Electrochemical energy storage power station safety debugging and detection method and system

Abstract

The application discloses a kind of electrochemical energy storage power station safety debugging and detection method and system, it is related to electrochemical energy storage power station operation safety and operation and maintenance detection field.The method is by electricity and safety confirmation, communication connection and equipment topology identification, low voltage isolation and insulation pre-inspection, high voltage controlled closing and double authorization, multi-source data acquisition and fusion diagnosis, step-by-step fault location and safety rollback, structured report generation and health record update, realize the safety, standardization and intelligent processing of energy storage power station first commissioning, periodic retest and fault location.The system is encapsulated in multifunctional safety inspection box, with electricity, measurement, acquisition, control and host computer collaborative function, can carry out comprehensive detection and linkage control to voltage, temperature, insulation, grounding and communication state, can improve abnormal identification precision, fault location efficiency and operation safety, and enhance life cycle management capability.

Description

Technical Field

[0001] This invention belongs to the field of operation safety and maintenance testing technology of electrochemical energy storage power stations, and involves high-voltage commissioning safety control, status detection and fault diagnosis of energy storage power stations. Specifically, it relates to a safety commissioning and testing method and system applied to the first commissioning, periodic retesting and fault location scenarios of electrochemical energy storage power stations and based on safety interlocking and integrated diagnosis. Background Technology

[0002] Electrochemical energy storage power stations are complex energy storage and release facilities composed of battery systems, converter systems, monitoring and protection systems, and auxiliary support systems. They have been widely used in scenarios such as frequency regulation and peak shaving on the power supply side, peak shaving and valley filling on the grid side, and backup power supply on the user side. With the rapid increase in the capacity of energy storage power stations and the continuous expansion of the scale of single stations, electrochemical energy storage systems operate continuously under conditions of high energy density, complex electrical environment, and thermal risk accumulation, which brings greater challenges to safety management. Therefore, it is of great significance to carry out standardized and systematic safety commissioning and testing of electrochemical energy storage power stations.

[0003] However, electrochemical energy storage power stations involve diverse equipment and complex architectures. The consistency of individual battery cell performance, the reliability of electrical insulation, and the stability of communication links all degrade and introduce potential faults over time, leading to a gradual decrease in operational safety margins. Therefore, periodic testing before initial commissioning and during operation is crucial for improving the long-term stability and fault prevention capabilities of energy storage systems. Unlike conventional primary equipment testing, energy storage power stations involve multiple collaborative safety elements, including high-voltage operation authorization management, interlocking control, and data communication quality verification. A lack of coordinated mechanisms during commissioning can lead to serious safety risks such as misoperation and failure to detect insulation anomalies in a timely manner. Furthermore, existing commissioning methods rely heavily on manual experience, with fragmented testing methods and difficulties in data fusion and analysis, hindering the prediction of risk trends and resulting in delayed anomaly identification and low fault location efficiency.

[0004] To address these challenges, the industry has explored relevant technologies. For example, Chinese invention patent application CN120408448A discloses a deep learning-based method for detecting operational anomalies in electrochemical energy storage power stations. This method constructs training data using historical operating data and real-world operating conditions, and trains a deep learning model to identify anomalies in real-time data. However, this approach focuses on software algorithm monitoring during operation and lacks systematic debugging and verification of physical safety factors such as insulation reliability and high-voltage operation interlocking before commissioning. CN117610955A discloses a debugging method, system, equipment, and medium for an electrochemical energy storage power station monitoring system. This method simulates the interlayer device through a debugging device to verify the consistency of information interaction between the monitoring system and the equipment. However, its application is limited to the information flow level and cannot adequately address the detection of the physical coupling characteristics of the battery itself. CN119936519A discloses a fault detection method, device, equipment and storage medium for an electrochemical energy storage power station. It uses a pre-trained model to determine the cause of the fault and generate maintenance guidance tutorials. However, it mainly serves the operation and maintenance after the fault is discovered, and does not adequately consider the pre-operational safety verification and safety constraints during the commissioning process in scenarios such as initial commissioning and periodic re-inspection.

[0005] In summary, existing technologies still have significant shortcomings in terms of safety, standardization, accuracy of anomaly identification, efficiency of fault location, and data traceability. To address the surge in the number of energy storage power stations and the need for full lifecycle safety management, it is urgent to build a safety commissioning and testing technology system with battery operation safety as the core and data acquisition and fusion, intelligent analysis and diagnosis, and interlocking control of the commissioning process as the foundation. This system will enable full-process safety control, efficient fault screening, and traceable data management during the equipment's grid connection and long-term operation phases, providing technical support for the safe and efficient operation of future energy storage power stations. Summary of the Invention

[0006] (a) Purpose of the invention The technical problem to be solved by this invention is to provide a method and system for safe commissioning and testing of electrochemical energy storage power stations. By constructing a safety interlocking commissioning and testing mechanism covering voltage verification, communication establishment and topology identification, low-voltage isolation and insulation pre-inspection, high-voltage controlled operation, multi-dimensional data acquisition, fusion diagnosis, and step-by-step fault location, this invention achieves efficient, standardized, and safe verification of energy storage power stations from initial commissioning and periodic re-inspection to fault location, and improves the ability to identify potential faults and abnormal states. This overcomes the problems of scattered commissioning processes, inconsistent testing standards, difficulty in ensuring the safety of high-voltage operation, and difficulty in timely identification of potential abnormal states in existing technologies.

[0007] (II) Technical Solution To achieve the objective of this invention and solve its technical problems, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for the safe commissioning and testing of an electrochemical energy storage power station, used for the initial commissioning, periodic retesting, and fault location of the electrochemical energy storage power station, comprising at least the following steps: SS1. Voltage Testing and Safety Confirmation: Perform voltage testing on the battery compartment door and exposed conductive parts, obtain the voltage testing results, and determine whether the electrical safety access conditions are met by combining the operator's position status and the initial state of high-voltage emergency stop. If any electrical safety access condition is not met, the subsequent process is prohibited. SS2. Communication Link Establishment and Device Topology Identification: Establish a communication link with the battery system and battery management system (BMS), identify the topology of the device under test and complete node address allocation, while monitoring the quality of the communication link, and performing node reconnection or bypass processing when the quality of the communication link does not meet the set conditions. SS3. Low-voltage isolation and insulation pre-inspection: Before performing high-voltage operation, the battery system is placed in low-voltage isolation state, and its insulation resistance and grounding status are checked. Based on this, it is determined whether the insulation pre-inspection conditions are met. If the insulation resistance is lower than the preset safety threshold and / or the grounding is abnormal, the subsequent high-voltage operation is locked. SS4. High-voltage controlled closing and dual authorization: Under the conditions that the electrical safety access conditions and insulation pre-inspection conditions are met and the communication link quality meets the set conditions, the high-voltage circuit is triggered to close in a controlled manner based on the operator's authorization confirmation and the system status authorization confirmation. If the dual authorization confirmation for closing is not obtained, it will fall back to the low-voltage isolation state. SS5. Multidimensional data acquisition and anomaly identification: After the high-voltage circuit enters the controlled commissioning state, the temperature, voltage, insulation resistance, grounding resistance and communication link quality data of the battery system are collected. A set of characteristic parameters characterizing voltage consistency, temperature rise change, insulation attenuation and communication stability are constructed and fusion diagnostic analysis is performed based on this. The results of abnormal unit identification, anomaly type determination and corresponding risk level are output. SS6. Step-by-step Fault Location and Safety Rollback: For identified abnormal units, step-by-step fault location is performed according to the hierarchical path of battery cluster, module, and cell. Based on the abnormality type and risk level, the located abnormal units are isolated or safety rollback is performed. The safety rollback takes controlled disconnection, test suspension, or alarm prompt actions according to the risk level. The isolated object, isolation reason, and corresponding timestamp are recorded simultaneously. Fault handling records are established according to the battery cluster, module, and cell levels.

[0008] SS7. Structured Commissioning Report Generation and Health Record Update (Preferred Step): Generate a structured commissioning report containing current test data, historical comparison differences, trend predictions, and maintenance recommendations, and store the data in association at the equipment level to update the power plant's full life cycle health record.

[0009] Secondly, the present invention provides a safety commissioning and testing system for an electrochemical energy storage power station, encapsulated in a multifunctional safety inspection box, and performing safety commissioning and testing of the electrochemical energy storage power station based on the above-mentioned method of the present invention. The system includes at least: The voltage detection and safety confirmation module is used to detect voltage on the battery compartment door and exposed conductive parts, and to determine whether the electrical safety access conditions are met by combining the operator's on-site status and the initial state of high-voltage emergency stop. The communication connection and device topology identification module is used to establish a communication connection with the battery system and battery management system, identify the topology of the device under test, complete node address allocation, and monitor communication quality. The low-voltage isolation and insulation pre-inspection module is used to place the battery system in a low-voltage isolation state, detect the insulation resistance and grounding status, and determine whether the insulation pre-inspection conditions are met. The high-voltage controlled closing and dual-authorization module is used to trigger the controlled closing of the high-voltage circuit based on operator authorization confirmation and system status authorization confirmation when electrical safety access conditions, insulation pre-inspection conditions and communication quality meet the set conditions. The multi-source data acquisition and fusion diagnostic module is used to acquire temperature, voltage, insulation resistance, grounding resistance and communication quality data, and output abnormal unit identification, abnormal type judgment and risk level; The step-by-step fault location and safety rollback module is used to locate the identified abnormal units step by step according to the battery cluster, module and cell level path, determine the branch isolation object or isolation range according to the abnormality type, and perform safety rollback processing according to the risk level. The structured report and archive update module is used to perform structured association storage of communication connection establishment results, insulation pre-inspection data, high-voltage controlled closing logs, multi-source detection data, fault location results, and safety rollback records, generate commissioning reports, and update the power plant's full life cycle health archive.

[0010] (III) Technical Effects Compared with the prior art, the electrochemical energy storage power station safety commissioning and testing method and system of the present invention have the following beneficial and significant technical effects: (1) By incorporating voltage verification, communication identification, low-voltage isolation, insulation pre-inspection, high-voltage controlled closing and dual-authorization control into the safety interlocking process, this invention achieves hierarchical verification of pre-commissioning conditions and controlled execution of high-voltage operations, which can significantly reduce the commissioning risks caused by misoperation, closing with faults and omission of safety conditions, and improve the safety of the first commissioning and periodic re-inspection process of the energy storage power station.

[0011] (2) This invention collects, extracts features and performs fusion diagnostic analysis on multi-source heterogeneous detection data such as voltage, temperature, insulation resistance, grounding resistance and communication quality in a unified manner, and combines rule discrimination, pattern recognition and abnormal evolution judgment to improve the accuracy of abnormal unit identification, the reliability of abnormal type judgment and the ability to suppress false alarms in low confidence scenarios, thereby enhancing the ability to identify complex faults.

[0012] (3) This invention constructs a step-by-step fault location mechanism at the battery cluster, module and cell levels, and links the risk level with branch isolation, safety rollback, structured debugging report and health record update, so as to realize closed-loop management of abnormal handling, result tracking and operation and maintenance decision-making, which can improve the efficiency of fault location, the rationality of re-inspection priority configuration and the full life cycle health management capability. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the overall process of the electrochemical energy storage power station safety commissioning and testing method of the present invention; Figure 2 A schematic diagram illustrating communication link establishment and automatic device topology identification; Figure 3 This is a schematic diagram of the low-voltage isolation and insulation pre-inspection process; Figure 4 This is a schematic diagram of the controlled closing and dual authorization mechanism for the high-voltage circuit; Figure 5 This is a schematic diagram of the multidimensional data acquisition and anomaly identification process; Figure 6 This is a schematic diagram of the step-by-step fault location and safety rollback strategy process; Figure 7 A schematic diagram illustrating the generation of structured commissioning reports and the updating of the power plant's life-cycle health records. Detailed Implementation

[0014] This invention aims to provide a method and system for the safe commissioning and testing of electrochemical energy storage power stations, used for the initial commissioning, periodic re-testing, and fault location of electrochemical energy storage power stations. To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. The described embodiments are some, but not all, embodiments of this invention, and are exemplary, intended to explain the invention, and should not be construed as limiting the invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0015] It should be noted that the battery cluster-module-cell level objects, the data collection objects such as voltage, temperature, insulation, grounding, and communication data, as well as the diagnostic outputs such as abnormal unit identification results, abnormal type determination results, abnormal confidence levels, and risk level determination results involved in the following embodiments are all centered around the specific scenario of commissioning and testing of electrochemical energy storage power stations. This is to address the technical problems of untimely anomaly detection, low fault location efficiency, and inaccurate safety interlocking control during high-voltage commissioning. For steps involving model construction, feature extraction, joint discrimination, and risk mapping, the algorithm processing is combined with on-site data collection, equipment topology relationships, and interlocking control actions.

[0016] Example 1: Safety Commissioning and Testing Methods for Electrochemical Energy Storage Power Stations As a specific example, such as Figure 1 As shown in the embodiment of the present invention, the specific implementation steps of the safety commissioning and testing method for electrochemical energy storage power stations are as follows: Step SS1: Voltage Testing and Safety Confirmation Perform an electrical test on the battery compartment door and exposed conductive parts, obtain the test results, and at least combine the operator's position status and the initial state of high-voltage emergency stop to determine whether the electrical safety conditions are met. If any electrical safety access condition is not met, the subsequent process is prohibited.

[0017] Specifically, when implementing step SS1, the voltage testing operation includes at least point-by-point voltage testing of the high-voltage interface inside the battery compartment door, the busbar connection, and exposed conductive components. The voltage testing results are then jointly verified with the operator's arrival confirmation information and the high-voltage emergency stop status information. Only when the voltage testing results indicate that there is no dangerous liveness, the operator's identity verification is passed, and it is confirmed that the operator has reached the designated safe operating position and the high-voltage emergency stop circuit is in the initial disconnected state, is it determined that the electrical safety access conditions are met. After generating the pre-commissioning safety confirmation record, the subsequent process begins. The pre-commissioning safety confirmation record preferably includes at least the voltage testing results, operator identity information, arrival confirmation results, high-voltage emergency stop status information, and confirmation time, so as to provide a basis for subsequent high-voltage controlled closing, anomaly tracing, and commissioning responsibility determination.

[0018] Step SS2: Communication Link Establishment and Device Topology Identification Establish a communication link with the battery system and battery management system (BMS), identify the topology of the device under test and complete node address allocation, monitor the quality of the communication link, and perform node reconnection or bypass processing when the quality of the communication link does not meet the set conditions.

[0019] Preferably, the system's host computer establishes a communication link with the battery cluster and battery management system via industrial Ethernet, while automatically identifying the power station equipment topology and node addresses, and monitoring the communication link quality in real time. When the communication reliability falls below a set threshold, the system automatically executes node reconnection or bypass strategies to ensure the stability and integrity of the entire commissioning network, and records the communication status in the system log. In this embodiment, as... Figure 2 As shown, the host computer control unit establishes a communication connection with the data acquisition unit through a network switch. The data acquisition unit is connected to the voltage detection unit, insulation withstand voltage measurement unit, communication measurement unit, grounding resistance measurement unit, high and low voltage safety control unit, voltage resistance measurement unit, and infrared thermal imaging unit, forming a hierarchical data acquisition architecture for multi-source detection and control execution. After establishing the communication link, the host computer control unit identifies the topology of the device under test based on the node response information fed back by the communication measurement unit and the data acquisition unit, completes node address mapping and address allocation, and writes the communication connection establishment, topology identification, and address allocation results into the system log.

[0020] More specifically, the communication link establishment supports negotiation and switching between multiple protocols including CAN, RS-485, and Industrial Ethernet. Industrial Ethernet is used for data transmission under high-speed data acquisition conditions, RS-485 is used for data transmission in compatibility scenarios, and CAN is switched to when communication interference exceeds a set threshold. The device topology identification is achieved by sending topology query messages to each node and receiving node responses. The identification content includes the battery cluster number and quantity, the module arrangement order and rated parameters within each battery cluster, and the number and nominal voltage of cells within each module. After topology identification is completed, The identified device topology is compared with historical topology records. Nodes with topology anomalies are marked, and node address correction or communication recovery is performed. After topology comparison and anomaly node processing, a three-level hierarchical topology mapping table for battery clusters, modules, and cells is established. Node address allocation is then performed based on this mapping table to achieve accurate addressing and hierarchical data archiving for subsequent detection tasks. The communication link quality is comprehensively judged based on link latency, message loss rate, number of consecutive erroneous frames, and heartbeat response status. When the communication link quality does not meet the set conditions, message retransmission verification, node address reallocation, or communication bypass processing is performed on the anomaly node. Furthermore, if the communication connection still fails to return to normal after reconnection and address correction, the system marks the corresponding node as a restricted communication node. In subsequent anomaly identification and fault location stages, a downgrade detection strategy is adopted for the branch corresponding to this node to avoid amplifying the impact of local communication anomalies on the overall debugging and judgment.

[0021] Step SS3: Low-voltage isolation and insulation pre-inspection Before performing high-voltage operations, the battery system is placed in a low-voltage isolation state, and its insulation resistance and grounding status are checked. Based on this, it is determined whether the insulation pre-inspection conditions are met. If the insulation resistance is lower than the preset safety threshold and / or the grounding is abnormal, subsequent high-voltage operations are locked.

[0022] In this embodiment, as Figure 3 As shown, after the communication link is established, the system first performs the low-voltage isolation operation in step S3-1, disconnecting the high-voltage main circuit of the battery system from the load side and recording the isolation execution time. Then, it proceeds to step S3-2 to measure the insulation resistance of the battery system. Insulation detection and grounding resistance detection can be performed in parallel according to a preset strategy to improve pre-inspection efficiency. The system first determines whether the insulation threshold requirement is met based on the insulation resistance detection result: when the insulation resistance is lower than the preset safety threshold, further debugging is prohibited, and an alarm is output through system log recording and audible / visual alarms. If the insulation resistance meets the threshold requirement, it proceeds to step S3-3 to further perform grounding resistance self-testing and compares and analyzes the grounding status change trend with historical detection results to determine if there is any grounding performance degradation. When the grounding resistance does not meet the set requirements, further debugging is also prohibited, and system log recording and audible / visual alarms are output. When the grounding resistance detection result is compliant, it proceeds to step SS4 to implement controlled high-voltage closing.

[0023] Furthermore, the low-voltage isolation state includes disconnecting the high-voltage main circuit from the load side, maintaining the control circuit in a detection-enabled state, and detecting the positive-to-ground insulation resistance, negative-to-ground insulation resistance, and protective grounding circuit continuity of the battery system in this state. The positive-to-ground and negative-to-ground insulation resistance values ​​characterize the battery system's insulation level to ground, while the protective grounding circuit continuity characterizes the integrity and reliability of the grounding path. The insulation pre-inspection conditions are jointly determined based on the insulation resistance threshold, the degree of insulation balance between the positive and negative electrodes to ground, and the grounding continuity detection results. The degree of insulation balance between the positive and negative electrodes to ground is characterized by the difference or ratio between the positive-to-ground insulation resistance and the negative-to-ground insulation resistance, reflecting whether the insulation distribution between the two electrodes of the battery system is balanced. When any indicator fails to meet the preset conditions, an insulation warning or insulation fault judgment is generated, and high-voltage operation permission is locked. For states that are only close to the threshold boundary but have not yet reached the fault judgment standard, an insulation warning judgment is generated and a review mark is retained. For states that have reached the fault judgment conditions, an insulation fault judgment is generated, and subsequent high-voltage commissioning procedures are prohibited.

[0024] Step SS4: Controlled closing of high-voltage circuit and dual authorization Under the conditions that the electrical safety access conditions and insulation pre-inspection conditions are met and the communication link quality meets the set conditions, the high-voltage circuit is triggered to close in a controlled manner based on the operator's authorization confirmation and the system status authorization confirmation. If the double authorization confirmation for closing is not obtained, it will revert to the low-voltage isolation state.

[0025] like Figure 4 As shown, in this embodiment, the controlled closing of the high-voltage circuit is implemented in the following order: system sends closing request, control module pre-verification, system electronic authorization, manual physical key authorization, issuance of closing command, and receipt of circuit breaker feedback. The control module pre-verification includes at least a review of electrical safety access conditions, insulation pre-inspection conditions, and communication status. The closing command can only be issued after both system electronic authorization and manual physical key authorization are completed. The controlled closing of the high-voltage circuit should receive circuit breaker closing feedback within a preset time limit. If no valid feedback is received within the preset time limit, the system reverts to the low-voltage isolation state and records the reason for failure to close. Under conditions of personnel authorization, high-voltage emergency stop confirmation, and insulation pre-inspection conditions being met, the system triggers the controlled closing operation of the high-voltage circuit. This process employs a dual authorization mechanism, including personnel confirmation input and system status interlock control. The operation must be completed within a limited time window; otherwise, the system will automatically revert to the low-voltage isolation state and issue an alarm to ensure the safety and controllability of the high-voltage operation.

[0026] Furthermore, after the control module pre-verification is completed and both electronic authorization and manual physical button authorization are satisfied, the controlled closing of the high-voltage circuit is implemented in the order of pre-charge control stage and main circuit closing stage. Specifically: in the pre-charge control stage, the voltage difference between the battery system side bus and the load side bus is detected, and the bus voltage is established by connecting the pre-charge branch. At the same time, the bus voltage establishment process, the operation status of the pre-charge contactor, and the pre-charge duration are monitored. When the bus voltage difference is lower than the set closing permission threshold and the pre-charge process meets the set timing requirements, the main circuit closing stage is entered. In the main circuit closing stage, a high-voltage closing command is issued, and the feedback status of the main circuit switching devices and the circuit response time are monitored. When the bus voltage establishment rate is abnormal, the contactor status feedback is inconsistent, or any stage exceeds the corresponding time limit threshold, the closing control is immediately canceled and the system is returned to the low-voltage isolation state.

[0027] Step SS5: Multidimensional Data Acquisition and Anomaly Identification After the high-voltage circuit enters the controlled commissioning state, the temperature, voltage, insulation resistance, grounding resistance and communication link quality data of the battery system are collected. A set of characteristic parameters characterizing voltage consistency, temperature rise, insulation attenuation and communication stability is constructed and fusion diagnostic analysis is performed based on this. The results of abnormal unit identification, abnormal type determination and corresponding risk level are output.

[0028] Preferably, the feature parameter set includes at least individual unit voltage deviation parameters, module temperature difference parameters, insulation resistance change rate parameters, grounding resistance fluctuation parameters, and communication connection stability parameters; the fusion diagnostic analysis includes standardizing the feature parameter set, extracting abnormal sensitive features, and jointly judging multi-source information to obtain abnormal unit identification results, abnormal type determination results, and risk levels, wherein the abnormal types include at least insulation abnormalities, voltage consistency abnormalities, thermal abnormalities, grounding abnormalities, and communication abnormalities.

[0029] Furthermore, the integrated diagnostic analysis also includes comparing historical debugging data with current detection data to construct longitudinal evolution parameters characterizing the state changes of the same battery cluster, module, or cell at different time points. These longitudinal evolution parameters include at least the state deviation increment, change slope, and abnormal duration. Based on these parameters, abnormal states are classified as new abnormalities, persistent abnormalities, or worsening abnormalities. Specifically: if the abnormal unit corresponding to the current detection data did not exhibit the same type of abnormality in the previous debugging cycle or a preset historical cycle, and the current state reaches the abnormality judgment threshold for the first time, it is judged as a new abnormality; if the abnormal unit corresponding to the current detection data already exhibited the same type of abnormality in the previous debugging cycle or a preset historical cycle, and the state deviation increment, change slope, and abnormal duration are all within a preset stable range, it is judged as a persistent abnormality; if the abnormal unit corresponding to the current detection data already exhibited the same type of abnormality in the previous debugging cycle or a preset historical cycle, and at least one of the current abnormality amplitude, change slope, or risk level exceeds the corresponding growth threshold, it is judged as a worsening abnormality. The risk level is jointly determined by the current abnormality degree, abnormality type, and abnormality evolution trend.

[0030] As another preferred implementation, such as Figure 5 As shown, the system collects multi-dimensional data such as temperature, voltage, insulation resistance, grounding resistance, and communication link quality, and performs fusion calculation and anomaly identification through a weighted support vector machine algorithm. It comprehensively judges voltage consistency, insulation resistance change rate, communication stability, and temperature rise trend, and finally outputs abnormal units and confidence levels. At the same time, it performs detection process downgrade processing on branches with poor communication quality to achieve accurate monitoring of the power plant status.

[0031] Specifically, when fusion diagnostic analysis is implemented based on a weighted support vector machine classification model, it includes: S501. Multi-source feature quantization and normalization: The voltage, temperature, communication, grounding resistance, and insulation withstand voltage test data are parsed, aligned with the same timestamp, repaired for missing values, and validated. Based on the physical mechanism of faults in electrochemical energy storage power stations, these data are quantified into diagnostic features representing voltage consistency, insulation attenuation, grounding fluctuation, communication stability, and temperature rise trend. These diagnostic features are then normalized to form standardized diagnostic feature vectors. In this embodiment, each diagnostic feature corresponds one-to-one with the measurement outputs of the voltage acquisition unit, insulation detection unit, grounding resistance test unit, communication quality monitoring unit, and infrared thermal imaging unit, thereby achieving an explicit correlation between the algorithm input and the on-site detection results. S502. Classification Model Construction for Energy Storage Scenarios: The training sample source is the historical data of voltage testing, temperature, communication, grounding resistance, and insulation withstand voltage collected during the initial commissioning, periodic re-inspection, and fault location of the energy storage power station. The corresponding abnormal units, abnormal types, and risk levels are used as sample labeling information to construct a weighted support vector machine multi-class classification model. The weights of each feature are determined based on the degree of influence of voltage deviation, temperature difference, insulation change, grounding fluctuation, and communication stability on fault identification. S503. Joint Judgment under Current Operating Conditions: The standardized diagnostic feature vector generated by the current detection is organized according to the battery cluster, module, and cell levels and their differences from historical benchmark values, and then input into a weighted support vector machine multi-class classification model. This model is then combined with rule-based threshold recognition results and pattern recognition results for joint judgment, yielding abnormal unit identification results, abnormal type determination results, and abnormal confidence levels. The preset safety criteria include at least the individual unit voltage deviation threshold, module temperature difference threshold, insulation resistance threshold, grounding resistance fluctuation threshold, and communication quality threshold. The pattern recognition result is the anomaly category and corresponding anomaly confidence level output by the weighted support vector machine multi-class classification model for the current standardized diagnostic feature vector. During joint discrimination, when the rule threshold recognition result is consistent with the pattern recognition result and the anomaly confidence level is not lower than the preset confidence threshold, the corresponding anomaly type judgment result and risk level judgment result are directly output. When the anomaly confidence level is lower than the preset confidence threshold, the corresponding anomaly unit is marked as a low-confidence object to be reviewed, and the risk level judgment result or inclusion in the manual review process is determined in conjunction with the rule threshold recognition result. When the recognition results are inconsistent and the anomaly confidence level is not lower than the preset confidence threshold, the judgment result is output according to the higher risk level and the corresponding anomaly unit is marked as an object to be reviewed. S504. Risk Mapping and Control Output: Generate an anomaly level determination result based on the anomaly type, anomaly confidence level, and current anomaly severity, and map the anomaly level determination result into alarm prompts, alarm records, fault location priority adjustments, or safety interlock control trigger conditions.

[0032] Preferably, between sub-steps S502 and S503, a feature weight adaptive allocation sub-step based on physical state parameters is further included: Based on the battery system health state parameters (SOH) reported in real time by the battery management system (BMS) and the ambient temperature measurement value obtained by the temperature acquisition unit, diagnostic weights are allocated to each feature dimension representing voltage consistency, insulation attenuation, grounding fluctuation, communication stability, and temperature rise trend in the standardized diagnostic feature vector according to a preset segmented mapping rule, forming a weight vector that satisfies normalization constraints; wherein, the preset segmented mapping rule includes at least the SOH normal zone, SOH degradation zone, and SOH severe degradation zone, as well as the ambient temperature normal zone, high temperature warning zone, and high temperature risk zone; when SOH is lower than a set degradation threshold, the diagnostic weights of the feature dimensions corresponding to insulation attenuation and grounding fluctuation are increased; when the ambient temperature is higher than a set high temperature threshold, the diagnostic weights of the feature dimension corresponding to temperature rise trend are increased; when both of the above conditions are met simultaneously, the adjusted feature dimension weights are re-normalized to construct a weighted diagnostic feature vector, and the weighted diagnostic feature vector is input into sub-step S503 for joint judgment of the current operating condition.

[0033] Step SS6: Step-by-step fault location and safety rollback For the identified abnormal units, fault location is performed step by step according to the hierarchical path of battery cluster, module, and cell. Based on the abnormality type and risk level, the located abnormal units are isolated or safety rollback is performed. The safety rollback takes controlled disconnection, test suspension or alarm prompt actions according to the risk level. The isolated object, isolation reason and corresponding timestamp are recorded simultaneously. Fault handling records are established according to the battery cluster, module and cell levels.

[0034] like Figure 6 As shown, based on the abnormal information identified in step SS5, the system narrows down the fault range step by step in the order of string, cluster, and module, and performs branch-level isolation operation on the identified faulty unit, while recording the isolation reason and timestamp; the safety rollback strategy selects controlled disconnection, test suspension or only warning action according to different risk levels, and records the whole process in the system log to achieve traceable management of fault location and safety rollback.

[0035] More specifically, the steps for hierarchical fault location are as follows: Using the anomaly identification results from the fusion diagnostics as input, the system first locates the battery cluster number where the anomaly is located in the three-level topology mapping table and performs a module-level check within the battery cluster. When the module-level check results meet the anomaly criteria, a cell-level check is further performed within the target anomaly module to confirm the faulty unit and record the anomaly cell identifier. When the module-level check results do not meet the anomaly criteria, the system continues to perform module-level verification or maintains the current monitoring status. The criteria for determining whether the module-level check results meet the anomaly criteria include at least the module voltage consistency deviation, module temperature difference, insulation resistance change trend, grounding resistance fluctuation amplitude, and communication status deviation. When any detection parameter in the target module exceeds the corresponding preset threshold, or the deviation from adjacent modules and historical benchmark values ​​exceeds the set judgment range, the target module is identified as an anomaly module and enters the cell-level check process. After confirming the faulty unit, the risk level of the faulty unit is determined, and controlled disconnection, test suspension or alarm prompt actions are executed according to high risk, medium risk and low risk respectively. At the same time, the fault isolation operation, isolation reason, corresponding timestamp and recovery condition prompt are output. The recovery condition prompt is generated based on the fault type, risk level, isolation processing method, current detection parameter recovery status and communication recovery status. When the voltage, temperature, insulation resistance, grounding resistance and communication status corresponding to the abnormal unit are all restored to the preset safety range and continue to meet the set maintenance time, the condition prompt that allows the isolation to be lifted, the detection process to be resumed or the high voltage operation to be resumed is output.

[0036] Preferably, a fault propagation criterion is constructed based on the equipment topology adjacency relationship, electrothermal coupling correlation, and anomaly propagation direction. The equipment topology adjacency relationship includes at least the electrical connection relationship and hierarchical position relationship between battery clusters, modules, and cells. The electrothermal coupling correlation includes at least the voltage deviation transmission relationship, temperature rise diffusion relationship, and insulation anomaly relationship between adjacent units. The anomaly propagation direction is determined based on the change gradient of the anomaly parameters between adjacent units and the propagation sequence. The fault propagation boundary is determined according to the fault propagation criterion, and the branch isolation range is determined accordingly to limit it to the smallest electrical unit that meets safety requirements, so as to balance the fault suppression effect and the continuity of commissioning.

[0037] Furthermore, when performing branch isolation or safety rollback, the system synchronously records the isolation reason, the isolated object, the actions taken, the status information before and after the actions, and the corresponding timestamps. It also establishes fault handling records according to the battery cluster, module, and cell levels for subsequent debugging report generation, health record updates, and fault tracing analysis.

[0038] Step SS7: Generate debug report and update health record Generate a structured commissioning report that includes current testing data, historical comparison differences, trend predictions, and maintenance recommendations. Store the data in association at the equipment level and update the power plant's full life cycle health record.

[0039] like Figure 7 As shown, after completing the step-by-step fault location and rollback operation, the system will store the collected test data in association according to the equipment level, generate a structured commissioning report, and the report content includes current test data, historical comparison differences, trend prediction and maintenance suggestions. It will also be synchronized to the power plant health management system for the whole life cycle health record update and long-term maintenance strategy formulation.

[0040] More specifically, the system takes the communication connection establishment results, low-voltage isolation and insulation pre-inspection data, high-voltage controlled closing logs, multi-source acquisition data and anomaly tags, and fault location and safety rollback records as inputs. First, it performs data cleaning and time series alignment, then rule-based discrimination and diagnostic model analysis to generate structured tags and risk tags representing the equipment status. The risk tags include at least normal tags, warning tags, and fault tags. Based on this, a structured commissioning report is generated according to the battery cluster, module, and cell levels, containing current detection data, historical comparison differences, trend analysis results, and maintenance suggestions. The system also simultaneously updates the power station's full lifecycle health record database.

[0041] In the above embodiments, the entire debugging and testing process supports automatic switching between initial debugging, periodic retesting and fault location modes, and can automatically adjust the test priority based on risk assessment, with high-risk equipment being tested first; a double confirmation mechanism is set for any manual operation step, and all key operations are recorded and tracked by system logs to ensure that the debugging process is safe, controllable and traceable.

[0042] Example 2: Safety Commissioning and Testing System for Electrochemical Energy Storage Power Station Based on Embodiment 1 above, Embodiment 2 further provides a safety commissioning and testing system for electrochemical energy storage power stations. This system is encapsulated in a multi-functional safety inspection box and, based on the aforementioned safety commissioning and testing method for electrochemical energy storage power stations, includes at least: The voltage detection and safety confirmation module is used to detect voltage on the battery compartment door and exposed conductive parts, and to determine whether the electrical safety access conditions are met by combining the operator's on-site status and the initial state of high-voltage emergency stop. The communication connection and device topology identification module is used to establish a communication connection with the battery system and battery management system, identify the topology of the device under test, complete node address allocation, and monitor communication quality. The low-voltage isolation and insulation pre-inspection module is used to place the battery system in a low-voltage isolation state, detect the insulation resistance and grounding status, and determine whether the insulation pre-inspection conditions are met. The high-voltage controlled closing and dual-authorization module is used to trigger the controlled closing of the high-voltage circuit based on operator authorization confirmation and system status authorization confirmation when electrical safety access conditions, insulation pre-inspection conditions and communication quality meet the set conditions. The multi-source data acquisition and fusion diagnostic module is used to acquire temperature, voltage, insulation resistance, grounding resistance and communication quality data of the battery system, and output abnormal unit identification results, abnormal type judgment results and risk level; The step-by-step fault location and safety rollback module is used to locate the identified abnormal units step by step according to the battery cluster, module and cell level path, determine the branch isolation object or isolation range according to the abnormality type, and perform safety rollback processing according to the risk level. The structured report and archive update module is used to perform structured association storage of communication connection establishment results, insulation pre-inspection data, high-voltage controlled closing logs, multi-source detection data, fault location results and safety back-off records, generate commissioning reports and update the power plant's full life cycle health archive; The above-mentioned functional modules can be implemented by hardware circuits, embedded control programs, host computer software programs, or a combination thereof. Among them, the functions of voltage detection, insulation / grounding detection, high and low voltage control and data acquisition are preferably executed by the field hardware units in the inspection box, while the functions of topology identification, joint diagnosis, step-by-step positioning, risk mapping and report generation are preferably executed by the host computer control unit, thus forming a layered system architecture in which field execution and upper-level analysis work together.

[0043] Preferably, the multi-functional safety inspection box also includes an electrical testing unit, a voltage and resistance measurement unit, a grounding resistance measurement power supply, an infrared thermal imaging unit, a data acquisition unit, a high and low voltage safety control unit, and a host computer control unit. Specifically: the electrical testing unit is located in the electrical testing and safety confirmation module and is used to perform electrical testing on the battery compartment door and exposed conductive components; the voltage and resistance measurement unit and the grounding resistance measurement power supply are located in the low-voltage isolation and insulation pre-inspection module and are used to acquire battery system voltage parameters and insulation resistance parameters, as well as to acquire grounding resistance test results; the infrared thermal imaging unit and the data acquisition unit are located in the multi-source data acquisition module. The integrated diagnostic module is used to acquire battery system temperature distribution information and local thermal anomaly image information, and to collect the output data of each detection unit to form a unified diagnostic input dataset. The high and low voltage safety control unit is set in the low voltage isolation and insulation pre-inspection module and the high voltage controlled closing and dual authorization module, and is used to perform low voltage isolation, high voltage controlled closing and safety backoff control. The communication connection and equipment topology identification module, the multi-source data acquisition and integrated diagnostic module and the structured report and file update module are set in the host computer control unit, and are used to realize debugging process control, diagnostic result output and health file update.

[0044] More preferably, the data acquisition unit is used to uniformly access, synchronize, convert formats, and cache multi-source heterogeneous data output from the voltage detection unit, voltage and resistance measurement unit, grounding resistance measurement power supply, communication measurement unit, and infrared thermal imaging unit, forming a unified input data stream for fusion diagnostic analysis. The host computer control unit performs topology mapping, feature construction, joint discrimination, risk mapping, and report generation based on the unified input data stream. The high and low voltage safety control unit executes high-voltage closing control, branch isolation control, and safety backoff control according to the interlocking control commands output by the host computer control unit. Through the collaboration between the above hardware units and software functional modules, this system can possess a complete closed-loop implementation capability from front-end detection, process interlocking, anomaly analysis to back-end archiving.

[0045] In a preferred implementation, the multi-source data acquisition and fusion diagnosis module includes a feature construction submodule, a weighted classification submodule, a joint discrimination submodule, and a risk mapping submodule. Specifically: the feature construction submodule is used to construct standardized diagnostic feature vectors from temperature, voltage, insulation, grounding, and communication data to characterize voltage consistency, temperature rise trend, insulation attenuation, grounding fluctuation, and communication stability; the weighted classification submodule is used to construct or call a weighted support vector machine multi-class classification model based on historically labeled samples; the joint discrimination submodule is used to combine rule-based threshold identification results and model identification results to output abnormal unit identification results, abnormality type determination results, and abnormality confidence levels; and the risk mapping submodule is used to generate risk level determination results based on abnormality type, abnormality degree, abnormality confidence, and evolution trend, and output interlocking control trigger conditions.

[0046] Example 3: Application Case Based on the above embodiments 1 and 2, the present invention will be further described below in conjunction with an exemplary application scenario of a 100MW / 200MWh electrochemical energy storage power station. This power station includes multiple battery compartments, each containing several battery clusters, modules, and cell-level units, and is connected to a BMS, station control system, thermal management system, and high and low voltage control equipment. For such large-scale, multi-level energy storage power stations, during initial commissioning, periodic re-inspection, and fault diagnosis, they typically face problems such as high high-voltage operation risks, intertwined anomaly sources, a large scope of manual troubleshooting, and insufficient utilization of historical data.

[0047] In initial commissioning scenarios, pre-emptive safety constraints are established through voltage verification, low-voltage isolation, insulation pre-inspection, dual-authorized controlled closing, and pre-charge control. Taking the commissioning of a battery compartment as an example, if the system detects that the positive electrode-to-ground insulation resistance has decreased from over 8MΩ in the historical normal range to 3.2MΩ, and the grounding resistance fluctuation exceeds the preset allowable range, this invention can provide early warning or fault determination before high-voltage closing and lock out subsequent high-voltage operations, preventing the defective state from entering the main circuit closing stage. Compared to methods that only detect anomalies after closing, this invention can move the risk interception point forward to before high-voltage operation, thereby significantly reducing the risk of commissioning with faults.

[0048] In periodic re-inspection scenarios, this invention, through unified collection and fusion diagnosis of voltage, temperature, insulation, grounding, and communication status, can identify complex anomalies that are difficult to detect using traditional single-threshold methods. Taking a target module as an example, when the individual unit voltage deviation reaches 28mV, the module temperature difference reaches 6.5℃, the insulation resistance change rate shows a downward trend for several consecutive detection cycles, and the communication message loss rate increases to over 4%, if each indicator has not yet reached its individual shutdown threshold, traditional methods may only issue scattered alarms. However, this invention, based on fusion diagnosis and longitudinal evolution analysis, can determine these as persistent or worsening anomalies and increase their re-inspection priority. This improves the ability to identify complex faults and reduces the probability of missed detections caused by relying solely on single-item limit judgments.

[0049] In fault location and handling scenarios, this invention progressively narrows the location range according to the battery cluster, module, and cell levels, and determines the isolation target and rollback action based on the anomaly type, risk level, and fault propagation criteria. Taking a localized thermal anomaly within a cluster as an example, when infrared thermal imaging shows that the highest local temperature point in a module reaches a temperature rise rate of 1.2℃ / min, and there is a temperature rise diffusion trend in adjacent units, the system can first lock the abnormal battery cluster, then locate the target module, and finally confirm the abnormal cell, and perform branch isolation or controlled disconnection according to the high-risk level. Compared with the crude method of directly shutting down the entire cluster, this invention can limit the isolation range to a smaller electrical unit that meets safety requirements, which is more conducive to balancing fault suppression effectiveness and commissioning continuity.

[0050] At the management level after commissioning, this invention can unify the communication establishment results, insulation pre-inspection data, high-voltage closing logs, multi-source detection results, fault location conclusions, and safety rollback records into a structured commissioning report and write it into a health record database. For energy storage power stations of this scale, this means that subsequent re-inspections and fault handling no longer rely on scattered records, but can be based on historical data managed hierarchically by battery clusters, modules, and cells for trend analysis, priority ranking, and maintenance strategy formulation. Therefore, this invention not only improves the safety and accuracy of individual commissioning but also enhances the systematicness and feasibility of the full life-cycle operation and maintenance management of large-scale electrochemical energy storage power stations.

[0051] The objectives of this invention have been fully and effectively achieved through the above embodiments. Those skilled in the art will understand that this invention includes, but is not limited to, the contents described in the accompanying drawings and the specific embodiments described above. Although the invention has been described with reference to what is currently considered the most practical and preferred embodiments, it should be understood that the invention is not limited to the disclosed embodiments, and any modifications that do not depart from the functional and structural principles of the invention will be included within the scope of the claims.

Claims

1. A method for safe commissioning and testing of an electrochemical energy storage power station, characterized in that, At least the following steps are included: SS1. Perform an electrical test on the battery compartment door and exposed conductive parts, obtain the test results, and determine whether the electrical safety access conditions are met by combining the operator's position status and the initial state of high-voltage emergency stop. If any electrical safety access condition is not met, the subsequent process is prohibited. SS2. Establish communication links with the battery system and BMS, identify the topology of the device under test and complete node address allocation, while monitoring the quality of the communication link, and performing node reconnection or bypass processing when the quality of the communication link does not meet the set conditions. SS3. Before performing high-voltage operation, place the battery system in a low-voltage isolation state, check its insulation resistance and grounding status, and determine whether the insulation pre-inspection conditions are met. If the insulation resistance is lower than the preset safety threshold and / or the grounding is abnormal, then lock out the subsequent high-voltage operation. SS4. Under the conditions that the electrical safety access conditions and insulation pre-inspection conditions are met and the communication link quality meets the set conditions, the high-voltage circuit is triggered to close in a controlled manner based on the operator's authorization confirmation and the system status authorization confirmation. If the double authorization confirmation for closing is not obtained, it will revert to the low-voltage isolation state. SS5. After the high-voltage circuit enters the controlled commissioning state, the temperature, voltage, insulation resistance, grounding resistance and communication link quality data of the battery system are collected. A set of characteristic parameters characterizing voltage consistency, temperature rise change, insulation attenuation and communication stability is constructed and fusion diagnostic analysis is performed based on this. The abnormal unit identification results, abnormal type judgment results and corresponding risk levels are output. SS6. For the identified abnormal units, fault location is performed according to the hierarchical path of battery cluster, module, and cell, and the abnormal units are isolated or safely rolled back based on the abnormality type and risk level.

2. The method according to claim 1, characterized in that, In step SS1, the voltage testing operation includes at least testing the high-voltage interface inside the battery compartment door, the busbar connection, and the exposed conductive parts point by point. The voltage testing results are then jointly verified with the operator's arrival confirmation information and the high-voltage emergency stop status information. Only when the voltage testing results show that there is no dangerous liveness, the operator's identity is verified, and the high-voltage emergency stop circuit is in the initial disconnected state, is it determined that the electrical safety access conditions are met.

3. The method according to claim 1, characterized in that, In step SS2, the establishment of the communication link supports negotiation and switching of multiple protocols including CAN, RS-485, and Industrial Ethernet. Industrial Ethernet is used for data transmission under high-speed data acquisition conditions, RS-485 is used for data transmission under compatibility scenarios, and CAN is switched to for data transmission when communication interference exceeds a set threshold. The identification of the device topology is achieved by sending topology query messages to each node and receiving node response information. The identification content includes the battery cluster number and quantity, the module arrangement order and rated parameters within each battery cluster, and the number and nominal voltage of cells within each module. After topology identification, the device topology is compared with historical topology records. Nodes with abnormal topologies are marked, and node address correction or communication recovery is performed. After topology comparison and abnormal node processing are completed, a three-level hierarchical topology mapping table for battery clusters, modules, and cells is established, and node address allocation is completed accordingly.

4. The method according to claim 1, characterized in that, In step SS3, the low-voltage isolation state includes disconnecting the high-voltage main circuit from the load side, keeping the control circuit in the detection enabled state, and then detecting the insulation resistance value of the positive electrode to ground, the insulation resistance value of the negative electrode to ground, and the continuity of the protective grounding circuit. The insulation pre-inspection conditions are determined based on the insulation resistance threshold, the insulation balance between the positive and negative electrodes to ground, and the grounding continuity detection results. The insulation balance between the positive and negative electrodes to ground is characterized by the difference or ratio between the insulation resistance value of the positive electrode to ground and the insulation resistance value of the negative electrode to ground. When any indicator does not meet the preset conditions, an insulation warning or insulation fault judgment is generated and the high-voltage operation permission is locked.

5. The method according to claim 1, characterized in that, In step SS4, the controlled closing of the high-voltage circuit is carried out in the following order: system sending closing request, control module pre-verification, system electronic authorization, manual physical key authorization, issuance of closing command, and receipt of circuit breaker feedback. The control module pre-verification includes at least a review of electrical safety access conditions, insulation pre-inspection conditions, and communication status. The closing command can only be issued after both system electronic authorization and manual physical key authorization are completed. Furthermore, the controlled closing of the high-voltage circuit should receive circuit breaker closing feedback within a preset time limit. If no valid feedback is received within the preset time limit, the circuit breaker will automatically revert to the low-voltage isolation state and record the reason for failure to close.

6. The method according to claim 5, characterized in that, In step SS4, the controlled closing of the high-voltage circuit is implemented in the order of the pre-charge control stage and the main circuit closing stage. Specifically: In the pre-charge control stage, the voltage difference between the battery system side bus and the load side bus is detected, and the bus voltage is established by connecting the pre-charge branch. At the same time, the bus voltage establishment process, the operation status of the pre-charge contactor, and the pre-charge duration are monitored. When the bus voltage difference is lower than the set closing allowable threshold and the pre-charge process meets the set timing requirements, the main circuit closing stage is entered and a high-voltage closing command is issued. The feedback status of the main circuit switching devices and the circuit response time are monitored. When the bus voltage establishment rate is abnormal, the contactor status feedback is inconsistent, or any stage exceeds the corresponding time limit threshold, the closing control is canceled and the circuit returns to the low-voltage isolation state.

7. The method according to claim 1, characterized in that, In step SS5, the feature parameter set includes at least individual unit voltage deviation parameters, module temperature difference parameters, insulation resistance change rate parameters, grounding resistance fluctuation parameters, and communication connection stability parameters; the fusion diagnostic analysis includes standardizing the feature parameter set, extracting abnormal sensitive features, and jointly judging multi-source information to obtain abnormal unit identification results, abnormal type determination results, and risk levels, wherein the abnormal types include at least insulation abnormalities, voltage consistency abnormalities, thermal abnormalities, grounding abnormalities, and communication abnormalities.

8. The method according to claim 7, characterized in that, In step SS5, the fusion diagnostic analysis further includes comparing the differences between historical debugging data and current detection data to construct longitudinal evolution parameters characterizing the state changes of the same battery cluster, module, or cell at different time points. The longitudinal evolution parameters include at least the state deviation increment, the change slope, and the duration of the abnormality. Based on the longitudinal evolution parameters, the abnormal state is determined as a new abnormality, a continuous abnormality, or a worsening abnormality. The risk level is jointly determined by the current degree of abnormality, the type of abnormality, and the trend of abnormality evolution.

9. The method according to claim 1, characterized in that, In step SS5, the fusion diagnostic analysis is implemented based on a weighted support vector machine classification model, and includes the following sub-steps: S501. Multi-source feature quantization and normalization processing: The voltage detection, voltage acquisition data, temperature acquisition data, communication acquisition data, grounding resistance test data, and insulation withstand voltage test data are parsed, aligned with the same timestamp, missing value repaired, and data verified, and respectively quantified into diagnostic feature quantities characterizing voltage consistency, insulation attenuation degree, grounding fluctuation degree, communication stability, and temperature rise trend. Then, the diagnostic feature quantities are normalized to form a standardized diagnostic feature vector. S502. Classification Model Construction for Energy Storage Scenarios: The training sample source is the historical data of voltage testing, temperature, communication, grounding resistance, and insulation withstand voltage collected during the initial commissioning, periodic re-inspection, and fault location of the energy storage power station. The corresponding abnormal units, abnormal types, and risk levels are used as sample labeling information to construct a weighted support vector machine multi-class classification model. The weights of each feature are determined based on the degree of influence of voltage deviation, temperature difference, insulation change, grounding fluctuation, and communication stability on fault identification. S503. Joint discrimination under current operating conditions: The standardized diagnostic feature vector formed by the current detection is organized according to the battery cluster, module and cell level and its difference from the historical benchmark value, and then input into the weighted support vector machine multi-class classification model. The joint discrimination is performed by combining the rule threshold recognition result and the pattern recognition result to obtain the abnormal unit identification result, the abnormal type determination result and the abnormal confidence level. S504. Risk Mapping and Control Output: Generate an anomaly level determination result based on the anomaly type, anomaly confidence level, and current anomaly severity, and map the anomaly level determination result into alarm prompts, alarm records, fault location priority adjustments, or safety interlock control trigger conditions.

10. The method according to claim 9, characterized in that, In step S503, the rule threshold identification result is obtained by comparing the current detection data based on preset safety criteria, wherein the preset safety criteria include at least the single-unit voltage deviation threshold, module temperature difference threshold, insulation resistance threshold, grounding resistance fluctuation threshold, and communication quality threshold; the pattern recognition result is the anomaly category and corresponding anomaly confidence level output by the weighted support vector machine multi-class classification model for the current standardized diagnostic feature vector; during joint discrimination, when the rule threshold identification result is consistent with the pattern recognition result and the anomaly confidence level is not lower than the preset confidence threshold, the corresponding anomaly type judgment result and risk level judgment result are directly output. When the anomaly confidence level is lower than the preset confidence threshold, the corresponding anomaly unit is marked as a low-confidence object to be reviewed, and the risk level judgment result is determined based on the rule threshold identification result or included in the manual review process. When the identification results are inconsistent and the anomaly confidence level is not lower than the preset confidence threshold, the judgment result is output according to the higher risk level and the corresponding abnormal unit is marked as an object to be reviewed.

11. The method according to claim 1, characterized in that, In step SS6, the specific steps for fault location at each level, following the path of battery cluster, module, and cell, are as follows: Using the anomaly identification results output by the fusion diagnostics as input, the battery cluster number where the anomaly is located is first locked in the three-level topology mapping table, and module-level checks are performed within the battery cluster. When the module-level check results meet the anomaly criteria, cell-level checks are further performed within the target anomaly module to confirm the faulty unit and record the anomaly cell identifier. When the module-level inspection results do not meet the abnormality criteria, the module-level review continues or the current monitoring status is maintained. The criteria for determining whether the module-level inspection results meet the abnormality criteria include at least the module voltage consistency deviation, module temperature difference, insulation resistance change trend, grounding resistance fluctuation amplitude, and communication status deviation. When any detection parameter in the target module exceeds the corresponding preset threshold, or the deviation from the adjacent module and historical benchmark value exceeds the set judgment range, the target module is identified as an abnormal module and enters the cell-level inspection process. After confirming the faulty unit, the risk level of the faulty unit is determined, and controlled disconnection, test pause or alarm prompt actions are executed according to high risk, medium risk and low risk respectively. At the same time, the fault isolation operation, isolation reason, corresponding timestamp and recovery condition prompt are output.

12. The method according to claim 1, characterized in that, In step SS6, a fault propagation criterion is constructed based on the device topology adjacency relationship, electrothermal coupling correlation, and anomaly propagation direction. The device topology adjacency relationship includes at least the electrical connection relationship and hierarchical position relationship between battery clusters, modules, and cells. The electrothermal coupling correlation includes at least the voltage deviation transmission relationship, temperature rise diffusion correlation, and insulation anomaly correlation between adjacent units. The anomaly propagation direction is determined based on the change gradient of the anomaly parameters between adjacent units and the propagation sequence. The fault propagation boundary is determined according to the fault propagation criterion, and the branch isolation range is determined accordingly to limit it to the smallest electrical unit that meets the safety requirements.

13. A safety commissioning and testing system for an electrochemical energy storage power station, characterized in that, The system is encapsulated in a multi-functional safety inspection box and performs safety commissioning and testing of an electrochemical energy storage power station based on the method described in any one of claims 1 to 12. The system includes at least: The voltage detection and safety confirmation module is used to detect voltage on the battery compartment door and exposed conductive parts, and to determine whether the electrical safety access conditions are met by combining the operator's on-site status and the initial state of high-voltage emergency stop. The communication connection and device topology identification module is used to establish communication connections with the battery system and BMS, identify the topology of the device under test, complete node address allocation, and monitor communication quality. The low-voltage isolation and insulation pre-inspection module is used to place the battery system in a low-voltage isolation state, detect the insulation resistance and grounding status, and determine whether the insulation pre-inspection conditions are met. The high-voltage controlled closing and dual-authorization module is used to trigger high-voltage controlled closing based on personnel authorization and system authorization when electrical safety access conditions, insulation pre-inspection conditions and communication quality meet the set conditions. The multi-source data acquisition and fusion diagnostic module is used to acquire temperature, voltage, insulation resistance, grounding resistance and communication quality data, and output abnormal unit identification, abnormal type judgment and risk level; The step-by-step fault location and safety rollback module is used to locate the identified abnormal units step by step according to the battery cluster, module and cell level path, determine the branch isolation object or isolation range according to the abnormality type, and perform safety rollback processing according to the risk level. The structured report and archive update module is used to perform structured association storage of communication connection establishment results, insulation pre-inspection data, high-voltage controlled closing logs, multi-source detection data, fault location results, and safety rollback records, generate commissioning reports, and update the power plant's full life cycle health archive.

Images