Self-adaptive source network load storage integrated intelligent scheduling system and method

Abstract

The invention discloses a self-adaptive source network load storage integrated intelligent scheduling system and method, and the system comprises a data collection module, a prediction module, a multi-objective optimization module, an optimization algorithm module, a self-adaptive parameter adjustment module, a constraint processing module, and a result output module. The future power demand is predicted through the power demand prediction model, the future renewable energy power generation amount is predicted through the renewable energy power generation prediction model, and a multi-target model with cost, benefit and greenhouse gas emission and model constraint conditions are constructed; the multi-target model is optimized through a hybrid optimization algorithm model combining a genetic algorithm and a particle swarm optimization algorithm, algorithm parameters are dynamically adjusted, a global optimal solution is searched, the optimized power generation equipment output power, the optimized energy storage equipment charging and discharging power and the optimized power grid interaction power are output, and a scheduling decision is guided; according to the method, the optimization efficiency and robustness are improved, and the adaptive capacity and the multi-target optimization capacity of a power grid system are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network dispatching, and more specifically, to an adaptive integrated intelligent dispatching system and method for source-network-load-storage integration. Background Art

[0002] Grid dispatching is the core of grid operation and accident handling, which is related to the safe, stable and economic operation of the grid. Any instability in details may pose a hidden danger to the safe and stable operation of the grid; the main task of grid dispatching is to command substation operation personnel to perform switching operations according to the grid operation mode and work needs. In the case of a grid accident, command relevant substation personnel to handle the grid accident and ensure the normal power supply of the grid as much as possible.

[0003] With the development of the energy Internet, the integrated intelligent dispatching of source-network-load-storage integration plays an increasingly important role in the power system. The system realizes the efficient, reliable and environmentally friendly operation of the power system by coordinating power generation, transmission, distribution, energy storage and load management.

[0004] However, there are some deficiencies in the existing dispatching systems, which are mainly manifested in the following aspects:

[0005] Single optimization algorithm: Most systems only rely on a single optimization algorithm and cannot make full use of the advantages of different algorithms; Insufficient adaptability: The existing dispatching systems have poor adaptability when facing complex and changeable power demands and environmental conditions, and it is difficult to adjust the optimization strategy in real time; Difficult multi-objective optimization: The optimization of the power system usually involves multiple objectives, such as minimizing costs, maximizing benefits and reducing environmental impacts. The existing systems often have poor effects when dealing with multi-objective optimization problems; Complex constraint conditions: There are various constraint conditions in the power system, such as power balance, equipment capacity, grid safety and regulatory standards. The existing systems are prone to optimization failures when dealing with these complex constraints.

[0006] Therefore, how to handle the complex constraint conditions of the source-network-load-storage system, realize the integrated intelligent dispatching of multi-objective optimization, improve the diversity and robustness of the optimization algorithm, so as to improve the ability to search for the global optimal solution, enhance the adaptive ability of the system, and enable the system to adapt to the changes of power demand and environmental conditions in real time is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0007] In view of this, the present invention provides an adaptive integrated intelligent dispatching system and method for source-network-load-storage integration to solve some of the technical problems mentioned in the background art.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] An adaptive integrated intelligent scheduling system for the source, grid, load, and energy storage, comprising:

[0010] A data acquisition module, configured to use sensors and measuring devices to collect environmental and power parameters in real time from power generation sources, power grids, load devices, and energy storage devices;

[0011] A prediction module, configured to predict future power demands through a power demand prediction model and predict future renewable energy power generation through a renewable energy power generation prediction model;

[0012] A multi-objective optimization module, configured to construct a multi-objective model with cost, benefit, and greenhouse gas emissions;

[0013] An optimization algorithm module, configured to optimize the multi-objective model through a hybrid optimization algorithm model GA-PSO that combines the genetic algorithm and the particle swarm optimization algorithm, and search for the global optimal solution;

[0014] An adaptive parameter adjustment module, configured to dynamically adjust algorithm parameters for the multi-objective optimization process of the optimization algorithm module;

[0015] A constraint handling module, configured to construct constraint conditions during the optimization process of the multi-objective model, including power balance constraints, equipment capacity constraints, grid security constraints, and greenhouse gas emission constraints;

[0016] A result output module, configured to output the output power of the optimized power generation equipment, the charge and discharge power of the energy storage equipment, and the grid interaction power to guide the scheduling decision-making.

[0017] Preferably, the adaptive integrated intelligent scheduling system for the source, grid, load, and energy storage further includes a data preprocessing module, configured to clean the collected data, including denoising, filling missing values, and handling outliers, and perform standardization and normalization processing on the data.

[0018] Preferably, the power demand prediction model includes an ARIMA model and an LSTM model established based on the time series model of power demand; the renewable energy power generation prediction model is a random forest model trained by using meteorological data and historical power generation data through machine learning.

[0019] Preferably, the ARIMA model is specifically:

[0020] (1 - φ 1 B - φ 2 B 2 - … - φ p B p )(1 - B) d y t =c + (1 + θ 1 B + θ 2 B 2 + … + θq B q )ε t

[0021] Among them, p is the number of autoregressive terms, d is the order of differencing, q is the number of moving average terms, and y t is the value of the time series at time, and φ 1 , φ 2 , …, φ p are autoregressive coefficients, ε t is the error term, c is the constant term, and θ 1 , θ 2 , …, θ q are moving average coefficients, B is the lag operator, that is, By t = y t-1 ;

[0022] The method for constructing the LSTM model is specifically as follows: Convert the time series data into a format suitable for the LSTM model, use Keras to construct the LSTM model, use the historical daily electricity demand data to train the LSTM model, and use the trained model to predict the daily electricity demand for the next 30 days.

[0023] Preferably, the renewable energy power generation prediction model takes the daily wind speed, light intensity, and temperature in the past 365 days as inputs and the renewable energy power generation data as outputs, and trains a random forest model to predict the daily renewable energy power generation for the next 30 days;

[0024] The final prediction value of the random forest model is the average of the prediction values of all decision trees:

[0025]

[0026] Among them, N is the number of decision trees, is the prediction value of the i-th decision tree;

[0027]

[0028] Among them, x is the input feature vector, and f i is the prediction function of the i-th decision tree.

[0029] Preferably, the multi-objective model is specifically:

[0030] F = αC - βR + γE

[0031] Among them, α, β, and γ are weight coefficients used to balance the importance of different objectives, C is the cost function, R is the benefit function, and E is the environmental impact function;

[0032]

[0033] Among them, Fg,t is the fuel cost of the g-th power generation equipment at time t, M g,t is the maintenance cost, P buy,t is the electricity purchase price, P g,t is the output power of the g-th power generation equipment at time t, P grid,t is the power of purchasing or selling electricity from / to the power grid; G is the number of power generation equipment, and T is the total number of time periods;

[0034]

[0035] wherein, P sell,t is the electricity selling price, B g,t is the subsidy coefficient;

[0036]

[0037] wherein, E g,t is the greenhouse gas emission per unit power generation of the g-th power generation equipment at time t.

[0038] Preferably, the power balance constraint is specifically:

[0039]

[0040] wherein, P storage (t) is the charge and discharge power of the energy storage equipment at time t, P g (t) is the power generation power of the g-th power generation equipment at time t, P grid (t) is the electricity quantity purchased from or sold to the power grid at time t, and D(t) is the power demand at time t;

[0041] The equipment capacity constraint is specifically:

[0042] 0 ≤ P g (t)\P g,max

[0043] 0 ≤ P storage (t) ≤ P storage,max

[0044] wherein, P g,max is the maximum output power of the g-th power generation equipment, P storage,max is the maximum output power of the energy storage equipment;

[0045] The power grid security constraint is specifically:

[0046] V min \V(t) ≤ V max

[0047] f min ≤ f(t) ≤ f max

[0048] Among them, V min and the sum is V max is the safety range of voltage, and f min and f max is the safety range of frequency; the greenhouse gas emission constraint is specifically:

[0049] E ≤ E max

[0050] Among them, E max is the maximum greenhouse gas emission limit.

[0051] Preferably, the hybrid optimization algorithm model GA - PSO optimizes the multi - objective model, and the specific content of searching for the global optimal solution is:

[0052] Initialize the population: Randomly generate an initial solution set, representing the output power configurations of different devices, and at the same time initialize the positions and velocities of the particle swarm;

[0053] Evaluate fitness: Calculate the fitness value of each solution according to the objective function;

[0054] Selection operation: Select individuals with high fitness according to fitness;

[0055] Crossover operation: Crossover the selected individuals to generate new offspring;

[0056] Mutation operation: Mutate the offspring with a certain probability to introduce a new solution space;

[0057] PSO update: For the new solutions generated in each generation, use the PSO algorithm to update the positions and velocities to further optimize the quality of the solutions;

[0058] Dynamically adjust parameters: Dynamically adjust the crossover rate and mutation rate according to the search process, and dynamically adjust the inertia weight and learning factor;

[0059] Termination condition: When the predetermined number of iterations is reached or other termination conditions are met, stop the optimization.

[0060] Preferably, the randomly generated initial solution set is Each solution x i includes the output power P of the power generation equipment g,t , the charge - discharge power P of the energy storage equipment storage,t and the power P of interaction with the power grid grid,t ;

[0061] The fitness value of each solution is calculated as:

[0062] f(x i ) = αC(x i ) - βR(x i ) + γE(xi )

[0063] Among them, α, β, and γ are the weights of cost, benefit, and environmental impact;

[0064] Use the roulette wheel selection method to select the next generation of population:

[0065]

[0066] Perform crossover operation on the selected individuals to generate new offspring:

[0067] x offspring = α′x 1 +(1 - α′)x 2

[0068] Among them, α′ is the crossover rate;

[0069] Perform mutation operation on the offspring to increase the diversity of solutions:

[0070] x mutated = x + σ·randn

[0071] Among them, σ is the mutation rate; randn is a function that generates random numbers with a standard normal distribution;

[0072] Update the position and velocity of the particles:

[0073] v i (t + 1)= ωv i (t)+ c 1 ·r 1 ·(p best,i - x i (t))+ c 2 ·r 2 ·(g best - x i (t))

[0074] x i (t + 1)= x i (t)+ v i (t + 1)

[0075] Among them, w is the inertia weight, c 1 and c 2 are the learning factors, r 1 and r 2 are random numbers, p best,i is the individual best position of the i-th particle, g best is the global best position.

[0076] Dynamically adjust the parameters according to the search process:

[0077]

[0078] Among them, t is the current iteration number, and T is the maximum iteration number.

[0079] An adaptive integrated intelligent scheduling method for source, grid, load, and energy storage, based on the described adaptive integrated intelligent scheduling system for source, grid, load, and energy storage, includes the following steps:

[0080] S1. From power generation sources, power grids, load devices, and energy storage devices, use sensors and measurement devices to collect environmental and power parameters in real time;

[0081] S2. Predict future power demands through a power demand prediction model, and predict future renewable energy power generation through a renewable energy power generation prediction model;

[0082] S3. Construct a multi-objective model for cost, benefit, and greenhouse gas emissions, as well as the constraints during the optimization process of the multi-objective model, including power balance constraints, equipment capacity constraints, grid security constraints, and greenhouse gas emission constraints;

[0083] S4. Optimize the multi-objective model through a hybrid optimization algorithm model GA-PSO that combines the genetic algorithm and the particle swarm optimization algorithm, and dynamically adjust the algorithm parameters for the multi-objective optimization process to search for the global optimal solution;

[0084] S5. Output the output power of the optimized power generation equipment, the charge and discharge power of the energy storage equipment, and the grid interaction power to guide the scheduling decision.

[0085] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses an adaptive integrated intelligent scheduling system and method for source, grid, load, and energy storage. By combining the global search ability of the genetic algorithm GA and the fast convergence characteristics of the particle swarm optimization PSO, a hybrid optimization algorithm is constructed, which can find the global optimal solution faster, improve the optimization efficiency, and enhance the optimization efficiency and robustness. At the same time, through dynamic parameter adjustment and a strict constraint processing mechanism, the adaptive ability and multi-objective optimization ability of the system are enhanced, and the optimal scheduling problem of the power system can be effectively solved in practical applications, realizing efficient, reliable, and environmentally friendly operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0087] Figure 1Schematic diagram of an adaptive source-grid-load-storage integrated intelligent scheduling system provided by the present invention;

[0088] Figure 2 Schematic diagram of the hybrid optimization algorithm model GA-PSO provided by the present invention;

[0089] Figure 3 Schematic diagram of an adaptive source-grid-load-storage integrated intelligent scheduling method provided by the present invention. Detailed implementation manners

[0090] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0091] An embodiment of the present invention discloses an adaptive source-grid-load-storage integrated intelligent scheduling system, as Figure 1 , including:

[0092] A data acquisition module, configured to use sensors and measurement devices to collect environmental and power parameters in real time from power generation sources, power grids, load devices, and energy storage devices;

[0093] A prediction module, configured to predict future power demands through a power demand prediction model and predict future renewable energy power generation through a renewable energy power generation prediction model;

[0094] A multi-objective optimization module, configured to build a multi-objective model with cost, benefit, and greenhouse gas emissions;

[0095] An optimization algorithm module, configured to optimize the multi-objective model through a hybrid optimization algorithm model GA-PSO that combines a genetic algorithm and a particle swarm optimization algorithm to search for the global optimal solution;

[0096] An adaptive parameter adjustment module, configured to dynamically adjust algorithm parameters for the multi-objective optimization process of the optimization algorithm module;

[0097] A constraint handling module, configured to build constraint conditions during the optimization process of the multi-objective model, including power balance constraints, equipment capacity constraints, grid security constraints, and greenhouse gas emission constraints;

[0098] A result output module, configured to output the output power of the optimized power generation equipment, the charge and discharge power of the energy storage equipment, and the grid interaction power to guide the scheduling decision.

[0099] In this embodiment, the sensors and measurement devices used include, but are not limited to, temperature sensors, humidity sensors, light intensity sensors, power meters, current transformers, voltage transformers, flow meters, pressure transmitters, etc. to collect environmental and power parameters in real time; standard communication protocols (such as Modbus, OPC UA, MQTT) are used to ensure the reliability and consistency of data transmission.

[0100] To further implement the above technical solution, an adaptive integrated intelligent scheduling system for the source, grid, load, and energy storage further includes a data preprocessing module for cleaning the collected data, including denoising, filling missing values, and handling outliers, as well as standardizing and normalizing the data.

[0101] To further implement the above technical solution, the power demand forecasting model includes an ARIMA model and an LSTM model established based on the time series model of power demand; the renewable energy power generation forecasting model is a random forest model trained using meteorological data and historical power generation data through machine learning.

[0102] To further implement the above technical solution, the ARIMA model is specifically:

[0103] (1 - φ 1 B - φ 2 B 2 -… - φ p B p )(1 - B) d y t =c+(1 + θ 1 B + θ 2 B 2 +… + θ q B q )ε t

[0104] where p is the number of autoregressive terms, d is the order of differencing, q is the number of moving average terms, y t is the value of the time series at time t, φ 1 , φ 2 ,…, φ p are autoregressive coefficients, ε t is the error term, c is the constant term, θ 1 , θ 2 ,…, θ q are moving average coefficients, B is the lag operator, i.e., By t =y t-1 ;

[0105] In this embodiment, the autoregressive part is:

[0106] y t =c + φ 1 yt-1 +φ 2 y t-2 +...+φ p y t-p +ε t

[0107] Differencing part:

[0108] Used to make the time series stationary through differencing operation;

[0109] For d - order differencing, the formula is:

[0110] y t ′ = y t -y t-1

[0111] Repeat for d times until the time series becomes stationary;

[0112] Moving average part:

[0113] ε t = θ 1 ε t-1 + θ 2 ε t-2 +,... + θ q ε t-q + η t

[0114] where η t is the white noise error term;

[0115] The specific method for constructing the LSTM model is as follows: Convert the time - series data into a format suitable for the LSTM model, build the LSTM model using Keras, train the LSTM model using historical daily power demand data, and use the trained model to predict the daily power demand for the next 30 days.

[0116] To further implement the above - mentioned technical solution, the renewable energy power generation prediction model takes the daily wind speed, light intensity, and temperature in the past 365 days as inputs and the renewable energy power generation data as outputs, and trains a random forest model for predicting the daily renewable energy power generation for the next 30 days;

[0117] The final prediction value of the random forest model is the average of all decision tree prediction values:

[0118]

[0119] where N is the number of decision trees, is the prediction value of the i - th decision tree;

[0120]

[0121] where x is the input feature vector, and f i is the prediction function of the i-th decision tree.

[0122] To further implement the above technical solution, the multi-objective model is specifically:

[0123] F = αC - βR + γE

[0124] where α, β, and γ are weight coefficients used to balance the importance of different objectives, C is the cost function, R is the benefit function, and E is the environmental impact function;

[0125]

[0126] where F g,t is the fuel cost of the g-th power generation equipment at time t, M g,t is the maintenance cost, P buy,t is the electricity purchase price, P g,t is the output power of the g-th power generation equipment at time t, P grid,t is the power of purchasing or selling electricity from / to the power grid; G is the number of power generation equipment, and T is the total number of time;

[0127]

[0128] where P sell,t is the electricity selling price, and B g,t is the subsidy coefficient;

[0129]

[0130] where E g,t is the greenhouse gas emission per unit power generation of the g-th power generation equipment at time t.

[0131] To further implement the above technical solution, the power balance constraint is specifically:

[0132]

[0133] where P storage (t) is the charge and discharge power of the energy storage equipment at time t, P g (t) is the power generation power of the g-th power generation equipment at time t, P grid (t) is the electricity purchased from or sold to the power grid at time t, and D(t) is the power demand at time t;

[0134] The equipment capacity constraint is specifically:

[0135] 0 ≤ P g (t) ≤ P g,max

[0136] 0 ≤ P storage (t) ≤ P storage,max

[0137] Wherein, P g,max is the maximum output power of the g-th power generation device, and P storage,max is the maximum output power of the energy storage device;

[0138] The power grid security constraints are specifically:

[0139] V min ≤ V(t) ≤ V max

[0140] f min ≤ f(t) ≤ f max

[0141] Wherein, V min and V max are the safety ranges of voltage, and f min and f max are the safety ranges of frequency;

[0142] The greenhouse gas emission constraints are specifically:

[0143] E ≤ E

[0144] max

[0145] Wherein, E max is the maximum greenhouse gas emission limit.

[0146] To further implement the above technical solution, such as Figure 2 , the hybrid optimization algorithm model GA-PSO optimizes the multi-objective model, and the specific content of searching for the global optimal solution is:

[0147] Initialize the population: Randomly generate an initial solution set, representing the output power configurations of different devices, and simultaneously initialize the positions and velocities of the particle swarm;

[0148] Evaluate the fitness: Calculate the fitness value of each solution according to the objective function;

[0149] Selection operation: Select the individuals with high fitness according to the fitness;

[0150] Crossover operation: Perform crossover on the selected individuals to generate new offspring;

[0151] Mutation operation: Mutate the offspring with a certain probability to introduce a new solution space;

[0152] PSO update: For the new solutions generated in each generation, use the PSO algorithm to update the positions and velocities to further optimize the quality of the solutions;

[0153] Dynamic parameter adjustment: Dynamically adjust the crossover rate and mutation rate according to the search process, and dynamically adjust the inertia weight and learning factor;

[0154] Termination condition: Stop the optimization when the predetermined number of iterations is reached or other termination conditions are met.

[0155] To further implement the above technical solution, the randomly generated initial solution set is Each solution x i includes the output power P of the power generation equipment g,t , the charge and discharge power P of the energy storage equipment storage,t and the grid interaction power P grid,t ;

[0156] Calculate the fitness value of each solution as:

[0157] f(x i ) = αC(x i ) - βR(x i ) + γE(x i )

[0158] where α, β, and γ are the weights of cost, benefit, and environmental impact;

[0159] Use the roulette wheel selection method to select the next generation population:

[0160]

[0161] Perform a crossover operation on the selected individuals to generate new offspring:

[0162] x offspring = α'x 1 + (1 - α')x 2

[0163] where α' is the crossover rate;

[0164] Perform a mutation operation on the offspring to increase the diversity of the solutions:

[0165] x mutated = x + α·randn

[0166] where σ is the mutation rate; randn is a function that generates random numbers with a standard normal distribution;

[0167] Update the position and velocity of the particles:

[0168] v i (t + 1) = ω·v i (t) + c 1 ·r 1 ·(p best,i - x i (t)) + c2 ·r 2 ·(g best -x i (t))

[0169] x i (t + 1) = x i (t) + v i (t + 1)

[0170] where w is the inertia weight, c 1 and c 2 are learning factors, r 1 and r 2 are random numbers, p best,i is the individual best position of the i-th particle, g best is the global best position.

[0171] Dynamically adjust the parameters according to the search process:

[0172]

[0173] where t is the current iteration number and T is the maximum iteration number.

[0174] Taking the power generation equipment of three generator sets and a set of renewable energy power generation equipment as an example, the system parameters and initial conditions are specifically:

[0175] The maximum output power P of generator set 1 of the power generation equipment g1,max = 1000 kW, the maximum output power P of generator set 2 g2,max = 1500 kW, the maximum output power P of generator set 3 g3,max = 2000 kW; the maximum output power P of the renewable energy power generation equipment re,max = 1000 kW; the maximum charge and discharge power of the energy storage equipment is P storage,max = 500 kW;

[0176] The power demand forecast results are:

[0177] Time 1: D 1 = 3000 kW

[0178] Time 2: D 2 = 3500 kW

[0179] Time 3: D 3 = 4000 kW

[0180] The renewable energy power generation forecast results are:

[0181] Time 1: P re,1 = 500 kW

[0182] Time 2: P re,2 = 700 kW

[0183] Time 3: P re,3 = 900 kW

[0184] The grid parameters are: the voltage range is 110 V to 130 V, and the frequency range is 58 Hz to 62 Hz;

[0185] The cost and benefit parameters are:

[0186] Fuel cost: F g1 = 0.1 yuan / kWh, F g2 = 0.15 yuan / kWh, F g3 = 0.2 yuan / kWh;

[0187] Maintenance cost: M g1 = 0.05 yuan / kWh, M g2 = 0.07 yuan / kWh, M g3 = 0.1 yuan / kWh;

[0188] Purchase electricity price: P buy = 0.12 yuan / kWh;

[0189] Selling electricity price: P sell = 0.15 yuan / kWh;

[0190] Subsidy coefficient: B g1 = 0.02 yuan / kWh, B g2 = 0.03 yuan / kWh, B g3 = 0.04 yuan / kWh;

[0191] Greenhouse gas emissions: E g1 = 0.5 kg CO2 / kWh, E g2 = 0.6 kg CO 2 / kWh, E g3 = 0.7 kg CO 2 / kWh;

[0192] Environmental standard: Maximum allowable greenhouse gas emissions: E max = 1000 kg CO 2 ;

[0193] Optimization algorithm and parameter settings:

[0194] Hybrid optimization algorithm: Population size N = 50, maximum number of iterations T = 100, crossover rate α = 0.5, mutation rate σ = 0.1, inertia weight: initial w = 0.7, minimum w min = 0.4, learning factor: initial c 1= 1.5, minimum c 1,min = 0.5, initial c 2 = 1.5, minimum c 2,min = 0.5;

[0195] Weight setting: cost weight α = 1.0, benefit weight: β = 1.0, environmental impact weight: γ = 1.0;

[0196] After 100 iterations, the optimal solution is as follows:

[0197] Output power of traditional power generation equipment at Time 1, Time 2, and Time 3:

[0198] Time 1: P g1,1 = 800 kW, P g2,1 = 1000 kW, P g3,1 = 1200 kW

[0199] Time 2: Pg 1,2 = 900 kW, P g2,2 = 1200 kW, P g3,2 = 1400 kW

[0200] Time 3: P g1,3 = 1000 kW, P g2,3 = 1500 kW, P g3,3 = 1500 kW

[0201] Output power of renewable energy power generation equipment at Time 1, Time 2, and Time 3: P re,1 = 500 kW, P re,2 = 700 kW, P re, 3 = 900 kW;

[0202] Charge and discharge power of energy storage equipment at Time 1, Time 2, and Time 3: P storage,1 = 0 kW, P storage,2 = 0 kW, P storage,3 = 0 kW;

[0203] Grid interaction power at Time 1, Time 2, and Time 3: P grid,1 = 0 kW, P grid,2 = 0 kW, P grid,3 = 0 kW;

[0204] Fitness values at Time 1, Time 2, and Time 3 are f(x 1 ), f(x 2 ), f(x 3 );

[0205] Based on the optimization results, the dispatching decision of the power system is guided as follows:

[0206] Traditional power generation equipment scheduling:

[0207] At time 1, generators 1, 2, and 3 are enabled, with outputs of 800 kW, 1000 kW, and 1200 kW respectively;

[0208] At time 2, generators 1, 2, and 3 are enabled, with outputs of 900 kW, 1200 kW, and 1400 kW respectively;

[0209] At time 3, generators 1, 2, and 3 are enabled, with outputs of 1000 kW, 1500 kW, and 1500 kW respectively;

[0210] Renewable energy power generation equipment scheduling:

[0211] At time 1, the renewable energy power generation equipment is enabled, with an output of 500 kW;

[0212] At time 2, the renewable energy power generation equipment is enabled, with an output of 700 kW;

[0213] At time 3, the renewable energy power generation equipment is enabled, with an output of 900 kW;

[0214] Energy storage equipment scheduling: At times 1, 2, and 3, the energy storage equipment does not perform charge and discharge operations;

[0215] Grid interaction scheduling: At times 1, 2, and 3, no electricity is purchased from or sold to the grid.

[0216] Through the above scheduling decisions, the efficient, reliable, and environmentally friendly operation of the power system is achieved, and the goals of minimizing costs, maximizing benefits, and reducing environmental impacts are reached.

[0217] This embodiment demonstrates how to use a hybrid optimization algorithm that combines genetic algorithm (GA) and particle swarm optimization (PSO), combines a power demand prediction model and a renewable energy power generation prediction model, realizes the multi-objective optimization of the source-grid-load-storage integrated intelligent scheduling system, and ensures the accuracy and feasibility of the optimization results through specific parameter settings and optimization processes, providing strong support for the efficient operation of the power system.

[0218] An adaptive source-grid-load-storage integrated intelligent scheduling method, based on an adaptive source-grid-load-storage integrated intelligent scheduling system, includes the following steps:

[0219] S1. From power generation sources, the grid, load equipment, and energy storage equipment, use sensors and measurement devices to collect environmental and power parameters in real time;

[0220] S2. Predict future power demand through a power demand prediction model, and predict future renewable energy power generation through a renewable energy power generation prediction model;

[0221] S3. Construct a multi-objective model considering cost, benefit, and greenhouse gas emissions, as well as the constraints during the optimization process of the multi-objective model, including power balance constraint, equipment capacity constraint, power grid security constraint, and greenhouse gas emission constraint;

[0222] S4. Optimize the multi-objective model through a hybrid optimization algorithm model GA-PSO that combines the genetic algorithm and the particle swarm optimization algorithm, and dynamically adjust the algorithm parameters during the multi-objective optimization process to search for the global optimal solution;

[0223] S5. Output the output power of the optimized power generation equipment, the charging and discharging power of the energy storage equipment, and the power interaction with the power grid to guide the dispatching decision-making.

[0224] In practical applications, the linkage between charging piles and air conditioners is one of the important means to realize the intelligent power grid and energy management system. An adaptive integrated intelligent dispatching method for the source-grid-load-storage can also be applied to the scenario of the linkage between charging piles and air conditioners to realize the intelligent dispatching of the load side of the source-grid-load-storage system, including dynamic load management, peak shaving and valley filling, renewable energy utilization, and real-time monitoring and feedback. Through dynamic load management, peak shaving and valley filling, renewable energy utilization, and real-time monitoring and feedback, the energy utilization efficiency can be effectively improved, the operation cost can be reduced, environmental pollution can be reduced, and the comfort and satisfaction of users can be improved. This linkage not only helps to optimize the operation of the power system but also promotes the development and application of sustainable energy.

[0225] Specifically:

[0226] Dynamic load management:

[0227] Charging pile management: Dynamically adjust the charging time and power of the charging pile according to the real-time state of the power grid and the power demand forecast; for example, during peak power demand periods, the charging power of the charging pile can be reduced or the charging time can be postponed, while during low-demand periods, the charging power can be increased or the charging can be advanced;

[0228] Air conditioner control: Dynamically adjust the working mode and power of the air conditioner according to the indoor and outdoor temperatures, the user comfort demand, and the power demand forecast. For example, during peak power demand periods, the set temperature of the air conditioner can be appropriately increased to reduce energy consumption; during low-demand periods, the set temperature can be appropriately reduced to improve user comfort.

[0229] Peak shaving and valley filling:

[0230] Peak shaving: During peak power demand periods, reduce the charging power of the charging pile and appropriately adjust the set temperature of the air conditioner to relieve the burden on the power grid and avoid overload;

[0231] Valley filling: During low-power demand periods, make full use of the redundant power resources to increase the charging power of the charging pile and the cooling / heating power of the air conditioner to improve the energy utilization efficiency.

[0232] Renewable energy utilization:

[0233] Charging piles and renewable energy: Combining renewable energy generation prediction, preferentially charging electric vehicles when the renewable energy generation is sufficient, reducing the dependence on traditional power generation equipment, and reducing carbon emissions;

[0234] Air conditioners and renewable energy: When the renewable energy generation is sufficient, appropriately increase the cooling / heating power of air conditioners, improve user comfort, and at the same time reduce the demand for traditional electricity.

[0235] Real-time monitoring and feedback:

[0236] Real-time monitoring: Through Internet of Things technology and sensors, real-time monitor the operating status of charging piles and air conditioners, collect data and analyze it;

[0237] Feedback mechanism: Adjust the scheduling strategy according to real-time data, respond to emergencies in a timely manner, and ensure the stability and reliability of the system.

[0238] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and reference can be made to the description of the method part for related parts.

[0239] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An adaptive source-grid-load-storage integrated intelligent dispatching system, characterized in that: include: Data acquisition module, used to collect environmental and power parameters in real time from power sources, power grids, load devices and energy storage devices using sensors and measuring devices; A prediction module, used to predict future electricity demand through an electricity demand prediction model, and to predict future renewable energy power generation through a renewable energy power generation prediction model; Multi-objective optimization module, used to build multi-objective models based on cost, benefit and greenhouse gas emissions; The optimization algorithm module is used to optimize the multi-objective model through the hybrid optimization algorithm model GA-PSO that combines the genetic algorithm and the particle swarm optimization algorithm to search for the global optimal solution; An adaptive parameter adjustment module is used to dynamically adjust algorithm parameters for the multi-objective optimization process of the optimization algorithm module; The constraint processing module is used to construct the multi-objective model for the constraints in the optimization process, including power balance constraints, equipment capacity constraints, grid security constraints and greenhouse gas emission constraints; The result output module is used to output the optimized output power of power generation equipment, charging and discharging power of energy storage equipment and grid interaction power to guide scheduling decisions.

2. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 1 is characterized in that: It also includes a data preprocessing module for cleaning the collected data, including denoising, filling missing values ​​and processing outliers, as well as standardizing and normalizing the data.

3. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 1 is characterized in that: The power demand forecasting model includes an ARIMA model and an LSTM model established based on a time series model of power demand; the renewable energy power generation forecasting model is a random forest model trained by machine learning using meteorological data and historical power generation data.

4. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 3 is characterized in that: The ARIMA model is expressed as: (1-φ1B-φ2B 2 -…-f p B p (1-B) d y t =c+(1+θ1B+θ2B 2 +…+θ q B q )e t Among them, p is the number of autoregressive terms, d is the difference order, q is the number of moving average terms, and y is the t is the value of the time series at time, φ1, φ2, …, φ p is the autoregressive coefficient, ε t is the error term, c is the constant term, θ1,θ2,…,θ q is the sliding average coefficient, B is the lag operator, that is, By t =y t-1 ; The LSTM model is constructed by converting time series data into a format suitable for the LSTM model, using Keras to construct the LSTM model, using historical daily electricity demand data to train the LSTM model, and using the trained model to predict daily electricity demand for the next 30 days.

5. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 3 is characterized in that: The renewable energy generation prediction model uses the daily wind speed, light intensity, and temperature of the past 365 days as input and the renewable energy generation data as output, and trains a random forest model to predict the daily renewable energy generation in the next 30 days; The final prediction of the random forest model is the average of all the decision tree predictions: Where N is the number of decision trees, is the predicted value of the i-th decision tree; Among them, x is the input feature vector, f i is the prediction function of the i-th decision tree.

6. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 1 is characterized in that: The multi-objective model is expressed as: F=αC-βR+γE Among them, α, β and γ are weight coefficients used to balance the importance of different goals, C is the cost function, R is the benefit function, and E is the environmental impact function; Among them, F g,t is the fuel cost of the g-th generating equipment at time t, M g,t is the maintenance cost, P buy,t is the electricity purchase price, P g,t is the output power of the g-th power generation equipment at time t, P grid,t is the power of electricity purchased or sold from the power grid; G is the number of power generation equipment, and T is the total time; Among them, P sell,t is the electricity price, B g,t is the subsidy coefficient; Among them, E g,t is the greenhouse gas emissions per unit of electricity generated by the g-th power generation equipment at time t.

7. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 1 is characterized in that: The power balance constraints are as follows: Among them, P storage (t) is the charging and discharging power of the energy storage device at time t, P g (t) is the power generated by the g-th power generation equipment at time t, P grid (t) is the amount of electricity purchased from or sold to the grid at time t, and D(t) is the electricity demand at time t; The equipment capacity constraints are as follows: 0≤P g (t)≤P g,max 0≤P storage (t)≤P storage,max Among them, P g,max is the maximum output power of the g-th power generation equipment, P storage,max is the maximum output power of the energy storage device; The specific grid security constraints are: V min ≤V(t)≤V max f min ≤f(t)≤f max Among them, V min and V max is the safe range of voltage, f min and f max is the safe range of frequency; The greenhouse gas emission constraints are as follows: E≤E max Among them, E max The maximum greenhouse gas emission limit.

8. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 1 is characterized in that: The hybrid optimization algorithm model GA-PSO optimizes the multi-objective model and searches for the global optimal solution as follows: Initialize the population: randomly generate an initial solution set, representing the output power configuration of different devices, and initialize the position and speed of the particle swarm; Evaluate fitness: Calculate the fitness value of each solution based on the objective function; Selection operation: select individuals with high fitness according to fitness; Crossover operation: crossover the selected individuals to produce new offspring; Mutation operation: mutate the offspring with a certain probability and introduce a new solution space; PSO update: For each new solution generated in each generation, the PSO algorithm is used to update the position and velocity to further optimize the quality of the solution; Dynamically adjust parameters: dynamically adjust the crossover rate and mutation rate according to the search process, as well as dynamically adjust the inertia weight and learning factor; Termination condition: Stop the optimization when the predetermined number of iterations is reached or other termination conditions are met.

9. The adaptive source-grid-load-storage integrated intelligent dispatching system according to claim 8 is characterized in that: The randomly generated initial solution set is Each solution x i Including the output power P of the power generation equipment g,t 、Charging and discharging power P of energy storage equipment storage,t Interaction power with the grid P grid,t ; The fitness value of each solution is calculated as: f(x i )=αC(x i )-βR(x i )+γE(x i ) Use roulette wheel selection to select the next generation of population: Perform a crossover operation on the selected individuals to generate new offspring: x offspring =α′x1+(1-α′)x2 Among them, α′ is the crossover rate; Perform mutation operations on the offspring to increase the diversity of solutions: x mutated =x+σ·randn Among them, σ is the mutation rate; randn is the function for generating standard normal distribution random numbers; Update the particle's position and velocity: v i (t+1)=ω·v i (t)+c1·r1·(p best,i -x i (t))+c2·r2·(g best -x i (t)) x i (t+1)=x i (t)+v i (t+1) Among them, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, p best,i is the individual optimal position of the ith particle, g best is the global best position; Dynamically adjust parameters based on the search progress: Among them, t is the current iteration number and T is the maximum iteration number.

10. An adaptive source-grid-load-storage integrated intelligent scheduling method, characterized in that: An adaptive source-grid-load-storage integrated intelligent dispatching system according to any one of claims 1 to 9 comprises the following steps: S1. Collect environmental and power parameters in real time from power sources, power grids, load devices and energy storage devices using sensors and measuring devices; S2. Predict future electricity demand through the electricity demand prediction model, and predict future renewable energy power generation through the renewable energy power generation prediction model; S3. Construct a multi-objective model based on cost, benefit and greenhouse gas emissions, as well as the constraints in the optimization process of the multi-objective model, including power balance constraints, equipment capacity constraints, grid security constraints and greenhouse gas emission constraints; S4. The multi-objective model is optimized by combining the hybrid optimization algorithm model GA-PSO with the genetic algorithm and the particle swarm optimization algorithm, and the algorithm parameters are dynamically adjusted in the multi-objective optimization process to search for the global optimal solution; S5. Output the optimized output power of power generation equipment, charging and discharging power of energy storage equipment and grid interaction power to guide scheduling decisions.

Images