Multi-area safety monitoring method and system for energy storage power station

Abstract

The invention discloses a multi-area safety monitoring method and system for an energy storage power station. Real-time data streams of temperature, smoke and voltage are obtained from a plurality of sensors of an energy storage system; real-time data streams of different communication protocols are converted into a unified format through a protocol conversion module, and standardized data streams are generated; processing the standardized data flow by adopting a data fusion algorithm, and calculating a weighted average fusion value according to the type of the sensor and historical data reliability; whether the weighted average fusion value exceeds a preset fire risk threshold value or not is judged, and if the weighted average fusion value exceeds the preset fire risk threshold value, the weighted average fusion value is input into a convolutional neural network for classification processing; and determining a fire risk level according to a classification result and generating an early warning signal. According to the invention, accuracy and real-time performance of fire risk assessment are improved, expansibility and adaptability of the system are enhanced, and effective guarantee is provided for safe operation of the energy storage system.

Description

Technical Field

[0001] The present invention relates to the technical field of fire safety of energy storage systems, and in particular discloses a multi-area safety monitoring method and system for an energy storage power station. Background Art

[0002] Energy storage system fire safety technology is a key area to ensure the stable operation of industrial and commercial energy storage power stations and the safety of people and property.

[0003] With the rapid development of the new energy industry, high-energy-density energy storage matrices are widely used in industrial and commercial scenarios, and their fire safety needs are becoming increasingly prominent.

[0004] Energy storage systems pose a significant fire risk due to their high voltage, chemical storage, and complex operating environments. Accidents can lead to significant economic losses and even threaten lives. Therefore, developing efficient and reliable fire monitoring technology has become a key issue for the energy storage industry.

[0005] Existing fire protection solutions have significant limitations when dealing with the complex demands of energy storage matrices:

[0006] 1. Traditional fire protection systems mostly rely on a single sensor, with decentralized data processing, making it difficult to achieve multi-parameter collaborative analysis, resulting in delayed early warning responses.

[0007] 2. The communication protocols between fire-fighting equipment and energy storage monitoring systems are incompatible with each other. For example, the separation of the fire-fighting bus and the industrial protocol hinders data sharing and linkage control.

[0008] 3. Faced with the dynamic expansion of the energy storage matrix, the current solution lacks modularity and is difficult to flexibly adapt to the fire protection needs of newly added energy storage units.

[0009] 4. The core challenge of energy storage matrix fire safety stems from the integration difficulties of multi-protocol communications.

[0010] 5. The protocol isolation between firefighting equipment and energy storage monitoring systems prevents real-time data interaction, which in turn limits the rapid identification and response to fire risks.

[0011] 6. Protocol isolation further increases the difficulty of multi-sensor data fusion. It is difficult to uniformly analyze multi-source signals such as temperature and smoke, which reduces the accuracy of early warning.

[0012] 7. Insufficient data fusion directly affects the realization of modular deployment. Due to the lack of a unified data processing framework, the fire protection module is difficult to achieve plug-and-play, which limits the scalability of the system.

[0013] Therefore, how to design a fire communication module that can integrate multi-protocol communications, realize efficient collaborative analysis of multi-sensor data, and support modular and flexible deployment has become a key issue in the field of energy storage matrix fire safety. Summary of the Invention

[0014] The present invention provides a multi-area safety monitoring method and system for energy storage power station, aiming to solve

[0015] One aspect of the present invention relates to a multi-area safety monitoring method for an energy storage power station, comprising the following steps:

[0016] Obtain real-time data streams on temperature, smoke, and voltage from multiple sensors in the energy storage system;

[0017] The protocol conversion module converts the real-time data streams of different communication protocols into a unified format to generate standardized data streams;

[0018] A data fusion algorithm is used to process the standardized data stream and a weighted average fusion value is calculated based on the type of sensor and the reliability of historical data;

[0019] Determine whether the weighted average fusion value exceeds the preset fire risk threshold. If the weighted average fusion value exceeds the preset fire risk threshold, input the weighted average fusion value into the convolutional neural network for classification processing;

[0020] Determine the fire risk level based on the classification results and generate early warning signals;

[0021] Transmit the warning signal to the fire control module through the modular communication interface to obtain control instructions;

[0022] Use a dynamic expansion interface to distribute control instructions to the fire protection module of the newly added energy storage unit;

[0023] Determine whether the control instructions are suitable for modular deployment requirements and output the execution results.

[0024] Furthermore, the steps of obtaining real-time data streams of temperature, smoke, and voltage from multiple sensors of the energy storage system include:

[0025] Obtain raw data streams of temperature, smoke, and voltage from multiple sensors in the energy storage system;

[0026] The data in the original data stream is denoised using a preset signal processing algorithm to obtain real-time data streams of temperature, smoke, and voltage.

[0027] Furthermore, the steps of converting the real-time data streams of different communication protocols into a unified format by the protocol conversion module to generate a standardized data stream include:

[0028] Acquire real-time data streams from input devices of multiple communication protocols, and use a preset protocol parser to decode data packets of the real-time data streams to obtain an initial data stream containing a timestamp and data format;

[0029] Determine whether the timestamps of the initial data stream are aligned. If the timestamps of the initial data stream are not aligned, sort the data packets of the initial data stream through the message queue mechanism to generate a time-synchronized data stream.

[0030] The time-synchronized data stream is subjected to format standardization processing according to preset conversion rules, and the time-synchronized data stream is repackaged in JSON format to obtain a standardized data stream.

[0031] Furthermore, the steps of using a data fusion algorithm to process the standardized data stream and calculating a weighted average fusion value according to the type of sensor and the reliability of historical data include:

[0032] Obtain standardized data streams from multiple data sources, classify the standardized data streams according to a preset sensor identification table, and obtain classified data streams;

[0033] For the classified data stream, historical data records are obtained and the integrity of the historical data is verified using a cyclic redundancy check algorithm. If the integrity is passed, the reliability score is calculated based on the sensor type and timestamp to obtain the reliability weight;

[0034] A weighted average algorithm is used to process the classified data stream, and weighted calculation is performed based on the reliability weight and data priority. The weighted calculation results are encapsulated in JSON format to obtain a fusion value data stream.

[0035] The fusion value data stream is time-synchronized according to the preset fusion frequency, and the data packets of the fusion value data stream are sorted using the message queue mechanism to generate a fusion value output stream with a timestamp and checksum tag.

[0036] Furthermore, determining whether the weighted average fusion value exceeds a preset fire risk threshold, if the weighted average fusion value exceeds the preset fire risk threshold, inputting the weighted average fusion value into a convolutional neural network for classification processing includes:

[0037] Obtain standardized data streams and use a cyclic redundancy check algorithm to verify the integrity of the standardized data streams. If the integrity passes, calculate the weight based on the sensor type and timestamp to obtain the reliability weight value;

[0038] Based on the reliability weight value and data priority, a weighted average algorithm is used to process the standardized data stream, and the calculation results are encapsulated in JSON format to obtain the fused value data stream;

[0039] Determine whether the fusion value in the fusion value data stream exceeds the preset fire risk threshold. If the fusion value exceeds the risk threshold, input the fusion value data stream into the convolutional neural network for classification to obtain a classification result;

[0040] The message queue mechanism is used to sort the classification results, and time synchronization is performed through timestamps and check tags to generate classification output streams.

[0041] Furthermore, the steps of determining the fire risk level and generating an early warning signal according to the classification result include:

[0042] Obtain the classification output stream, process it using a time series analysis algorithm, and generate a fire risk level sequence;

[0043] According to the fire risk level sequence, determine whether the level value exceeds the preset warning threshold. If it exceeds, generate a warning signal data stream;

[0044] Prioritize the warning signal data stream through the message queue mechanism to obtain the sorted signal stream;

[0045] Use data verification algorithm to verify the integrity of the sorted signal stream and generate a signal stream that passes the verification;

[0046] According to the signal stream that has passed the verification, the time synchronization mechanism is used to align the timestamps of the signal stream that has passed the verification to obtain a synchronized signal stream;

[0047] The synchronous signal stream is classified through a preset signal mapping table to determine the early warning signal.

[0048] Furthermore, the step of transmitting the warning signal to the fire control module through the modular communication interface to obtain a control instruction includes:

[0049] Obtain the warning signal data stream, use the message queue mechanism to prioritize the data stream, and obtain the sorted signal stream;

[0050] The integrity of the sorted signal stream is verified by a preset signal verification algorithm to obtain a signal stream that passes the verification;

[0051] Use the time synchronization mechanism to align the timestamps of the signal stream that has passed the verification to obtain a synchronized signal stream;

[0052] The synchronous signal stream is transmitted to the fire control module through the modular communication interface to generate control instructions.

[0053] Furthermore, the steps of distributing the control instructions to the fire protection module of the newly added energy storage unit using the dynamic expansion interface include:

[0054] Receive a registration request message carrying the unique identifier of a newly added energy storage unit. The registration request message is sent when the newly added energy storage unit is connected to the dynamic expansion interface. Use a preset unit registration protocol to identify the unit and obtain a list of registered energy storage units.

[0055] According to the list of registered energy storage units, the status monitoring data of each unit is obtained, and the message queue mechanism is used to prioritize the status monitoring data to obtain the sorted status data stream;

[0056] For the sorted state data stream, if the state data meets the preset firefighting task threshold, the corresponding control instructions are generated through the dynamic extension interface to obtain the instruction set to be assigned;

[0057] The instruction set to be assigned is transmitted to the target fire protection module through the module communication protocol, and the instruction verification algorithm is used to perform integrity verification on the instruction set to be assigned to determine the assigned control instructions.

[0058] Furthermore, the steps of determining whether the control instruction is adapted to the modular deployment requirements and outputting the execution result include:

[0059] Obtain the deployment requirement data of the target module from the preset module configuration database, classify the deployment requirement data using the rule engine, and obtain a classified requirement data set;

[0060] The control instruction set to be judged is received through the communication protocol. For the classified demand data set, if the parameters of the control instruction set meet the preset threshold parameters, a logic matching algorithm is used to generate an adaptation identifier to obtain an adapted instruction set;

[0061] The message queue mechanism is used to prioritize the adapted instruction set, and the sorting order is adjusted according to the real-time status monitoring data of the target module to obtain the sorted instruction sequence;

[0062] The sorted instruction sequence is transmitted through the instruction execution interface of the target module, and the sorted instruction sequence is verified using an integrity check algorithm to determine whether the execution result is generated.

[0063] Another aspect of the present invention relates to a multi-area safety monitoring system for an energy storage power station, which is used to implement the above-mentioned multi-area safety monitoring method for an energy storage power station. The multi-area safety monitoring system for an energy storage power station includes:

[0064] A first acquisition module is used to obtain real-time data streams of temperature, smoke and voltage from multiple sensors of the energy storage system;

[0065] A first generating module is used to convert the real-time data streams of different communication protocols into a unified format through a protocol conversion module to generate a standardized data stream;

[0066] The calculation module is used to process the standardized data stream using the data fusion algorithm and calculate the weighted average fusion value according to the type of sensor and the reliability of historical data;

[0067] The first judgment module is used to judge whether the weighted average fusion value exceeds the preset fire risk threshold. If the weighted average fusion value exceeds the preset fire risk threshold, the weighted average fusion value is input into the convolutional neural network for classification processing;

[0068] The second generation module is used to determine the fire risk level according to the classification results and generate an early warning signal;

[0069] A second acquisition module is used to transmit the warning signal to the fire control module through the modular communication interface to obtain a control instruction;

[0070] A distribution module for distributing control instructions to the fire protection modules of the newly added energy storage units using a dynamic expansion interface;

[0071] The second judgment module is used to judge whether the control instruction is adapted to the modular deployment requirements and output the execution result.

[0072] The beneficial effects achieved by the present invention are:

[0073] The present invention provides a multi-area safety monitoring method and system for an energy storage power station, which realizes real-time monitoring of parameters such as temperature, smoke and voltage through multi-sensor data acquisition, protocol conversion and data fusion. A weighted average algorithm is used to process the standardized data stream, and a convolutional neural network is combined to classify the fire risk level and generate an early warning signal. The present invention also designs a modular communication interface and a dynamic expansion interface, which can flexibly allocate control instructions to the fire protection module of the newly added energy storage unit to meet the needs of modular deployment. The multi-area safety monitoring method and system for an energy storage power station provided by the present invention not only improves the accuracy and real-time performance of fire risk assessment, but also enhances the scalability and adaptability of the system, providing effective protection for the safe operation of the energy storage system. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 The figure is a flow chart of an embodiment of a multi-area safety monitoring method for an energy storage power station according to the present invention. DETAILED DESCRIPTION

[0075] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0076] like Figure 1 As shown, the first embodiment of the present invention provides a multi-area safety monitoring method for an energy storage power station, comprising the following steps:

[0077] Step S100: Acquire real-time data streams of temperature, smoke, and voltage from multiple sensors of the energy storage system.

[0078] An energy storage system (ESS) is a system that stores energy in a specific form and releases it on demand when needed. The core function of an ESS is to address the mismatch between energy supply and demand in time and space, improving energy efficiency and system stability.

[0079] Real-time data stream refers to a data sequence generated and transmitted in a continuous and dynamic manner. The core characteristics of real-time data stream are the immediacy, high timeliness and continuous transmission capability of data.

[0080] Step S200: converting the real-time data streams of different communication protocols into a unified format through a protocol conversion module to generate a standardized data stream.

[0081] The protocol conversion module is a dedicated hardware or software component used to resolve differences in communication protocols between heterogeneous networks or devices. Its core function is to achieve compatible conversion of data formats, transmission rules and semantics between different protocols.

[0082] Standardized data flow refers to the systematic process of normalizing continuously generated and dynamically transmitted data sequences. The goal of standardized data flow is to eliminate the processing barriers caused by the heterogeneity of data sources and achieve efficient integration and analysis across systems.

[0083] Step S300: Use a data fusion algorithm to process the standardized data stream, and calculate a weighted average fusion value based on the type of sensor and the reliability of historical data.

[0084] Data fusion algorithms are a collection of processing methods that use mathematical models and computational techniques to integrate, correlate, and optimize multi-source, heterogeneous data to generate more accurate, complete, or advanced information. Their core goal is to eliminate data redundancy and conflict, thereby enhancing decision support capabilities.

[0085] Historical data reliability refers to the ability of historical data sets to maintain their accuracy, completeness, consistency and traceability during long-term storage and use.

[0086] Weighted average fusion is a multi-source data integration method based on weighted distribution. By assigning different weights to different data sources or features and calculating their weighted average, it produces a more reliable or representative fusion result. The core goal of weighted average fusion is to reduce the impact of noise and improve the robustness of the fusion result by quantifying the confidence, accuracy, or importance of the data sources.

[0087] Step S400: Determine whether the weighted average fusion value exceeds a preset fire risk threshold. If the weighted average fusion value exceeds the preset fire risk threshold, input the weighted average fusion value into a convolutional neural network for classification processing.

[0088] Fire risk thresholds are critical values or thresholds used to define acceptable risk in fire risk assessments. Their core function is to quantify risk tolerance and provide a basis for decision-making in risk management. When the actual risk level exceeds the threshold, emergency intervention measures are initiated. When it falls below the threshold, the risk is considered manageable.

[0089] A convolutional neural network (CNN) is a deep learning model designed specifically for processing grid-like topological data (such as images, video, and audio). Its core approach is to achieve efficient feature extraction and pattern recognition through convolution, pooling, and hierarchical feature abstraction. CNN significantly reduces model complexity through local receptive fields and parameter sharing, making it suitable for tasks such as image classification, object detection, and semantic segmentation.

[0090] Classification refers to the process of extracting features and identifying patterns from input data using an algorithm or model, ultimately classifying it into pre-defined categories. In convolutional neural networks (CNNs), this process relies on a hierarchical network structure to achieve step-by-step abstraction from low-level features to high-level semantics, outputting classification results based on a probability distribution.

[0091] Step S500: Determine the fire risk level according to the classification result and generate an early warning signal.

[0092] Fire risk rating is a tiered assessment system for locations, regions, or environments based on factors such as the potential hazard of a fire, its speed of spread, the difficulty of extinguishing it, and the severity of its consequences. It aims to quantify risk levels and guide prevention and control measures. Its definition is typically based on a comprehensive assessment of multiple parameters, including casualty risk, property loss threshold, combustible material characteristics, and environmental conditions.

[0093] Early warning signals are information-based instructions that use monitoring systems or assessment models to identify and warn of potential risks or disasters in advance, aiming to trigger emergency responses and reduce the risk of loss. Essentially, they represent a visual representation of risk quantification, often combining threshold determination, dynamic data tracking, and multi-source information fusion to form a tiered early warning mechanism.

[0094] Step S600: Transmit the warning signal to the fire control module through the modular communication interface to obtain a control instruction.

[0095] A modular communication interface is a hardware or software interaction specification designed according to standardized protocols, enabling data exchange and functional collaboration between modules. Its core principle is to ensure independent development and expansion of each module through decoupling while maintaining the overall functional integrity of the system. This interface definition covers physical connection specifications, data format standards, and control instruction sets, forming a unified language for communication between modules.

[0096] The fire control module is the core component of the fire linkage control system. It is an intelligent control unit that receives fire alarm signals and coordinates external firefighting equipment to execute safety operations such as fire extinguishing and evacuation. Through hardware interfaces and software protocols, it enables centralized monitoring and coordinated response of the fire protection system. The fire control module's core functions include signal analysis, equipment control, and status feedback, forming the key hub for automatic fire alarm and emergency response.

[0097] Control commands are standardized operational commands issued by the fire linkage controller to external firefighting equipment. They are used to start, stop, or adjust the equipment's operating status, ensuring that fire emergency response is executed according to pre-set logic. Essentially, they are digital or electrical signals containing the target device address, action type, and parameter thresholds. These signals enable precise interaction between devices via a bus or direct lines.

[0098] Step S700: Use the dynamic expansion interface to distribute the control instructions to the fire protection module of the newly added energy storage unit.

[0099] The Dynamic Expansion Interface is a standardized interaction protocol that allows the flexible integration of new functional modules or devices during system runtime (without downtime or restart). Through plug-and-play mechanisms and adaptive resource allocation, it enables seamless integration and functional expansion of hardware or software components. Its core goal is to break through system boundaries, support on-demand adjustments to scale and capabilities, and adapt to the rapid iteration requirements of complex scenarios.

[0100] New energy storage units are independent, dynamically expanded units within an existing energy storage system. These units consist of a battery pack, a battery management system (BMS), and a power conversion system (PCS). They offer comprehensive energy storage and release capabilities and seamlessly integrate with the main system to enable on-demand capacity or power expansion. Their core goal is to enhance the flexibility and scalability of the energy storage system to accommodate load fluctuations or upgrade scenarios.

[0101] The fire protection module is the core functional unit in the automatic fire alarm system that implements device linkage control. It receives controller commands, drives external fire protection equipment, and simultaneously provides feedback on the equipment's operating status to the main system, forming a complete fire emergency response chain. Essentially, it is a standardized component that combines signal acquisition, logic processing, and electrical control.

[0102] Step S800: determine whether the control instruction is adapted to the modular deployment requirements and output the execution result.

[0103] Modular deployment is a technical requirement in system architecture design for functional decoupling, independent configuration, and elastic expansion. It emphasizes breaking complex systems into standardized functional modules with high cohesion and low coupling. These modules support on-demand assembly, independent upgrades, and dynamic loading to reduce operational costs and improve scenario adaptability. Its core value lies in ensuring business continuity and efficient resource utilization through standardized interfaces and automated management.

[0104] Output execution results are a key process in the fire linkage control system, where the output module converts logical instructions into physical actions. By receiving control signals from the controller, it drives external firefighting equipment to execute preset operations (such as activating alarms and opening smoke exhaust valves), and then feeds back the action status to the main system to form a closed-loop control link. Essentially, this process maps and converts weak electrical signals into strong electrical control capabilities, ensuring the timeliness and reliability of fire emergency response.

[0105] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S100:

[0106] Step S110: Obtaining raw data streams of temperature, smoke, and voltage from multiple sensors of the energy storage system.

[0107] The technical topics of sensor data processing for energy storage systems involve raw data stream acquisition, sensor fusion, time synchronization, signal processing, and denoising. The following will be expanded through specific examples and analysis, focusing on the implementation and effectiveness of technologies in energy storage system scenarios.

[0108] The acquisition of raw data streams is fundamental. Temperature, smoke, and voltage sensors in energy storage systems monitor the status of battery modules. For example, a temperature sensor collects data once per second, with typical values between 25 and 60 degrees Celsius.

[0109] The smoke sensor detects the concentration of particulate matter in the air and outputs an analog signal from 0 to 500 micrograms per cubic meter; the voltage sensor monitors the battery cell voltage in the range of 3.2 to 4.2 volts.

[0110] Ideally, these sensors are connected to a data acquisition unit via high-speed sampling circuitry to ensure a continuous data flow. The acquisition frequency is uniformly set at 1 Hz to avoid subsequent synchronization difficulties. This ensures the integrity of multi-source data and lays the foundation for subsequent processing.

[0111] Sensor fusion technology aims to integrate data from multiple sources to improve information reliability. In one embodiment, an energy storage system is deployed in a new energy power plant, with multiple sensors working together within the battery compartment. The fusion process must account for differences in sensor accuracy, such as a temperature sensor with an accuracy of ±0.5 degrees Celsius and a smoke sensor with an accuracy of ±10 micrograms per cubic meter.

[0112] For example, the fusion algorithm can use a weighted averaging approach to assign higher weights to high-precision sensor data, generating a unified fused data stream. This approach improves data consistency and reduces the risk of single sensor failure.

[0113] Step S120: De-noising the data in the original data stream using a preset signal processing algorithm to obtain real-time data streams of temperature, smoke, and voltage.

[0114] Signal processing algorithms are key to removing noise from raw data streams. Specifically, energy storage system data is often affected by electromagnetic interference or environmental noise, such as the superposition of 50 Hz power frequency interference on voltage signals.

[0115] Preferably, a low-pass filter is used to remove high-frequency noise and retain valid signals from 0 to 10 Hz. For example, the original voltage data fluctuates between 3.8 and 4.0 volts and stabilizes at around 3.9 volts after denoising.

[0116] In another embodiment, median filtering can be combined to remove sudden spikes. For example, if smoke concentration jumps to 600 micrograms per cubic meter due to instantaneous airflow disturbance, it will return to the normal value of 450 micrograms per cubic meter after filtering. This denoising process improves data reliability and provides an accurate basis for real-time monitoring.

[0117] As you can see, these technologies work together to generate real-time data streams on temperature, smoke, and voltage, supporting the safe operation of the energy storage system. For example, this real-time data stream can be fed into an anomaly detection model to trigger an alarm when the temperature exceeds 60 degrees Celsius or the smoke concentration exceeds 500 micrograms per cubic meter. This high-precision, low-noise data stream significantly improves the system's response to overheating or fire risks, ensuring battery safety and stable power station operation.

[0118] Each technical link has strict logic and supports each other to jointly achieve the goal of efficient data processing.

[0119] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S200:

[0120] Step S210: Acquire real-time data streams from input devices of multiple communication protocols, and use a preset protocol parser to decode data packets of the real-time data streams to obtain an initial data stream including a timestamp and a data format.

[0121] In energy storage systems, acquiring real-time data streams from input devices using multiple communication protocols is fundamental to building an efficient data processing architecture. For example, sensor devices in energy storage systems may transmit data via Modbus, CAN, or MQTT. For example, a temperature sensor sends a data packet containing the temperature value and device ID via Modbus, while a voltage sensor transmits battery cell voltage data via CAN.

[0122] Preferably, the system deploys a unified protocol parser that supports dynamic identification and decoding of multiple protocols. Specifically, after receiving a Modbus data packet, the parser extracts the temperature value of 25 degrees Celsius and the timestamp; for a CAN data packet, it parses the voltage value of 3.8 volts.

[0123] The parser uses predefined protocol templates to decode data packets of different formats into initial data streams containing timestamps and unified data formats. This approach ensures data compatibility among multi-protocol devices.

[0124] Preferably, the queue uses a priority sorting algorithm to process packets with earlier timestamps first. This mechanism avoids packet disorder and ensures time consistency in subsequent processing.

[0125] Step S220: determine whether the timestamps of the initial data stream are aligned. If the timestamps of the initial data stream are not aligned, sort the data packets of the initial data stream through the message queue mechanism to generate a time-synchronized data stream.

[0126] It should be noted that the timestamps of the initial data stream may be misaligned due to device clock skew. In one embodiment, the system uses a message queue mechanism to sort data packets. For example, if the temperature sensor data packet has a timestamp of 10:00:01 and the voltage sensor data packet has a timestamp of 10:00:02, the message queue will rearrange the data packets according to the timestamp order, generating a time-synchronized data stream.

[0127] Step S230: Apply format standardization processing to the time-synchronized data stream according to a preset conversion rule, re-encapsulate the time-synchronized data stream in JSON format, and obtain a standardized data stream.

[0128] For time-synchronized data streams, format standardization is a key step. Understandably, energy storage systems require a unified format to facilitate data integration. For example, the system maps temperature, voltage, and other data to JSON format based on pre-set conversion rules. For example, raw temperature data of 25 degrees Celsius is encapsulated as {"type":"temperature","value":25,"unit":"Celsius"}.

[0129] In one possible implementation, the JSON format clearly expresses data types and units through a key-value pair structure, making it easier to transmit and parse across systems. This standardized processing simplifies the complexity of data interaction.

[0130] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S300:

[0131] Step S310: Obtain standardized data streams from multiple data sources, classify the standardized data streams according to a preset sensor identification table, and obtain classified data streams.

[0132] In energy storage systems, acquiring and classifying standardized data streams from multiple sources is key to efficient data processing. For example, sensors in an energy storage system may include temperature sensors, current sensors, and humidity sensors, all of which transmit data streams in a unified format.

[0133] The system categorizes data streams by sensor type based on a pre-set sensor identification table. For example, temperature sensor data is categorized as a "temperature data stream," containing the device ID and temperature value; current sensor data is categorized as a "current data stream." In one possible implementation, the identification table uses a key-value pair structure to store sensor ID and type mappings, such as {"ID001":"temperature"}, facilitating rapid classification.

[0134] Step S320: For the classified data stream, obtain historical data records and verify the integrity of the historical data through a cyclic redundancy check algorithm. If the integrity passes, calculate the reliability score based on the sensor type and timestamp to obtain the reliability weight.

[0135] For classified data streams, verifying the integrity of historical data is a crucial step in ensuring data reliability. Specifically, the system uses a cyclic redundancy check (CRC) algorithm to check historical data records. For example, a temperature data stream record contains a timestamp and a numerical value. The CRC algorithm generates a checksum. If it matches the stored checksum, the data is confirmed to be untampered with.

[0136] Preferably, the verification process is automatically triggered when the data is read, ensuring that the data for subsequent analysis is reliable. When calculating the reliability score, the system assesses data reliability based on sensor type and timestamp. In one embodiment, a temperature sensor is assigned a higher initial weight, such as 0.9, due to its long-term stable operation; a newly installed humidity sensor is assigned a weight of 0.7. Data with a more recent timestamp is given a slightly higher weight due to its higher real-time nature. For example, temperature data at 10:00:01 has a weight 0.05 higher than data at 10:00:00.

[0137] It should be noted that the weight calculation takes into account the sensor calibration frequency and operating environment to ensure that the score reflects the actual reliability.

[0138] Step S330: Use a weighted average algorithm to process the classified data stream, perform weighted calculation according to the reliability weight and data priority, encapsulate the weighted calculation result in JSON format, and obtain a fusion value data stream.

[0139] For weighted average calculations, the system processes categorized data streams based on reliability weights and data priority. For example, a temperature data stream contains two sets of data: 25 degrees Celsius and 26 degrees Celsius, with weights of 0.9 and 0.8, respectively, and equal priority. After weighted calculations, a fused value, such as 25.53 degrees Celsius, is generated and encapsulated in JSON format: {"type":"temperature","value":25.53,"unit":"Celsius"}. In one possible implementation, the JSON structure facilitates cross-system transmission and simplifies data interaction.

[0140] Step S340: Time synchronization processing is performed on the fused value data stream according to the preset fusion frequency, and the data packets of the fused value data stream are sorted using a message queue mechanism to generate a fused value output stream with a timestamp and a check tag.

[0141] Time synchronization processing ensures the time consistency of the fused value data stream.

[0142] It is understandable that the energy storage system needs to output data at a fixed frequency. For example, the system adjusts the data packet timestamp at a fusion frequency of once per second.

[0143] The message queue mechanism sorts packets by timestamp, such that the temperature fused value at 10:00:01 precedes the current fused value at 10:00:02.

[0144] Preferably, the queue adopts a priority sorting algorithm to give priority to processing data packets with high real-time requirements.

[0145] Finally, the fused value output stream is appended with a checksum tag, such as {"timestamp":"10:00:01","value":25.53","checksum":"ABC123"}, to facilitate verification by downstream systems.

[0146] Through the above method, the system realizes the complete process of data classification, verification, fusion and synchronization.

[0147] Each step supports each other, ensuring the efficiency and reliability of data flow and providing a solid foundation for the stable operation of the energy storage system.

[0148] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S400:

[0149] Step S410: Obtain a standardized data stream, and use a cyclic redundancy check algorithm to verify the integrity of the standardized data stream. If the integrity passes, calculate the weight according to the sensor type and timestamp to obtain a reliability weight value.

[0150] In energy storage systems, obtaining standardized data streams is the basis for efficient fire risk monitoring.

[0151] Exemplarily, the sensors in the energy storage system include a temperature sensor, a smoke sensor, and a voltage sensor, and the data stream is transmitted in a unified format, including a sensor ID, a timestamp, and a value.

[0152] Specifically, the data stream format generated by the temperature sensor is {“ID”: “T001”, “timestamp”: “2025-05-02 10:00:01”, “value”: 25, “unit”: “Celsius”}, which is convenient for the system to identify and process.

[0153] Using a cyclic redundancy check algorithm to verify the integrity of the data stream is the key to ensuring data reliability. In one embodiment, the system generates a check code for each data record, for example, the check code for the temperature data stream is "XYZ789".

[0154] When reading data, the system compares the checksum and confirms that the data has not been tampered with if they match. Preferably, the checksum process is automatically triggered to reduce manual intervention and improve efficiency.

[0155] Calculating weights based on sensor type and timestamp is the core of subsequent processing. It's important to note that the temperature sensor's initial weight is set to 0.9 due to its long-term stable operation; the smoke sensor's weight is 0.8 due to its high environmental sensitivity. Data with more recent timestamps is given a slightly higher weight. For example, temperature data at 10:00:02 has a 0.05 higher weight than data at 10:00:01. This design ensures that more timely data receives a greater weight in the fusion process.

[0156] Step S420: Based on the reliability weight value and data priority, a weighted average algorithm is used to process the standardized data stream, and the calculation result is encapsulated in JSON format to obtain a fusion value data stream.

[0157] In the weighted averaging algorithm, the system generates a fused value based on reliability weights and data priority. For example, a temperature data stream contains two sets of data: 25 degrees Celsius and 26 degrees Celsius, with weights of 0.9 and 0.8, respectively, and the same priority. The fused value is calculated through weighted calculation to generate 25.55 degrees Celsius, which is encapsulated in JSON format: {"type":"temperature","value":25.55","unit":"Celsius"}. The JSON structure facilitates cross-system transmission and simplifies data interaction.

[0158] Step S430: Determine whether the fusion value in the fusion value data stream exceeds a preset fire risk threshold. If the fusion value exceeds the risk threshold, input the fusion value data stream into a convolutional neural network for classification to obtain a classification result.

[0159] Determining whether the fused value exceeds the fire risk threshold is key to risk monitoring. In one possible implementation, the temperature fusion threshold is set at 50 degrees Celsius, and the smoke concentration threshold is set at 200 ppm. If the temperature fusion value reaches 55 degrees Celsius, exceeding the threshold, the subsequent classification process is triggered. This mechanism can quickly identify potential risks.

[0160] Feeding fused data streams exceeding a threshold into a convolutional neural network for classification demonstrates refined risk assessment. Specifically, the network outputs classification results, such as "high risk" or "low risk," based on multi-dimensional features like temperature and smoke. For example, a combination of 55 degrees Celsius and a smoke concentration of 250 ppm is classified as "high risk." The network, trained on historical data, effectively distinguishes risk levels.

[0161] Step S440: sort the classification results using a message queue mechanism, perform time synchronization processing using timestamps and verification tags, and generate a classification output stream.

[0162] A message queue mechanism is used to sort the classification results to ensure the efficiency of the output stream. Preferably, the queue is sorted by timestamp, for example, the classification result at 10:00:02 takes precedence over the result at 10:00:03. Results with high real-time requirements are processed first to improve response speed. Time synchronization processing is achieved through timestamps and checksum tags. For example, the format of the classification output stream is {"timestamp":"2025-05-0210:00:02","result":"high_risk","checksum":"DEF456"}, and the downstream system can verify the data integrity through the checksum tag.

[0163] It can be understood that synchronous processing ensures the consistency of multi-source data on the timeline, providing a reliable basis for fire risk monitoring.

[0164] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S500:

[0165] Step S510: Obtain the classification output stream, process the classification output stream using a time series analysis algorithm, and generate a fire risk level sequence.

[0166] In the field of fire risk monitoring for energy storage systems, processing the classified output stream is crucial for ensuring the accuracy of early warnings. For example, the classified output stream contains high-risk or low-risk labels for multiple time points, such as {"timestamp":"2025-05-02 10:00:02","result":"high_risk","checksum":"DEF456"}.

[0167] These data are processed using time series analysis algorithms, aiming to identify trends and cyclical changes in risk levels.

[0168] Specifically, the system uses a sliding window mechanism to analyze classification results over consecutive time periods, for example, taking a set of data every 5 seconds, to generate a risk level sequence. This sequence reflects the changing trend of risk over time, facilitating subsequent early warning and assessment.

[0169] Step S520: According to the fire risk level sequence, determine whether the level value exceeds the preset warning threshold value. If it exceeds, generate a warning signal data stream.

[0170] It should be noted that determining whether the risk level sequence exceeds the warning threshold is a key step. In one embodiment, the warning threshold is set to three consecutive high-risk labels. If it appears in the sequence, it triggers the generation of the warning signal data stream. Exemplarily, the format of the warning signal data stream is {"timestamp":"2025-05-0210:00:05","alert":"fire_risk","priority":1}, where the priority field facilitates subsequent sorting. This design ensures that high-risk conditions are captured in a timely manner.

[0171] Step S530: Prioritize the warning signal data stream through the message queue mechanism to obtain a sorted signal stream.

[0172] Prioritizing the warning signal data stream through a message queue mechanism can effectively improve processing efficiency. In one possible implementation, the queue is sorted based on the priority field and timestamp. For example, signals with priority 1 take precedence over signals with priority 2, and signals with newer timestamps are sorted first. Preferably, a signal with a time stamp of 10:00:05 takes precedence over a signal with a time stamp of 10:00:06. This sorting mechanism ensures that urgent signals are quickly delivered to downstream systems.

[0173] Step S540: Use a data verification algorithm to verify the integrity of the sorted signal stream and generate a signal stream that passes the verification.

[0174] Using a data verification algorithm to verify the integrity of the sorted signal stream is an important measure to ensure data reliability. For example, the system generates a SHA-256 checksum for each signal, such as "ABC123".

[0175] When reading the signal, the system compares the checksum to confirm that the data has not been tampered with. The signal stream format of the signal that passes the checksum remains unchanged, but a checksum tag is added, such as {"timestamp":"2025-05-02 10:00:05","alert":"fire_risk","checksum":"ABC123"}.

[0176] Step S550: Based on the verified signal stream, a time synchronization mechanism is used to align the timestamps of the verified signal stream to obtain a synchronized signal stream.

[0177] The time synchronization mechanism aligns timestamps on validated signal streams, ensuring consistency across the timeline of multi-source signals. For example, the system calibrates timestamps with millisecond accuracy, for example, aligning signals at 10:00:05.123 and 10:00:05.125 to 10:00:05.100. This alignment facilitates subsequent classification processing.

[0178] Step S560: Classify the synchronization signal stream using a preset signal mapping table to determine a warning signal.

[0179] By classifying the synchronous signal stream through a preset signal mapping table, the type of early warning signal can be refined.

[0180] It is understandable that the mapping table defines the correspondence between signals and warning types, for example, "fire_risk" is mapped to "emergency warning" and "smoke_risk" is mapped to "secondary warning".

[0181] Specifically, the synchronized signal stream {"timestamp":"2025-05-02 10:00:05","alert":"fire_risk"} is classified as an emergency alert and output to downstream systems. This classification mechanism enables rapid response to different levels of fire risk.

[0182] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S600:

[0183] Step S610: Acquire the warning signal data stream, use the message queue mechanism to prioritize the data stream, and obtain a sorted signal stream.

[0184] In the field of fire risk monitoring of energy storage systems, the processing of early warning signal data streams is a key link in ensuring the efficiency of fire control.

[0185] The following analysis and examples focus on technical topics such as obtaining warning signal data streams, message queue prioritization, signal verification, time synchronization, and modular communication interface transmission, focusing on a single scenario. For example, the acquisition of warning signal data streams is based on real-time collection of environmental data by a sensor network. For example, a temperature sensor in an energy storage system detects abnormally high temperatures and generates a data stream {"timestamp":"2025-05-0210:00:10","alert":"overheat_risk","priority":2}. Specifically, the system aggregates multi-source data through distributed acquisition nodes to ensure that the signal stream covers multiple risk types, such as high temperature and smoke. This approach supports comprehensive monitoring and improves signal diversity.

[0186] In one possible implementation, a message queue mechanism prioritizes the warning signal data stream. Preferably, the queue is dynamically adjusted based on the priority field and timestamp. For example, signal A is {"timestamp":"2025-05-0210:00:10","alert":"overheat_risk","priority":2}, and signal B is {"timestamp":"2025-05-0210:00:11","alert":"fire_risk","priority":1}. After sorting, signal B with priority 1 is ranked before signal A, and the signal with the newer timestamp takes precedence. This mechanism ensures that emergency signals are processed quickly and improves response efficiency.

[0187] Step S620: perform integrity verification on the sorted signal stream using a preset signal verification algorithm to obtain a signal stream that passes the verification.

[0188] It should be noted that the signal verification algorithm verifies the integrity of the sorted signal stream. In one embodiment, the system uses the MD5 algorithm to generate a checksum for each signal. For example, the signal {"timestamp":"2025-05-0210:00:11","alert":"fire_risk"} generates a checksum "XYZ789". When reading, the system compares the checksum to confirm that the data has not been tampered with. Signals that pass the verification are appended with a verification tag, such as {"timestamp":"2025-05-0210:00:11","alert":"fire_risk","checksum":"XYZ789"}. This verification method ensures data reliability.

[0189] Step S630: Use a time synchronization mechanism to align timestamps on the signal stream that has passed verification to obtain a synchronized signal stream.

[0190] As you can understand, the time synchronization mechanism aligns the timestamps of verified signal streams. For example, if signal C is 10:00:11.150 and signal D is 10:00:11.170, the system calibrates the timestamps in 100-millisecond increments to 10:00:11.100. This alignment ensures consistency across multiple signal sources on the timeline, supporting accurate subsequent analysis.

[0191] Step S640: Transmit the synchronization signal stream to the fire control module through the modular communication interface to generate a control instruction.

[0192] Specifically, the modular communication interface transmits the synchronization signal stream to the fire control module. In one embodiment, the interface uses the MQTT protocol to ensure low-latency transmission. For example, the synchronization signal {"timestamp":"2025-05-02 10:00:11","alert":"fire_risk"} is transmitted to the control module via the interface, generating the command "Start the sprinkler system." This modular design supports system expansion and enhances the flexibility of command generation.

[0193] For example, these mechanisms work together to ensure an efficient and reliable process from signal acquisition to command generation. Prioritization accelerates emergency signal processing, verification and synchronization ensure data accuracy, and modular interfaces enhance system compatibility. These links support each other, forming a rigorous fire risk response chain.

[0194] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S700:

[0195] Step S710: Receive a registration request message carrying a unique identifier of a newly added energy storage unit. The registration request message is sent when the newly added energy storage unit accesses the dynamic expansion interface. Use a preset unit registration protocol to perform unit identification and obtain a list of registered energy storage units.

[0196] In the field of fire risk monitoring for energy storage systems, dynamic registration of energy storage units and processing of their status data are core components of fire control. For example, upon receiving a registration request message carrying the unique identifier of a newly added energy storage unit, the system captures the request through a dynamic extension interface. For example, energy storage unit E issues a registration request carrying the identifier "UNIT_E_001."

[0197] Using a pre-defined unit registration protocol, such as one based on JSON format, the request content is parsed, the identifier is extracted, and its uniqueness is verified. Once verified, unit E is added to the list of registered energy storage units, such as "UNIT_E_001, UNIT_F_002, UNIT_G_003." This registration mechanism supports dynamic system expansion and facilitates the rapid integration of new units.

[0198] Step S720: According to the list of registered energy storage units, the status monitoring data of each unit is obtained, and the status monitoring data is prioritized using a message queue mechanism to obtain a sorted status data stream.

[0199] Specifically, based on the list of registered energy storage units, the system obtains status monitoring data for each unit. In one embodiment, this status data is collected via sensors and covers temperature, voltage, and smoke concentration. For example, the status data for unit E is "temperature: 45°C, voltage: 3.7V, smoke: 0ppm," while that for unit F is "temperature: 50°C, voltage: 3.6V, smoke: 10ppm." This data is collected through periodic polling or event-triggered methods to ensure real-time performance.

[0200] After collection, data enters a message queue mechanism for prioritization. Preferably, the queue assigns priority based on the severity of the anomaly in the data. For example, data with a temperature exceeding 48°C or a smoke concentration exceeding 5 ppm is marked as high priority. Cell F's data, due to temperature and smoke anomalies, is assigned priority 1, while cell E is assigned priority 2. After sorting, the status data stream prioritizes priority 1 data, ensuring that abnormal conditions are handled first.

[0201] Step S730: For the sorted state data stream, if the state data meets the preset firefighting task threshold, a corresponding control instruction is generated through the dynamic extension interface to obtain an instruction set to be allocated.

[0202] It should be noted that the system determines whether the sorted status data stream meets the preset firefighting task threshold. In one possible implementation, the firefighting task threshold is set to "temperature > 48°C or smoke > 5 ppm." When the data from unit F triggers the threshold, the system generates a control instruction through the dynamic expansion interface, such as "Start the cooling system in zone 1." These instructions form a command set to be assigned, for example, "Command 1: Cooling, Command 2: Alarm."

[0203] Step S740: The instruction set to be assigned is transmitted to the target fire protection module through the module communication protocol, and the instruction verification algorithm is used to perform integrity verification on the instruction set to be assigned to determine the assigned control instructions.

[0204] After the instruction set is generated, it is transmitted to the target fire protection module via the module communication protocol. For example, the CoAP protocol is used for lightweight transmission, ensuring efficient communication in a low-power environment. After the instruction set is transmitted to the fire protection module, an instruction verification algorithm is used for integrity verification. For example, the system generates a SHA-256 checksum for each instruction. For example, the instruction "Start cooling system" corresponds to the checksum "ABC123."

[0205] After receiving the command, the fire protection module compares the checksum to confirm that the data has not been tampered with. The command that passes the check is marked as the assigned control command, such as "Command 1: Cooling, checksum: ABC123".

[0206] It's easy to understand that the above process forms a complete fire control chain through dynamic registration, status monitoring, prioritization, threshold determination, and command generation and verification. Dynamic expansion interfaces support system flexibility, prioritization optimizes emergency response, and verification mechanisms ensure command reliability. These links support each other and collectively improve the efficiency of the energy storage system's firefighting missions.

[0207] Furthermore, the multi-area safety monitoring method for an energy storage power station proposed in this embodiment includes step S800:

[0208] Step S810: Obtain deployment requirement data of the target module from a preset module configuration database, and classify the deployment requirement data using a rule engine to obtain a classified requirement data set.

[0209] In the field of fire risk monitoring for energy storage systems, a preset module configuration database stores deployment requirement data for target modules, covering hardware specifications, communication protocols, and operating environments. For example, the database records the deployment requirements for fire protection module M1 as "support for the CoAP protocol, operating temperature range of -10°C to 60°C, and response time <100ms." This data is categorized using a rules engine to generate a requirements dataset. For example, the rules engine categorizes modules into "CoAP support" and "MQTT support" based on protocol type, with M1 classified as CoAP support and M2 as MQTT support. This categorized dataset facilitates subsequent instruction matching.

[0210] Step S820: Receive the control instruction set to be judged through the communication protocol. For the classified demand data set, if the parameters of the control instruction set meet the preset threshold parameters, use the logic matching algorithm to generate an adaptation identifier to obtain an adapted instruction set.

[0211] Specifically, the system receives a set of control instructions to be judged via a communication protocol. For example, the instruction set includes "start the M1 cooling system" and "trigger the M2 alarm." These instructions are transmitted in JSON format via the CoAP protocol, with content such as "{action: cooling, module: M1}."

[0212] For the classified demand data set, the system checks whether the instruction parameters meet the threshold. For example, the threshold requirement for M1 is "the cooling instruction must be triggered when the temperature is greater than 45°C."

[0213] Assuming that the real-time temperature of M1 is 50°C and the instruction parameters meet the threshold, the system uses a logical matching algorithm to generate an adaptation identifier and marks the instruction as "Adapt M1".

[0214] Step S830: Use the message queue mechanism to prioritize the adapted instruction set, adjust the sorting order according to the real-time status monitoring data of the target module, and obtain a sorted instruction sequence.

[0215] In one possible implementation, the adapted instruction set enters a message queue for priority sorting. The message queue adjusts the order based on module status monitoring data. For example, M1's temperature is 50°C and the smoke concentration is 8 ppm; M2's temperature is 40°C and the smoke concentration is 2 ppm.

[0216] The priority rule is set as "temperature > 48°C or smoke > 5ppm is high priority", so the command priority of M1 is 1 and M2 is 2. After sorting, the command sequence is "cool M1, alarm M2", ensuring that urgent tasks are handled first.

[0217] Step S840: The sorted instruction sequence is transmitted through the instruction execution interface of the target module, and the sorted instruction sequence is verified using an integrity check algorithm to determine whether an execution result is generated.

[0218] Preferably, the sorted instruction sequence is transmitted through the instruction execution interface of M1 and M2. Before transmission, the system verifies the instruction using an integrity check algorithm. For example, the instruction "cool M1" generates the MD5 checksum "XYZ789".

[0219] After receiving the instruction, the target module compares the verification code to confirm that the instruction has not been tampered with. If the verification passes, M1 performs the cooling operation and M2 triggers the alarm.

[0220] It should be noted that if the verification fails, the system records the error log and resends the command.

[0221] It can be understood that the above process ensures efficient execution of instructions by classifying requirements, instruction adaptation, priority sorting and verification.

[0222] Classification improves matching accuracy, priority optimizes response speed, verification ensures reliability, and all links work together to improve system stability.

[0223] The present invention relates to a multi-area safety monitoring system for an energy storage power station, which is used to implement the above-mentioned multi-area safety monitoring method for an energy storage power station. The multi-area safety monitoring system for an energy storage power station includes a first acquisition module, a first generation module, a calculation module, a first judgment module, a second generation module, a second acquisition module, a distribution module, and a second judgment module. The first acquisition module is used to obtain real-time data streams of temperature, smoke, and voltage from multiple sensors of the energy storage system; the first generation module is used to convert real-time data streams of different communication protocols into a unified format through a protocol conversion module to generate a standardized data stream; the calculation module is used to process the standardized data stream using a data fusion algorithm, and the data is processed according to the type and history of the sensor. The data reliability calculates the weighted average fusion value; the first judgment module is used to judge whether the weighted average fusion value exceeds the preset fire risk threshold. If the weighted average fusion value exceeds the preset fire risk threshold, the weighted average fusion value is input into the convolutional neural network for classification processing; the second generation module is used to determine the fire risk level according to the classification result and generate an early warning signal; the second acquisition module is used to transmit the early warning signal to the fire control module through the modular communication interface to obtain the control instruction; the distribution module is used to use the dynamic expansion interface to distribute the control instruction to the fire module of the newly added energy storage unit; the second judgment module is used to judge whether the control instruction is adapted to the modular deployment requirements and output the execution result.

[0224] Compared with the existing technology, the multi-area safety monitoring method and system for energy storage power stations provided in this embodiment realize real-time monitoring of parameters such as temperature, smoke and voltage through multi-sensor data acquisition, protocol conversion and data fusion. A weighted average algorithm is used to process the standardized data stream, and a convolutional neural network is combined to classify the fire risk level and generate an early warning signal. This embodiment also designs a modular communication interface and a dynamic expansion interface, which can flexibly allocate control instructions to the fire protection module of the newly added energy storage unit to meet the needs of modular deployment. The multi-area safety monitoring method and system for energy storage power stations provided in this embodiment not only improve the accuracy and real-time performance of fire risk assessment, but also enhance the scalability and adaptability of the system, providing effective protection for the safe operation of the energy storage system.

[0225] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the invention. Thus, the present invention is intended to include such changes and modifications as fall within the scope of the claims and their equivalents.

Claims

1. A multi-area safety monitoring method for an energy storage power station, characterized in that: The following steps are involved: Obtain real-time data streams on temperature, smoke, and voltage from multiple sensors in the energy storage system; The protocol conversion module converts the real-time data streams of different communication protocols into a unified format to generate standardized data streams; Processing the standardized data stream using a data fusion algorithm, and calculating a weighted average fusion value based on the type of sensor and the reliability of historical data; Determining whether the weighted average fusion value exceeds a preset fire risk threshold, and if the weighted average fusion value exceeds the preset fire risk threshold, inputting the weighted average fusion value into a convolutional neural network for classification processing; Determine the fire risk level based on the classification results and generate early warning signals; Transmitting the warning signal to the fire control module through a modular communication interface to obtain control instructions; The dynamic expansion interface is used to distribute the control instructions to the fire protection module of the newly added energy storage unit; Determine whether the control instruction is adapted to the modular deployment requirement and output the execution result.

2. The multi-area safety monitoring method for an energy storage power station according to claim 1, characterized in that: The step of obtaining real-time data streams of temperature, smoke, and voltage from multiple sensors of the energy storage system includes: Obtain raw data streams of temperature, smoke, and voltage from multiple sensors in the energy storage system; The data in the original data stream is denoised using a preset signal processing algorithm to obtain real-time data streams of temperature, smoke and voltage.

3. The multi-area safety monitoring method for an energy storage power station according to claim 1, wherein: The step of converting the real-time data streams of different communication protocols into a unified format by the protocol conversion module to generate a standardized data stream includes: Acquire real-time data streams from input devices of multiple communication protocols, and use a preset protocol parser to decode data packets of the real-time data streams to obtain an initial data stream including a timestamp and a data format; It is determined whether the timestamps of the initial data stream are aligned. If the timestamps of the initial data stream are not aligned, the data packets of the initial data stream are sorted by a message queue mechanism to generate a time-synchronized data stream. The time-synchronized data stream is subjected to format standardization processing according to preset conversion rules, and the time-synchronized data stream is repackaged in JSON format to obtain a standardized data stream.

4. The multi-area safety monitoring method for an energy storage power station according to claim 1, wherein: The step of using a data fusion algorithm to process the standardized data stream and calculating a weighted average fusion value according to the type of sensor and the reliability of historical data includes: Acquire standardized data streams from multiple data sources, and classify the standardized data streams according to a preset sensor identification table to obtain classified data streams; For the classified data stream, historical data records are obtained and the integrity of the historical data is verified using a cyclic redundancy check algorithm. If the integrity passes, a reliability score is calculated based on the sensor type and timestamp to obtain a reliability weight. A weighted average algorithm is used to process the classified data stream, a weighted calculation is performed according to the reliability weight and data priority, and the weighted calculation result is encapsulated in JSON format to obtain a fusion value data stream; The fused value data stream is subjected to time synchronization processing according to a preset fusion frequency, and the data packets of the fused value data stream are sorted using a message queue mechanism to generate a fused value output stream with a timestamp and a check tag.

5. The multi-area safety monitoring method for an energy storage power station according to claim 1, wherein: The step of determining whether the weighted average fusion value exceeds a preset fire risk threshold, and if the weighted average fusion value exceeds the preset fire risk threshold, inputting the weighted average fusion value into a convolutional neural network for classification processing includes: Obtaining a standardized data stream, verifying the integrity of the standardized data stream using a cyclic redundancy check algorithm, and if the integrity passes, calculating a weight based on the sensor type and timestamp to obtain a reliability weight value; According to the reliability weight value and data priority, a weighted average algorithm is used to process the standardized data stream, and the calculation result is encapsulated in JSON format to obtain a fusion value data stream; Determining whether a fusion value in the fusion value data stream exceeds a preset fire risk threshold, and if the fusion value exceeds the risk threshold, inputting the fusion value data stream into a convolutional neural network for classification to obtain a classification result; The classification results are sorted using a message queue mechanism, and time synchronization is performed using the timestamp and verification tag to generate a classification output stream.

6. The multi-area safety monitoring method for an energy storage power station according to claim 1, characterized in that: The steps of determining the fire risk level according to the classification results and generating an early warning signal include: Obtaining a classification output stream, processing the classification output stream using a time series analysis algorithm to generate a fire risk level sequence; According to the fire risk level sequence, determining whether the level value exceeds a preset warning threshold, and if so, generating a warning signal data stream; Prioritizing the warning signal data stream through a message queue mechanism to obtain a sorted signal stream; Use data verification algorithm to verify the integrity of the sorted signal stream and generate a signal stream that passes the verification; According to the signal stream that has passed the verification, the time synchronization mechanism is used to align the timestamps of the signal stream that has passed the verification to obtain a synchronized signal stream; The synchronization signal stream is classified by a preset signal mapping table to determine the early warning signal.

7. The multi-area safety monitoring method for an energy storage power station according to claim 1, wherein: The step of transmitting the warning signal to the fire control module through the modular communication interface to obtain a control instruction includes: Obtaining a warning signal data stream, and prioritizing the data stream using a message queue mechanism to obtain a sorted signal stream; The integrity of the sorted signal stream is verified by a preset signal verification algorithm to obtain a signal stream that passes the verification; Use the time synchronization mechanism to align the timestamps of the signal stream that has passed the verification to obtain a synchronized signal stream; The synchronization signal stream is transmitted to the fire control module through the modular communication interface to generate a control instruction.

8. The multi-area safety monitoring method for an energy storage power station according to claim 1, wherein: The steps of using the dynamic expansion interface to distribute the control instruction to the fire protection module of the newly added energy storage unit include: Receive a registration request message carrying a unique identifier of a newly added energy storage unit, the registration request message being sent when the newly added energy storage unit is connected to the dynamic expansion interface, perform unit identification using a preset unit registration protocol, and obtain a list of registered energy storage units; According to the list of registered energy storage units, the status monitoring data of each unit is obtained, and the status monitoring data is prioritized using a message queue mechanism to obtain a sorted status data stream; For the sorted state data stream, if the state data meets the preset firefighting task threshold, the corresponding control instructions are generated through the dynamic extension interface to obtain the instruction set to be assigned; The instruction set to be assigned is transmitted to the target fire protection module through the module communication protocol, and the instruction verification algorithm is used to perform integrity verification on the instruction set to be assigned to determine the assigned control instructions.

9. The multi-area safety monitoring method for an energy storage power station according to claim 1, wherein: The steps of determining whether the control instruction is adapted to the modular deployment requirement and outputting the execution result include: Obtaining deployment requirement data of a target module from a preset module configuration database, and classifying the deployment requirement data using a rule engine to obtain a classified requirement data set; The control instruction set to be judged is received through the communication protocol. For the classified demand data set, if the parameters of the control instruction set meet the preset threshold parameters, a logic matching algorithm is used to generate an adaptation identifier to obtain an adapted instruction set; The message queue mechanism is used to prioritize the adapted instruction set, and the sorting order is adjusted according to the real-time status monitoring data of the target module to obtain the sorted instruction sequence; The sorted instruction sequence is transmitted through the instruction execution interface of the target module, and the sorted instruction sequence is verified using an integrity check algorithm to determine whether the execution result is generated.

10. A multi-area safety monitoring system for an energy storage power station, used to implement the multi-area safety monitoring method for an energy storage power station according to any one of claims 1 to 9, characterized in that: The multi-area safety monitoring system of the energy storage power station includes: A first acquisition module is used to obtain real-time data streams of temperature, smoke and voltage from multiple sensors of the energy storage system; A first generating module is used to convert the real-time data streams of different communication protocols into a unified format through a protocol conversion module to generate a standardized data stream; a calculation module, configured to process the standardized data stream using a data fusion algorithm and calculate a weighted average fusion value based on the type of sensor and the reliability of historical data; a first judgment module, configured to judge whether the weighted average fusion value exceeds a preset fire risk threshold, and if the weighted average fusion value exceeds the preset fire risk threshold, inputting the weighted average fusion value into a convolutional neural network for classification processing; The second generation module is used to determine the fire risk level according to the classification results and generate an early warning signal; A second acquisition module is used to transmit the warning signal to the fire control module through the modular communication interface to obtain a control instruction; A distribution module, configured to distribute the control instructions to the fire protection module of the newly added energy storage unit using the dynamic expansion interface; The second judgment module is used to judge whether the control instruction is adapted to the modular deployment requirement and output an execution result.

Images