CN-121984052-A - Multi-mode coupling industrial park energy management and distribution method and system

CN121984052ACN 121984052 ACN121984052 ACN 121984052ACN-121984052-A

Abstract

The invention discloses a management and distribution method and a system for multi-mode coupling industrial park energy, which belong to the technical field of operation and control of electric power systems, and comprise the steps of collecting operation state data of all elements of charge storage and charge release of an endogenous network of an industrial park, identifying equipment types and real-time states by utilizing a non-intrusive load identification algorithm, the method comprises the steps of inputting long-term and short-term memory network prediction photovoltaic power and total load, outputting maximum interruptible load capacity through fuzzy reasoning, calculating a power mutual-aid reference value by taking direct-current bus voltage as constraint, solving a collaborative optimization model based on a quadratic programming algorithm, and outputting optimal control vector to drive adjustable equipment and flexible interconnection converters. According to the invention, PID control parameters are adaptively adjusted according to the equipment type, so that the response lag of industrial load is compensated, the power mutual compensation is performed by utilizing the low-voltage flexible direct current interconnection system, the power unbalance degree of a transformer area is eliminated, the voltage deviation is restrained, and the collaborative optimization and the accurate distribution of all elements of the charge storage and the charge storage of the source network of the industrial park are realized.

Inventors

LI XIAOQIANG
GAO WENLONG
DOU AIHUA
MIAO XIN
LIU WEI
YANG BAOHUA
HOU ZHIXUE
YING ZHAOJIE
TAN ZHENG
WANG WENZHE
LI DEZHONG
LI ZIKAI
CAO SHOUGUO

Assignees

国网山东省电力公司临沂供电公司
国网山东省电力公司费县供电公司

Dates

Publication Date: 20260505
Application Date: 20260124

Claims (10)

1. The management and distribution method of the multi-mode coupling industrial park energy is characterized by comprising the following steps of: acquiring and preprocessing running state data of all elements of the charging, storing, charging and discharging of the endogenous network in the industrial park, acquiring voltage deviation, power unbalance degree of the platform area and load adjustment potential, and identifying equipment types and real-time states of adjustable equipment by adopting a non-intrusive load identification algorithm, wherein the running state data at least comprises direct current bus voltage of a low-voltage flexible direct current interconnection system and tide data of a flexible interconnection converter; inputting the preprocessed running state data into a source load trend prediction model based on a long-term and short-term memory network, and predicting photovoltaic power and total load; according to the equipment type, the load regulation potential and the real-time state, outputting the maximum interruptible load capacity at the current moment through fuzzy reasoning; taking the power unbalance degree of the transformer area and the voltage of the direct current bus as constraints, constructing a transformer area mutual-aid evaluation model based on a self-adaptive dynamic allocation strategy, and calculating a power mutual-aid reference value; And constructing an objective function comprising voltage deviation, full-element operation cost and power mutual-aid reference value deviation items, taking predicted photovoltaic power and total load as input, taking maximum interruptible load capacity as constraint, solving a collaborative optimization model based on a quadratic programming algorithm, outputting an optimal control vector, and driving adjustable equipment and a flexible interconnection converter.
2. The method for managing and distributing energy to a multi-mode coupled industrial park of claim 1, wherein driving the adjustable devices and the flexible interconnect inverter comprises: performing physical boundary out-of-limit verification on the optimal control vector, and issuing the optimal control vector to a bottom execution unit after the optimal control vector passes the verification; the driving field energy controller tracks and executes an optimal control vector by adopting a multi-target optimization PID control algorithm, wherein the PID control algorithm adaptively adjusts a proportional coefficient, an integral coefficient and a differential coefficient according to the identified equipment type so as to compensate response lag of different types of industrial loads; the bottom layer executing unit executes actions including controlling the on-off of a wire outlet switch, adjusting the opening and closing of an intelligent circuit breaker, issuing a load guiding interaction signal, executing frequency modulation on the frequency conversion equipment and controlling the flexible interconnection converter; Generating a photovoltaic absorption analysis report, a load characteristic analysis report and a power consumption optimization suggestion by combining the predicted photovoltaic power and the real-time state of the identified adjustable equipment; And driving the digital twin interaction interface, rendering and mapping a virtual scene of a physical entity of the industrial park based on the preprocessed running state data, and dynamically displaying the predicted trend of the voltage fluctuation of the direct current bus, the photovoltaic power and the total load and the tide data of the flexible interconnection converter.
3. The method for managing and distributing energy to a multi-mode coupled industrial park according to claim 1, wherein the process of collecting and preprocessing the operation state data of all the elements of the network charge storage and release in the industrial park comprises: A multifunctional intelligent ammeter and an electric energy quality monitoring terminal are deployed at the low-voltage bus side of each power supply station area of an industrial park, non-intrusive load sensing equipment is connected in series at the main inlet line of the industrial park and the inlet line of each production workshop, a data acquisition gateway is deployed at the output end of a distributed photovoltaic inverter, a communication port of a battery management system of an energy storage device and a port of a charging pile controller of an electric automobile, a low-voltage flexible direct current interconnection system is deployed between alternating current low-voltage buses of adjacent power supply stations, the low-voltage flexible direct current interconnection system consists of flexible interconnection converters connected with all alternating current low-voltage buses and direct current buses connected with the direct current sides of all flexible interconnection converters, and a voltage transformer and a current transformer are respectively arranged at the alternating current sides and the direct current sides of the flexible interconnection converters, and all the data acquisition terminals are connected to a field energy controller through optical fibers and an industrial Ethernet; synchronously acquiring analog quantity and state quantity of all elements of the charge storage, charge discharge of the endogenous network of the industrial park by using a multifunctional intelligent ammeter, non-intrusive load sensing equipment and a data acquisition gateway at a preset sampling frequency; The analog quantity comprises three-phase voltage, three-phase current, three-phase active power, three-phase reactive power, high-frequency voltage and current data acquired by non-intervening load sensing equipment of each power supply station area, direct-current bus voltage of a low-voltage flexible direct-current interconnection system and power flow data of a flexible interconnection converter, wherein the power flow data comprise direct-current side current and alternating-current side power of the flexible interconnection converter; The state quantity comprises a switch position signal, the charge state of the energy storage device and the connection state of the electric automobile; The field energy controller performs outlier rejection and time sequence alignment preprocessing on the acquired analog quantity and state quantity of the full-element charge storage and charge discharge of the industrial park endogenous network, and obtains preprocessed running state data.
4. The method of claim 3, wherein identifying the device type and real-time status of the adjustable device comprises: according to the three-phase voltage and the three-phase active power in the preprocessed running state data, calculating the voltage deviation and the power unbalance of the station area at the current moment; dividing the difference value between the three-phase voltage acquired at the current moment and the rated voltage value of the system by the rated voltage value of the system to obtain voltage deviation; Dividing the difference value between the maximum value and the minimum value in the three-phase active power by the average value of the three-phase active power to obtain the power unbalance of the station area; invoking a non-intrusive load identification algorithm to process high-frequency voltage and current data uploaded by non-intrusive load sensing equipment, constructing a voltage-current track image, and extracting geometric characteristic parameters of the voltage-current track image as load characteristic vectors; inputting the load characteristic vector into a pre-trained load characteristic library for matching, and identifying the equipment type and the real-time state of electric equipment accessed to a power grid at the current moment, wherein the equipment type comprises adjustable equipment and non-adjustable equipment, the adjustable equipment comprises inductive load equipment and variable frequency load equipment, and the real-time state comprises start-stop state and running power; And calculating the load adjustment potential according to the identified equipment type and the real-time state and the rated power of the adjustable equipment.
5. The method of managing and distributing energy to a multi-mode coupled industrial park of claim 1, wherein predicting photovoltaic power and total load comprises: Extracting historical photovoltaic power and historical total load from the preprocessed running state data, executing maximum and minimum normalization processing to obtain a photovoltaic power normalization value and a total load normalization value, intercepting the photovoltaic power normalization value and the total load normalization value by utilizing a sliding time window, and constructing an input feature matrix, wherein the input feature matrix is composed of input feature vectors arranged in time sequence, and each input feature vector comprises the photovoltaic power normalization value and the total load normalization value at corresponding moments; Inputting an input feature matrix into a source charge trend prediction model based on a long-term and short-term memory network, wherein the source charge trend prediction model sequentially processes each input feature vector contained in the input feature matrix according to time steps by using a forgetting gate, an input gate and an output gate, and sets a hidden state vector at the last moment and a cell state vector at the last moment to adopt a preset initial zero vector when processing the input feature vector at the first time step; The forgetting gate calculates the value of the forgetting gate by performing linear weighting and S-shaped function activation operation on the hidden state vector at the previous moment and the input feature vector at the current moment; The input gate calculates the value of the input gate by performing linear weighting and S-shaped function activation operation on the hidden state vector at the previous moment and the input feature vector at the current moment; The source charge trend prediction model generates candidate cell states by performing linear weighting and hyperbolic tangent function activation operation on the hidden state vector at the previous moment and the input feature vector at the current moment; Performing element-wise multiplication on the cell state vector and the forgetting gate value at the previous moment, performing element-wise multiplication on the candidate cell state and the input gate value, adding the results of the two multiplication operations to obtain a cell state vector at the current moment, wherein the cell state vector at the current moment comprises time sequence characteristic information of the screened historical photovoltaic power and the historical total load; The source load trend prediction model calculates an output gate value by performing linear weighting and S-shaped function activation operation on the hidden state vector at the previous moment and the input feature vector at the current moment; Performing hyperbolic tangent function activation operation on the cell state vector at the current moment, and performing element-wise multiplication operation on an operation result and an output gate value to obtain a hidden state vector at the current moment; inputting the hidden state vector at the current moment into a fully-connected output layer of a source charge trend prediction model; Performing linear transformation on the hidden state vector by using the weight matrix and the bias vector of the fully-connected output layer to obtain a normalized prediction result; And executing inverse normalization processing on the normalization prediction result, and outputting the photovoltaic power and the total load in a preset time period in the future.
6. The method for managing and distributing energy to multi-mode coupled industrial parks according to claim 4, wherein the process of outputting the maximum interruptible load capacity at the current time by fuzzy inference comprises: screening electric equipment in an operating state according to the identified equipment type and the real-time state; extracting the running power of each electric equipment from the real-time state, and calculating the sum of the running power of the electric equipment in the running state to obtain the total load power at the current moment; Screening out equipment types belonging to a set of adjustable equipment from electric equipment in an operation state, and calculating the sum of the operation power of the adjustable equipment to obtain the adjustable load power at the current moment; Dividing the adjustable load power by the total load power, and calculating to obtain an adjustable load duty ratio; Establishing an input language variable and an output language variable of a fuzzy inference system; Setting the load adjustment potential and the adjustable load duty ratio as input language variables of the fuzzy inference system, and setting the maximum interruptible load capacity as output language variables of the fuzzy inference system; Establishing a fuzzy language value set of an input language variable and an output language variable, wherein the fuzzy language value set comprises five fuzzy language values of minus small, zero, plus small, median and plus big; Constructing a membership function, and mapping a clear numerical value of the load adjustment potential and a clear numerical value of the adjustable load duty ratio into membership of a corresponding fuzzy language value by using the membership function; Constructing a fuzzy rule base, wherein the fuzzy rule base consists of fuzzy rules, and each fuzzy rule is used for defining a fuzzy language value in which the maximum interruptible load capacity is supposed to be when the load adjustment potential and the adjustable load duty ratio are in a fuzzy state; calculating triggering strength by adopting fuzzy intersection operation aiming at each fuzzy rule, wherein the triggering strength is determined by the minimum value of membership degrees of two input language variables in each fuzzy rule; According to the membership function of the fuzzy language value corresponding to the triggering strength and the fuzzy rule, executing fuzzy implication operation to obtain an output fuzzy set deduced by the fuzzy rule; aggregating the output fuzzy sets deduced by each fuzzy rule in the fuzzy rule base by utilizing fuzzy union operation to obtain a total output fuzzy set; and executing the deblurring processing of the total output fuzzy set by adopting a gravity center method to obtain the maximum interruptible load capacity at the current moment.
7. The method for managing and distributing energy of a multi-mode coupling industrial park according to claim 4, the method is characterized in that the process of calculating the power mutual aid reference value comprises the following steps: Extracting the DC bus voltage of the low-voltage flexible DC interconnection system from the preprocessed running state data, and obtaining the power unbalance degree of the platform region and the average value of three-phase active power used when calculating the power unbalance degree of the platform region, wherein the average value is used for calculating the theoretical mutual power requirement required by eliminating the power unbalance degree of the platform region; constructing a dynamic response gain function in a platform region mutual-aid evaluation model based on a self-adaptive dynamic allocation strategy, and mapping the platform region power unbalance degree into a dynamic response coefficient with a value range of zero to one by using the dynamic response gain function; constructing a direct-current voltage safety constraint function in a platform region mutual-aid evaluation model based on a self-adaptive dynamic allocation strategy, and mapping the direct-current bus voltage into a voltage safety factor with a value range of zero to one by utilizing the direct-current voltage safety constraint function; and calculating a power mutual-aid reference value by integrating the theoretical mutual-aid power requirement, the dynamic response coefficient and the voltage safety factor.
8. The method of claim 7, wherein constructing an objective function comprising voltage bias, full-factor operating cost, and power-mutual reference bias terms comprises: Setting a decision variable vector of a collaborative optimization model, wherein the decision variable vector comprises power grid interaction power, energy storage charge and discharge power, electric vehicle V2G charge and discharge power, interrupt load execution power and flexible interconnection converter mutual power at the current moment; Constructing an objective function of the collaborative optimization model, wherein the objective function is formed by weighting and summing voltage deviation items, full-element operation cost items and mutual-aid power reference value deviation items; Selecting alternating-current low-voltage buses of each power supply station area as key nodes, pre-calculating voltage-power sensitivity coefficients of the key nodes based on network topology structures and line impedance parameters of an industrial park, and converting voltage deviation of the key nodes into quadratic functions related to decision variable vectors by utilizing the voltage-power sensitivity coefficients to obtain voltage deviation items; The full-element operation cost item comprises power grid interaction cost, interrupt load compensation cost, energy storage loss cost and electric vehicle battery life loss cost, wherein the power grid interaction cost is determined by power grid time-sharing electricity price, the interrupt load compensation cost is determined by preset interrupt load compensation unit price, and the energy storage loss cost and the electric vehicle battery life loss cost are respectively modeled as secondary functions of energy storage charge and discharge power and electric vehicle V2G charge and discharge power so as to represent nonlinear accelerated aging influence of high-rate charge and discharge on battery life; and calculating the square of Euclidean distance between the mutual power of the flexible interconnection converter and the power mutual reference value to obtain a mutual power reference value deviation term.
9. The method for managing and distributing energy of a multi-mode coupling industrial park according to claim 8, the method is characterized in that the process of outputting the optimal control vector comprises the following steps: Constructing constraint conditions of a collaborative optimization model, wherein the constraint conditions comprise power balance constraint, equipment physical constraint and load adjustment constraint; Establishing power balance constraint, wherein the sum of the power grid interaction power and the predicted photovoltaic power is equal to the predicted total load minus the interrupt load execution power, and the energy storage charge-discharge power, the electric vehicle V2G charge-discharge power and the flexible interconnection converter mutual power are added; Establishing equipment physical constraint, setting upper and lower limits of energy storage charging and discharging power, electric vehicle V2G charging and discharging power and flexible interconnection converter mutual power, and limiting the state of charge of an energy storage device to be in a preset safety interval; establishing load regulation constraint, wherein the interruption load execution power is required to be in a section which is more than or equal to zero and less than or equal to the maximum interruptible load capacity; converting the objective function and the constraint condition into a standard quadratic programming mathematical form; solving a collaborative optimization model based on a quadratic programming algorithm by using one of an interior point method and an effective set method to obtain an optimal solution vector for minimizing an objective function; and outputting the optimal solution vector as an optimal control vector, and driving the adjustable equipment and the flexible interconnection converter to execute corresponding power adjustment actions.
10. A system for managing and distributing energy to a multi-mode coupled industrial park for implementing the method according to any one of the preceding claims 1-9, comprising: The sensing and identifying module is used for acquiring and preprocessing the running state data of all elements of the charge storage and charge release of the endogenous network in the industrial park, acquiring voltage deviation, power unbalance degree of the platform area and load adjustment potential, and identifying the equipment type and real-time state of the adjustable equipment by adopting a non-intrusive load identification algorithm, wherein the running state data at least comprises the direct current bus voltage of the low-voltage flexible direct current interconnection system and the tide data of the flexible interconnection converter; the trend prediction module is used for inputting the preprocessed running state data into a source load trend prediction model based on a long-term and short-term memory network to predict photovoltaic power and total load; the capacity evaluation module is used for outputting the maximum interruptible load capacity at the current moment through fuzzy reasoning according to the equipment type, the load adjustment potential and the real-time state; the mutual-aid evaluation module is used for constructing a platform mutual-aid evaluation model based on a self-adaptive dynamic allocation strategy by taking the platform power unbalance degree and the direct current bus voltage as constraints, and calculating a power mutual-aid reference value; The collaborative optimization module is used for constructing an objective function containing voltage deviation, full-element operation cost and power mutual-compensation reference value deviation items, taking predicted photovoltaic power and total load as input, taking maximum interruptible load capacity as constraint, solving a collaborative optimization model based on a quadratic programming algorithm, outputting an optimal control vector, and driving the adjustable equipment and the flexible interconnected converter.

Description

Multi-mode coupling industrial park energy management and distribution method and system Technical Field The invention belongs to the technical field of operation and control of power systems, and particularly relates to a management and distribution method and system for energy sources of a multi-mode coupling industrial park. Background Along with the progress of industrialization, the industrial park has evolved into a complex energy utilization system integrating multiple elements of 'source network charge storage and discharge', distributed photovoltaics, energy storage equipment, electric automobile charging piles and various industrial production loads are deployed in the industrial park, management and distribution of energy sources of the industrial park are realized, physical connection and digital technology are essentially utilized to monitor and schedule the flow direction and flow of electric energy in the industrial park in real time, management and distribution work is mainly based on production energy requirements of different enterprises, real-time output conditions of the distributed photovoltaics and peak-valley electricity price time periods of a power grid, photovoltaic power generation consumption, energy storage equipment charge and discharge, electric automobile charging and production line load operation are coordinated, and electric energy is accurately conveyed from a supply end to a consumption end and a storage end through unified allocation of multiple-end resources, so that orderly flow and on-demand supply of the electric energy and the heat energy between different voltage levels and different power supply areas are realized, and diversified energy requirements in the industrial park are continuously met. In order to solve the problems of unbalanced distribution of power and reduced voltage quality of a station area caused by double fluctuation of source load in an industrial park, the prior art adopts alternating current contactor switching interconnection and a mode of executing fixed parameter control based on total load monitoring data for processing, but the problems that the mutual power cannot be smoothly and continuously regulated, the influence of inertia characteristics of inductive loads and response characteristics of variable frequency loads on the execution effect of control instructions is delayed can be ignored, and further the problems that the fluctuation amplitude of the voltage of a direct current bus is large, the tracking control instruction precision of adjustable equipment is low, the power unbalance of the station area cannot be accurately eliminated and the operation stability of an energy system of the industrial park is reduced can be caused. Disclosure of Invention The invention aims to provide a management and distribution method and a system for multi-mode coupling industrial park energy, which are used for compensating industrial load response lag by adaptively adjusting PID control parameters according to equipment types, executing power mutual aid by utilizing a low-voltage flexible direct current interconnection system, eliminating power unbalance of a platform region and inhibiting voltage deviation, and realizing collaborative optimization and accurate distribution of all elements of storage, charging and discharging of a source network of the industrial park. In order to solve the technical problems, the invention adopts the following technical scheme: the management and distribution method of the multi-mode coupling industrial park energy is characterized by comprising the following steps of: acquiring and preprocessing running state data of all elements of the charging, storing, charging and discharging of the endogenous network in the industrial park, acquiring voltage deviation, power unbalance degree of the platform area and load adjustment potential, and identifying equipment types and real-time states of adjustable equipment by adopting a non-intrusive load identification algorithm, wherein the running state data at least comprises direct current bus voltage of a low-voltage flexible direct current interconnection system and tide data of a flexible interconnection converter; inputting the preprocessed running state data into a source load trend prediction model based on a long-term and short-term memory network, and predicting photovoltaic power and total load; according to the equipment type, the load regulation potential and the real-time state, outputting the maximum interruptible load capacity at the current moment through fuzzy reasoning; taking the power unbalance degree of the transformer area and the voltage of the direct current bus as constraints, constructing a transformer area mutual-aid evaluation model based on a self-adaptive dynamic allocation strategy, and calculating a power mutual-aid reference value; And constructing an objective function comprising voltage deviation, full-element operation cost and power mutu