Non-step-in liquid cooling phosphoric acid and all-vanadium liquid flow energy storage power station cooperative control system in butt joint with 220KV transformer substation

Abstract

The invention relates to the technical field of electric power engineering, in particular to a non-step-in liquid cooling phosphoric acid and all-vanadium liquid flow energy storage power station cooperative control system in butt joint with a 220KV transformer substation. Comprising a hybrid energy storage module, a non-stepping intelligent liquid cooling module, a multi-strategy cooperative control module and a substation deep cooperative interface module. The multi-strategy cooperative control module is used for constructing a multi-stage cooperative control architecture, generating control strategies adapted to different operation scenes by fusing multi-source data, and realizing multi-module cooperative operation optimization; lithium iron phosphate and all-vanadium liquid flow energy storage units are connected in parallel through a direct-current bus, and a short-time high-power response-long-time deep energy storage characteristic complementary system is constructed; through cooperative operation of the hybrid energy storage module, limitation of a single battery type on power or capacity is avoided; a'monitoring-evaluation-response 'totally-closed active safety chain is formed, and the temperature can be kept stable and potential safety hazards can be eliminated without manual intervention in the whole process.

Description

Technical Field

[0001] This invention relates to the field of power engineering technology, and more specifically, to a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow energy storage power station collaborative control system for 220KV substation connection. Background Technology

[0002] As the power grid's demands for both power response speed and energy storage duration increase, single-type energy storage systems (such as pure lithium iron phosphate or vanadium redox flow storage) struggle to simultaneously achieve the complementary characteristics of "short-term high-power response" and "long-term deep energy storage." Traditional energy storage systems employ coarse-grained temperature control strategies for thermal management, failing to meet the differentiated requirements of high-power heat dissipation in lithium iron phosphate and long-term stable operation in vanadium redox flow storage, and lacking a fully enclosed, unmanned safety protection mechanism. Furthermore, the coordinated control of energy storage systems and substations generally suffers from non-standardized data interaction and delayed command response, resulting in a lack of real-time synchronization of grid status information and hindering the efficiency of coordinated operation between energy storage systems and the grid.

[0003] For example, Chinese patent CN202311526974.9 discloses a method and system for analyzing the configuration of energy storage connected to a substation. This includes obtaining the power / load time-series simulation coefficients for each side of each main transformer in the substation throughout the year based on historical data; performing time-series simulation of the planned power / load of the substation based on these coefficients to obtain the power / load time-series simulation power for each side of each main transformer; considering the full-power charging and discharging operation scenario of the energy storage station and the substation's operating limits, calculating the energy storage power and capacity that can be connected to each side of each main transformer based on the power and load time-series simulation power; and combining the power and capacity of the energy storage station, determining the voltage level, energy storage power, and capacity allocation for the energy storage station's connection to the substation based on the energy storage power and capacity that can be connected to each side of each main transformer and the substation's connection conditions. This application ensures that energy storage power stations can be connected to appropriate voltage levels with suitable power and capacity. Chinese patent CN202011383161.5 discloses an energy storage power station control system, which includes a production monitoring module, a production control module, a scheduling module, and a cloud platform. The production monitoring module is used to monitor each energy storage station and transmit monitoring data. The production control module is used to receive and process the monitoring data from each production monitoring module and to receive control commands from the scheduling module and the cloud platform. The scheduling module is used to schedule and control the operation of each energy storage power station. The production monitoring module is communicatively connected to the production control module, and the production control module is communicatively connected to both the scheduling module and the cloud platform. It has the ability to interact with the cloud platform, and can use the cloud platform for data storage and analysis, thereby improving the system's operating efficiency, increasing the scheduling response speed of each energy storage power station, achieving real-time monitoring, and reducing costs.

[0004] While the aforementioned existing technologies each have their own advantages, they also suffer from the following technical shortcomings: First, the energy storage architecture is too simplistic: existing technologies all employ a single type of energy storage unit (such as pure lithium iron phosphate or an undefined hybrid energy storage architecture), failing to construct a complementary system of "short-term high-power response - long-term deep energy storage." For example, Chinese patent CN202311526974.9 only addresses power and capacity calculations during the energy storage access planning phase, without addressing the coordinated control of different energy storage units during the operation phase; the control system of Chinese patent CN202011383161.5 is only adapted to a single energy storage site, lacking a complementary mechanism between the high and low characteristics of lithium iron phosphate and vanadium redox flow batteries, resulting in the system being unable to simultaneously meet the grid's dual requirements for power response speed and energy storage duration; Second, the thermal management is not sufficiently refined: existing solutions do not design differentiated temperature control strategies for the thermal characteristics of different batteries. For example, Chinese patent CN202011383161.5 does not mention a fully enclosed liquid cooling architecture or dynamic temperature control zoning. Its "production monitoring module" only realizes routine temperature monitoring and lacks differentiated thermal isolation and flow regulation capabilities for the high power heat dissipation requirements of lithium iron phosphate and the long-term stable operation of all vanadium liquid flow. It is difficult to avoid safety risks (such as thermal runaway) caused by uneven temperature field, and the response mechanism that relies on manual intervention is not timely enough. Thirdly, the depth of grid collaboration is lacking: the interaction between existing technologies and substations is limited to the power calculation level of data monitoring or planning. The "access conditions" of Chinese patent CN202311526974.9 do not involve real-time data synchronization and rapid command response; although the "scheduling module" of Chinese patent CN202011383161.5 has communication capabilities, it lacks standardized interfaces and intelligent algorithm support, and cannot realize real-time analysis of grid transient information (such as frequency changes and voltage drops) and rapid power support of energy storage systems, resulting in collaborative control lagging behind the dynamic needs of the grid. In order to address the above deficiencies, we propose a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow energy storage power station collaborative control system for 220kV substation connection. Summary of the Invention

[0005] The purpose of this invention is to provide a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow energy storage power station collaborative control system for 220KV substation connection, so as to solve the problems of the above-mentioned background technology, such as the single energy storage architecture, insufficient thermal management and lack of grid collaboration depth.

[0006] To solve the above-mentioned technical problems, the present invention aims to provide a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station collaborative control system for 220kV substation connection, comprising: The hybrid energy storage module connects lithium iron phosphate energy storage units and vanadium redox flow batteries in parallel via a DC bus to construct a complementary energy storage system with high and low characteristics of lithium iron phosphate batteries and vanadium redox flow batteries, supporting the coordinated operation of short-term high-power response and long-term deep energy storage. The non-walk-in intelligent liquid cooling module is used to implement fully enclosed differentiated thermal management and active safety protection for hybrid energy storage modules. It monitors the battery pack temperature and abnormal operating conditions in real time through multi-source sensors, and achieves stable temperature maintenance and safety response without human intervention. The multi-strategy collaborative control module is used to build a multi-level collaborative control architecture. It generates control strategies that adapt to different operating scenarios by integrating multi-source data, thereby achieving multi-module collaborative operation optimization. The substation deep collaboration interface module is used to build a standardized data interaction and command response channel with 220kV substations, supporting seamless integration between energy storage systems and substation automation systems, and enabling real-time synchronization of grid status information and rapid response to control commands.

[0007] As a further improvement to this technical solution, the lithium iron phosphate energy storage unit includes a battery module subunit and a battery management subunit, wherein: The battery module subunit consists of N parallel battery modules, and each battery module contains M lithium iron phosphate battery cells connected in series. The battery management subunit uses the Kalman filter algorithm to estimate the state of charge (SOC).

[0008] As a further improvement to this technical solution, the all-vanadium redox flow storage unit includes an electrolyte circulation subunit and a stack management subunit, wherein: The electrolyte circulation subunit includes an electrolyte storage tank, a circulation pump, and a heat exchanger; The stack management subunit is used to collect stack voltage and current in real time and communicate with the multi-strategy collaborative control module through the ModbusTCP protocol; when the voltage difference between adjacent single cells exceeds a preset threshold, the electrolyte distribution ratio is dynamically adjusted based on the fuzzy control algorithm.

[0009] As a further improvement to this technical solution, the non-walk-in intelligent liquid cooling module includes a temperature control chamber unit and a liquid cooling circulation unit, wherein: The temperature-controlled chamber unit uses a metal frame to construct an independent temperature-controlled space. Through thermal insulation partitions, it forms a dynamic temperature-controlled chamber adapted to the high-power heat dissipation of lithium iron phosphate and a steady-state temperature-controlled chamber adapted to the long-term stability of vanadium liquid flow. The two zones are thermally isolated by a three-layer composite wall (inner layer of thermally conductive silicone sheet, middle layer of nano-aerogel felt, and outer layer of aluminum alloy plate). The chamber interface is sealed with double O-ring seals (protection level IP68), supporting wide temperature range operation (-40℃-60℃), providing a fully enclosed temperature-controlled environment without human intervention for the hybrid energy storage module. The liquid cooling circulation unit includes a dual-loop control subunit and a flow regulation subunit, wherein: The dual-loop control subunit includes a lithium iron phosphate unit loop and a vanadium redox flow unit loop. The lithium iron phosphate unit loop uses a centrifugal liquid-cooled pump, a plate heat exchanger, and a solenoid valve, and uses ethylene glycol aqueous solution as coolant to quickly remove instantaneous heat through high-frequency flow regulation. The vanadium redox flow unit loop uses a screw circulating pump, an electric heating rod, and a solenoid valve, and uses deionized water as coolant to maintain electrolyte stability through low-fluctuation temperature control. The flow regulation subunit dynamically adjusts the opening of the solenoid valve based on temperature data from multiple sources and a fuzzy PID algorithm, thereby achieving reasonable control of the temperature difference between the lithium iron phosphate battery and the vanadium redox flow battery.

[0010] As a further improvement to this technical solution, the non-walk-in intelligent liquid cooling module also includes a safety protection unit, which comprises a multi-dimensional monitoring subunit, a rapid response subunit, and an intelligent decision-making subunit, wherein: The multi-dimensional monitoring subunit monitors the temperature field in real time through distributed optical fiber sensors laid in a serpentine pattern along the battery module, and scans the battery surface with an infrared thermal imager to identify abnormal temperature gradients. The multi-dimensional monitoring subunit also monitors the gas concentration inside the battery box (H2 detection limit ≤ 5ppm, CO detection limit ≤ 10ppm) through an integrated gas sensor array, and monitors the battery leakage current through a micro-current sensor to provide early warning of insulation faults when leakage occurs. The fast response subunit uses a pipeline pressure sensor to monitor the pressure in real time. When the detected value exceeds the threshold, the coolant circuit is simultaneously cut off based on a redundancy control algorithm. The intelligent decision-making subunit constructs a Bayesian network risk assessment model based on multi-source sensor data. When the predicted risk of thermal runaway reaches a preset threshold, it is automatically triggered. The fire extinguishing system is isolated from the battery cluster and sends an emergency power adjustment command to the 220KV substation through the substation deep collaboration interface module.

[0011] As a further improvement to this technical solution, the multi-level collaborative control architecture in the multi-strategy collaborative control module includes a grid response unit, an energy storage coordination unit, and a device execution unit, wherein: The grid response unit, based on an improved model predictive control algorithm, predicts the peak-valley electricity price curve and load fluctuation trend of the grid within the next 48 hours through a rolling optimization window, and generates a charging and discharging strategy for the energy storage system. The objective function of the improved model predictive control algorithm is: ; in, Indicates control input Optimize the energy storage power control signal to minimize the objective function. To achieve the optimal solution for energy storage charging and discharging strategies; This indicates a time-domain summation prediction. This indicates control over time-domain summation; Indicates control variables; Relative to the current time The number of future prediction steps; To predict the length of the time domain, To control the length of the time domain; Represented as the predicted time The load power; For the predicted time The charging and discharging power of the energy storage system; For the predicted time The planned power output of the power grid; This represents the change in energy storage power between adjacent time points, and ,in Indicates the current moment Predicting the future Energy storage capacity at any given time Indicates the current moment Predicting the future Energy storage capacity at any given time; ; , The weighting coefficient is dynamically adjusted using an adaptive fuzzy inference algorithm, and its calculation formula is as follows: ; in, For the first The error of each control target; The learning rate; Adjust the step size for weights; It is Each weight coefficient at the current time The value; It is The control objective at the current moment The error; The energy storage coordination unit is based on the multi-objective particle swarm optimization algorithm (MOPSO) to construct a three-dimensional objective space that includes grid dispatch tracking, cycle life loss of lithium iron phosphate energy storage units, and entropy change of electrolyte in vanadium redox flow energy storage units. It solves the Pareto optimal solution set to minimize the comprehensive operating cost of the energy storage system (including battery life decay cost and electrolyte maintenance cost) under the constraint of satisfying grid power command. The device execution unit uses a distributed consensus algorithm to allocate power to the hybrid energy storage module, dynamically adjusts the output weight of each energy storage unit, and achieves synergy between the power allocation strategy and the multi-objective optimization results of the energy storage coordination unit to achieve a consistent optimization goal and improve system control accuracy.

[0012] As a further improvement to this technical solution, the multi-strategy collaborative control module also includes an adaptive learning unit. The adaptive learning unit constructs a value function of the energy storage system's operating state through a deep Q-network and dynamically adjusts the control parameters based on historical operating data and real-time monitoring data. The adaptive learning unit includes a transfer learning subunit, which generates a new control strategy through knowledge transfer when a change in the power grid operation mode is detected.

[0013] As a further improvement to this technical solution, the power grid response unit also includes an abnormal event detection subunit, which identifies abnormal events in real time based on the isolated forest algorithm; When an abnormal event is detected, the abnormal event detection subunit achieves rapid power support for the energy storage system through an improved droop control algorithm. The power allocation formula of the improved droop control algorithm is as follows: ; in, The real-time output power (kW) of the energy storage system, positive for discharging and negative for charging. This is a reference value (kW) for energy storage power when the power grid is operating normally. The droop control coefficient is dynamically adjusted to... ,in These are the initial coefficients. For adaptive gain (0.1-0.3), = The power grid frequency deviation (Hz); The real-time frequency of the power grid (Hz); The rated frequency of the power grid (e.g., 50Hz); The droop control coefficient Adaptive update via gradient descent: ; in Let the learning rate be (0.01-0.05), so that... Dynamically optimizes based on frequency deviation to improve frequency adjustment accuracy and response speed; It is the partial derivative sign, which supports the adaptive parameter optimization of the improved droop control algorithm, enabling the energy storage system to quickly and accurately adjust the power when the grid is abnormal, and meet the high requirements of 220kV substation for frequency stability (such as millisecond-level response and high-precision adjustment). This indicates the next moment.

[0014] As a further improvement to this technical solution, the substation deep collaboration interface module includes a power grid status synchronization unit, a control command parsing unit, and a security authentication unit, wherein: The power grid status synchronization unit is used to acquire the power grid status information of the substation in real time, realize the synchronization of power grid transient information, and provide real-time power grid operation data support for the collaborative control of the energy storage system. It collects the power grid status data (such as voltage, current, frequency, etc.) of the substation in real time, and after preprocessing (filtering, noise suppression, etc.), it synchronizes it to the energy storage system with millisecond-level accuracy, providing real-time power grid operation parameters for the multi-strategy collaborative control module, supporting the optimization of its dynamic control strategy, and ensuring that the energy storage system can make collaborative control decisions based on accurate power grid status information.

[0015] The control command parsing unit constructs a command mapping model based on a deep neural network, learns the relationship between historical commands and energy storage responses, parses dispatch commands, and converts them into power control signals, improving the accuracy of complex command parsing and ensuring rapid response. This model can parse various dispatch commands issued by the substation (such as AGC / AVC commands, emergency power adjustment commands, etc.) and convert them into power control signals executable by the energy storage system. In this way, the accuracy of parsing complex control commands is improved, ensuring that the energy storage system can quickly and accurately understand and execute the substation's dispatch commands, thereby further enhancing the overall system's collaborative control capabilities and achieving deep collaborative cooperation between the energy storage system and the substation.

[0016] The security authentication unit is used to establish a two-way authentication mechanism between the energy storage system and the substation, providing security protection for data transmission, ensuring the security of data interaction, and preventing unauthorized command intrusion and data leakage. This mechanism ensures the security of data interaction, effectively prevents unauthorized command intrusion and data leakage, ensures the reliability and security of the system during operation, and provides a secure and reliable communication environment for deep collaboration between the energy storage system and the substation.

[0017] As a further improvement to this technical solution, the power grid state synchronization unit includes a power grid state estimation subunit. The power grid state estimation subunit is used to process the acquired power grid state data and output key parameters (such as the equivalent impedance of the power grid and the phase angle of the node voltage) through the Kalman filtering algorithm, so as to provide input for the optimization of the control strategy of the energy storage system.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a complementary system of "short-term high-power response and long-term deep energy storage" by connecting lithium iron phosphate and vanadium redox flow batteries in parallel on a DC bus. The lithium iron phosphate battery provides rapid response to sudden changes in grid load (such as second-level power regulation) with its high power density, while the vanadium redox flow battery achieves hour-level stable storage of renewable energy thanks to its long cycle life. The two operate collaboratively through a hybrid energy storage module, avoiding the limitations of a single battery type in terms of power or capacity. This design enhances the comprehensive service capabilities of the energy storage system for grid "peak-valley frequency regulation" and "energy time shifting". 2. The non-walk-in intelligent liquid cooling module in this invention adopts an independent temperature control chamber design. A thermally insulated partition divides the module into a dynamic temperature control chamber adapted for high-power heat dissipation of lithium iron phosphate batteries and a steady-state temperature control chamber adapted for long-term stability of vanadium redox flow batteries. Combined with a dual-loop liquid cooling cycle (high-frequency adjustment for rapid heat dissipation in the lithium iron phosphate unit loop and low-fluctuation temperature control in the vanadium redox flow unit loop to maintain electrolyte stability), precise adaptation of the thermal characteristics of lithium iron phosphate and vanadium redox flow batteries is achieved. Simultaneously, a distributed fiber optic sensor, infrared thermal imager, and gas sensor array form a multi-dimensional monitoring network, scanning the temperature field, gas concentration, and leakage current in real time. This, combined with a Bayesian network risk assessment model, automatically warns of thermal runaway risks and triggers [further details needed]. The fire extinguishing system and the battery cluster isolation device form a fully enclosed active safety chain of "monitoring-assessment-response", which can maintain temperature stability and eliminate safety hazards without human intervention throughout the process; 3. This invention employs a multi-strategy collaborative control module to construct a three-tiered architecture of "grid response—energy storage coordination—equipment execution": The grid response unit uses an improved model predictive control algorithm to continuously optimize the charging and discharging strategy for the next 48 hours, generating optimal power commands by combining peak-valley electricity prices and load fluctuation trends; the energy storage coordination unit utilizes a multi-objective particle swarm optimization algorithm to balance grid dispatch tracking, battery life loss, and electrolyte stability, forming a Pareto optimal solution set; simultaneously, the equipment execution unit dynamically adjusts the output weights of each energy storage unit through a distributed consensus algorithm, ensuring that power allocation and optimization objectives converge collaboratively. Meanwhile, the substation deep collaborative interface module uses a standardized communication architecture to synchronize grid transient information (such as frequency mutations and voltage drops) in real time and quickly parse dispatch commands. It leverages a deep neural network model to improve command parsing accuracy and employs a security authentication mechanism to ensure data interaction security. Ultimately, this achieves seamless integration between the energy storage system and the 220kV substation in terms of status monitoring, power regulation, and security protection, thereby enhancing grid stability and renewable energy consumption efficiency. Attached Figure Description

[0019] Figure 1 This is a system framework diagram of the present invention; The meanings of the labels in the diagram are as follows: 100. Hybrid energy storage module; 110. Lithium iron phosphate energy storage unit; 111. Battery module subunit; 112. Battery management subunit; 120. Vanadium redox flow battery energy storage unit; 121. Electrolyte circulation subunit; 122. Stack management subunit; 200. Non-walk-in intelligent liquid cooling module; 210. Temperature control chamber unit; 220. Liquid cooling circulation unit; 221. Dual-loop control subunit; 222. Flow regulation subunit; 230. Safety protection unit; 231. Multi-dimensional monitoring subunit; 232. Rapid response subunit; 233. Intelligent decision-making subunit; 300. Multi-strategy collaborative control module; 310. Power grid response unit; 311. Abnormal event detection subunit; 320. Energy storage coordination unit; 330. Equipment execution unit; 340. Adaptive learning unit; 341. Transfer learning subunit; 400. Substation Deep Collaboration Interface Module; 410. Power Grid Status Synchronization Unit; 411. Power Grid Status Estimation Subunit; 420. Control Command Parsing Unit; 430. Security Authentication Unit. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] like Figure 1 As shown, this embodiment provides a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station collaborative control system for 220KV substation connection, including: The hybrid energy storage module 100 connects the lithium iron phosphate energy storage unit 110 and the vanadium redox flow battery 120 in parallel via a DC bus to construct a complementary energy storage system with high and low characteristics of lithium iron phosphate batteries and vanadium redox flow batteries, supporting the coordinated operation of short-term high-power response and long-term deep energy storage; the hybrid energy storage module 100 adopts a redundant industrial Ethernet architecture to realize internal communication, and the communication protocol is ProfinetIORT.

[0022] In this embodiment, the lithium iron phosphate energy storage unit 110 includes a battery module subunit 111 and a battery management subunit 112, wherein: The battery module subunit 111 consists of N parallel battery modules, each of which contains M series-connected lithium iron phosphate battery cells. Battery management subunit 112 uses the Kalman filter algorithm to estimate SOC.

[0023] In this embodiment, the all-vanadium redox flow storage unit 120 includes an electrolyte circulation subunit 121 and a stack management subunit 122, wherein: Electrolyte circulation subunit 121 includes an electrolyte storage tank, a circulation pump, and a heat exchanger; The stack management subunit 122 is used to collect stack voltage and current in real time and communicate with the multi-strategy collaborative control module 300 through the ModbusTCP protocol; when the voltage difference between adjacent single cells exceeds the preset threshold, the electrolyte distribution ratio is dynamically adjusted based on the fuzzy control algorithm.

[0024] As a further explanation of this embodiment, the electrolyte circulation subunit 121 in this embodiment forms a closed loop consisting of an electrolyte storage tank, a circulation pump, and a heat exchanger. The electrolyte storage tank is connected to the inlet of the circulation pump via a pipeline to provide electrolyte reserves for the circulation system. The outlet of the circulation pump is connected to the inlet of the heat exchanger, driving the electrolyte to flow through the heat exchanger for temperature regulation. The outlet of the heat exchanger is branched through pipelines to various branches of the vanadium redox flow stack, and after reaction in the stack, it flows back to the electrolyte storage tank, forming a circulation path. The system controls the electrolyte flow rate through pipeline valves, and the components are connected by sealed flanges to ensure the airtightness of the loop.

[0025] Furthermore, in the liquid cooling circuit, an ethylene glycol-deionized water composite solution can be used as the coolant. Considering the application environment (taking North China as an example, with an extreme minimum temperature of -12℃) and the heating characteristics of the energy storage unit (lithium iron phosphate peak heating power of 15kW / m³, and vanadium redox flow steady-state heating power of 3kW / m³), the freezing point and specific heat capacity at different ethylene glycol mass fractions were tested using differential scanning calorimetry. Taking into account antifreeze redundancy (freezing point must be lower than the minimum ambient temperature) and heat dissipation requirements, an ethylene glycol mass fraction of 30% was selected. The deionized water underwent two-stage reverse osmosis treatment, controlling the conductivity to not exceed 5μS / cm, and sodium molybdate was added to adjust the pH to 6.8-7.2 to inhibit pipeline electrochemical corrosion.

[0026] As a further explanation of this embodiment, N and M are configurable design parameters that can be flexibly adjusted according to the capacity requirements, voltage level, and power characteristics of the energy storage system. M represents the number of lithium iron phosphate battery cells connected in series within a single battery module. Depending on the system voltage requirements, M typically ranges from 20 to 40 (e.g., when M=24, the module voltage is 76.8V; when M=30, the module voltage is 96V). N represents the number of modules connected in parallel within a battery module sub-unit. Depending on the system capacity requirements, N typically ranges from 10 to 100 (e.g., when N=48, a megawatt-level energy storage unit can be constructed).

[0027] As a further explanation of this embodiment, the lithium iron phosphate energy storage unit 110 and the vanadium redox flow energy storage unit 120 in this embodiment adopt droop control to achieve power distribution, and the distribution formula is as follows: ; in, Allocate power to each unit; Total power demand The charge state of each unit; The equivalent internal resistance of each unit is 0.003Ω for lithium iron phosphate energy storage units and 0.005Ω for all vanadium redox flow storage units. Indicates the unit number or index.

[0028] As a further explanation of this embodiment, the battery management subunit 112 and the stack management subunit 122 in this embodiment achieve data fusion and collaborative control through an edge computing gateway, and the edge computing gateway can be configured as follows: CPU: IntelXeonE3-1245v6, 3.7GHz; Memory: 16GB DDR4 ECC; Storage: 512GB SSD + 2TB HDD; Operating system: Linux Ubuntu 20.04 LTS; Software platform: Docker containerized deployment, supporting Python and C++ algorithm development; The gateway enables energy management and optimization control of the hybrid energy storage system and has the ability to operate autonomously locally. Therefore, it can maintain safe operation of the system for ≥72 hours in the event of communication interruption.

[0029] Furthermore, the fuzzy control algorithm configured in the stack management subunit 122 is specifically designed to address the potential issue of unbalanced single-cell voltage during the operation of the vanadium redox flow battery stack. When the voltage difference between adjacent single cells exceeds a preset threshold (e.g., ±0.05V), the stack management subunit 122 dynamically adjusts the electrolyte distribution ratio based on the fuzzy control algorithm. The algorithm takes the voltage difference between adjacent single cells and its rate of change as input, and outputs an electrolyte distribution ratio correction coefficient (range 0.8-1.2). This is achieved by adjusting the opening of the solenoid valve to dynamically optimize the flow rate of each branch, with a control cycle of 500ms. This strategy can stabilize the single-cell voltage difference, adapt to electrolyte temperatures ranging from 15℃ to 45℃ and concentration fluctuations, and improve the stack's operational balance and lifespan.

[0030] It should be added that the battery management subunit 112 uses the Kalman filter algorithm to achieve high-precision estimation of the battery's state of charge (SOC). This algorithm, based on the principles of probability theory and mathematical statistics, fuses the battery model's predicted values ​​with actual measured values ​​through recursive operations of the system state equation and observation equation, dynamically adjusting the weighting coefficients to achieve the optimal estimate of SOC. The basic principle of the Kalman filter algorithm is as follows: The Kalman filter algorithm uses a recursive approach to estimate the system state, and its core consists of two stages: Prediction phase: Based on the SOC estimate of the previous moment and the battery equivalent circuit model, combined with the current current input, predict the SOC state at the current moment.

[0031] Update phase: The predicted value is fused with the actual measured voltage, current and other data, and the weights of the two are dynamically adjusted by Kalman gain to obtain the optimal SOC estimate at the current moment.

[0032] The algorithm continuously iterates through the above process to gradually reduce the estimation error and achieve real-time and accurate tracking of the SOC.

[0033] In the lithium iron phosphate energy storage unit 110, high-precision SOC estimation provides crucial data support for system control: Power distribution optimization: Based on real-time SOC status, the collaborative control system can accurately allocate the charging and discharging power of the lithium iron phosphate energy storage unit 110 and the vanadium redox flow energy storage unit 120 to avoid overcharging and over-discharging.

[0034] Lifetime prediction and management: Combining SOH (State of Health) estimation can provide a basis for decision-making on battery pack balancing control and preventive maintenance.

[0035] Fault diagnosis: Through residual analysis of Kalman filtering (the deviation between predicted and measured values), battery anomalies can be detected in real time, providing early warning of potential safety risks.

[0036] The non-walk-in intelligent liquid cooling module 200 is used to implement fully enclosed differentiated thermal management and active safety protection for the hybrid energy storage module 100. It monitors the battery pack temperature and abnormal operating conditions in real time through multi-source sensors, achieving stable temperature maintenance and safe response without manual intervention. The non-walk-in intelligent liquid cooling module 200, through its fully enclosed differentiated thermal management and active safety protection design, enables manual temperature control and risk prevention for the hybrid energy storage module 100. The system uses a metal frame to construct an independent temperature control space, integrating multi-source sensors and intelligent control algorithms to support stable operation over a wide temperature range.

[0037] In this embodiment, the non-walk-in intelligent liquid cooling module 200 includes a temperature control chamber unit 210 and a liquid cooling circulation unit 220, wherein: The temperature control chamber unit 210 uses a metal frame to construct an independent temperature control space. It forms a dynamic temperature control chamber adapted to the high-power heat dissipation of lithium iron phosphate and a steady-state temperature control chamber adapted to the long-term stability of vanadium redox flow through heat insulation partitions. The two chambers are thermally isolated by a three-layer composite wall. The chamber interface is sealed with double O-ring seals, which supports wide temperature range operation and provides a fully enclosed temperature control environment for the hybrid energy storage module 100 without human intervention. The liquid cooling circulation unit 220 includes a dual-loop control subunit 221 and a flow regulation subunit 222, wherein: The dual-loop control subunit 221 includes a lithium iron phosphate unit loop and a vanadium redox flow unit loop. The lithium iron phosphate unit loop uses a centrifugal liquid-cooled pump, a plate heat exchanger, and a solenoid valve, and uses ethylene glycol aqueous solution as coolant to quickly remove instantaneous heat through high-frequency flow regulation. The vanadium redox flow unit loop uses a screw circulating pump, an electric heating rod, and a solenoid valve, and uses deionized water as coolant to maintain electrolyte stability through low-fluctuation temperature control. The flow regulation subunit 222 dynamically adjusts the opening of the solenoid valve based on temperature data from multiple sources and combined with a fuzzy PID algorithm to achieve reasonable control of the temperature difference between the lithium iron phosphate battery and the vanadium redox flow battery.

[0038] In this embodiment, the non-walk-in intelligent liquid cooling module 200 further includes a safety protection unit 230, which includes a multi-dimensional monitoring subunit 231, a rapid response subunit 232, and an intelligent decision-making subunit 233, wherein: The multi-dimensional monitoring subunit 231 monitors the temperature field in real time through distributed optical fiber sensors laid in a serpentine pattern along the battery module, and scans the battery surface with an infrared thermal imager to identify abnormal temperature gradients. The multi-dimensional monitoring subunit 231 also monitors the gas concentration inside the battery box through an integrated gas sensor array, and monitors the battery leakage current through a micro-current sensor to provide early warning of insulation faults when leakage occurs. The fast response subunit 232 uses a pipeline pressure sensor to monitor the pressure in real time. When the detected value exceeds the threshold, it synchronously cuts off the coolant circuit based on a redundant control algorithm. The intelligent decision-making subunit 233 constructs a Bayesian network risk assessment model based on multi-source sensor data. When the predicted risk of thermal runaway reaches a preset threshold, it automatically triggers... The fire suppression system is isolated from the battery cluster and sends an emergency power adjustment command to the 220KV substation through the substation deep coordination interface module 400.

[0039] As a further explanation of this embodiment, the lithium iron phosphate unit circuit is equipped with a centrifugal liquid-cooled pump, a plate heat exchanger, and a high-frequency response solenoid valve, using ethylene glycol aqueous solution as the coolant, and the flow rate can be dynamically adjusted. When entering high-power operation, the flow rate is increased through a fuzzy PID algorithm to quickly remove instantaneous heat and control the battery temperature difference within a reasonable range. The vanadium redox flow unit circuit uses a screw-type circulating pump, an electric heating rod, and a low-fluctuation solenoid valve, using deionized water as the coolant to maintain a stable electrolyte temperature and compensate for heat loss in low-temperature environments.

[0040] As a further explanation of this embodiment, the temperature control chamber is constructed with a high-strength metal frame to create an independent sealed space, and the interior is divided into a dynamic temperature control chamber and a steady-state temperature control chamber by a heat-insulating partition: The dynamic temperature control chamber is adapted to the high power heat dissipation requirements of the lithium iron phosphate energy storage unit 110, and adopts a liquid cooling pipeline with optimized flow channel design to quickly dissipate instantaneous high heat. The steady-state temperature control chamber is adapted to the long-term stable operation of the vanadium redox flow storage unit 120, and maintains the optimal reaction temperature of the electrolyte through a low-fluctuation temperature control strategy.

[0041] The two zones employ a three-layer composite wall to achieve efficient thermal insulation, and the chamber interfaces are sealed with double fluororubber O-rings to ensure an IP67 protection level, effectively blocking external environmental interference with temperature and humidity.

[0042] As a further explanation of this embodiment, the flow regulation subunit 222 employs a fuzzy PID algorithm to achieve dynamic control of the coolant flow. This algorithm is an intelligent control strategy that combines fuzzy control logic with traditional PID control. It adjusts the proportional (Kp), integral (Ki), and derivative (Kd) parameters of the PID controller in real time through a fuzzy inference mechanism to adapt to the nonlinear characteristics and time-varying operating conditions of the system. The core principles of the fuzzy PID algorithm include: Fuzzy logic decision-making: Using real-time temperature data of lithium iron phosphate and vanadium redox flow units collected by multi-source sensors as input, the temperature deviation (the difference between the actual value and the set value) and the deviation change rate are calculated, and the optimal PID parameter combination is automatically matched through a preset fuzzy rule base. PID dynamic adjustment: Based on real-time optimized Kp, Ki, and Kd parameters, output solenoid valve opening control signal to accurately adjust coolant flow and quickly suppress temperature fluctuations.

[0043] In this embodiment, considering the difference in instantaneous heat generation during high-power operation of lithium iron phosphate and the temperature stability during long-term operation of all-vanadium liquid flow, the algorithm uses a dynamic matching flow regulation strategy: When the lithium iron phosphate cell enters a high-power charge / discharge state (e.g., >1C), the flow rate of the corresponding circuit is automatically increased to quickly dissipate peak heat. When the vanadium liquid flow unit is in steady-state operation, maintain low-fluctuation flow control to avoid temperature overshoot or under-adjustment.

[0044] This strategy allows the temperature difference between lithium iron phosphate batteries and vanadium redox flow batteries to be controlled within the design range, ensuring temperature uniformity and system efficiency of the hybrid energy storage module under a wide range of operating conditions.

[0045] As a further explanation of this embodiment, the fast response subunit 232 adopts a redundancy control algorithm to achieve safety protection. This algorithm constructs a redundant architecture by configuring master and slave dual controllers (such as PLC or microprocessor) to ensure that the system can still operate reliably when a single controller fails.

[0046] The core principles of redundancy control algorithms include: Redundant configuration: The main controller collects pipeline pressure sensor data in real time and mirrors the main controller status synchronously from the controller. The two maintain data consistency through a high-speed communication link (such as ProfinetRT). Fault detection and switching: The system periodically detects the heartbeat signal of the main controller. When the main controller malfunctions (such as communication interruption or calculation timeout), the controller automatically and seamlessly takes over control to avoid single point of failure. Emergency response logic: Regardless of whether it is a master or slave controller, when the pressure exceeds the threshold (such as too high or too low), the same control command is triggered to synchronously cut off the solenoid valve of the coolant circuit, ensuring the consistency of the actions performed.

[0047] Its function in this embodiment is: To address pressure anomalies caused by potential pipe leaks or pump malfunctions in liquid cooling systems, a redundant control algorithm employs dual hardware and software logic redundancy to achieve the following: High reliability response: The control cycle from pressure anomaly detection to loop disconnection is ≤20ms, which far exceeds the response speed of a single controller system; Fault tolerance capability: Controller-level redundancy design can reduce the probability of single point of failure and meet the high reliability requirements of long-term unmanned operation of energy storage systems; Consistency assurance: By synchronizing control command output, execution chaos caused by controller differences is avoided, ensuring the accuracy of safety protection actions.

[0048] This technical solution effectively improves the system's fault tolerance in the high-pressure liquid cooling circuit, providing dual protection for the safe operation of the hybrid energy storage module 100.

[0049] As a further illustration of this embodiment, a Bayesian network risk assessment model is constructed based on multi-source sensor data. When the predicted risk of thermal runaway exceeds a preset threshold, a three-level response is initiated sequentially: Level 1 Response: Release Fire extinguishing agents suppress localized thermal runaway; Level 2 response: Disconnect the electrical connection of the faulty battery cluster using a pneumatic isolation device; Level 3 response: An emergency power adjustment command is sent through the substation deep coordination interface module, and the grid side is coordinated to adjust the operation strategy within 1 second, forming a safety response closed loop between the energy storage system and the grid.

[0050] Furthermore, to address the potential thermal runaway risks of vanadium redox flow batteries and lithium iron phosphate batteries in hybrid energy storage systems, this embodiment constructs a Bayesian network risk assessment model, defining three observation nodes (temperature, voltage change rate, and internal resistance growth rate) and a risk decision node, wherein: Temperature nodes are classified according to the battery surface temperature as follows: The data is collected in real time by an infrared thermal imager. Voltage change rate nodes are classified according to the rate of change of terminal voltage. It is calculated online through the battery management system (BMS); Internal resistance growth rate nodes are classified according to AC impedance identification results as follows: It relies on the impedance analyzer for periodic updates; Risk nodes are defined as {low risk (probability < 0.7), high risk (probability ≥ 0.7)}, and are used as... The triggering conditions for fire extinguishing and battery cluster isolation are based on a conditional probability table trained on historical thermal runaway experimental data. The core logic is that when temperature, voltage change rate, and internal resistance growth simultaneously exhibit abnormal characteristics (such as high temperature + positive mutation + high growth), the probability of thermal runaway increases significantly. This enables dynamic risk assessment and ensures the timeliness and accuracy of safety response.

[0051] It should be added that the multi-dimensional monitoring subunit 231 in this embodiment realizes full-state perception of the hybrid energy storage module through a distributed sensor network. The sensor layout and connection relationship are as follows: Distributed fiber optic sensors are laid in a serpentine pattern along the surface of the lithium iron phosphate module and the vanadium redox flow stack, and are closely attached to the contact area of ​​the liquid cooling pipeline (temperature measurement nodes are set every meter). They are connected to the optical signal processing equipment through the fiber optic interface, and after data conversion, they communicate with the main controller of the safety protection unit to upload temperature field data in real time. Infrared thermal imagers are evenly arranged on the top of the temperature control chamber (covering the entire surface of the battery module), and are connected to the switch via industrial Ethernet and communicate with the main controller using the standard Modbus TCP protocol to periodically scan the temperature gradient on the battery surface. The gas sensor array has gas detection probes (covering CO, CO2 ... (and smoke concentration monitoring), which is connected to the control module via analog signal cable and converted into a digital signal before being transmitted to the main controller to provide real-time warning of abnormal gas concentration; A micro-current sensor is connected in series in the negative electrode circuit of the battery pack (each battery cluster is equipped with an independent monitoring unit), and is connected to the main controller through a dedicated communication interface to collect leakage current data in real time and assess the insulation status. The pipeline pressure sensor is installed at the pump outlet, heat exchanger inlet and outlet, and front end of each branch solenoid valve in the liquid cooling circulation loop. It is rigidly connected to the liquid cooling pipeline through the pressure interface and outputs analog signals to the control module to provide real-time feedback on the system pressure status.

[0052] All sensors are connected to a master-slave redundant controller in the security protection unit via a redundant industrial Ethernet ring network. The system employs a dual-switch redundant configuration and supports industrial communication protocols such as ProfinetRT. The master controller collects sensor data in real time and the slave controller synchronizes and mirrors the status, forming a dual-machine hot standby redundancy mechanism to ensure uninterrupted operation of the monitoring system.

[0053] The multi-strategy collaborative control module 300 is used to construct a multi-level collaborative control architecture. By integrating multi-source data, it generates control strategies adapted to different operating scenarios, achieving optimized collaborative operation of multiple modules. The multi-strategy collaborative control module 300 adopts a three-level collaborative control architecture, using intelligent algorithms to achieve dynamic interaction optimization between the energy storage system and the power grid. The system supports industrial-grade communication protocols, meeting the requirements of real-time grid dispatch and collaborative control of energy storage devices.

[0054] In this embodiment, the multi-level cooperative control architecture in the multi-strategy cooperative control module 300 includes a grid response unit 310, an energy storage coordination unit 320, and a device execution unit 330, wherein: The grid response unit 310, based on an improved model predictive control algorithm, predicts the peak-valley electricity price curve and load fluctuation trend of the power grid within the next 48 hours through a rolling optimization window, and generates a charging and discharging strategy for the energy storage system. The objective function of the improved model predictive control algorithm is: ; in, Indicates control input The energy storage power control signal is optimized to minimize the objective function. To achieve the optimal solution for energy storage charging and discharging strategies; This indicates a time-domain summation prediction. This indicates control over time-domain summation; Indicates control variables; Relative to the current time The number of future prediction steps; To predict the length of the time domain, To control the length of the time domain; Represented as the predicted time The load power; For the predicted time The charging and discharging power of the energy storage system; For the predicted time The planned power output of the power grid; This represents the change in energy storage power between adjacent time points, and ,in Indicates the current moment Predicting the future Energy storage capacity at any given time Indicates the current moment Predicting the future Energy storage capacity at any given time; Indicates the current moment; , This represents the weighting coefficient; it is dynamically adjusted through an adaptive fuzzy inference algorithm, and its calculation formula is: ; in, For the first The error of each control target; The learning rate; Adjust the step size for weights; It is the first Each weight coefficient at the current time The value; It is the first The control objective at the current moment The error; Furthermore, the error Defined as: when hour, , For power grid dispatching, The actual output power of the energy storage represents the power tracking error. when hour, , To estimate the number of cycles for the battery management system, The actual number of cycles represents the error in battery life loss. when hour, , To predict entropy change for the model, The vanadium ion concentration is calculated using a thermodynamic formula, which is: , The gas constant characterizes the entropy change error of the electrolyte. This indicates the concentration of +3 vanadium ions in the electrolyte; This indicates the presence of +4 valent vanadium ions in the electrolyte; Furthermore, in the formula for the adaptive fuzzy inference algorithm, its denominator... This is used for normalization processing to ensure that the weight update is within a reasonable range, ultimately achieving dynamic weight adaptation for power tracking, lifetime loss, and entropy change control, thereby improving the robustness of multi-objective cooperative control strategies.

[0055] The energy storage coordination unit 320 is based on a multi-objective particle swarm optimization algorithm. It constructs a three-dimensional objective space that includes grid dispatch tracking, cycle life loss of lithium iron phosphate energy storage unit 110, and entropy change of electrolyte in vanadium redox flow energy storage unit 120. It solves the Pareto optimal solution set to minimize the overall operating cost of the energy storage system under the constraint of satisfying the grid power command. The device execution unit 330 implements the power allocation of the hybrid energy storage module 100 through a distributed consensus algorithm, dynamically adjusts the output weight of each energy storage unit, and realizes the synergistic effect of the power allocation strategy and the multi-objective optimization results of the energy storage coordination unit 320 to achieve a consistent optimization goal.

[0056] In this embodiment, the multi-strategy collaborative control module 300 further includes an adaptive learning unit 340. The adaptive learning unit 340 constructs a value function of the energy storage system's operating state through a deep Q-network and dynamically adjusts the control parameters based on historical operating data and real-time monitoring data. The adaptive learning unit 340 includes a transfer learning subunit 341, which generates a new control strategy through knowledge transfer when a change in the power grid operation mode is detected.

[0057] In this embodiment, the power grid response unit 310 further includes an abnormal event detection subunit 311, which identifies abnormal events such as power grid frequency changes and voltage drops in real time based on the isolated forest algorithm. When an abnormal event is detected, the abnormal event detection subunit 311 achieves rapid power support for the energy storage system through an improved droop control algorithm. The power allocation formula of the improved droop control algorithm is as follows: ; in, Provides real-time power output for the energy storage system; This is a reference value for energy storage power during normal grid operation; This is the droop control coefficient; For the real-time frequency of the power grid; The rated frequency of the power grid; Sag control coefficient Adaptive update via gradient descent: ; in For the learning rate, make Dynamically optimizes based on frequency deviation; It is the sign of a partial derivative; This indicates the next moment.

[0058] As a further explanation of this embodiment, the energy storage coordination unit 320 uses a multi-objective particle swarm optimization algorithm (MOPSO) to construct a three-dimensional optimization objective space, including: Grid dispatch tracking accuracy: Ensure the consistency between energy storage power output and grid commands; Cycle life loss of lithium iron phosphate energy storage units: Reduce battery aging by optimizing charge and discharge depth and rate; Entropy change of electrolyte in vanadium redox flow storage unit: Maintaining electrolyte temperature and concentration within the optimal range to reduce operation and maintenance costs.

[0059] This algorithm simulates the foraging behavior of bird flocks to search for Pareto optimal solutions in a three-dimensional target space, achieving collaborative optimization of multiple objectives. In this embodiment, the algorithm is used to generate an operating strategy that balances grid dispatch requirements with the health status of energy storage devices. For example, while meeting frequency regulation power requirements, it prioritizes the use of vanadium redox flow cells for long-term energy storage tasks to extend the lifespan of lithium iron phosphate cells.

[0060] As a further explanation of this embodiment, the device execution unit 330 implements dynamic power allocation of the hybrid energy storage module based on a distributed consensus algorithm. This algorithm gradually reaches a consensus on the global power allocation strategy through information exchange between energy storage unit nodes (such as SOC status and equivalent internal resistance), ensuring that the output weight of each unit is consistent with the upper-level optimization results. Its core principle is to iteratively update local variables so that the state of the distributed nodes converges to a unified target. In this embodiment, the algorithm dynamically adjusts the output ratio based on the real-time status of the lithium iron phosphate and vanadium redox flow units (such as the high power response capability of lithium iron phosphate and the long-term energy storage advantages of vanadium redox flow).

[0061] As a further explanation of this embodiment, in the adaptive learning unit 340, when the power grid operation mode changes (such as from "peak regulation as the main mode" to "frequency regulation as the main mode"), the historical control strategies under similar modes are reused through knowledge transfer technology to quickly generate optimized parameters under the new operating conditions, reduce the time cost of retraining, and improve the system's adaptability to complex scenarios.

[0062] It should be added that the units in this embodiment achieve data interaction through industry standard communication protocols (such as OPCUA): The grid response unit 310 sends charging and discharging strategy instructions to the energy storage coordination unit; The energy storage coordination unit 320 sends the optimized power allocation target to the device execution unit; The device execution unit 330 feeds back the actual power output and device status to the grid response unit; The adaptive learning unit 340 collects operational data from the entire module, updates control parameters, and synchronizes them to each unit, forming a closed loop of "perception-decision-execution-learning".

[0063] Furthermore, in this embodiment, the isolated forest algorithm is an anomaly detection technique based on unsupervised learning. It evaluates the degree of anomaly of data points by constructing a random forest composed of multiple "isolated trees".

[0064] The core principles of the Isolation Forest algorithm include: Sample isolation: Randomly sample power grid operation data (such as frequency and voltage) and gradually isolate each data point through random feature segmentation; Path length assessment: The shorter the path of a data point in the tree (i.e., the faster it is isolated), the higher its anomaly score, and the more likely it is to be judged as an anomalous event. Integrated decision-making: The final anomaly detection threshold is determined by the voting results of multiple trees.

[0065] Its function in this embodiment is: Real-time anomaly identification: Continuously scans real-time data such as power grid frequency and voltage to quickly identify abnormal events that exceed the normal fluctuation range (such as frequency mutations and voltage drops). Triggering rapid response: When an anomaly is detected, a trigger signal is sent to the grid response unit to activate the improved droop control algorithm, enabling the energy storage system to provide power support in a short time and suppress grid condition deterioration; Unsupervised adaptation: It does not rely on a large number of labeled samples, and is suitable for complex and ever-changing abnormal scenarios in power grid operation, improving the system's robustness to unknown faults.

[0066] This algorithm provides crucial preliminary judgment for the rapid response mechanism of energy storage systems through its efficient anomaly detection capabilities, ensuring grid stability and the safety of energy storage devices.

[0067] The Substation Deep Collaboration Interface Module 400 is used to build a standardized data interaction and command response channel with 220KV substations, supporting seamless integration between energy storage systems and substation automation systems, and enabling real-time synchronization of grid status information and rapid response to control commands.

[0068] In this embodiment, the substation deep collaboration interface module 400 includes a power grid status synchronization unit 410, a control command parsing unit 420, and a security authentication unit 430, wherein: The power grid status synchronization unit 410 is used to acquire the power grid status information of the substation in real time, realize the synchronization of power grid transient information, and provide real-time power grid operation data support for the coordinated control of the energy storage system. The power grid status synchronization unit 410 is connected to the substation monitoring system SCADA through a dedicated communication link, and collects key power grid operation data such as bus voltage, line current, frequency, active / reactive power, etc. in real time. Based on the IEEE1588v2 precise clock synchronization protocol, it realizes nanosecond-level time synchronization between the energy storage system and the substation, ensuring the consistency of data sampling time.

[0069] The control command parsing unit 420 constructs a command mapping model based on a deep neural network, learns the relationship between historical commands and energy storage response, parses scheduling commands and converts them into power control signals. The security authentication unit 430 is used to establish a two-way identity authentication mechanism between the energy storage system and the substation to provide security protection for data transmission.

[0070] In this embodiment, the grid state synchronization unit 410 includes a grid state estimation subunit 411, which processes the acquired grid state data and outputs key parameters through a Kalman filter algorithm to provide input for the optimization of the control strategy of the energy storage system.

[0071] As a further explanation of this embodiment, the substation deep collaboration interface module 400 in this embodiment provides a standardized communication interface. The power grid side supports protocols such as IEC61850 and DL / T634.5104, and realizes high-speed data transmission through fiber optic channels. The energy storage side adopts the Modbus series protocol and connects with the multi-strategy collaborative control module 300 to realize bidirectional interaction of commands and status data.

[0072] As a further explanation of this embodiment, the Kalman filter algorithm is used to denoise and estimate the state of real-time measurement data. The Kalman filter algorithm is a recursive filtering technique based on a state-space model. Its core principle is to construct the system state transition equation and measurement equation, and recursively calculate the optimal state estimate at the current moment using the state estimate value at the previous moment and the current measurement value, effectively suppressing measurement noise from sensors such as CT / PT. In this embodiment, the algorithm is used to improve the reliability of grid state parameters (such as node voltage and power flow distribution), provide high-precision input data for model predictive control (MPC) of the energy storage system, and ensure the accuracy of the charging and discharging strategy.

[0073] As a further explanation of this embodiment, the control command parsing unit 420 in this embodiment constructs a command mapping model based on a deep neural network (DNN). This model achieves intelligent parsing of substation commands by learning the correspondence between historical scheduling commands (such as power regulation and charging / discharging time period plans) and energy storage system response data (such as power curves and SOC changes). Its principle is to utilize the nonlinear mapping capability of multi-layer neural networks to convert abstract scheduling commands into executable power control signals for the energy storage system (such as charging / discharging power values ​​and timing nodes). In this embodiment, this unit can automatically parse the "peak-valley charging / discharging commands" issued by the substation into specific control parameters including power amplitude and time windows. For example, it can convert the "18:00-22:00 discharge" command into a real-time power allocation curve for lithium iron phosphate and vanadium redox flow units, thereby improving the accuracy and efficiency of command execution.

[0074] It should be added that the security authentication unit 430 establishes a two-way authentication mechanism between the energy storage system and the substation. Based on the X.509 digital certificate and the SSL / TLS protocol, it verifies the identities of both parties before communication and negotiates an encrypted session key. During data transmission, the AES-256 encryption algorithm is used to encrypt instructions and status information, and the HMAC-SHA256 algorithm is used for integrity verification to prevent data tampering or illegal interception. In this embodiment, this unit ensures that the interaction between the energy storage system and the substation complies with the power system information security protection requirements. For example, when receiving peak-shaving instructions, it first verifies the validity of the substation certificate, and then decrypts and verifies the instruction content to avoid system anomalies caused by malicious instructions and ensure the safe operation of the power grid and energy storage equipment.

[0075] Those skilled in the art will understand that the process of implementing all or part of the steps of the above embodiments can be carried out by hardware or by a program instructing the relevant hardware.

[0076] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A non-walk-in collaborative control system for a 220kV substation-connected liquid-cooled phosphoric acid and vanadium redox flow storage power station, characterized in that, include: The hybrid energy storage module (100) connects a lithium iron phosphate energy storage unit (110) and a vanadium redox flow battery (120) in parallel via a DC bus to construct a complementary energy storage system with high and low characteristics of lithium iron phosphate batteries and vanadium redox flow batteries, supporting the coordinated operation of short-term high-power response and long-term deep energy storage. The non-walk-in intelligent liquid cooling module (200) is used to implement fully enclosed differentiated thermal management and active safety protection for the hybrid energy storage module (100). It monitors the battery pack temperature and abnormal operating conditions in real time through multi-source sensors, and achieves stable temperature maintenance and safety response without human intervention. The multi-strategy collaborative control module (300) is used to build a multi-level collaborative control architecture. It generates control strategies that adapt to different operating scenarios by integrating multi-source data, and realizes multi-module collaborative operation optimization. The substation deep collaboration interface module (400) is used to build a standardized data interaction and command response channel with the 220KV substation, supporting seamless connection between the energy storage system and the substation automation system, and realizing real-time synchronization of power grid status information and rapid response to control commands.

2. The collaborative control system for the non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station connected to the 220KV substation according to claim 1, characterized in that, The lithium iron phosphate energy storage unit (110) includes a battery module subunit (111) and a battery management subunit (112), wherein: The battery module subunit (111) consists of N parallel battery modules, each of which contains M series-connected lithium iron phosphate battery cells. The battery management subunit (112) performs SOC estimation based on the Kalman filter algorithm.

3. The collaborative control system for the non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station connected to the 220KV substation according to claim 1, characterized in that: The vanadium redox flow storage unit (120) includes an electrolyte circulation subunit (121) and a stack management subunit (122), wherein: The electrolyte circulation subunit (121) includes an electrolyte storage tank, a circulation pump, and a heat exchanger; The stack management subunit (122) is used to collect stack voltage and current in real time and communicate with the multi-strategy collaborative control module (300) through the ModbusTCP protocol; when the voltage difference between adjacent single cells exceeds the preset threshold, the electrolyte distribution ratio is dynamically adjusted based on the fuzzy control algorithm.

4. The collaborative control system for a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station connected to a 220kV substation according to claim 1, characterized in that, The non-walk-in intelligent liquid cooling module (200) includes a temperature-controlled chamber unit (210) and a liquid cooling circulation unit (220), wherein: The temperature control chamber unit (210) uses a metal frame to construct an independent temperature control space. It forms a dynamic temperature control chamber adapted to the high-power heat dissipation of lithium iron phosphate and a steady-state temperature control chamber adapted to the long-term stability of vanadium liquid flow through heat insulation partitions. The two chambers are thermally isolated by a three-layer composite wall. The chamber interface is sealed with double O-ring seals, which supports wide temperature range operation and provides a fully enclosed temperature control environment without human intervention for the hybrid energy storage module (100). The liquid cooling circulation unit (220) includes a dual-loop control subunit (221) and a flow regulation subunit (222), wherein: The dual-loop control subunit (221) includes a lithium iron phosphate unit loop and a vanadium redox flow unit loop. The lithium iron phosphate unit loop uses a centrifugal liquid-cooled pump, a plate heat exchanger, and a solenoid valve, and uses ethylene glycol aqueous solution as coolant to quickly remove instantaneous heat through high-frequency flow regulation. The vanadium redox flow unit loop uses a screw circulating pump, an electric heating rod, and a solenoid valve, and uses deionized water as coolant to maintain electrolyte stability through low-fluctuation temperature control. The flow regulation subunit (222) dynamically adjusts the opening of the solenoid valve based on temperature data from multiple sources and combined with a fuzzy PID algorithm, thereby achieving reasonable control of the temperature difference between the lithium iron phosphate battery and the vanadium redox flow battery.

5. The collaborative control system for the non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station connected to the 220KV substation according to claim 4, characterized in that, The non-walk-in intelligent liquid cooling module (200) also includes a safety protection unit (230), which includes a multi-dimensional monitoring subunit (231), a rapid response subunit (232), and an intelligent decision-making subunit (233), wherein: The multi-dimensional monitoring subunit (231) monitors the temperature field in real time through distributed optical fiber sensors laid in a serpentine pattern along the battery module, and scans the battery surface with an infrared thermal imager to identify abnormal temperature gradients; the multi-dimensional monitoring subunit (231) also monitors the gas concentration inside the battery box through an integrated gas sensor array, and monitors the battery leakage current through a micro current sensor to realize early warning of insulation faults when leakage occurs. The fast response subunit (232) uses a pipeline pressure sensor to monitor the pressure in real time. When the detected value exceeds the threshold, the coolant circuit is cut off synchronously based on a redundancy control algorithm. The intelligent decision-making subunit (233) constructs a Bayesian network risk assessment model based on multi-source sensor data. When the predicted risk of thermal runaway reaches a preset threshold, it automatically triggers the CO2 fire extinguishing system and the battery cluster isolation device, and sends an emergency power adjustment command to the 220KV substation through the substation deep collaborative interface module (400).

6. The collaborative control system for a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station connected to a 220kV substation according to claim 1, characterized in that, The multi-strategy collaborative control module (300) includes a multi-level collaborative control architecture comprising a grid response unit (310), an energy storage coordination unit (320), and a device execution unit (330), wherein: The grid response unit (310) is based on an improved model predictive control algorithm. It predicts the peak-valley electricity price curve and load fluctuation trend of the grid within the next 48 hours through a rolling optimization window, generating a charging and discharging strategy for the energy storage system. The objective function of the improved model predictive control algorithm is: ； in, Indicates the control input Optimize and minimize the objective function To achieve the optimal solution for energy storage charging and discharging strategies; This indicates a time-domain summation prediction. This indicates a time-domain summation control. Indicates control variables; Relative to the current time The number of future prediction steps; To predict the length of the time domain, To control the length of the time domain; Represented as the predicted time The load power; For the predicted time The charging and discharging power of the energy storage system; For the predicted time The planned power output of the power grid; This represents the change in energy storage power between adjacent time points; Indicates the current moment; , Indicates the weighting coefficient; The energy storage coordination unit (320) is based on a multi-objective particle swarm optimization algorithm to construct a three-dimensional target space that includes grid dispatch tracking, cycle life loss of lithium iron phosphate energy storage unit (110), and entropy change of electrolyte in vanadium redox flow energy storage unit (120). The Pareto optimal solution set is used to solve the problem and minimize the overall operating cost of the energy storage system under the constraint of grid power command. The device execution unit (330) implements the power allocation of the hybrid energy storage module (100) through a distributed consensus algorithm, dynamically adjusts the output weight of each energy storage unit, and realizes the synergistic effect of the power allocation strategy and the multi-objective optimization results of the energy storage coordination unit (320) to achieve a consistent optimization goal.

7. The non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station collaborative control system for 220KV substation connection according to claim 6, characterized in that: The multi-strategy collaborative control module (300) also includes an adaptive learning unit (340), which constructs a value function of the energy storage system's operating state through a deep Q-network and dynamically adjusts control parameters based on historical operating data and real-time monitoring data. The adaptive learning unit (340) includes a transfer learning subunit (341), which generates a new control strategy through knowledge transfer when a change in the power grid operation mode is detected.

8. The non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station collaborative control system for 220KV substation connection according to claim 6, characterized in that: The power grid response unit (310) further includes an abnormal event detection subunit (311), which identifies abnormal events in real time based on the isolated forest algorithm; When an abnormal event is detected, the abnormal event detection subunit (311) achieves rapid power support for the energy storage system through an improved droop control algorithm. The power allocation formula of the improved droop control algorithm is as follows: ； in, Provides real-time power output for the energy storage system; This is a reference value for energy storage power during normal grid operation; This is the droop control coefficient; For the real-time frequency of the power grid; The rated frequency of the power grid; Sag control coefficient Adaptive update via gradient descent: ； in For the learning rate, make Dynamically optimizes based on frequency deviation; It is the sign of a partial derivative; This indicates the next moment.

9. The collaborative control system for a non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station connected to a 220kV substation according to claim 1, characterized in that, The substation deep collaboration interface module (400) includes a power grid status synchronization unit (410), a control command parsing unit (420), and a security authentication unit (430), wherein: The power grid status synchronization unit (410) is used to acquire the power grid status information of the substation in real time, realize the synchronization of power grid transient information, and provide real-time power grid operation data support for the coordinated control of the energy storage system. The control command parsing unit (420) constructs a command mapping model based on a deep neural network, learns the relationship between historical commands and energy storage response, parses the scheduling command and converts it into a power control signal; The security authentication unit (430) is used to establish a two-way identity authentication mechanism between the energy storage system and the substation to provide security protection for data transmission.

10. The non-walk-in liquid-cooled phosphoric acid and vanadium redox flow storage power station collaborative control system for 220KV substation connection according to claim 9, characterized in that: The grid state synchronization unit (410) includes a grid state estimation subunit (411), which processes the acquired grid state data and outputs key parameters through a Kalman filter algorithm to provide input for the optimization of the control strategy of the energy storage system.

Images