtering, Bayesian estimation and deep learning model processing steps, to accurately locate the leakage area;

[0035] The hierarchical response and control module is used to perform hierarchical response on the stack module based on the leakage information provided by the data processing and fusion module, including hierarchical isolation, cooling, load reduction and shutdown operations, and implement the optimal fault management strategy according to the fault level;

[0036] Data storage and analysis module, used to store real-time and historical leakage data, and to perform trend prediction and parameter optimization on data based on time series analysis, cluster analysis and deep learning models, optimize system response parameters and adaptively adjust thresholds;

[0037] The redundant control module is used to detect the failure of the main module and automatically switch to the redundant module to ensure the continuous and stable operation of the system.

[0038] The present invention provides a method for detecting and collecting leakage of a flow battery energy storage system stack. It has the following beneficial effects:

[0039] 1. The present invention adopts an intelligent hierarchical response and control technical solution, and realizes hierarchical adaptive response of isolation, cooling, load reduction and shutdown operations by dynamically analyzing factors such as leakage diffusion and environmental conditions. Compared with the prior art solution of only using a single isolation or shutdown to deal with leakage events, the present invention can optimize the response strategy according to actual conditions, and effectively solves the problems of resource waste and excessive system interruption caused by a single response method.

[0040] 2. The present invention adopts a multi-source data acquisition and fusion technology solution, and realizes high-precision real-time monitoring and positioning of liquid leakage by integrating a multi-dimensional sensor array of humidity, gas, temperature, vision and pressure signals. Compared with the monitoring solutions of a single sensor or a few data sources in the prior art, the present invention can provide comprehensive and integrated environmental information, solving the problem of insufficient sensitivity and accuracy in liquid leakage monitoring.

[0041] 3. The present invention ensures the continuous and stable operation of the system by designing a data storage and analysis module and a redundant control module, and adopts a technical solution of trend prediction and adaptive optimization. Compared with the solutions in the prior art that rely on fixed parameters and manual adjustment, it realizes automatic fault prediction and dynamic parameter optimization, and solves the shortcomings of low system adaptability and delayed fault response in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a flow chart of the method of the present invention;

[0043] Figure 2 It is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0044] 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.

[0045] Please see attached Figure 1 - Attachment Figure 2 The embodiment of the present invention provides a method for detecting and collecting leakage of a flow battery energy storage system stack, comprising the following steps:

[0046] Humidity sensors, gas sensors, temperature sensors, visual sensors, and pressure sensors are deployed in key areas of the fuel cell stack to form a high-density detection network with multi-source data.

[0047] Collect sensor data and transmit it to the control center in real time. Use Kalman filtering, Bayesian estimation and deep learning models to fuse multi-source data and obtain characteristic information of leakage.

[0048] Based on the feature information of multi-source data fusion, multi-point interpolation method and three-dimensional spatial modeling are used to locate the leakage area and generate a three-dimensional heat map to intuitively display the leakage diffusion area and its concentration distribution, and provide the spatial location information of the leakage.

[0049] According to the real-time update of the three-dimensional heat map, the system analyzes the diffusion speed, concentration changes and environmental factors of the leak. Based on the diffusion trend and dynamic change information, the system identifies potential high-risk areas and possible paths of the leak, and predicts the diffusion range, providing data support for adjusting the detection strategy;

[0050] Adaptively adjust the sensor's detection threshold and sampling frequency according to environmental conditions, divide the battery stack into multiple independent modules, and perform intelligent isolation operations on the detected leakage area, including local isolation, regional isolation and complete isolation modes;

[0051] Activate the bypass circuit of the isolation module to achieve bidirectional power supply, ensure the continuity of data and power after isolation, and perform graded response according to the fault level. Response measures include cooling, exhaust, load reduction and shutdown.

[0052] The hierarchical response includes the following steps:

[0053] According to the leakage volume, diffusion speed and environmental parameters, the risk level of the fault is calculated through the analytic hierarchy process and grey relational analysis;

[0054] When the fault risk level is level I, only a warning signal is issued and data monitoring is started;

[0055] When the fault risk level is II, local isolation operation is initiated to reduce the power of the affected modules;

[0056] When the fault risk level is level III, regional isolation and cooling operations are initiated;

[0057] When the fault risk level is IV, the fault area is completely isolated and the surrounding modules are unloaded;

[0058] When the fault risk level is V, the system will be completely shut down and the exhaust device will be started to reduce the internal pressure.

[0059] Specifically, humidity, gas, temperature, visual and pressure sensors are deployed in key areas of the fuel cell stack to form a high-density multi-source data detection network. Sensors are placed in leak-prone locations such as sealing interfaces, valves, pipelines, high-temperature areas and confined spaces to collect multi-dimensional data such as humidity, gas concentration, temperature, visual images and pressure signals in real time, providing basic support for comprehensive monitoring of leaks. The collected data is transmitted to the control center via the data bus.

[0060] At the control center, the system uses Kalman filtering, Bayesian estimation and deep learning models to fuse multi-source data. Kalman filtering is used to smooth various sensor data and reduce the impact of environmental noise; Bayesian estimation performs probabilistic fusion of observation results from different sensors to generate high-confidence information on leakage characteristics; deep learning models (such as convolutional neural networks (CNN)) further extract features from visual images, identify the microscopic features of leakage, including droplet morphology and wetting traces, and generate comprehensive feature information of leakage.

[0061] Based on the fused multi-source feature information, the system uses multi-point interpolation and three-dimensional spatial modeling technology to locate the leakage area and generate a three-dimensional heat map to intuitively display the diffusion area of ​​the leakage and its concentration distribution. This three-dimensional heat map provides real-time spatial location information, enabling the system to accurately understand the occurrence point, diffusion direction and concentration of the leakage, providing decision support for subsequent response.

[0062] Based on the real-time update of the three-dimensional heat map, the diffusion rate of the leakage, concentration changes and environmental factors are analyzed. By evaluating the diffusion trend and dynamic change information, the system identifies potential high-risk areas and possible paths of leakage spread, predicts the spread range, and provides data support for subsequent detection strategies and isolation measures adjustments.

[0063] According to the analyzed environmental conditions, the system adaptively adjusts the detection threshold and sampling frequency of the sensor to ensure the sensitivity and accuracy of monitoring. After detecting the leakage, the battery stack is divided into multiple independent modules, and intelligent isolation operations are performed on the detected leakage area, including local isolation, regional isolation and complete isolation modes, to reduce the impact of fault spread on other modules of the system.

[0064] After the isolation is completed, the system automatically activates the bypass circuit of the isolation module, and ensures the continuity of data transmission and power supply after isolation through a bidirectional power supply design. The bypass circuit enables the system to maintain the normal operation of other modules during the isolation period. For detected leakage events, the system responds in stages according to the fault level, including cooling, exhaust, load reduction and shutdown response measures, to ensure that the best response strategy is adopted according to the actual situation.

[0065] The characteristic information of the multi-source data fusion includes multi-dimensional data of humidity, gas concentration, temperature, image information and pressure signals. After smoothing and noise filtering through Kalman filtering and Bayesian estimation, the deep learning model is further used to extract features from the fused data to generate comprehensive leakage features.

[0066] The three-dimensional heat map generated by the multi-point interpolation method combined with the three-dimensional space modeling is used to display the leakage diffusion area, and provides visual information on the dynamic diffusion of the leakage by updating the diffusion model and the heat map changes in real time.

[0067] Specifically, Kalman filtering is used to smooth various sensor data to reduce environmental noise interference and make the data more stable and accurate. Kalman filtering adjusts the estimated value according to the error between the current state estimate and the observed value, providing dynamic and continuous data optimization. Secondly, the system uses Bayesian estimation to fuse the observation results of different sensors, and uses probability distribution to calculate high-confidence data of leakage events. The Bayesian formula is used to integrate multi-source information to make the data fusion result more credible. After the fused multi-source data is noise filtered and smoothed, the system further uses deep learning models for feature extraction, especially through convolutional neural networks (CNN) to analyze the image data collected by visual sensors, so as to identify the microscopic features of the leakage, including droplet morphology, wetting marks and other details, and generate comprehensive leakage feature information.

[0068] Multi-dimensional data fusion and leakage location

[0069] Data fusion processing

[0070] The data processing and fusion module fuses multi-source data using Kalman filtering, Bayesian estimation and deep learning models. The specific steps are as follows: Kalman filtering: Smooths various sensor data to reduce noise interference and improve data stability. The recursive formula is:

[0071] x k|k =x k|k-1 +K k (z k -Hx k|k-1 )

[0072] Among them, x k|k represents the current posterior estimate, K k is the Kalman gain, z k is the observed value, H is the observation matrix, and Bayesian estimation: According to the observation results of each sensor, the probability distribution of leakage is calculated using the Bayesian formula, and multi-source data is integrated into high-confidence fusion data. The formula is:

[0073]

[0074] Among them, P(A|B) represents the posterior probability of event A under condition B. Deep learning feature extraction: Use convolutional neural network (CNN) to process image data collected by visual sensors to identify microscopic features of leakage (such as droplet morphology and wetting traces) to further improve the sensitivity of detection.

[0075] Leak location

[0076] The fused data is used to accurately locate the leakage. Through multi-point interpolation and three-dimensional space modeling technology, a three-dimensional heat map is generated to show the diffusion area of ​​the leakage, which can be accurately located to the millimeter level. Interpolation formula:

[0077]

[0078] Among them, f(x,y,z) represents the interpolation result in three-dimensional space, w i is the weight factor, f i The leakage location algorithm will dynamically update the heat map based on the real-time sensor data to form a complete view of the leakage area.

[0079] In terms of positioning, the multi-point interpolation method is combined with the three-dimensional space modeling technology to generate a three-dimensional heat map. The multi-point interpolation method calculates the distribution of leakage characteristics in three-dimensional space through the detection data of each sensor to form a continuous three-dimensional scene. Then, the data is intuitively displayed as a three-dimensional heat map through three-dimensional modeling technology to reflect the area and concentration distribution of the leakage diffusion. The three-dimensional heat map has the ability to dynamically update. The system continuously updates the diffusion model and heat map according to the real-time collected data to generate visualization information of the dynamic diffusion of the leakage. This real-time heat map not only provides intuitive spatial location information, but also supports the system to monitor in real time and dynamically adjust the response strategy when the leakage spreads, providing a scientific basis for precise isolation, cooling and exhaust operations.

[0080] The intelligent isolation operation includes isolating the detected leakage area and automatically activating the bypass circuit after the fault is isolated to ensure the normal operation of other modules of the system. The bypass circuit is provided with a bidirectional power supply design to support bidirectional data transmission and continuous power supply during isolation, ensuring that isolation does not affect the overall operation of the system.

[0081] The bypass circuit is switched by a fast electronic switch, which adopts a control mechanism to intelligently adjust the current flow to prevent instantaneous high current from affecting the data transmission and power supply of other modules, ensuring seamless power and data transmission between modules.

[0082] Specifically, the intelligent isolation operation includes automatically isolating the faulty area after detecting a leakage area to prevent the leakage from spreading to other areas. After the isolation is completed, the system will automatically activate the bypass circuit and continue to provide data and power support to other normally operating modules through the bypass design to ensure that the isolated area does not affect the normal operation of the system as a whole. The bypass circuit has a bidirectional power supply design, which can realize bidirectional data transmission and continuous power supply during the isolation process, and maintain communication and power supply stability between modules of the system.

[0083] The switching of the bypass circuit is achieved through a fast electronic switch with a switching delay of less than 1 millisecond, which can achieve instantaneous response and ensure that the system is not disturbed during the circuit switching process. In addition, the electronic switch adopts an intelligent current regulation mechanism to monitor and automatically adjust the current flow in real time to prevent the instantaneous high current generated during the switching process from affecting the power supply or data transmission of other modules. Through this design, the system can maintain seamless power and data connection between modules while isolating the fault area, further improving the stability and fault tolerance of the system.

[0084] The collection method further comprises the following steps:

[0085] The leakage data, isolation information and response data are stored in multiple levels, and the historical data are analyzed and predicted based on time series analysis, cluster analysis and deep learning models to optimize the adaptive parameter configuration of the system;

[0086] The data analysis and feedback optimization includes time series analysis and deep learning model prediction of historical data, and automatically updating the system's response threshold and sensor sampling frequency according to the data change trend. The deep learning model is a long short-term memory network LSTM, which is used to identify changes in leakage patterns in complex environments.

[0087] Specifically, time series analysis and deep learning model-based prediction are performed on the system's historical data to achieve accurate prediction of leakage events and adaptive adjustment of system response parameters. First, time series analysis is used to detect trends in historical data and identify periodic and trend changes in leakage data. By analyzing the changing trends of data, the system can predict potential leakage risks and take preventive measures in advance.

[0088] On this basis, the system uses the long short-term memory network (LSTM) deep learning model to learn and model historical data. Due to its special structure, the LSTM network can effectively handle long-term dependencies in time series data and is suitable for capturing changes in leakage patterns in complex environments. During the training process, the LSTM model gradually learns the leakage characteristics and patterns in different environments, thereby identifying potential factors and signal characteristics that may cause leakage. This predictive ability enables the system to identify and adapt to leakage patterns caused by changes in environmental conditions in real time.

[0089] Based on time series analysis and LSTM prediction results, the system can automatically update the response threshold and the sampling frequency of the sensor. When the data trend shows an increased risk of leakage, the system automatically lowers the response threshold to improve detection sensitivity, while increasing the sampling frequency of the sensor to obtain more detailed environmental data and ensure rapid response. When the trend is stable, the system appropriately relaxes the threshold and reduces the sampling frequency to reduce energy consumption. Through this dynamic feedback mechanism, the system achieves adaptive optimization, so that it can maintain a high monitoring sensitivity and response efficiency under various environmental conditions.

[0090] The adaptive parameter configuration includes adjusting the sensor's detection threshold, sampling frequency, and response strength of isolation, cooling, and exhaust operations:

[0091] Sensor threshold and sampling frequency adjustment to adapt to changing environmental conditions;

[0092] Isolation mode response intensity adjustment, including automatic selection of local isolation or complete isolation based on the detected leakage diffusion speed to reduce the impact on other modules of the system;

[0093] The adaptive configuration of cooling and exhaust operation adjusts the startup frequency and power of cooling and exhaust equipment in real time according to the leakage location and temperature.

[0094] Specifically, the sensor threshold and sampling frequency adjustment are used to adapt to the dynamic changes in environmental conditions. Specifically, when external environmental factors (such as temperature and humidity) fluctuate, the system automatically adjusts the detection threshold and sampling frequency of various sensors according to real-time monitoring data to ensure the accuracy and sensitivity of leakage detection. For example, in a high temperature or high humidity environment, the system will lower the detection threshold of the humidity sensor and increase the sampling frequency to capture tiny leakage change signals more quickly, thereby responding to potential leakage risks in a timely manner. On the contrary, in a stable environment, the system will appropriately relax the detection threshold and reduce the sampling frequency to save energy.

[0095] The isolation mode response strength adjustment automatically selects the optimal isolation mode based on the leakage diffusion speed, including local isolation and complete isolation. Specifically, when the system detects that the leakage spreads slowly and the affected area is small, the system starts local isolation and only isolates the module where the leakage occurs to reduce the impact on surrounding modules and maintain the overall functional operation of the system; when the leakage spreads rapidly or the affected area is large, the system starts complete isolation to completely disconnect the faulty area from other modules to prevent the leakage from further spreading to other areas, thereby maximizing the protection of the system's safety and integrity.

[0096] The adaptive configuration of cooling and exhaust operations adjusts the starting frequency and power of the cooling and exhaust equipment in real time according to the leakage location and temperature changes. When the system detects that the temperature in the leakage area has risen, it will dynamically control the starting frequency and air volume of the cooling equipment according to the temperature conditions to ensure that the temperature can be quickly reduced and stabilized within a safe range. At the same time, if it is detected that the gas concentration generated by the leakage increases, the system automatically adjusts the power and discharge speed of the exhaust equipment to maintain the balance of internal pressure in the system and avoid risks caused by gas accumulation. Through the above-mentioned adaptive cooling and exhaust configuration, the system can flexibly adjust various parameters in complex environments to ensure the efficiency of leakage management and the stability of system operation.

[0097] The leakage detection and collection system described below and the leakage detection and collection method for the flow battery energy storage system stack described above can be referenced to each other.

[0098] 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.

[0099] 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 detecting and collecting leakage of a flow battery energy storage system stack, characterized in that: The following steps are involved: Humidity sensors, gas sensors, temperature sensors, visual sensors, and pressure sensors are deployed in key areas of the fuel cell stack to form a high-density detection network with multi-source data. Collect sensor data and transmit it to the control center in real time. Use Kalman filtering, Bayesian estimation and deep learning models to fuse multi-source data and obtain characteristic information of leakage. Based on the feature information of multi-source data fusion, multi-point interpolation method and three-dimensional spatial modeling are used to locate the leakage area and generate a three-dimensional heat map to intuitively display the leakage diffusion area and its concentration distribution, and provide the spatial location information of the leakage.

2. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 1, characterized in that: The collection method further comprises the following steps: According to the real-time update of the three-dimensional heat map, the system analyzes the diffusion speed, concentration changes and environmental factors of the leak. Based on the diffusion trend and dynamic change information, the system identifies potential high-risk areas and possible paths of the leak, and predicts the diffusion range, providing data support for adjusting the detection strategy; Adaptively adjust the sensor's detection threshold and sampling frequency according to environmental conditions, divide the battery stack into multiple independent modules, and perform intelligent isolation operations on the detected leakage area, including local isolation, regional isolation and complete isolation modes; Activate the bypass circuit of the isolation module to achieve bidirectional power supply, ensure the continuity of data and power after isolation, and perform graded response according to the fault level. Response measures include cooling, exhaust, load reduction and shutdown.

3. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 2, characterized in that: The hierarchical response includes the following steps: According to the leakage volume, diffusion speed and environmental parameters, the risk level of the fault is calculated through the analytic hierarchy process and grey relational analysis; When the fault risk level is level I, only a warning signal is issued and data monitoring is started; When the fault risk level is II, local isolation operation is initiated to reduce the power of the affected modules; When the fault risk level is level III, regional isolation and cooling operations are initiated; When the fault risk level is IV, the fault area is completely isolated and the surrounding modules are unloaded; When the fault risk level is V, the system will be completely shut down and the exhaust device will be started to reduce the internal pressure.

4. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 1, characterized in that: The characteristic information of the multi-source data fusion includes multi-dimensional data of humidity, gas concentration, temperature, image information and pressure signals. After smoothing and noise filtering through Kalman filtering and Bayesian estimation, the deep learning model is further used to extract features from the fused data to generate comprehensive leakage features.

5. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 1, characterized in that: The three-dimensional heat map generated by the multi-point interpolation method combined with the three-dimensional space modeling is used to display the leakage diffusion area, and provides visual information on the dynamic diffusion of the leakage by updating the diffusion model and the heat map changes in real time.

6. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 2, characterized in that: The intelligent isolation operation includes isolating the detected leakage area and automatically activating the bypass circuit after the fault is isolated to ensure the normal operation of other modules of the system. The bypass circuit is provided with a bidirectional power supply design to support bidirectional data transmission and continuous power supply during isolation, ensuring that isolation does not affect the overall operation of the system.

7. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 2, characterized in that: The bypass circuit is switched by a fast electronic switch, which adopts a control mechanism to intelligently adjust the current flow to prevent instantaneous high current from affecting the data transmission and power supply of other modules, ensuring seamless power and data transmission between modules.

8. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 1, characterized in that: The collection method further comprises the following steps: The leakage data, isolation information and response data are stored in multiple levels, and the historical data are analyzed and predicted based on time series analysis, cluster analysis and deep learning models to optimize the adaptive parameter configuration of the system; The data analysis and feedback optimization includes time series analysis and deep learning model prediction of historical data, and automatically updating the system's response threshold and sensor sampling frequency according to the data change trend. The deep learning model is a long short-term memory network LSTM, which is used to identify changes in leakage patterns in complex environments.

9. The method for collecting leakage detection of a flow battery energy storage system stack according to claim 2, characterized in that: The adaptive parameter configuration includes adjusting the sensor's detection threshold, sampling frequency, and response strength of isolation, cooling, and exhaust operations: Sensor threshold and sampling frequency adjustment to adapt to changing environmental conditions; Isolation mode response intensity adjustment, including automatic selection of local isolation or complete isolation based on the detected leakage diffusion speed to reduce the impact on other modules of the system; The adaptive configuration of cooling and exhaust operation adjusts the startup frequency and power of cooling and exhaust equipment in real time according to the leakage location and temperature.

10. A leakage detection and collection system, the leakage detection and collection method for a flow battery energy storage system stack according to claims 1-9, characterized in that: include: Multi-source data acquisition module, which is used to collect humidity, gas, temperature, visual and pressure signals from a multi-source sensor array in real time and send the collected environmental data to the control center; A data processing and fusion module, a control center connected to the multi-source data acquisition module, is used to perform multi-source data fusion and feature extraction operations, specifically including Kalman filtering, Bayesian estimation and deep learning model processing steps, to accurately locate the leakage area; The hierarchical response and control module is used to perform hierarchical response on the stack module based on the leakage information provided by the data processing and fusion module, including hierarchical isolation, cooling, load reduction and shutdown operations, and implement the optimal fault management strategy according to the fault level; Data storage and analysis module, used to store real-time and historical leakage data, and to perform trend prediction and parameter optimization on data based on time series analysis, cluster analysis and deep learning models, optimize system response parameters and adaptively adjust thresholds; The redundant control module is used to detect the failure of the main module and automatically switch to the redundant module to ensure the continuous and stable operation of the system.

Images