Distributed energy storage station monitoring system

Abstract

A distributed energy storage station monitoring system relates to the technical field of energy storage station control and comprises an equipment layer and a plurality of energy storage units, and each unit is provided with an intelligent terminal and an edge computing unit. The communication layer is used for hybrid networking and supports three-channel redundancy switching and quantum key distribution; in the station control layer, an Acrel-1000DP monitoring host is deployed, and an IEC-61850 protocol conversion module and a Hyperledger Fabric block chain evidence storage system are integrated; the optimization engine is used for generating a dynamic charging and discharging strategy based on mixed integer quadratic programming and reinforcement learning; the optimization engine is based on the mixed integer quadratic programming and reinforcement learning algorithm, the optimal charging and discharging strategy is dynamically generated by integrating multiple factors, the charging and discharging power is flexibly adjusted, the purposes of peak load shifting and the like are achieved, and the operation efficiency and economic benefits of the energy storage system are improved; the station control layer monitoring host has a comprehensive monitoring alarm function, masters the operation state of the energy storage station in real time, deals with abnormity in time, and guarantees safe operation.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage station control, and in particular to a distributed energy storage station monitoring system. Background Art

[0002] With the accelerating global energy transition, the penetration rate of distributed energy resources (DER) such as solar energy and wind energy, which are renewable energy sources, in the power system is continuously increasing. The intermittency and uncertainty of these distributed energy sources pose huge challenges to the stable operation of the traditional power grid. As an effective solution, the distributed energy storage system (DES) has emerged. The distributed energy storage station can realize the flexible storage and release of energy, effectively suppress the fluctuations of distributed energy, improve the power quality, and enhance the reliability and flexibility of the power grid.

[0003] In the distributed energy storage system, the monitoring system plays a crucial role. It is like the "intelligent brain" of the energy storage station, which real-time controls the operating status of each component in the energy storage station, coordinates and controls the charging and discharging processes of the energy storage units, and ensures the safe, stable and efficient operation of the energy storage system. Some insurmountable defects have gradually emerged in the functions and performances of traditional energy storage monitoring systems. For example, the data acquisition accuracy is insufficient and cannot meet the requirements for accurately evaluating the battery status; the communication reliability is poor and is easily interfered, often resulting in data transmission interruption or errors; the optimization strategy is too simple and it is difficult to balance multiple objectives such as grid cost and battery life.

[0004] In recent years, the rapid development of cutting-edge technologies such as edge computing, quantum communication, and reinforcement learning has brought new opportunities for the upgrading of distributed energy storage station monitoring systems. Edge computing effectively reduces the dependence on cloud computing resources by processing data near the data source, significantly reduces the data transmission delay, and improves the real-time response ability of the system. Quantum communication, with its unconditional security feature, provides an indestructible security barrier for the transmission of energy storage station monitoring data, eliminating the risk of data being stolen or tampered with. As an advanced machine learning method, reinforcement learning enables the monitoring system to autonomously learn and optimize decision-making strategies in complex operating environments, realizing the refined management and control of the energy storage system. How to organically integrate these advantages is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In order to overcome the deficiencies in the background art, the present invention discloses a distributed energy storage station monitoring system.

[0006] To achieve the above invention objective, the present invention adopts the following technical solutions:

[0007] A distributed energy storage station monitoring system, comprising:

[0008] Device layer, multiple energy storage units, each equipped with an intelligent terminal and an edge computing unit;

[0009] Communication layer, hybrid networking, supporting three-channel redundant switching and quantum key distribution;

[0010] Station control layer, deploying an Acrel-1000DP monitoring host, integrating an IEC-61850 protocol conversion module and a Hyperledger Fabric blockchain evidence storage system;

[0011] Optimization engine, generating a dynamic charge and discharge strategy based on mixed integer quadratic programming and reinforcement learning, with the decision variable being the charge and discharge power P of each energy storage unit i (i = 1, …, N), and the objective function is:

[0012]

[0013]

[0014] Among them, is the grid interaction cost, including time-of-use electricity price and demand charge; is the battery health degradation rate, calculated based on electrochemical impedance spectroscopy (EIS); is the total communication and computing delay; is the charge and discharge power tracking error; is the grid interaction cost weight coefficient, with a value range of [0.1, 1.0], used to balance the priority between grid cost and battery life degradation; is the battery health degradation rate weight coefficient, with a value range of [0.01, 0.2], reflecting the impact of battery life degradation on the total cost; is the communication and computing delay weight coefficient, with a value range of [0.001, 0.01], in seconds, used to quantify the impact of delay on system real-time performance; is the power deviation penalty coefficient, with a value range of [0.0001, 0.001], in kW, used to constrain the deviation between the charge and discharge power and the planned value.

[0015] Preferably, the intelligent terminal is built-in with an electrochemical impedance spectroscopy EIS chip, with a sampling frequency of 1 kHz and an SOC estimation error of ±1%; the edge computing unit is an NVIDIA Jetson AGX Xavier, supporting local model training and real-time data preprocessing; wavelet threshold denoising and Kalman filtering are used for data preprocessing, and the signal-to-noise ratio is increased by 20 dB after denoising.

[0016] Preferably, the hybrid networking includes a LoRaWAN channel, a 5G URLLC slice, and a fiber optic ring network, with a time slot allocation accuracy of ±0.5 μs and a primary / backup channel switching delay of <5 ms; the quantum communication uses the decoy state BB84 protocol through QKD equipment, the light source is a single photon emitter, the key generation rate is 2.1 Mbps, and the bit error rate is <10 -12 .

[0017] Preferably, the dynamic charge and discharge strategy adopts a two-charge and two-discharge mode, charging from 0:00 to 6:00, discharging from 9:00 to 11:00, charging from 14:00 to 16:00, and discharging from 18:00 to 20:00.

[0018] Due to the adoption of the above-mentioned technical solutions, the present invention has the following beneficial effects:

[0019] A distributed energy storage station monitoring system disclosed by the present invention, the intelligent terminal at the device layer is built-in with an EIS chip, which collects battery data at a high sampling frequency of 1 kHz, and preprocesses it by combining wavelet threshold denoising and Kalman filtering technologies to ensure high-precision and high-reliability of the data. The SOC estimation error is controlled within the range of ±1%, providing a data basis for the precise control and management of the energy storage system. The hybrid networking at the communication layer supports three-channel redundant switching, and combines quantum key distribution technology to ensure efficient and secure data transmission. According to the data transmission requirements in different network environments, the communication channels are flexibly selected and switched to ensure the stable operation of the system. The optimization engine is based on the mixed integer quadratic programming and reinforcement learning algorithms, dynamically generates the optimal charge and discharge strategy by comprehensively considering multiple factors, flexibly adjusts the charge and discharge power, realizes goals such as peak shaving and valley filling, and improves the operation efficiency and economic benefits of the energy storage system. The monitoring host at the station control layer has a comprehensive monitoring and alarm function, can real-time master the operation status of the energy storage station, process anomalies in a timely manner, and ensure safe operation. The intuitive monitoring interface helps the operation personnel understand the operation parameters and equipment status, and improves the management efficiency and accuracy. Detailed implementation manners

[0020] The present invention can be explained in detail through the following embodiments. The purpose of disclosing the present invention is to protect all technical improvements within the scope of the present invention. In the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "front", "rear", "left", "right", etc. indicating the orientation or position relationship, it is for the convenience of describing the present invention, rather than indicating or implying that the device or element referred to must have a specific orientation.

[0021] Embodiment 1, a distributed energy storage station monitoring system, includes:

[0022] The device layer, which serves as the foundation of the entire system, consists of multiple energy storage units. Each energy storage unit is equipped with an intelligent terminal and an edge computing unit. The intelligent terminal is built-in with an Electrochemical Impedance Spectroscopy (EIS) chip, which can accurately estimate the battery state at a high sampling frequency of 1 kHz. The SOC estimation error is strictly controlled within the range of ±1%, providing accurate data support for battery charge and discharge management. The edge computing unit adopts the NVIDIA Jetson AGX Xavier platform, supporting local model training and real-time data preprocessing. It uses wavelet threshold denoising and Kalman filtering technologies to effectively remove data noise interference. After denoising, the signal-to-noise ratio is significantly increased by 20 dB, ensuring high-quality and highly reliable data, and laying a solid foundation for the intelligent decision-making of the subsequent monitoring system.

[0023] The communication layer is responsible for data transmission. It adopts a hybrid networking method including LoRaWAN channels, 5G URLLC slices, and fiber optic ring networks, supporting three-channel redundant switching. The time slot allocation accuracy reaches ±0.5 μs, and the switching delay between the primary and backup channels is less than 5 ms, ensuring efficient data transmission and high reliability. At the same time, the quantum key distribution technology is introduced. Encrypted communication is carried out through the QKD device using the decoy state BB84 protocol. The light source is a single-photon emitter, and the key generation rate is 2.1 Mbps, with a bit error rate lower than 10 -12 , providing a high level of security for data transmission, ensuring that data is accurately transmitted between the device layer and the station control layer, and at the same time accurately transmitting the control instructions issued by the station control layer and the optimization engine to the device layer, realizing remote and precise control of the energy storage units.

[0024] The station control layer deploys the Acrel-1000DP monitoring host, integrating the IEC-61850 protocol conversion module and the Hyperledger Fabric blockchain evidence storage system to achieve centralized monitoring and management of the energy storage station. It receives the data transmitted by the communication layer, comprehensively monitors the operation status of the energy storage station, and can grasp in real time information such as device status, fault alarms, and operation parameters. The station control layer integrates and analyzes the data, provides decision-making support for the optimization engine, and also converts the dynamic charge and discharge strategies generated by the optimization engine into specific control instructions, which are sent to the energy storage units of the device layer through the communication layer to ensure the effective execution of the charge and discharge strategies.

[0025] The optimization engine, as the core of intelligent decision-making, is based on the mixed-integer quadratic programming and reinforcement learning algorithms. Taking the charge and discharge power P i (i = 1, …, N) of each energy storage unit as decision variables, a target function is constructed, aiming to minimize the weighted sum of the grid interaction cost, the battery health degradation rate, the total communication and computing delay, and the charge and discharge power tracking error. The specific target function is:

[0026]

[0027]

[0028] Among them, is the grid interaction cost, including time-of-use electricity price and demand charge; is the battery health degradation rate, calculated based on Electrochemical Impedance Spectroscopy (EIS); is the total communication and computing delay; is the charge and discharge power tracking error; is the grid interaction cost weight coefficient, with a value range of [0.1, 1.0], used to balance the priority between grid cost and battery life degradation; is the battery health degradation rate weight coefficient, with a value range of [0.01, 0.2], reflecting the impact of battery life degradation on the total cost; is the communication and computing delay weight coefficient, with a value range of [0.001, 0.01], in seconds, used to quantify the impact of delay on system real-time performance; is the power deviation penalty coefficient, with a value range of [0.0001, 0.001], in kW, used to constrain the deviation between the charge and discharge power and the planned value. Specifically, the dynamic charge and discharge strategy adopts a two-charge and two-discharge mode, charging from 0:00 to 6:00, discharging from 9:00 to 11:00, charging from 14:00 to 16:00, and discharging from 18:00 to 20:00. This strategy is transmitted to the device layer through the station control layer and the communication layer to guide the charge and discharge behavior of the energy storage unit, achieving peak shaving and valley filling, reducing the grid cost, while extending the battery life and ensuring system real-time performance.

[0029] The operation process is as follows:

[0030] System initialization:

[0031] When the system starts, it is initialized. The intelligent terminal and edge computing unit at the device layer perform self-checks and initialize communication parameters, and establish a connection with the communication layer; the communication layer detects and configures the communication channel to ensure stable network connectivity; the station control layer Acrel-1000DP monitoring host starts the monitoring program, initializes the database and parameters, and prepares to receive and process data; the optimization engine loads the preset model algorithm, initializes parameters such as weight coefficients, and prepares for strategy generation.

[0032] Data acquisition and preprocessing:

[0033] The intelligent terminal at the device layer collects battery data of the energy storage unit in real time at a sampling frequency of 1 kHz, including information such as voltage, current, and temperature. The collected data is preprocessed in the edge computing unit, and the wavelet threshold denoising technology is used to remove high-frequency noise, and then the Kalman filtering algorithm is used for further smoothing processing to improve the accuracy and reliability of the data. The preprocessed data is timestamped, encapsulated in a certain data format, and transmitted to the station control layer through the communication layer.

[0034] Data Transmission and Reception:

[0035] The communication layer automatically selects the optimal communication channel for data transmission according to the current network conditions and communication quality, and the quantum key distribution technology ensures the security of data transmission. The monitoring host in the station control layer receives data through the communication interface, unpacks and analyzes it, and stores it in the local database to provide data support for subsequent monitoring optimization.

[0036] Real-time Monitoring and Alarms:

[0037] The monitoring host in the station control layer analyzes and processes data in real time, and displays the operation status of the energy storage station through the monitoring interface, including information such as the SOC, SOH, voltage, current, and temperature of each energy storage unit. At the same time, the monitoring host detects and judges abnormal situations according to the preset alarm rules, such as overcharging, over-discharging, too high temperature, or communication interruption of the battery, immediately triggers the alarm mechanism, emits audible and visual alarm signals, and notifies relevant personnel through text messages and emails to ensure the safe operation of the energy storage station.

[0038] Strategy Optimization and Instruction Issuance:

[0039] The optimization engine formulates the optimal charge and discharge strategy according to the real-time and historical data provided by the station control layer, combines the dynamic information of the power market and the user's needs, and uses the mixed integer quadratic programming and reinforcement learning algorithms. With the goal of minimizing the grid interaction cost, etc., it adjusts the charge and discharge power of each energy storage unit and formulates the charge and discharge plan. The generated strategy is transmitted to the energy storage unit in the device layer through the station control layer and the communication layer to guide its charge and discharge behavior and achieve optimization goals such as peak shaving and valley filling.

[0040] Execution and Feedback:

[0041] The energy storage unit in the device layer adjusts the charge and discharge power according to the received charge and discharge instructions and performs the corresponding operations. During the execution, the intelligent terminal collects battery data in real time and feeds back the execution results and the current status to the edge computing unit. The edge computing unit processes and analyzes the feedback data, judges the execution situation and effect of the charge and discharge operations, and reports to the station control layer and the optimization engine in time when abnormal, so as to adjust and optimize the strategy to ensure the stable operation of the system.

[0042] Data Storage and Analysis:

[0043] The monitoring host in the station control layer stores the collected data and strategy information and establishes a historical database. Through data analysis and mining, it understands the operation rules of the energy storage station and the change trend of battery performance, providing a basis for system maintenance and optimization. Using the analysis results, it improves the algorithm model of the optimization engine, improves the accuracy and effectiveness of strategy generation, and enhances the intelligent level of the monitoring system performance.

[0044] The parts not detailed in the present invention are prior arts. For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, and all changes falling within the meaning and scope of the equivalent elements are intended to be embraced by the present invention.

Claims

1. A distributed energy storage station monitoring system, characterized in that it includes: Equipment layer: multiple energy storage units, each equipped with a smart terminal and edge computing unit; Communication layer, hybrid networking, supports three-channel redundant switching and quantum key distribution; At the station control layer, the Acrel-1000DP monitoring host is deployed, integrating the IEC-61850 protocol conversion module and the Hyperledger Fabric blockchain evidence storage system; The optimization engine generates dynamic charging and discharging strategies based on mixed integer quadratic programming and reinforcement learning. The decision variables are the charging and discharging power P of each energy storage unit. i (i=1,…,N), the objective function is: in, The grid interaction cost includes time-of-use electricity price and demand electricity charge; is the battery health decay rate, calculated based on electrochemical impedance spectroscopy (EIS); is the total delay of communication and computation; is the charge and discharge power tracking error; is the grid interaction cost weight coefficient, with a value range of [0.1, 1.0], which is used to balance the priority of grid cost and battery life attenuation; is the weight coefficient of battery health decay rate, ranging from [0.01, 0.2], reflecting the impact of battery life decay on total cost; is the communication and computing delay weight coefficient, with a value range of [0.001, 0.01], in seconds, used to quantify the impact of delay on the real-time performance of the system; It is the power deviation penalty coefficient, with a value range of [0.0001, 0.001], in kW, and is used to constrain the deviation of the charging and discharging power from the planned value.

2. The distributed energy storage station monitoring system according to claim 1, characterized in that: The smart terminal has a built-in electrochemical impedance spectroscopy EIS chip with a sampling frequency of 1kHz and a SOC estimation error of ±1%. The edge computing unit is NVIDIA Jetson AGXXavier, which supports local model training and real-time data preprocessing. The data preprocessing uses wavelet threshold denoising and Kalman filtering, and the signal-to-noise ratio is improved by 20dB after denoising.

3. The distributed energy storage station monitoring system according to claim 1, characterized in that: The hybrid network includes LoRaWAN channels, 5G URLLC slices and fiber ring networks, with a time slot allocation accuracy of ±0.5μs and a main-standby channel switching delay of <5ms; the quantum communication adopts the decoyed state BB84 protocol through the QKD device, the light source is a single photon transmitter, the key generation rate is 2.1Mbps, and the bit error rate is <10 -12 .

4. The distributed energy storage station monitoring system according to claim 1, characterized in that: The dynamic charge and discharge strategy adopts a two-charge and two-discharge mode, charging from 0:00 to 6:00, discharging from 9:00 to 11:00, charging from 14:00 to 16:00, and discharging from 18:00 to 20:00.

Images