Energy coordination control system and scheduling method for multi-energy complementary new energy storage

Abstract

The invention provides an energy coordination control system and scheduling method for multi-energy complementary new energy storage, and belongs to the technical field of new energy storage and energy technology management. The data acquisition and processing module is connected with a current and voltage sensor, a power transmitter, a meteorological station sensor and a battery management system sensor, and is used for acquiring voltage, current, power, state of charge and environmental data of photovoltaic power, wind power, energy storage and load through the sensors; and sliding window filtering and wavelet de-noising processing are carried out through the embedded processor. The method has the beneficial effects that the uncertainty of wind and light output is effectively processed by establishing a layered distributed optimization architecture and adopting a method of combining multi-objective optimization and robust optimization, and the consumption rate of renewable energy sources can be remarkably improved, the net load fluctuation can be reduced and the energy consumption can be reduced on the premise of meeting various system operation constraints. Therefore, dependence on conventional fossil energy is reduced, and the economic operation level and the environmental protection benefit of the whole energy system are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of new energy storage and energy technology management, and particularly relates to an energy coordination control system and a scheduling method for a multi-energy complementary new energy storage. BACKGROUND

[0002] With the transformation of global energy structure to clean energy, the installed capacity of renewable energy such as wind power and photovoltaic is rapidly increasing, but such energy has the problems of large output fluctuation and strong randomness due to natural conditions, which brings challenges to the safe and stable operation of the power grid. To solve this problem, a multi-energy complementary new energy storage system emerges as the times require, which integrates multiple energy and storage forms and utilizes the complementary characteristics of different energies to improve the reliability of the system. However, the energy management of the existing multi-energy complementary system still has the problems of insufficient coordination control precision, poor adjustment strategy adaptability and low market mechanism integration, so an energy coordination control system and a scheduling method for a multi-energy complementary new energy storage are proposed to solve the above problems. SUMMARY

[0003] The technical problem to be solved by the application is how to improve the coordination control precision and improve the adaptability of the adjustment strategy. In view of the deficiencies of the prior art, an energy coordination control system and a scheduling method for a multi-energy complementary new energy storage are provided.

[0004] To solve the above technical problems, the technical solution adopted by the application is: an energy coordination control system for a multi-energy complementary new energy storage, comprising: A data acquisition and processing module, the data acquisition and processing module is connected with a current-voltage sensor, a power transmitter, a weather station sensor and a battery management system sensor, collects voltage, current, power, state of charge and environmental data of photovoltaic, wind power, energy storage and load through the sensors, and performs sliding window filtering and wavelet denoising processing through an embedded processor; An energy prediction module, the energy prediction module is in communication connection with the data acquisition and processing module, and is used to receive the filtered and denoised physical state data signal, outputs point prediction values, confidence intervals and extreme weather warnings of wind and light output through a hybrid prediction model containing a long short-term memory network and a physical mechanism; An optimization scheduling engine module, which is in communication connection with the data acquisition and processing module and the energy prediction module respectively, establishes a hierarchical distributed optimization architecture, which includes a centralized optimization layer at the upper layer and a distributed collaborative control layer at the lower layer; the centralized optimization layer, based on the prediction results and the prediction confidence interval received from the energy prediction module and the factual system state received from the data acquisition and processing module, takes the system operation economy and stability as the objective function, comprehensively considers the prediction uncertainty and equipment operation constraints, and adopts a rolling horizon optimization method to obtain the globally optimal energy storage unit reference power plan in the future scheduling period; the distributed collaborative control layer is deployed in each local controller, used to receive the reference power plan, and based on a consistency algorithm or a distributed quadratic programming algorithm, combined with the locally measured bus frequency deviation or voltage deviation data, to quickly calculate and generate accurate energy storage charging and discharging power set value instructions; An energy storage management module, which is in communication connection with the data acquisition and processing module and the optimization scheduling engine module respectively, is configured to perform charging and discharging collaborative control and state management on different types of energy storage devices according to the power set value instructions; A communication and network module, which is in communication connection with the data acquisition and processing module, the energy prediction module, the optimization scheduling engine module and the energy storage management module, is used to provide heterogeneous protocol conversion, data encryption and edge computing services, and realize safe data interaction between modules; A control execution module, which is in communication connection with the optimization scheduling engine module and the communication and network module, is configured to drive the power converter to perform operations according to the power set value instructions, for realizing smooth control of new energy power fluctuation and anti-peaking compensation.

[0005] Further, a market strategy and collaborative control module and a state monitoring and intelligent operation and maintenance module are further included; the market strategy and collaborative control module is in communication connection with the optimization scheduling engine module and the communication and network module, used to receive system real-time power balance state, energy storage state of charge and short-term power prediction information, based on dynamic price signals and system operation constraints, to generate bidding strategies for participating in power market transactions and adjustment instructions for controllable loads by solving a multi-time scale economic optimization model; the state monitoring and intelligent operation and maintenance module is in communication connection with each of the above modules, used to collect equipment operation characteristic data in real time and input into a pre-trained neural network fault diagnosis model for analysis, and output equipment health state evaluation results and early fault warning; at the same time, the module integrates a visual human-computer interaction unit, used to dynamically display system topology, real-time energy flow distribution, key alarms and performance indicators.

[0006] Further, the data acquisition and processing module comprises a data acquisition unit, a data processing unit and a data quality evaluation and output unit; The data acquisition unit is connected with the current sensor, the voltage sensor, the power transmitter, the weather station sensor and the battery management system (BMS) sensor through an isolated signal conditioning circuit, and is used for synchronously collecting the multi-channel analog and digital signals of the photovoltaic array, the wind turbine generator, the energy storage unit and the load at a sampling frequency of not less than 1 kHz, and performing amplitude calibration and format unification on the collected signals; the data processing unit is in communication connection with the data acquisition unit, and is internally provided with a multi-core embedded processor, which is configured to: firstly perform moving average filtering or median filtering based on a fixed-length sliding window on the collected time series data stream, so as to suppress impulse noise; then perform multi-resolution decomposition on the filtered data by using a Daubechies wavelet basis function, process detail coefficients by using a soft threshold function or a hard threshold function to suppress Gaussian noise, and perform wavelet reconstruction; the data quality evaluation and output unit is connected with the data processing unit, and is used for calculating the signal-to-noise ratio and the root mean square error of the denoised data, and outputting the data with qualified quality to the energy prediction module and the optimization scheduling engine module through a standard industrial bus interface after time stamping.

[0007] Further, the optimization scheduling engine module comprises: The centralized optimization layer is connected with the data acquisition and processing module and the energy prediction module, and is used for receiving real-time system state data, wind and light power point prediction values and confidence intervals thereof; the centralized optimization layer establishes an upper-layer optimization model based on a model predictive control (MPC) framework, and the upper-layer optimization model takes minimization of system comprehensive operation cost and minimization of power fluctuation rate as objective functions, and the expression is as follows: Meanwhile, the following constraints are met: wherein, represents the exchange cost with the external power grid, represents the energy storage loss cost, represents a power balance penalty factor, , , and respectively represent the grid interaction power, the energy storage output, the load power and the new energy power at the t period; the centralized optimization layer solves the optimization problem on line by using an interior point method, and rolls out the reference power plans of the energy storage units in a future scheduling period; A distributed cooperative control layer is deployed in each local controller of the energy storage unit, connected to the centralized optimization layer through a signal network, receives the reference power plan issued by the upper layer, and performs distributed cooperative calculation based on a consistency algorithm: each local controller exchanges state information with its communication neighbors, updates the local state through iteration, so that the processing instructions of all energy storage units converge to consistency, and quickly compensate for real-time power differences. The distributed cooperative control layer finally generates accurate energy storage charging and discharging power set value instructions, and issues them to the energy storage management module.

[0008] Further, the energy storage management module comprises: A multi-objective power allocation unit is connected to the optimization scheduling engine module for receiving the accurate energy storage charging and discharging power set value instructions; based on the dynamic characteristics of different types of energy storage devices, this unit performs separation calculation of high-frequency power components and low-frequency power components, and dynamically calculates the optimal power allocation ratio of each energy storage unit according to the equivalent internal resistance, remaining cycle life and instantaneous power response capability of lithium ion batteries, flow batteries and super capacitors using an adaptive weighting algorithm; A multi-constraint charging and discharging control unit is connected to the multi-objective power allocation unit and the data acquisition and processing module for receiving the allocated power instructions and real-time state data of each energy storage unit; this unit has built-in charging and discharging control logic, which is configured to: according to the received power instruction value and the real-time state of charge SOC, temperature and state of health SOH data uploaded by the battery management system BMS, adopt a three-stage charging curve or a constant power discharging strategy, and generate charging and discharging current limit value instructions and voltage protection threshold values in real time through table lookup or interpolation calculation; when it is monitored that the SOC of any energy storage unit exceeds the preset safety interval, the temperature exceeds the working limit value or the port voltage is abnormal, the hardware protection logic is triggered immediately to generate a power reduction or shutdown instruction; A pulse width modulation signal generation unit is connected to the multi-constraint charging and discharging control unit for converting the stable charging and discharging current instructions into corresponding pulse width modulation PWM duty cycle signals; this unit integrates a carrier generator and a comparator circuit, which accurately controls the on and off time of insulated gate bipolar transistors IGBT or metal-oxide semiconductor field effect transistors MOSFET in the energy storage converter PCS by changing the duty cycle of the PWM wave, thereby driving the energy storage converter to execute the power instructions and realize physical control of the battery pack charging and discharging process.

[0009] Further, the control execution module comprises: An instruction analysis and signal conversion unit is connected with the optimization scheduling engine module through a high-speed data bus, and is configured to receive the accurate energy storage charge and discharge power setting value instruction; a digital signal processor (DSP) is built in the unit, and is configured to convert the power instruction value into corresponding current loop reference instruction or voltage loop reference instruction through a preset lookup table or real-time calculation according to the direct-current bus voltage level and the current switching frequency of the power conversion system (PCS); A pulse width modulation wave generation unit is connected with the instruction analysis and signal conversion unit, and is configured to convert the current or voltage reference instruction into a pulse width modulation (PWM) wave with a corresponding duty cycle; the unit integrates a triangular carrier generator and a digital comparator, and generates a plurality of complementary PWM drive signals with dead time through a space vector pulse width modulation (SVPWM) or a sine pulse width modulation (SPWM) algorithm; A drive signal amplification and isolation unit is connected with the pulse width modulation wave generation unit, and is configured to perform power amplification and electrical isolation on the generated PWM logic signal; the unit includes a gate drive integrated circuit and an isolation optocoupler or a magnetic coupling element, amplifies the voltage and current of the generated drive signal to a level sufficient to drive the power switching device, and provides several kilovolts of electrical isolation, and finally outputs to the power unit of the power conversion system (PCS); A power unit execution and state feedback unit is directly connected with the drive signal amplification and isolation unit and an insulated gate bipolar transistor (IGBT) or silicon carbide MOSFET power module in the power conversion system (PCS); the unit controls the conduction and turn-off of the power switching device according to the amplified drive signal, converts direct-current electric energy into alternating-current electric energy and connects to a power grid or performs a charging operation; at the same time, the unit samples the voltage and current on the output side of the power conversion system (PCS) in real time through a high-speed analog-to-digital converter (ADC), forms a feedback signal required for closed-loop control, and sends the signal back to the instruction analysis and signal conversion unit, to form a closed-loop control loop, realize accurate and rapid tracking control of the output power, and complete millisecond-level smoothing and anti-peaking compensation for new energy power fluctuations.

[0010] Further, the market strategy and cooperative control module includes: A multi-time scale collaborative optimization unit is connected with the optimization scheduling engine module and the data acquisition and processing module through a high-speed data bus, and is used to receive real-time power balance state, energy storage state of charge (SOC), and ultra-short-term power prediction information. A high-performance computing core is built in the unit, which is configured to establish a multi-time scale economic optimization model based on dynamic price signals, system operation constraints, and network transmission losses, with the goal of minimizing total operation cost and maximizing renewable energy consumption. The model uses stochastic optimization or robust optimization methods to handle the uncertainty of wind and solar output and market electricity prices, and calls an interior point method or a mixed integer programming solver for online solution, rolling out a global optimal economic dispatch plan for the future 24 hours to 15 minutes time scale. A controllable load regulation instruction generation unit is connected with the multi-time scale collaborative optimization unit. The unit receives the economic dispatch plan and calculates the optimal load regulation strategy through a model predictive control (MPC) framework based on the user's preset interruptible load priority and comfort constraint. The strategy is converted into specific direct load control (DLC) instructions or price-based demand response (PR-DR) signals and issued to each intelligent terminal device through the communication and network module. A cross-regional collaborative control interface unit is connected with the multi-time scale collaborative optimization unit and the communication and network module. It is used for real-time data exchange with adjacent regional energy management systems. Based on a consistency algorithm or a distributed optimization method, the interface calculates the optimal power reference value for inter-regional power exchange and securely encrypts the interaction data for transmission through a hardware encryption module. A real-time power balance correction unit is connected with the controllable load regulation instruction generation unit and the cross-regional collaborative control interface unit. The unit collects grid frequency deviation and voltage phase information through a high-precision phase measurement unit (PMU). When power imbalance or frequency limit is detected, a real-time power correction algorithm based on sensitivity analysis is immediately started to generate fast regulation instructions for energy storage units and controllable loads, which are sent to the optimization scheduling engine module for execution through the communication and network module.

[0011] Further, the energy coordination control and dispatch method of multi-energy complementary new energy storage includes the following steps: S1, data acquisition and preprocessing, collecting photovoltaic irradiance, environmental temperature, wind speed, energy storage state of charge (SOC), load power, and grid dispatching instruction signals through a sensor network, and preprocessing the collected raw data using a combination of sliding window filtering and wavelet threshold denoising. S2, uncertainty and risk quantification, based on the wind and solar power point prediction values and their confidence intervals output by the energy prediction module, the uncertainty range of new energy output at different time scales in the future is quantified; for extreme weather warning signals, the potential impact risk on system power balance is evaluated, and probabilistic input representing prediction uncertainty and extreme risk is provided for subsequent optimal scheduling; S3, optimal scheduling decision, taking the minimum net load variance and the maximum clean energy consumption as the upper target, and taking the minimum conventional power regulation cost as the lower target, using multi-objective optimization algorithm and robust optimization method, combining system operation constraints and prediction uncertainty, formulating global optimal scheduling plan and determining the output of each unit; S4, energy storage collaborative control, according to the scheduling plan, the hybrid control of active balancing and passive balancing is performed on the battery and super capacitor hybrid energy storage system; Based on the real-time internal short circuit detection algorithm and insulation resistance online monitoring data, start hierarchical safety protection and redundant control logic; And according to the power demand and system operation cost of different time scales, the charge and discharge power of each energy storage unit is dynamically allocated to complete the rapid response and accurate execution of power command; S5, communication and instruction transmission, with the help of the communication and network module, at least 20 kinds of industrial protocols such as Modbus, BACnet and CAN are compatible, through 5G+ optical fiber hybrid networking, using SM4 algorithm encryption, the optimal scheduling instruction is transmitted from the optimal scheduling engine module to the control execution module of each energy equipment safely; S6, control execution and feedback, the control execution module executes power smoothing control according to the received instruction, controls the wind and solar power output within 7% with 15-minute fluctuation rate, maintains the grid frequency stable at 50±0.2Hz through millisecond level wide area coordinated control, and at the same time monitors and feeds back the control execution result to the data acquisition and processing module, providing data support for subsequent optimal scheduling; S7, adjustable load control and regional coordination, according to the system real-time power shortage and prediction information, generate load regulation instructions, and directly act on interruptible load control unit through hardwiring or special communication channel; Energy routing equipment adjusts the working point of its power conversion unit according to the power interaction instruction between regions, realizes power exchange and dynamic support with adjacent regions; S8, state monitoring and closed-loop regulation, real-time acquisition of energy storage SOC, equipment temperature, voltage harmonic distortion rate and frequency deviation data, input into the pre-trained neural network model to calculate the battery health state SOH and generate fault warning signals; According to the diagnosis results and warning levels, trigger the corresponding device protection logic or system operation parameter adjustment instruction to complete the closed-loop regulation.

[0012] Further, the S4 is combined with the S4, and the S4 is combined with the S4. The active equalization circuit based on the switch matrix is used for energy transfer between the battery clusters, and the passive equalization resistor is used for absorbing the small differences between the single batteries in the battery pack; the dynamic allocation of the charge and discharge power of each energy storage unit is based on the instantaneous high-power response characteristics of the super capacitor and the high energy density characteristics of the lithium battery, and the high-frequency component in the power instruction is allocated to the super capacitor, and the low-frequency component and the steady-state power support task are allocated to the lithium battery.

[0013] Further, the S8 is combined with the S8, and the S8 is combined with the S8. The real-time collected battery charge and discharge voltage, current curve, temperature and historical cycle number data are input as input features into the pre-trained one-dimensional convolutional neural network model, the model outputs the current health state SOH quantitative value and the remaining useful life RUL estimation value of the battery; the trigger corresponding device protection logic includes: when the SOH value is lower than the preset threshold or the RUL estimation value reaches the warning limit value, the automatic de-rating operation instruction or the maintenance alarm signal is generated.

[0014] Compared with the prior art, the beneficial effects of the present application are as follows: 1. The present application establishes a hierarchical distributed optimization architecture, and adopts a method combining multi-objective optimization and robust optimization, effectively handles the uncertainty of wind and light output, can significantly improve the consumption rate of renewable energy under the premise of meeting various system operation constraints, reduce the net load fluctuation, thereby reducing the dependence on conventional fossil energy, and improving the economic operation level and environmental protection benefit of the whole energy system.

[0015] 2. The present application decomposes the power instruction according to the time scale and dynamically allocates it to different types of energy storage devices, combines the safety redundancy control based on real-time internal short circuit detection and insulation monitoring, and the millisecond-level wide-area fast power regulation, realizes the multi-level and fine collaborative control of wind and light power fluctuation, ensures the stability of power grid frequency, and greatly improves the safety and reliability of power system under high proportion of new energy access. BRIEF DESCRIPTION OF DRAWINGS

[0016] The present application will be further described in detail below with reference to the accompanying drawings.

[0017] Figure 1 : The structure schematic diagram of each system connection of the present application; Figure 2 : The scheduling method flowchart of energy coordination management of the present application.

[0018] Wherein, 1, data acquisition and processing module; 2, energy prediction module; 3, optimization scheduling engine module; 4, energy storage management module; 5, communication and network module; 6, control execution module; 7, market strategy and collaborative control module; 8, state monitoring and intelligent operation and maintenance module. DETAILED DESCRIPTION

[0019] In order to better understand the present application, the following further clearly sets forth the content of the present application in conjunction with the examples and drawings, but the protection scope of the present application is not limited to the following examples. In the following description, a large number of specific details are given in order to provide a more thorough understanding of the present application. However, it is obvious to those skilled in the art that the present application can be implemented without one or more of these details.

[0020] The present application adopts the following technical scheme: a multi-energy complementary new energy energy storage energy coordination control system, comprising: The data acquisition and processing module 1 is connected with the current and voltage sensor, the power transmitter, the weather station sensor and the battery management system sensor, collects the voltage, current, power, state of charge and environmental data of photovoltaic, wind power, energy storage and load through the sensor, and performs sliding window filtering and wavelet denoising processing through the embedded processor; The energy prediction module 2 is in communication connection with the data acquisition and processing module 1, and is used for receiving the filtered and denoised physical state data signal, outputting the point prediction value, confidence interval and extreme weather warning of wind and light output through a hybrid prediction model containing a long short-term memory network and a physical mechanism; The optimization scheduling engine module 3 is in communication connection with the data acquisition and processing module 1 and the energy prediction module 2 respectively, establishes a hierarchical distributed optimization architecture, which includes a centralized optimization layer at the upper layer and a distributed collaborative control layer at the lower layer; the centralized optimization layer takes the system operation economy and stability as the objective function based on the prediction results and prediction confidence interval received from the energy prediction module and the factual system state received from the data acquisition and processing module, comprehensively considers the prediction uncertainty and equipment operation constraints, adopts a rolling time domain optimization method, and solves to obtain the globally optimal energy storage unit reference power plan in the future scheduling period; the distributed collaborative control layer is deployed in each local controller, is used for receiving the reference power plan, and performs fast calculation based on a consistency algorithm or a distributed quadratic programming algorithm in combination with the bus frequency deviation or voltage deviation data measured in real time to generate accurate energy storage charging and discharging power set value instructions; The energy storage management module 4 is in communication connection with the data acquisition and processing module 1 and the optimization scheduling engine module 3 respectively, and is configured to perform charging and discharging collaborative control and state management on different types of energy storage devices according to the power set value instructions; The communication and network module 5 is in communication connection with the data acquisition and processing module 1, the energy prediction module 2, the optimal scheduling engine module 3 and the energy storage management module 4 respectively, and is configured to provide heterogeneous protocol conversion, data encryption and edge computing services, and to realize safe data interaction between the modules. The control execution module 6 is in communication connection with the optimal scheduling engine module 3 and the communication and network module 5, and is configured to drive the power converter to perform operation according to the power set value instruction, so as to realize smooth control of new energy power fluctuation and anti-peaking compensation.

[0021] It should be noted that the data acquisition and processing module 1 adopts an industrial-grade multi-core ARM processor with a model of TI AM5708 as a core, and is externally connected with a 16-bit high-precision ADC chip (such as ADS8568) for collecting sensor analog signals. The current and voltage sensor adopts a closed-loop Hall sensor of the LEM company, the power transmitter adopts a Honeywell PPT series, the weather station sensor includes a Kipp&Zonen CMP22 type total radiation table and a Vaisala WXT536 multifunctional weather sensor, and the battery management system (BMS) sensor collects battery cell voltage, temperature and total voltage and current signals. All sensor signals are connected to the terminal board card of the data acquisition and processing module 1 through shielded twisted pair wires.

[0022] The upper layer centralized optimization layer of the energy prediction module 2 and the optimal scheduling engine module 3 is deployed on a Dell PowerEdge R740xd server, which is configured with two Intel Xeon Silver 4216 processors, 256GB DDR4 memory and a NVIDIA T4 GPU acceleration card, and is used to run a machine learning prediction model and a large-scale optimization solving algorithm. The lower layer distributed collaborative control layer of the optimal scheduling engine module 3 is deployed in a local controller near each energy storage unit, and the local controller adopts a Guangqian GQ3288 industrial-grade ARM controller.

[0023] The energy storage management module 4 adopts an NXP MPC5744P multi-core microcontroller, which communicates with the BMS master of each energy storage unit through a CAN bus and controls the MOSFET switch array in the active equalization circuit through a PWM output interface.

[0024] The communication and network module 5 adopts Moxa UC-8100A-ME-T industrial Internet of Things gateway, which is pre-installed with TIA (Totally Integrated Automation) protocol conversion software, supports mutual conversion of more than 20 kinds of industrial protocols such as Modbus TCP / RTU, IEC 61850, Profinet, BACnet / IP, etc. The gateway is built-in with a hardware encryption chip, supporting SM4 and AES-256 encryption algorithms. The 5G communication adopts Huawei MU739 industrial module, and the optical fiber network adopts Hirschmann MACH4000 series switch.

[0025] The core of the control execution module 6 is the controller of the energy storage converter PCS, which adopts TI TMS320F28379D dual-core DSP chip. The DSP receives power setting value instructions, executes space vector pulse width modulation (SVPWM) algorithm, generates 6-way PWM drive signals, and drives Infineon FF450R17ME4 type IGBT power module after amplification by an isolation drive chip.

[0026] Technical effects: Through the combination of millisecond-level fast power control and multi-time scale optimization scheduling, the fluctuation rate of wind and light output can be effectively controlled within 7%, and the grid frequency deviation is strictly stabilized within the range of 50±0.2Hz, which greatly smooths the randomness and intermittency of renewable energy on the grid, guarantees the continuity and quality of power supply, and significantly improves the operation stability and power quality of the grid under high proportion of new energy access; at the same time, through the optimization model of the upper layer aiming at economic operation and maximum consumption, the prediction uncertainty, equipment constraints and network loss are comprehensively considered to generate the globally optimal scheduling plan. This makes the system maximize the consumption of wind power and photovoltaic power generation under the premise of meeting safety constraints, reduce abandoned wind and light, and reduce dependence on expensive peak shaving power, thereby significantly reducing the overall operation cost of the system, realizing economic optimization operation at the system level, and greatly improving the utilization rate of clean energy.

[0027] Optionally, a market strategy and collaborative control module 7 and a state monitoring and intelligent operation module 8 are further included; the market strategy and collaborative control module 7 is in communication connection with the optimization scheduling engine module 3 and the communication and network module 5 respectively, for receiving system real-time power balance state, energy storage state of charge and short-term power prediction information, generating bidding strategies for participating in power market transactions and adjustment instructions for controllable loads based on dynamic price signals and system operation constraints by solving a multi-time scale economic optimization model; the state monitoring and intelligent operation module 8 is in communication connection with each of the above modules, for collecting equipment operation characteristic data in real time and inputting the same into a pre-trained neural network fault diagnosis model for analysis, outputting equipment health state evaluation results and early fault warning; at the same time, the module integrates a visual human-computer interaction unit for dynamically displaying system topology, real-time energy flow distribution, key alarms and performance indicators.

[0028] It is worth mentioning that the market strategy and collaborative control module 7 is deployed on another independent virtual machine instance in the Dell PowerEdge R740xd server cluster, ensuring the isolation and efficient parallelism of computing resources. The module has a built-in market strategy core algorithm library, implemented using a hybrid programming of Python and Julia. The module obtains real-time system power balance status, energy storage SOC, and short-term power prediction information from the optimization scheduling engine module 3 through the OPC UA client of the communication and network module 5; and receives dynamic electricity price signals of day-ahead and intraday markets from the third-party power trading platform interface following the IEC 62351 security standard API. Its core computing unit calls the mixed integer linear programming MILP solver CPLEX based on linear programming and integer programming, aiming to maximize the total revenue of the system in the next 24 hours, considering the operation constraints, and rolling to generate the optimal market bidding price curve and controllable load grouping regulation timing instruction set. These instructions are issued to the user-side interruptible load management system and cross-regional energy routing equipment through the secure channel of the communication and network module 5; the state monitoring and intelligent operation and maintenance module 8 adopts a edge-cloud collaborative architecture, with its edge side integrated into the Moxa UC-8100A-ME-T gateway, using the built-in computing power of the gateway to collect and cache device state raw data at a frequency of 1 minute. Its cloud analysis part is deployed on a server with an NVIDIA Jetson AGX Orin embedded module for running high-computing-load diagnostic models. The module has a pre-trained one-dimensional convolutional neural network 1D-CNN model, which uses historical fault data for training and can take multi-dimensional time series data such as vibration, temperature, and current harmonics as input, and output classification results (such as "normal", "capacitor aging", "IGBT junction temperature anomaly") and fault confidence. The diagnostic results and health status SOH evaluation results are transmitted to the human-computer interaction unit through a data push service based on the WebSocket protocol. The human-computer interaction unit is developed using Siemens WinCC OA configuration software and runs on an independent industrial touch screen, providing system topology, real-time energy flow animation, health status radar chart, fault warning pop-up window, and historical data query interface for operation and maintenance personnel.

[0029] At the same time, the market strategy and collaborative control module 7 uploads the generated market bidding strategy to the trading platform and sends the corresponding internal dispatching plan instructions to the upper centralized optimization layer of the optimization scheduling engine module 3 as an important input of its economic objective function, forming a "market-dispatching" closed-loop optimization to maximize the overall system revenue; Technical effects: Through the market strategy and synergy control module 7, the external dynamic price signal is deeply integrated with the internal running state, and the optimal bidding and transaction strategy is automatically generated based on the rigorous mathematical optimization model MILP. This makes the system not just an energy consumer or producer, but a "virtual power plant" VPP that can actively participate in the market and flexibly adjust the operation strategy to maximize economic benefits. This significantly improves the project's return on investment, fully realizes the economic value of clean energy, maximizes the economic benefits of the system, and enhances the survival and competitiveness in the market environment.

[0030] Optionally, the data acquisition and processing module 1 comprises a data acquisition unit, a data processing unit, and a data quality evaluation and output unit. The data acquisition unit is connected with the current sensor, voltage sensor, power transmitter, weather station sensor and battery management system BMS sensor through the isolated signal conditioning circuit, and is used to synchronously acquire the multi-channel analog and digital signals of the photovoltaic array, wind turbine, energy storage unit and load at a sampling frequency not less than 1kHz, and to perform amplitude calibration and format unification on the acquired signals; the data processing unit is in communication connection with the data acquisition unit, and is internally provided with a multi-core embedded processor, and the processor is configured to: firstly perform moving average filtering or median filtering based on a fixed length sliding window on the acquired time series data stream to suppress impulse noise; then perform multi-resolution decomposition on the filtered data using a Daubechies wavelet basis function, process the detail coefficients by a soft threshold or hard threshold function to suppress Gaussian noise, and perform wavelet reconstruction; the data quality evaluation and output unit is connected with the data processing unit, and is used to calculate the signal-to-noise ratio and root mean square error of the denoised data, and to output the data that meets the quality evaluation to the energy prediction module and optimization scheduling engine module after time stamping.

[0031] The data acquisition unit adopts a high-performance synchronous acquisition card, fully meets the requirement of 1kHz sampling frequency, and ensures that the sampling synchronization error between multiple channels is less than 1 microsecond; the data processing unit uses an i.MX 8 series multi-core ARM processor (such as i.MX 8M Plus) of NXP as the core computing unit, which integrates a dedicated neural network processing unit (NPU) and an image processing unit GPU, and can be used to accelerate the filtering algorithm.

[0032] Technical effects: The two-stage progressive filtering scheme of "sliding window filtering (mean / median) + wavelet threshold denoising" effectively suppresses impulse noise and Gaussian white noise. This processing method greatly eliminates the measurement noise caused by the complex electromagnetic environment and the sensor itself, extracts the effective signal reflecting the real running state of the equipment, and makes the subsequent decision and execution of all modules based on "clean" data, significantly improving the robustness and reliability of the entire system in a noisy environment.

[0033] Optionally, the optimization scheduling engine module 3 comprises: a centralized optimization layer connected with the data acquisition and processing module 1 and the energy prediction module 2, for receiving real-time system state data, wind and light power point prediction values and their confidence intervals; the centralized optimization layer establishes an upper layer optimization model based on a model predictive control (MPC) framework, and the upper layer optimization model takes the minimization of system comprehensive operation cost and power fluctuation rate as objective functions, and its expression is: while meeting the following constraints: wherein, represents the exchange cost with the external power grid, represents the energy storage loss cost, represents the power balance penalty factor, , , and represent the grid interaction power, energy storage output, load power and new energy power at time t, respectively; the centralized optimization layer solves the optimization problem online by using an interior point method, and outputs the reference power plan of each energy storage unit in the future scheduling period; a distributed collaborative control layer deployed in the local controller of each energy storage unit and connected with the centralized optimization layer through a signal network, which receives the reference power plan issued by the upper layer and performs distributed collaborative calculation based on a consistency algorithm: each local controller exchanges state information with its communication neighbors, updates the local state through iteration, so that the processing instructions of all energy storage units converge to be consistent, and quickly compensate for real-time power differences. The distributed collaborative control layer finally generates accurate energy storage charging and discharging power set value instructions and issues them to the energy storage management module 4.

[0034] It is necessary to supplement that the centralized optimization layer is deployed on a high-performance server cluster in the central control room. The server is configured with two Intel Xeon Platinum 8360Y processors (64 cores in total), 512 GB of DDR4 memory, and installed with MATLAB, Python (including the Pyomo optimization modeling library), and Gurobi 9.5 commercial mathematical optimization solver software, providing sufficient computing power for handling large-scale, high-complexity nonlinear optimization problems; the model construction and solving process is as follows: (1) data input: through the communication network, real-time receive system state data such as load power and new energy power generation, energy storage SOC from the data acquisition and processing module, and wind and light power and their upper and lower confidence intervals from the energy prediction module; (2) optimization model establishment: at the beginning of each rolling optimization period (such as every 15 minutes), based on the model predictive control (MPC) framework, a random optimization model for the next 24 hours (96 15-minute periods) is established; (3) model solving: call the Interior-Point Method of the Gurobi solver to solve this large-scale nonlinear programming problem. After solving, output the reference power plan of each energy storage unit in the future scheduling period; (4) output: the reference power plan obtained by solving, as well as the corresponding planned purchase and sale of electricity power, is transmitted to each distributed collaborative control layer through high-speed optical fiber network.

[0035] The distributed collaborative control layer is deployed in the industrial-grade edge computing controller at each energy storage power station site, such as GQ3288 ARM controller or Intel Atom processor of Guangqian Technology, which has powerful local computing capability and various industrial communication interfaces.

[0036] Technical effects: The high-precision mathematical model established by the centralized optimization layer considers multiple factors such as purchase cost, equipment loss, and power balance, and can generate a scheduling plan with the highest global economic benefit and the friendliest to the grid (smallest fluctuation) under the premise of meeting all safety constraints, laying a foundation for economic operation and safety and stability of the system from the strategic level, and realizing the global optimization of economy and stability; the distributed collaborative control layer realizes the "plug and play" and autonomous collaboration of energy storage through a consistency algorithm. Without central instructions, local energy storage can respond to power shortages in a very short time according to neighbor status and local frequency measurement, achieving a primary frequency modulation function for wind and light fluctuations and load changes, compensating for the slow response of centralized optimization, forming a perfect complement of "slow optimization and fast control", and firmly controlling the grid frequency deviation within the standard range.

[0037] Optionally, the energy storage management module 4 comprises: A multi-objective power distribution unit connected with the optimization scheduling engine module 3, used to receive accurate energy storage charge and discharge power setting value instructions; based on the dynamic characteristics of different types of energy storage devices, this unit performs separation calculation of high-frequency power components and low-frequency power components, and according to the equivalent internal resistance, remaining cycle life and instantaneous power response capability of lithium ion batteries, flow batteries and super capacitors, adopts an adaptive weighting algorithm to dynamically calculate the optimal power distribution ratio of each energy storage unit; A multi-constraint charge and discharge control unit connected with the multi-objective power distribution unit and the data acquisition and processing module 1, used to receive the distributed power instructions and real-time state data of each energy storage unit; this unit is built-in with charge and discharge control logic, which is configured to: according to the received power instruction value and the real-time state of charge SOC, temperature and state of health SOH data uploaded by the battery management system BMS, adopt a three-stage charging curve or a constant power discharge strategy, and generate charge and discharge current limit value instructions and voltage protection threshold through table lookup or interpolation calculation; when it is monitored that the SOC of any energy storage unit exceeds the preset safety interval, the temperature exceeds the working limit value or the port voltage is abnormal, the hardware protection logic is triggered immediately to generate a power reduction or shutdown instruction; A pulse width modulation signal generation unit connected with the multi-constraint charge and discharge control unit, used to convert the stabilized charge and discharge current instructions into corresponding pulse width modulation PWM duty cycle signals; this unit integrates a carrier generator and a comparator circuit, which accurately controls the on and off time of the insulated gate bipolar transistor IGBT or metal-oxide semiconductor field effect transistor MOSFET in the energy storage converter PCS by changing the duty cycle of the PWM wave, thereby driving the energy storage converter to execute the power instructions and realize physical control of the battery pack charge and discharge process.

[0038] Technical effects: Through the multi-objective power distribution unit, based on the inherent characteristics of batteries and capacitors (high energy density of batteries, high power density of capacitors), the frequency separation technology (such as high-pass / low-pass filter) is used to distribute the high-frequency pulsating component in the total power instruction to the super capacitor response, and the low-frequency and steady-state component to the lithium battery response. This "strength and weakness" distribution strategy fully utilizes the advantages of various types of energy storage, not only meets the demand of the system for fast power response, but also avoids the accelerated aging of lithium batteries due to frequent response to high-frequency fluctuations, thereby significantly prolonging the overall cycle life of the battery and reducing the degree of electric cost throughout the life cycle.

[0039] Optionally, the control execution module 6 comprises: An instruction analysis and signal conversion unit is connected with the optimization scheduling engine module 3 through a high-speed data bus, and is configured to receive accurate energy storage charge and discharge power setting value instructions; a digital signal processor (DSP) is built in the unit, and is configured to convert the power instruction value into corresponding current loop reference instructions or voltage loop reference instructions through a preset lookup table or real-time calculation according to the DC bus voltage level and the current switching frequency of the energy storage converter (PCS); A pulse width modulation wave generation unit is connected with the instruction analysis and signal conversion unit, and is configured to convert the current or voltage reference instructions into pulse width modulation (PWM) waves with corresponding duty cycles; the unit integrates a triangular carrier generator and a digital comparator, and generates multiple complementary PWM drive signals with dead time through a space vector pulse width modulation (SVPWM) or a sine pulse width modulation (SPWM) algorithm; A drive signal amplification and isolation unit is connected with the pulse width modulation wave generation unit, and is configured to perform power amplification and electrical isolation on the generated PWM logic signals; the unit includes a gate drive integrated circuit and an isolation optocoupler or a magnetic coupling element, amplifies the voltage and current of the generated drive signals to a level sufficient to drive the power switching devices, and provides several kilovolts of electrical isolation, and finally outputs to the power unit of the energy storage converter (PCS); A power unit execution and state feedback unit is directly connected with the drive signal amplification and isolation unit and an insulated gate bipolar transistor (IGBT) or a silicon carbide MOSFET power module in the energy storage converter (PCS); the unit controls the conduction and turn-off of the power switching devices according to the amplified drive signals, converts the direct current power into alternating current power and connects to the grid or performs charging operation; at the same time, the unit samples the voltage and current on the output side of the converter in real time through a high-speed analog-to-digital converter (ADC), forms a feedback signal required for closed-loop control, and sends the signal back to the instruction analysis and signal conversion unit, to form a closed-loop control loop, realize accurate and rapid tracking control of the output power, and complete millisecond-level smoothing and anti-peaking compensation for new energy power fluctuations.

[0040] Optionally, the market strategy and collaborative control module 7 includes: A multi-time scale collaborative optimization unit is connected with the optimization scheduling engine module 3 and the data acquisition and processing module 1 through a high-speed data bus, and is configured to receive real-time power balance states, energy storage state of charge (SOC), and ultra-short-term power prediction information of the system; a high-performance computing core is built in the unit, and is configured to establish a multi-time scale economic optimization model with the minimum total operation cost and the maximum renewable energy consumption as the targets based on dynamic price signals, system operation constraints, and network transmission losses; the model uses a stochastic optimization or robust optimization method to process the uncertainty of wind and light output and market electricity prices, and calls an interior point method or a mixed integer programming solver for online solution, to rollingly generate a global optimal economic dispatch plan in a time scale of 24 hours to 15 minutes in the future. a controllable load regulation instruction generation unit connected with the multi-time scale collaborative optimization unit; the unit receives an economic dispatch plan, and calculates an optimal load regulation strategy according to a user preset interruptible load priority and comfort constraint through a model prediction control (MPC) framework; the strategy is converted into specific direct load control (DLC) instructions or price-based demand response (PR-DR) signals, and is issued to each intelligent terminal device through a communication and network module; a cross-regional collaborative control interface unit connected with the multi-time scale collaborative optimization unit and the communication and network module; the unit is used for real-time data exchange with adjacent regional energy management systems; the interface calculates optimal power reference values of inter-regional power exchange based on a consistency algorithm or a distributed optimization method, and performs secure encryption transmission of interactive data through a hardware encryption module; a real-time power balance correction unit connected with the controllable load regulation instruction generation unit and the cross-regional collaborative control interface unit; the unit collects power grid frequency deviation and voltage phase information through a high-precision phase measurement unit (PMU), and when power imbalance or frequency overrun is detected, a real-time power correction algorithm based on sensitivity analysis is immediately started to generate fast regulation instructions of the energy storage unit and the controllable load, and the instructions are sent to the optimization dispatch engine module for execution through the communication and network module.

[0041] Optionally, an energy coordination control and dispatch method of a multi-energy complementary new energy energy storage system includes the following steps: S1, data acquisition and preprocessing, photovoltaic irradiance, environmental temperature, wind speed, state of charge (SOC) of the energy storage unit, load power and grid dispatch instruction signals are collected through a sensor network, and the collected raw data is preprocessed by combining sliding window filtering and wavelet threshold denoising; S2, uncertainty and risk quantification, based on the wind and light output point prediction values and their confidence intervals output by the energy prediction module 2, the uncertainty range of new energy output in different time scales in the future is quantified; for extreme weather warning signals, the potential impact risk on system power balance is evaluated, and probabilistic inputs representing prediction uncertainty and extreme risk are provided for subsequent optimization dispatch; S3, optimization dispatch decision, taking the minimum net load variance and the maximum clean energy consumption as upper layer targets, and taking the minimum conventional power regulation cost as a lower layer target, a multi-objective optimization algorithm and a robust optimization method are used to develop a global optimization dispatch plan in combination with system operation constraints and prediction uncertainty, and the output of each unit is determined; S4, energy storage collaborative control, according to the scheduling plan, the battery and super capacitor hybrid energy storage system executes the mixed control of active balancing and passive balancing; based on the real-time internal short circuit detection algorithm and the insulation resistance online monitoring data, the hierarchical safety protection and redundancy control logic is started; and according to the power demand of different time scales and the system operation cost, the charging and discharging power of each energy storage unit is dynamically allocated, and the fast response and accurate execution of the power instruction are completed; S5, communication and instruction transmission, by means of the communication and network module 5, the communication mode compatible with Modbus, BACnet, CAN and at least 20 kinds of industrial protocols is used, 5G+ optical fiber hybrid networking is adopted, SM4 algorithm encryption is used, and the optimized scheduling instruction is safely transmitted from the optimized scheduling engine module to the control execution module of each energy equipment; S6, control execution and feedback, the control execution module 6 executes power smoothing control according to the received instruction, controls the wind and light output within 7% with 15 minute fluctuation rate, maintains the grid frequency stable at 50±0.2Hz through millisecond level wide area coordinated control, and at the same time, the control execution result is monitored and fed back to the data acquisition and processing module, providing data support for subsequent optimization scheduling; S7, adjustable load control and regional collaboration, according to the system real-time power shortage and prediction information, the load adjustment instruction is generated, and the interruptible load control unit is directly acted on through hardwiring or special communication channel; the energy routing equipment adjusts the working point of its power conversion unit according to the power interaction instruction between regions, realizes the power exchange and dynamic support with adjacent regions; S8, state monitoring and closed-loop regulation, real-time acquisition of energy storage SOC, equipment temperature, voltage harmonic distortion rate and frequency deviation data, input into the pre-trained neural network model to calculate the battery health state SOH and generate fault warning signals; according to the diagnosis results and warning levels, the corresponding device protection logic or system operation parameter adjustment instruction is triggered to complete the closed-loop regulation.

[0042] Technical effects: through accurate capacity prediction and optimization scheduling, the wind and light curtailment rate is reduced from 15% to less than 5%, the market transaction and strategy adjustment greatly improve the energy storage peak valley arbitrage income, thereby improving the energy utilization efficiency, the data preprocessing improves the signal-to-noise ratio of the sampling data, thereby improving the data quality and communication security, combined with the response of the market transaction and strategy module to the dynamic price and auxiliary service market, the annual operation income of the system can be improved, the economic value of clean energy is fully released, and the operation efficiency and economy of the multi-energy complementary system are significantly improved.

[0043] Optionally, the mixed control of active balancing and passive balancing of the battery and super capacitor hybrid energy storage system in S4 is specifically: The active equalization circuit based on the switch matrix is used for energy transfer between battery clusters, and the passive equalization resistor is used to absorb the small differences between the single batteries in the battery pack; the dynamic allocation of the charge and discharge power of each energy storage unit is based on the instantaneous high-power response characteristics of the super capacitor and the high energy density characteristics of the lithium battery, and the high-frequency component in the power command is allocated to the super capacitor, and the low-frequency component and the steady-state power support task is allocated to the lithium battery, and through the hybrid mode of "active equalization (inter-cluster) + passive equalization (intra-cluster)", the energy management is refined. The energy transfer efficiency of the active equalization circuit based on the switch matrix is usually > 85% to efficiently transfer energy from the battery cluster with high SOC to the battery cluster with low SOC, which fundamentally solves the "barrel effect" caused by the difference between the parallel clusters, and avoids overcharging or overdischarging of some battery clusters. At the same time, the passive equalization resistor can eliminate the small voltage differences between the single batteries in the battery pack at a low cost, and this combination strategy greatly improves the working consistency of the battery pack and slows down the capacity decay rate, thereby significantly prolonging the overall service life of the battery pack.

[0044] Optionally, the calculation of the battery state of health SOH in S8 is specifically as follows: The real-time collected battery charge and discharge voltage, current curve, temperature and historical cycle number data are input as input features into the pre-trained one-dimensional convolutional neural network model, the model outputs the current health state SOH quantitative value and the remaining useful life RUL estimation value of the battery; triggering the corresponding device protection logic includes: when the SOH value is lower than the preset threshold or the RUL estimation value reaches the warning limit value, automatically generating a derating operation instruction or issuing a maintenance warning signal.

[0045] Technical effects: By outputting the estimation value of the remaining useful life RUL, the system can predict the failure time of the battery in advance. This makes the operation and maintenance personnel change from the traditional and fixed "replacement after a certain time or cycle number" mode to the "condition-based maintenance" mode based on the actual health state of the equipment. The maintenance window and budget can be accurately planned in advance before the battery performance drops to the critical point, spare parts can be ordered in advance, and the system unplanned downtime and huge power outage economic losses caused by sudden battery failure can be avoided, which greatly improves the forward-looking and economy of operation and maintenance.

[0046] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it, and other modifications or equivalent replacements of the technical solutions of the present application made by those skilled in the art should be covered in the scope of the claims of the present application as long as they do not deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. An energy coordination and control system for multi-energy complementary new energy storage, characterized in that: include: The data acquisition and processing module is connected to current and voltage sensors, power transmitters, weather station sensors, and battery management system sensors. It collects voltage, current, power, state of charge, and environmental data of photovoltaic, wind power, energy storage, and load through the sensors, and performs sliding window filtering and wavelet denoising processing through the embedded processor. An energy prediction module is communicatively connected to the data acquisition and processing module and is used to receive the filtered and denoised physical state data signal. Through a hybrid prediction model that includes a long short-term memory network and physical mechanisms, it outputs the point prediction value of wind and solar power output, confidence interval, and extreme weather warning. An optimized scheduling engine module is established, which is communicatively connected to the data acquisition and processing module and the energy prediction module, respectively, to create a hierarchical distributed optimization architecture. This architecture includes an upper centralized optimization layer and a lower distributed collaborative control layer. The centralized optimization layer, based on the prediction results and prediction confidence intervals received from the energy prediction module, and the actual system state received from the data acquisition and processing module, uses system operating economy and stability as objective functions, comprehensively considers prediction uncertainty and equipment operating constraints, and employs a rolling time-domain optimization method to solve for the globally optimal energy storage unit reference power plan within the future scheduling cycle. The distributed collaborative control layer, deployed in each local controller, receives the reference power plan and, based on a consensus algorithm or a distributed quadratic programming algorithm, combines locally measured real-time bus frequency deviation or voltage deviation data to quickly calculate and generate accurate energy storage charging and discharging power setpoint instructions. An energy storage management module is communicatively connected to the data acquisition and processing module and the optimization scheduling engine module, and is configured to perform charging and discharging coordinated control and status management of different types of energy storage devices according to the power setpoint command; The communication and network module is connected to the data acquisition and processing module, energy prediction module, optimization scheduling engine module, and energy storage management module, respectively, and is used to provide heterogeneous protocol conversion, data encryption and edge computing services to realize secure data interaction between the modules. The control execution module is communicatively connected to the optimization scheduling engine module and the communication and network module, and is configured to drive the power converter to perform operations according to the power setpoint command, so as to realize smooth control and anti-peak compensation for the power fluctuation of new energy sources.

2. The energy coordination control system for multi-energy complementary new energy storage as described in claim 1, characterized in that: It also includes a market strategy and collaborative control module and a status monitoring and intelligent operation and maintenance module. The market strategy and collaborative control module is communicatively connected to the optimization scheduling engine module and the communication and network module, respectively. It is used to receive real-time power balance status, energy storage charge status and short-term power forecast information of the system. Based on dynamic electricity price signals and system operation constraints, it generates bidding strategies for participating in electricity market transactions and adjustment instructions for controllable loads by solving multi-time-scale economic optimization models. The status monitoring and intelligent operation and maintenance module is communicatively connected to the above modules. It is used to collect equipment operation characteristic data in real time and input it into a pre-trained neural network fault diagnosis model for analysis, outputting equipment health status assessment results and early fault warnings. At the same time, this module integrates a visualization human-computer interaction unit to dynamically display system topology, real-time energy flow distribution, key alarms and performance indicators.

3. The energy coordination control system for multi-energy complementary new energy storage as described in claim 1, characterized in that: The data acquisition and processing module includes a data acquisition unit, a data processing unit, and a data quality assessment and output unit. The data acquisition unit is connected to the current sensor, voltage sensor, power transmitter, weather station sensor, and battery management system (BMS) sensor via an isolated signal conditioning circuit. It is used to synchronously acquire multiple analog and digital signals from the photovoltaic array, wind turbine, energy storage unit, and load at a sampling frequency of not less than 1 kHz, and to perform amplitude calibration and format standardization on the acquired signals. The data processing unit is communicatively connected to the data acquisition unit and has a built-in multi-core embedded processor. The processor is configured to first perform moving average filtering or median filtering based on a fixed-length sliding window on the acquired time-series data stream to suppress impulse noise. Subsequently, the Daubechies wavelet basis function is used to perform multi-resolution decomposition on the filtered data. Detail coefficients are processed using soft or hard thresholding functions to suppress Gaussian noise, and wavelet reconstruction is performed. The data quality assessment and output unit, connected to the data processing unit, is used to calculate the signal-to-noise ratio and root mean square error of the denoised data. After timestamping the data that has passed the quality assessment, it is output to the energy prediction module and the optimization scheduling engine module through a standard industrial bus interface.

4. The energy coordination control system for multi-energy complementary new energy storage as described in claim 1, characterized in that: The optimized scheduling engine module includes: A centralized optimization layer, connected to the data acquisition and processing module and the energy prediction module, is used to receive real-time system status data, predicted wind and solar power points, and their confidence intervals. The centralized optimization layer establishes an upper-level optimization model based on the Model Predictive Control (MPC) framework. This upper-level optimization model takes minimizing the overall system operating cost and minimizing power fluctuation rate as its objective functions, and its expression is: The following constraints must be satisfied simultaneously: in, This represents the cost of exchanging data with an external power grid. Indicates the cost of energy storage losses. Represented as the power balance penalty factor, and These are respectively represented as the grid interaction power, energy storage output, load power, and new energy power generation power during time period t; the centralized optimization layer uses the interior point method to solve the optimization problem online and outputs the reference power plan for each energy storage unit in future scheduling cycles; The distributed collaborative control layer is deployed in the local controllers of each energy storage unit and is connected to the centralized optimization layer via a communication network. The distributed collaborative control layer receives the reference power plan issued by the centralized optimization layer and performs distributed collaborative calculation based on a consensus algorithm: each local controller exchanges status information with its communication neighbors, iteratively updates the local status, so that the processing instructions of all energy storage units converge to consistency, and quickly compensates for real-time power differences. The distributed collaborative control layer finally generates accurate energy storage charging and discharging power setpoint instructions and issues them to the energy storage management module.

5. The energy coordination control system for multi-energy complementary new energy storage as described in claim 1, characterized in that: The energy storage management module includes: A multi-objective power allocation unit, connected to the optimization scheduling engine module, is used to receive the precise energy storage charging and discharging power setting command. Based on the dynamic characteristics of different types of energy storage devices, this unit performs separate calculations of high-frequency power components and low-frequency power components, and dynamically calculates the optimal power allocation ratio of each energy storage unit using an adaptive weighting algorithm based on the equivalent internal resistance, remaining cycle life, and instantaneous power response capability of lithium-ion batteries, flow batteries, and supercapacitors. A multi-constraint charge / discharge control unit, connected to the multi-target power allocation unit and data acquisition and processing module, is used to receive the allocated power command and real-time status data of each energy storage unit. This unit has built-in charge / discharge control logic, which is configured to: based on the received power command value and the real-time state of charge (SOC), temperature, and state of health (SOH) data uploaded by the battery management system (BMS), adopt a three-stage charging curve or constant power discharge strategy, and generate charge / discharge current limit commands and voltage protection thresholds in real time through lookup table method or interpolation calculation; when it is detected that the SOC of any energy storage unit exceeds the preset safety range, the temperature exceeds the operating limit, or the port voltage is abnormal, the hardware protection logic is immediately triggered to generate a power reduction or shutdown command. A pulse width modulation (PWM) signal generation unit, connected to the multi-constraint charge / discharge control unit, is used to convert the stabilized charge / discharge current command into a corresponding PWM duty cycle signal. This unit integrates a carrier generator and a comparator circuit, and precisely controls the on and off times of the insulated gate bipolar transistor (IGBT) or metal-oxide-semiconductor field-effect transistor (MOSFET) in the energy storage converter PCS by changing the duty cycle of the PWM wave, thereby driving the energy storage converter to execute power commands and realizing physical control of the battery pack charging and discharging process.

6. The energy coordination control system for multi-energy complementary new energy storage as described in claim 1, characterized in that: The control execution module includes: The instruction parsing and signal conversion unit is connected to the optimized scheduling engine module via a high-speed data bus and is used to receive the precise energy storage charging and discharging power setting value instruction. The unit has a built-in digital signal processor (DSP) and is configured to convert the power instruction value into the corresponding current loop reference instruction or voltage loop reference instruction based on the DC bus voltage level of the energy storage converter PCS and the current switching frequency through a preset lookup table or real-time calculation. The pulse width modulation wave generation unit, which is connected to the instruction parsing and signal conversion unit, is used to convert current or voltage reference instructions into pulse width modulation (PWM) waves with corresponding duty cycles. This unit integrates a triangular carrier generator and a digital comparator, and generates multiple complementary PWM drive signals with dead time through space vector pulse width modulation (SVPWM) or sinusoidal pulse width modulation (SPWM) algorithms. A drive signal amplification and isolation unit, which is connected to the pulse width modulation wave generation unit, is used to amplify and electrically isolate the generated PWM logic signal. The unit includes a gate driver integrated circuit and an isolation optocoupler or magnetic coupling element, which amplifies the voltage and current of the generated drive signal to a level sufficient to drive the power switching device and provides electrical isolation of several kilovolts, and finally outputs it to the power unit of the energy storage converter PCS. The power unit execution and status feedback unit is directly connected to the drive signal amplification and isolation unit and the insulated gate bipolar transistor (IGBT) or silicon carbide MOSFET power module in the energy storage converter PCS. This unit controls the on and off of the power switching devices according to the amplified drive signal, converting DC power into AC power for grid connection or performing charging operations. At the same time, this unit samples the voltage and current on the output side of the converter in real time through a high-speed analog-to-digital converter (ADC) to form the feedback signal required for closed-loop control, and sends the signal back to the instruction parsing and signal conversion unit to form a closed-loop control loop, realizing precise and rapid tracking control of the output power, and completing millisecond-level smoothing and anti-peak compensation for new energy power fluctuations.

7. The energy coordination control system for multi-energy complementary new energy storage as described in claim 2, characterized in that: The market strategy and collaborative control module includes: The multi-timescale collaborative optimization unit is connected to the optimization scheduling engine module and data acquisition and processing module via a high-speed data bus. It is used to receive real-time power balance status, energy storage state of charge (SOC), and ultra-short-term power forecast information. This unit has a built-in high-performance computing core and is configured to establish a multi-timescale economic optimization model based on dynamic electricity price signals, system operation constraints, and network transmission losses, with the objectives of minimizing total operating costs and maximizing renewable energy consumption. The model uses stochastic optimization or robust optimization methods to handle the uncertainties of wind and solar power output and market electricity prices, and calls interior-point methods or mixed-integer programming solvers for online solving, continuously generating globally optimal economic scheduling plans for the next 24 hours to 15 minutes. A controllable load adjustment instruction generation unit is connected to the multi-time-scale collaborative optimization unit. This unit receives the economic scheduling plan and calculates the optimal load adjustment strategy through the model predictive control (MPC) framework based on the user-preset interruptible load priority and comfort constraints. The strategy is converted into specific direct load control (DLC) instructions or price-based demand response (PR-DR) signals and sent to each smart terminal device through the communication and network module. A cross-regional collaborative control interface unit is connected to the multi-timescale collaborative optimization unit and the communication and network module; it is used to exchange data with the energy management system of adjacent regions in real time; the interface is based on a consensus algorithm or a distributed optimization method to calculate the optimal power reference value for power exchange between regions, and uses a hardware encryption module to securely encrypt and transmit the interactive data. The real-time power balance correction unit is connected to the controllable load adjustment command generation unit and the cross-regional collaborative control interface unit. This unit collects grid frequency deviation and voltage phase information through a high-precision phase measurement unit (PMU). When a power imbalance or frequency over-limit is detected, it immediately starts a real-time power correction algorithm based on sensitivity analysis to generate fast adjustment commands for energy storage units and controllable loads, and sends them to the optimization scheduling engine module for execution through the communication and network module.

8. A method for coordinated control and scheduling of energy from multi-energy complementary new energy storage, characterized in that: Includes the following steps: S1. Data Acquisition and Preprocessing: Photovoltaic irradiance, ambient temperature, wind speed, state of charge (SOC) of energy storage units, load power, and grid dispatch command signals are collected through a sensor network. The collected raw data is preprocessed using a combination of sliding window filtering and wavelet threshold denoising. S2. Uncertainty and Risk Quantification: Based on the predicted values ​​of wind and solar power output points and their confidence intervals output by the energy prediction module, the uncertainty range of new energy power output at different time scales in the future is quantified; for extreme weather warning signals, the potential impact risk on the system power balance is assessed, and probabilistic inputs characterizing the prediction uncertainty and extreme risks are provided for subsequent optimized scheduling. S3. Optimize scheduling decisions, with the upper-level objectives of minimizing net load variance and maximizing clean energy consumption, and the lower-level objective of minimizing conventional power supply regulation costs. Employ multi-objective optimization algorithms and robust optimization methods, combined with system operation constraints and prediction uncertainties, to formulate a global optimization scheduling plan and determine the output of each unit. S4. Energy storage collaborative control: According to the scheduling plan, the battery and supercapacitor hybrid energy storage system is subjected to hybrid control that combines active and passive balancing; based on the real-time internal short-circuit detection algorithm and online insulation resistance monitoring data, the hierarchical safety protection and redundant control logic are activated; and the charging and discharging power of each energy storage unit is dynamically allocated according to the power demand and system operating cost at different time scales to complete the rapid response and precise execution of power commands. S5. Communication and command transmission: With the help of the communication and network module, the system operates a communication mode compatible with at least 20 industrial protocols such as Modbus, BACnet, and CAN. Through 5G+fiber hybrid networking and encryption using the national cryptographic SM4 algorithm, the optimized scheduling commands are securely transmitted from the optimized scheduling engine module to the control execution module of each energy device. S6. Control Execution and Feedback: The control execution module performs power smoothing control according to the received instructions, keeping the wind and solar power output within 7% fluctuation rate over 15 minutes. It maintains the grid frequency stable at 50±0.2Hz through millisecond-level wide-area coordinated control. At the same time, it monitors the control execution results in real time and feeds them back to the data acquisition and processing module to provide data support for subsequent optimized scheduling. S7. Adjustable load control and regional coordination: Based on the real-time power deficit and forecast information of the system, load adjustment commands are generated and directly applied to the interruptible load control unit through hard wiring or dedicated communication channels; the energy routing equipment adjusts the operating point of its power conversion unit according to the power exchange commands between regions to realize power exchange and dynamic support with adjacent regions. S8, Status Monitoring and Closed-Loop Control: Real-time acquisition of energy storage SOC, equipment temperature, voltage harmonic distortion rate and frequency deviation data, input into a pre-trained neural network model to calculate battery health status SOH and generate fault warning signals; based on the diagnostic results and warning level, triggering corresponding equipment protection logic or system operating parameter adjustment commands to complete closed-loop control.

9. The energy coordination control and scheduling method for multi-energy complementary new energy storage as described in claim 8, characterized in that: The specific implementation of hybrid control combining active and passive balancing for the battery and supercapacitor hybrid energy storage system in S4 is as follows: An active balancing circuit based on a switch matrix is ​​used to transfer energy between battery clusters, while a passive balancing resistor is used to eliminate the slight differences between individual cells within the battery pack. The dynamic allocation of charging and discharging power of each energy storage unit is based on the instantaneous high-power response characteristics of supercapacitors and the high energy density characteristics of lithium batteries. The high-frequency components in the power command are allocated to the supercapacitors, while the low-frequency components and steady-state power support are allocated to the lithium batteries.

10. The energy coordination control and scheduling method for multi-energy complementary new energy storage as described in claim 8, characterized in that: The calculation of the battery health state (SOH) in S8, which is input into the pre-trained neural network model, specifically involves: The real-time collected battery charging and discharging voltage, current curves, temperature, and historical cycle count data are used as input features and input into a pre-trained one-dimensional convolutional neural network model. The model outputs the current state of health (SOH) quantification value and the estimated remaining service life (RUL) value of the battery. The triggering of the corresponding equipment protection logic includes: when the SOH value is lower than a preset threshold or the RUL estimate reaches a warning limit, automatically generating a derating operation command or issuing a maintenance alarm signal.

Images