A dynamic optimization method and equipment for wind-solar-electricity coordinated power supply in oilfields

Abstract

This application discloses a dynamic optimization method and equipment for wind-solar-electric coordinated power supply in oilfields, relating to the field of oil and gas field development technology. The method includes: based on the target oilfield well site... T Historical demand sequences of the power grid at all times are used to construct target oilfield well sites. T Real-time power grid demand forecasting model; targeting oilfield well sites T The historical demand sequence of the power grid at any given time is input into the demand forecasting model to obtain... T The application constructs a multi-objective optimization model for wind-solar-power coordinated energy supply based on the predicted demand values ​​of the power grid at various times. It then uses the dung beetle optimization algorithm to solve the model and obtain the optimal power grid strategy. T The real-time power grid demand prediction model and the multi-objective optimization model for wind-solar-electricity coordinated power supply have been used to achieve dynamic optimization of wind-solar-electricity coordinated power supply in oilfields, thereby reducing environmental pollution while improving the stability of the power supply system.

Description

Technical Field

[0001] This application relates to the field of oil and gas field development technology, and in particular to a dynamic optimization method and equipment for wind-solar-electricity coordinated power supply in oil fields. Background Technology

[0002] Traditional oil and gas field development processes demand large amounts of energy, often relying directly on the power grid for power supply. Thermal power remains the primary source of grid power, but its emissions of air pollutants contribute to environmental pollution. With the increasing maturity of clean energy generation technologies such as wind and solar power, multi-energy complementarity and coordinated power supply have become key measures to promote the green transformation and upgrading of oil and gas field development. Therefore, achieving localized consumption of clean energy to support clean energy use in oil and gas development is urgently needed. Thus, a dynamic optimization method for wind-solar-electricity coordinated power supply at oilfield sites is urgently required to reduce environmental pollution while improving the stability of the on-site power supply system, achieving energy conservation, emission reduction, cost reduction, and efficiency improvement during oilfield development, and promoting the green transformation of the oil and gas industry. Summary of the Invention

[0003] The purpose of this application is to provide a dynamic optimization method and equipment for wind-solar-electricity coordinated power supply in oilfields, which can improve the stability of the power supply system while reducing environmental pollution and solving the energy consumption problem in the oil and gas field development process.

[0004] To achieve the above objectives, this application provides the following solution:

[0005] Firstly, this application provides a dynamic optimization method for wind-solar-electricity coordinated power supply in oilfields, including: acquiring the target oilfield well site... T Historical demand sequence of the power grid at any given time.

[0006] Based on the target oilfield well site T Historical demand sequences of the power grid at all times are used to construct target oilfield well sites. T A time-based power grid demand forecasting model.

[0007] Target oilfield well site T The historical demand sequence of the power grid at any given time is input into the demand prediction model to obtain... T The forecast value of the power grid demand at any given time.

[0008] Based on the predicted demand, a multi-objective optimization model for wind-solar-electricity coordinated energy supply is constructed.

[0009] The dung beetle optimization algorithm is used to solve the multi-objective optimization model of wind-solar-electricity coordinated energy supply to obtain the optimal strategy of the power grid; the dung beetle optimization algorithm is a dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode.

[0010] The number of wind turbines, photovoltaic panels, and fossil fuel generators operating in the target oilfield well site is adjusted based on the aforementioned optimal power grid strategy.

[0011] Optionally, after adjusting the number of operating wind turbines, photovoltaic panels, and fossil fuel generators at the target oilfield well site based on the optimal grid strategy, the method further includes:

[0012] Acquire target oilfield well site T Measured values ​​of power grid demand at any given time;

[0013] The measured demand value is added as the last element to the historical demand sequence to obtain the target oilfield well site. T+1 Historical demand sequence of the power grid at any given time;

[0014] make T = T + 1 And return to the step "Target oilfield well site" T The historical demand sequence of the power grid at any given time is input into the demand prediction model to obtain... T "The demand forecast of the power grid at any given time."

[0015] Optionally, based on the target oilfield well site T Historical demand sequences of the power grid at all times are used to construct target oilfield well sites. T The demand forecasting model for the power grid at any given time includes:

[0016] Will STL Decomposition model, seasonal autoregressive integral moving average model, LSTM - Attention Model and error weight allocation: Model initialization, all model parameters are assigned a value of 0; STL The parameters of the decomposition model include the seasonal difference step size. S The parameters of the seasonal autoregressive integral moving average model include the order of the non-seasonal autoregression. p Period-by-period difference order d non-seasonal moving average order q Seasonal autoregression order P Seasonal difference order D and seasonal moving average order Q , LSTM - Attention The model parameters include window length. h The parameters of the error weighting model include the weights of the seasonal autoregressive integral moving average model and the weights of the residual correction model. The residual correction model is composed of... LSTM - Attention Obtained through model training and optimization;

[0017] Input the historical demand sequence into the initialized data. STL Decompose the model to obtain the seasonal difference step size. S Seasonal series and residual series;

[0018] The seasonal difference step size S The seasonal autoregressive integral moving average model initialized with the seasonal sequence input is used to determine the parameters of the seasonal autoregressive integral moving average model using the Akaike information criterion, thus obtaining the parameter-determined seasonal autoregressive integral moving average model.

[0019] Sure LSTM - Attention Model window length h Slide the window sequentially along the residual sequence, with the values ​​within the window... h The training takes one set of data as input and the first data outside the window as output. LSTM - Attention The model utilizes the stochastic gradient descent algorithm to... LSTM - Attention The model hyperparameters were optimized to obtain the residual correction model;

[0020] The seasonal autoregressive integral moving average model, with its seasonal series input parameters determined, yields the target oilfield well site. T Initial forecast of grid demand at any given time;

[0021] Inputting the residual sequence into the residual correction model yields the target oilfield well site. T The residual forecast value of the power grid demand at any given time;

[0022] The initial demand forecast, the residual demand forecast, the seasonal series, and the residual series are input into the error weight allocation model to obtain the weights of the seasonal autoregressive integral moving average model with determined parameters, as well as the weights of the residual correction model.

[0023] The target oilfield well site is obtained by multiplying the weights of the seasonal autoregressive integral moving average model determined by the parameters by the initial demand forecast. T The initial weighted forecast value of the power grid at any given time is used to multiply the weights of the residual correction model by the demand residual forecast value to obtain the target oilfield well site. T The residual weighted prediction value of the power grid at any given time is obtained by adding the initial weighted prediction value and the residual weighted prediction value. T Forecasted demand for the power grid at any given time;

[0024] Sure STL The decomposition model, the seasonal autoregressive integral moving average model with defined parameters, the residual correction model, and the error weight allocation model are used for the target oilfield well site. T A time-based power grid demand forecasting model.

[0025] Optionally, the target oilfield well site T The historical demand sequence of the power grid at any given time is input into the demand prediction model to obtain... T The demand forecast for the power grid at any given time includes:

[0026] Target oilfield well site T Historical demand sequences of the power grid at any given time are input into... STL Decompose the model to obtain the seasonal series and residual series;

[0027] The seasonal data series is input into a seasonal autoregressive integral moving average model with defined parameters to obtain the target oilfield well site. T Initial forecast of grid demand at any given time;

[0028] Inputting the residual sequence into the residual correction model yields the target oilfield well site. T The residual forecast value of the power grid demand at any given time;

[0029] The initial demand forecast, the residual demand forecast, the seasonal series, and the residual series are input into the error weight allocation model to obtain the weights of the seasonal autoregressive integral moving average model with determined parameters, as well as the weights of the residual correction model.

[0030] The target oilfield well site is obtained by multiplying the weights of the seasonal autoregressive integral moving average model determined by the parameters by the initial demand forecast. T The initial weighted forecast value of the power grid at any given time is used to multiply the weights of the residual correction model by the demand residual forecast value to obtain the target oilfield well site. T The residual weighted prediction value of the power grid at any given time is obtained by adding the initial weighted prediction value and the residual weighted prediction value. T The forecast value of the power grid demand at any given time.

[0031] Optionally, the wind-solar-electricity coordinated energy supply multi-objective optimization model includes an economic benefit objective function, an equivalent emission reduction objective function, optimization quantity constraints, equivalent quantity constraints, demand constraints, and integer constraints.

[0032] Optionally, the objective function for economic benefits is:

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] In the formula, EB For economic benefits, c o At the current oil price, C o This represents the cumulative oil production of the target oilfield well site within a unit of time. E wThis refers to the power generation per unit time of a single wind turbine generator. N w The number of wind turbine generators. E s This refers to the power generation per unit time of a single photovoltaic panel. N s The number of photovoltaic panels. E f This refers to the power generation per unit time of a single fossil fuel generator unit. N f The number of fossil fuel generator sets. P t Time-of-use electricity pricing for the target oilfield well site. C w Let α be the installation cost of a single wind turbine generator set, and α be the depreciation period of the wind turbine generator set. OP w The unit time operation and maintenance cost of a single wind turbine generator. C s β represents the installation cost of a single photovoltaic panel, and β represents the depreciation period of the photovoltaic generator set. OP s The unit time operation and maintenance cost of a single photovoltaic panel. C f γ represents the installation cost of a single fossil fuel generator set, and γ represents the depreciation period of the fossil fuel generator set. OP f The unit time operation and maintenance cost of a single fossil fuel generator unit. v 1 represents the cut-in speed of the wind turbine generator per unit time. v 2 represents the cut-out speed of the wind turbine generator per unit time. p ( v For wind turbine generator sets at wind speeds v Power generation at that time f ( v ( ) represents the probability distribution of wind speed at the hub height of a wind turbine generator. H A This represents the total solar radiation on a horizontal surface. P AZ The installation capacity of a single photovoltaic panel. E A1 The irradiance under standard conditions at the end of the previous unit of time. E A2 The irradiance under the initial standard conditions for the next unit time. f ( E A () is the irradiance variation function under standard conditions. K The overall efficiency coefficient, q The calorific value of coal per kilogram.N 1 This represents the coal consumption at the end of the previous unit of time. N 2 This represents the initial coal consumption for the next unit of time. f ( N ) is a coal consumption function used to describe the amount of coal consumed per unit time;

[0038] The objective function for the equivalent emission reduction is:

[0039] ;

[0040] ;

[0041] In the formula, EQE For equivalent emission reductions, C This refers to the carbon dioxide emissions produced per unit time by a single fossil fuel power generation unit. for T The number of fossil fuel generators that generate all of the grid's demand at any given time. Car The carbon content, a basic element of coal. OF The carbon oxidation rate of coal;

[0042] The optimized quantity constraints include optimizing the number of wind turbine generators, photovoltaic generators, and fossil fuel generators.

[0043] The optimization quantity constraint is:

[0044] ;

[0045] In the formula, N wmax It refers to the number of existing wind turbine generators at the target oilfield well site. N smax It refers to the number of existing photovoltaic panels at the target oilfield well site. N fmax It refers to the number of existing fossil fuel generator units at the target oilfield well site;

[0046] The equivalent quantity constraint for the fossil fuel generator sets is:

[0047] ;

[0048] The power generation demand constraint at time T is:

[0049] ;

[0050] In the formula, This represents the predicted demand of the power grid at time T.

[0051] The integer constraint is:

[0052] ;

[0053] In the formula, It is a positive integer.

[0054] Optionally, the dung beetle optimization algorithm is used to solve the multi-objective optimization model of wind-solar-power coordinated energy supply to obtain the optimal grid strategy, including:

[0055] Set the population size N and the maximum number of iterations (Iteration);

[0056] The objective functions for economic benefits and equivalent emission reductions are determined as fitness functions.

[0057] Let the number of iterations t=1;

[0058] Based on the population size N, the dung beetle population is initialized using chaotic mapping, and the initialized population is used as the dung beetle population at the t-th iteration.

[0059] Calculate the fitness value of individuals in the dung beetle population at the t-th iteration;

[0060] Calculate the non-dominated solutions in the dung beetle population at the t-th iteration and store them in the Pareto front;

[0061] Calculate the crowding distance for each non-dominated solution in the Pareto front to obtain multiple crowding distances;

[0062] The dung beetle population at iteration t+1 is updated using dancing, rolling, foraging, and stealing strategies.

[0063] Calculate the fitness value of individuals in the undetermined dung beetle population at the (t+1)th iteration;

[0064] The dung beetle population at the (t+1)th iteration is determined based on the fitness values ​​of individuals in the undetermined dung beetle population at the (t+1)th iteration and the fitness values ​​of individuals in the dung beetle population at the (t)th iteration.

[0065] Calculate the non-dominated solutions in the dung beetle population at the (t+1)th iteration and store them in the Pareto front;

[0066] Calculate the crowding distance for each nondominated solution in the Pareto front, and prioritize retaining nondominated solutions with larger crowding distances;

[0067] Determine whether the iteration number t is greater than or equal to the maximum iteration number Iteration, and obtain the determination result;

[0068] If the judgment result is negative, then let t = t + 1, and return to the step "update the dung beetle population at the (t+1)th iteration using the dancing strategy, rolling strategy, foraging strategy, and stealing strategy";

[0069] If the judgment result is yes, then the Pareto front is output as the optimal strategy for the power grid.

[0070] Optionally, the foraging strategy is:

[0071] Calculate the oviposition area and oviposition location of a dung beetle population:

[0072] ;

[0073] ;

[0074] ;

[0075] In the formula, For the current number t The local optimum position of the dung beetle population in the next iteration; This marks the lower boundary of the spawning area. This is the upper boundary of the spawning area; R The boundary convergence factor; Lb This serves as the lower bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. Ub This serves as the upper bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. f b Fitness value of an individual in a dung beetle population The maximum value, f w Fitness value of an individual in a dung beetle population The minimum value;

[0076] The position of dung beetle egg masses is updated using a dynamic sine and cosine update strategy:

[0077] ;

[0078] ;

[0079] In the formula, B i ( t+1 ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t+1 The location of spawning in the next iteration; B i ( t ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t The location of spawning in the next iteration; b 1 and b2 represents two sizes, 1× D Independent random vectors; D Dimensions for optimizing the wind-solar-power coordinated energy supply model; l A random number in the range [-1, 1]. r For dynamic sine and cosine shape parameters; m A random number in the range [-1, 1]. n The constants that determine the shape of the dynamic sine and cosine waves;

[0080] Dung beetle eggs hatch into baby dung beetles; the optimal foraging area for the baby dung beetles is calculated.

[0081] ;

[0082] ;

[0083] In the formula, X b for t The position of the global optimum in the next iteration; Lb b This is the lower boundary of the optimal foraging area; Ub b This is the upper limit of the optimal foraging area;

[0084] The position of the dung beetle is updated using an iterative strategy based on the golden section line.

[0085] ;

[0086] In the formula, x i ( t+ 1) is the [number]th species in the dung beetle population. i The little dung beetle t+ The position of the first iteration; x i ( t ) is the [number]th [number] species in the dung beetle population. i The little dung beetle t The position of the next iteration; R 1 is Random numbers; R 2 is Random numbers; C 1 and C 2 represents the coefficient introduced by the golden ratio;

[0087] ;

[0088] .

[0089] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned dynamic optimization method for wind-solar-electricity coordinated power supply in oilfields.

[0090] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0091] This application provides a dynamic optimization method and equipment for wind-solar-electric coordinated energy supply in oilfields, which obtains the historical demand sequence of the target oilfield well site and predicts... T The demand of the power grid at any time, according to T An optimization model is constructed based on the real-time grid demand and the corresponding conditions of the target oilfield well site. The dung beetle optimization algorithm, based on a high-dimensional swarm intelligence search model, is used to solve the optimization model, thereby adjusting the number of wind turbines, photovoltaic panels, and fossil fuel generators at the target oilfield well site. The adjusted model yields... T The system continuously monitors the actual demand of the power grid and updates the historical demand sequence of the target oilfield well site to achieve dynamic optimization.

[0092] This application is based on the historical demand sequence of the target oilfield well site and predictions. T Real-time grid demand, and obtained after adjustment T The system continuously monitors the actual demand of the power grid and updates the historical demand sequence of the target oilfield well site. This achieves dynamic optimization while simultaneously increasing the length of the historical demand sequence for the target oilfield well site, thus improving prediction accuracy. T The accuracy of real-time grid demand is constantly improving. This paper utilizes the dung beetle optimization algorithm based on a high-dimensional swarm intelligent search model to optimize the number of operating wind turbines, photovoltaic panels, and fossil fuel generators in a target oilfield well site. The high-dimensional swarm intelligent search model introduces chaotic mapping initialization, adaptive population fitness convergence factors, dynamic sine and cosine update strategies, and golden tangent iteration strategies, achieving efficient global optimization while effectively avoiding getting trapped in local optima. Attached Figure Description

[0093] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0094] Figure 1 This is a flowchart of a dynamic optimization method for wind-solar-electric coordinated energy supply in an oilfield, as described in one embodiment of this application.

[0095] Figure 2This is a framework diagram of a power grid demand prediction model at time T in one embodiment of this application.

[0096] Figure 3 This is a flowchart of a dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode in one embodiment of this application.

[0097] Figure 4 This is a stacked bar chart showing the historical demand of the target oilfield well site, the power generation per unit time of fossil fuel generator sets, wind turbine generator sets, and photovoltaic panels in one embodiment of this application.

[0098] Figure 5 This is a line graph showing the historical demand and predicted value of the target oilfield well site in one embodiment of this application.

[0099] Figure 6 This is a Pareto front diagram obtained by the dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode in one embodiment of this application. Detailed Implementation

[0100] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0101] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0102] In one exemplary embodiment, such as Figure 1 As shown, a dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields is provided, including:

[0103] Step 101: Obtain the target oilfield well site T Historical demand sequence of the power grid at any given time.

[0104] Step 102: Based on the target oilfield well site T Historical demand sequences of the power grid at all times are used to construct target oilfield well sites. T A time-based power grid demand forecasting model.

[0105] Step 103: Target oilfield well site T The historical demand sequence of the power grid at any given time is input into the demand forecasting model to obtain... T The forecast value of the power grid demand at any given time.

[0106] Step 104: Based on the demand forecast, construct a multi-objective optimization model for wind-solar-electricity coordinated energy supply.

[0107] Step 105: Solve the multi-objective optimization model of wind-solar-electricity coordinated energy supply using the dung beetle optimization algorithm to obtain the optimal power grid strategy. The dung beetle optimization algorithm is a dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode.

[0108] Step 106: Adjust the number of wind turbine generators, photovoltaic panels and fossil fuel generators operating in the target oilfield well site based on the optimal grid strategy.

[0109] Step 107: Obtain the target oilfield well site T The measured value of the power grid demand at any given time.

[0110] Step 108: Add the measured demand value as the last element to the historical demand sequence to obtain the target oilfield well site. T+1 Historical demand sequence of the power grid at any given time.

[0111] Step 109: Let T = T + 1 Then return to the step "Input the historical demand sequence of the power grid at time T of the target oilfield well site into the demand prediction model to obtain the predicted demand value of the power grid at time T".

[0112] Step 102 includes:

[0113] Step 102-1: Apply the seasonal and trend decomposition algorithm based on local weighted regression ( STL Decomposition model), seasonal autoregressive integral moving average model, LSTM - Attention Model and error weight allocation: Model initialization, all model parameters are assigned a value of 0. STL The parameters of the decomposition model include the seasonal difference step size. S The parameters of the seasonal autoregressive integral moving average model include the order of the non-seasonal autoregression. p Period-by-period difference order d non-seasonal moving average order q Seasonal autoregression order P Seasonal difference order D and seasonal moving average order Q , LSTM - Attention The model parameters include window length. h The parameters of the error weighting model include the weights of the seasonal autoregressive integral moving average model and the weights of the residual correction model. The residual correction model is composed of... LSTM - Attention Obtained through model training and optimization.

[0114] Step 102-2: Input the historical demand sequence into the initialized data. STL Decompose the model to obtain the seasonal difference step size. S Seasonal series and residual series.

[0115] Step 102-3: Adjust the seasonal difference step size S The seasonal autoregressive integral moving average model, initialized with seasonal series input, is used to determine the parameters of the seasonal autoregressive integral moving average model using the Akaike information criterion, resulting in a parameter-determined seasonal autoregressive integral moving average model.

[0116] Step 102-4: Confirm LSTM - Attention Model window length h Slide the window sequentially along the residual sequence, and within the window... h The training takes one set of data as input and the first data outside the window as output. LSTM - Attention The model utilizes the stochastic gradient descent algorithm to... LSTM - Attention The model hyperparameters were optimized to obtain the residual correction model.

[0117] Step 102-5: Input the seasonal series data into the seasonal autoregressive integral moving average model with determined parameters to obtain the target oilfield well site. T Initial forecast of the power grid demand at any given time.

[0118] Step 102-6: Input the residual sequence into the residual correction model to obtain the target oilfield well site. T The residual forecast value of the power grid demand at any given time.

[0119] Step 102-7: Input the initial demand forecast, the demand residual forecast, the seasonal series, and the residual series into the error weight allocation model to obtain the weights of the seasonal autoregressive integral moving average model with determined parameters, as well as the weights of the residual correction model.

[0120] Step 102-8: Multiply the weights of the seasonal autoregressive integral moving average model with the determined parameters by the initial demand forecast to obtain the target oilfield well site. T The initial weighted forecast value of the power grid at any given time is used to multiply the weights of the residual correction model by the demand residual forecast value to obtain the target oilfield well site. T The residual weighted prediction value of the power grid at any given time is obtained by adding the initial weighted prediction value and the residual weighted prediction value. T The forecast value of the power grid demand at any given time.

[0121] Step 102-9: Confirm STL The decomposition model, the seasonal autoregressive integral moving average model with defined parameters, the residual correction model, and the error weight allocation model are used for the target oilfield well site. T A time-based power grid demand forecasting model.

[0122] Step 103 includes:

[0123] Step 103-1: Target oilfield well siteT Historical demand sequences of the power grid at any given time are input into... STL Decompose the model to obtain the seasonal series and residual series;

[0124] Step 103-2: Input the seasonal series into a seasonal autoregressive integral moving average model with defined parameters to obtain the target oilfield well site. T Initial forecast of grid demand at any given time;

[0125] Step 103-3: Input the residual sequence into the residual correction model to obtain the target oilfield well site. T The residual forecast value of the power grid demand at any given time;

[0126] Step 103-4: Input the initial demand forecast, the demand residual forecast, the seasonal series and the residual series into the error weight allocation model to obtain the weights of the seasonal autoregressive integral moving average model with determined parameters, and the weights of the residual correction model.

[0127] Step 103-5: Multiply the weights of the seasonal autoregressive integral moving average model determined by the parameters by the initial demand forecast to obtain the target oilfield well site. T The initial weighted forecast value of the power grid at any given time is used to multiply the weights of the residual correction model by the demand residual forecast value to obtain the target oilfield well site. T The residual weighted prediction value of the power grid at any given time is obtained by adding the initial weighted prediction value and the residual weighted prediction value. T The forecast value of the power grid demand at any given time.

[0128] The multi-objective optimization model for wind-solar-electricity coordinated energy supply includes an economic benefit objective function, an equivalent emission reduction objective function, optimization quantity constraints, equivalent quantity constraints, demand constraints, and integer constraints.

[0129] The objective function for economic benefits is:

[0130] .

[0131] .

[0132] .

[0133] .

[0134] In the formula, EB For economic benefits, c o At the current oil price, C o This represents the cumulative oil production of the target oilfield well site within a unit of time. E wThis refers to the power generation per unit time of a single wind turbine generator. N w The number of wind turbine generators. E s This refers to the power generation per unit time of a single photovoltaic panel. N s The number of photovoltaic panels. E f This refers to the power generation per unit time of a single fossil fuel generator unit. N f The number of fossil fuel generator sets. P t Time-of-use electricity pricing for the target oilfield well site. C w Let α be the installation cost of a single wind turbine generator set, and α be the depreciation period of the wind turbine generator set. OP w The unit time operation and maintenance cost of a single wind turbine generator. C s β represents the installation cost of a single photovoltaic panel, and β represents the depreciation period of the photovoltaic generator set. OP s The unit time operation and maintenance cost of a single photovoltaic panel. C f γ represents the installation cost of a single fossil fuel generator set, and γ represents the depreciation period of the fossil fuel generator set. OP f The unit time operation and maintenance cost of a single fossil fuel generator unit. v 1 represents the cut-in speed of the wind turbine generator per unit time. v 2 represents the cut-out speed of the wind turbine generator per unit time. p ( v For wind turbine generator sets at wind speeds v Power generation at that time f ( v ( ) represents the probability distribution of wind speed at the hub height of a wind turbine generator. H A This represents the total solar radiation on a horizontal surface. P AZ The installation capacity of a single photovoltaic panel. E A1 The irradiance under standard conditions at the end of the previous unit of time. E A2 The irradiance under the initial standard conditions for the next unit time. f ( E A () is the irradiance variation function under standard conditions. K The overall efficiency coefficient, q The calorific value of coal per kilogram.N 1 This represents the coal consumption at the end of the previous unit of time. N 2 This represents the initial coal consumption for the next unit of time. f ( N ) is a coal consumption function used to describe the amount of coal consumed per unit time.

[0135] The objective function for equivalent emission reduction is:

[0136] .

[0137] .

[0138] In the formula, EQE For equivalent emission reductions, C This refers to the carbon dioxide emissions produced per unit time by a single fossil fuel power generation unit. for T The number of fossil fuel generators that generate all of the grid's demand at any given time. Car The carbon content, a basic element of coal. OF The carbon oxidation rate of coal.

[0139] Optimizing quantity constraints includes optimizing the number of wind turbine generators, photovoltaic generators, and fossil fuel generators.

[0140] The optimized quantity constraint is:

[0141] ;

[0142] In the formula, N wmax It refers to the number of existing wind turbine generators at the target oilfield well site. N smax It refers to the number of existing photovoltaic panels at the target oilfield well site. N fmax It refers to the number of existing fossil fuel generator units at the target oilfield well site.

[0143] The equivalent quantity constraint for fossil fuel power generation units is:

[0144] .

[0145] The power generation demand constraint at time T is:

[0146] .

[0147] In the formula, Let T be the predicted demand of the power grid at time T.

[0148] Integer constraints are:

[0149] .

[0150] In the formula, It is a positive integer.

[0151] Step 105 includes:

[0152] Step 105-1: Set the population size N and the maximum number of iterations (Iteration).

[0153] Step 105-2: Determine the fitness function as the objective function for economic benefits and the objective function for equivalent emission reduction.

[0154] Step 105-3: Let the iteration number t=1.

[0155] Step 105-4: Based on the population size N, initialize the dung beetle population using chaotic mapping, and use the initialized population as the dung beetle population at the t-th iteration.

[0156] Step 105-5: Calculate the fitness value of individuals in the dung beetle population at the t-th iteration.

[0157] Step 105-6: Calculate the non-dominated solutions in the dung beetle population at the t-th iteration and save them in the Pareto front.

[0158] Step 105-7: Calculate the crowding distance for each non-dominated solution in the Pareto front to obtain multiple crowding distances.

[0159] Steps 105-8: Update the dung beetle population at iteration t+1 using the dancing, rolling, foraging, and stealing strategies.

[0160] Step 105-9: Calculate the fitness value of individuals in the undetermined dung beetle population at the (t+1)th iteration.

[0161] Steps 105-10: Determine the dung beetle population at the (t+1)th iteration based on the fitness values ​​of individuals in the undetermined dung beetle population at the (t+1)th iteration and the fitness values ​​of individuals in the dung beetle population at the (t)th iteration.

[0162] Steps 105-11: Calculate the non-dominated solutions in the dung beetle population at the (t+1)th iteration and save them in the Pareto front.

[0163] Step 105-12: Calculate the crowding distance of each nondominated solution in the Pareto front, and prioritize retaining nondominated solutions with larger crowding distances.

[0164] Step 105-13: Determine whether the iteration number t is greater than or equal to the maximum iteration number Iteration, and obtain the determination result.

[0165] Step 105-14: If the result is negative, let t = t + 1 and return to step 105-8.

[0166] Step 105-15: If the judgment result is yes, then output the Pareto front as the optimal strategy for the power grid.

[0167] The convergence factor for the adaptive population fitness is:

[0168] .

[0169] In the formula, R is the adaptive population fitness convergence factor; f b Fitness value of an individual in a dung beetle population The maximum value; f w Fitness value of an individual in a dung beetle population The minimum value.

[0170] Foraging strategies are as follows:

[0171] Calculate the oviposition area of ​​the dung beetle population and record the oviposition.

[0172] .

[0173] .

[0174] .

[0175] In the formula, For the current number t The local optimum position of the dung beetle population in the next iteration; This marks the lower boundary of the spawning area. This is the upper boundary of the spawning area; R This is the boundary convergence factor (its value is the adaptive population fitness convergence factor). Lb This serves as the lower bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. Ub This serves as the upper bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. f b Fitness value of an individual in a dung beetle population The maximum value, f w Fitness value of an individual in a dung beetle population The minimum value.

[0176] The position of dung beetle egg masses is updated using a dynamic sine and cosine update strategy:

[0177] .

[0178] .

[0179] In the formula, B i ( t+1 ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t+1 The location of spawning in the next iteration; B i ( t ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t The location of spawning in the next iteration; b 1 and b 2 represents two sizes, 1× D Independent random vectors; D Dimensions for optimizing the wind-solar-power coordinated energy supply model; l A random number in the range [-1, 1]. r For dynamic sine and cosine shape parameters; m A random number in the range [-1, 1]. n The constants that determine the shape of the dynamic sine and cosine.

[0180] Dung beetle eggs hatch into baby dung beetles; the optimal foraging area for the baby dung beetles is calculated.

[0181] .

[0182] .

[0183] In the formula, X b for t The position of the global optimum in the next iteration; Lb b This is the lower bound of the optimal foraging area (the optimal area for offspring); Ub b This is the upper limit of the optimal foraging area.

[0184] The position of the dung beetle is updated using an iterative strategy based on the golden section line.

[0185] .

[0186] In the formula, x i ( t+ 1) is the [number]th species in the dung beetle population. i The little dung beetle t+ The position of the first iteration; x i ( t ) is the [number]th [number] species in the dung beetle population. i The little dung beetle t The position of the next iteration; R 1 is Random numbers; R 2 is Random numbers; C 1 and C 2 represents the coefficient introduced by the golden ratio.

[0187] .

[0188] .

[0189] The embodiment provides a dynamic optimization method for wind-solar-electric synergistic energy supply in oilfields, which can quickly and dynamically adjust the number of wind turbine generators, photovoltaic panels and fossil fuel power generation equipment to achieve a win-win situation for economic and environmental benefits.

[0190] In one exemplary embodiment, a dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields is provided, comprising:

[0191] Step 1: Obtain the historical demand sequence of the target oilfield well site and establish... T Training a demand forecasting model for the power grid at any time. T A time-based power grid demand forecasting model was developed to predict the target oilfield well site. T Real-time grid demand, such as Figure 2 .

[0192] S11. Obtain the historical demand sequence of the target oilfield well site. .

[0193] S12. Establish T A model for predicting grid demand at any given time.

[0194] S121. Input the historical demand sequence of the target oilfield well site. ,pass STL Decomposition determines the seasonal difference step size in the seasonal autoregressive integral moving average model (i.e., the SARIMA model). S Seasonal sequence and residual sequence .

[0195] seasonal difference step size S and seasonal sequence Input the seasonal autoregressive integral moving average model, and use the Akaike Information Criterion to determine other parameters of the seasonal autoregressive integral moving average model, thus obtaining the seasonal autoregressive integral moving average model with determined parameters.

[0196] Sure LSTM - Attention Model window length h Slide the window sequentially along the residual sequence, and within the window... hThe training takes one set of data as input and the first data outside the window as output. LSTM - Attention The model utilizes the stochastic gradient descent algorithm to... LSTM - Attention The model hyperparameters were optimized to obtain the residual correction model.

[0197] Seasonal sequence Given a seasonal autoregressive integral moving average model with defined input parameters, we obtain... T Initial forecast of grid demand at any time .

[0198] residual sequence Input the residual correction model and obtain T Real-time grid demand residual forecast .

[0199] Will T Initial forecast of grid demand at any time , T Real-time grid demand residual forecast Seasonal sequence and residual sequence The input error weighting model yields the weights of the seasonal autoregressive integral moving average model with defined parameters. Weights of the residual correction model The error weight allocation model sets different weights according to the second formula. Substituting back into the first formula of the error weight allocation model, the error function is calculated. When R is minimized, the weights of the two models are obtained. The error weight allocation model is as follows:

[0200] .

[0201] .

[0202] In the formula, Let be the error function. weight Lr The weights of the seasonal autoregressive integral moving average model for which parameters are determined. weight Err These are the weights for the residual correction model.

[0203] Sure STL The decomposition model, the seasonal autoregressive integral moving average model with defined parameters, the residual correction model, and the error weight allocation model are used for the target oilfield well site. T A time-based power grid demand forecasting model.

[0204] S13. Constructed using S12 T The time-based power grid demand forecasting model takes historical demand sequences as input and generates forecasts. T Real-time grid demand .

[0205] Step 2: Based on the predictions obtained in Step 1 T Real-time grid demand We set the objective function and constraints, and constructed a multi-objective optimization model for wind-solar-electricity coordinated energy supply.

[0206] S21. Specifically, the objective functions of the wind-solar-electricity coordinated energy supply multi-objective optimization model are the economic benefit objective function and the equivalent emission reduction objective function.

[0207] S211. Specifically, the economic benefits are as follows:

[0208] ;

[0209] in, ;

[0210] ;

[0211] ;

[0212] In the formula, EB For economic benefits, c o At the current oil price, C o This represents the cumulative oil production of the target oilfield well site within a unit of time. E w This refers to the power generation per unit time of a single wind turbine generator. N w The number of wind turbine generators. E s This refers to the power generation per unit time of a single photovoltaic panel. N s The number of photovoltaic panels. E f This refers to the power generation per unit time of a single fossil fuel generator unit. N f The number of fossil fuel generator sets. P t Time-of-use electricity pricing for the target oilfield well site. C w Let α be the installation cost of a single wind turbine generator set, and α be the depreciation period of the wind turbine generator set. OP w The unit time operation and maintenance cost of a single wind turbine generator. C s β represents the installation cost of a single photovoltaic panel, and β represents the depreciation period of the photovoltaic generator set. OP sThe unit time operation and maintenance cost of a single photovoltaic panel. C f γ represents the installation cost of a single fossil fuel generator set, and γ represents the depreciation period of the fossil fuel generator set. OP f The unit time operation and maintenance cost of a single fossil fuel generator unit. v 1 represents the cut-in speed of the wind turbine generator per unit time. v 2 represents the cut-out speed of the wind turbine generator per unit time. p ( v For wind turbine generator sets at wind speeds v Power generation at that time f ( v ( ) represents the probability distribution of wind speed at the hub height of a wind turbine generator. H A This represents the total solar radiation on a horizontal surface. P AZ The installation capacity of a single photovoltaic panel. E A1 The irradiance under standard conditions at the end of the previous unit of time. E A2 The irradiance under the initial standard conditions for the next unit time. f ( E A () is the irradiance variation function under standard conditions. K The overall efficiency coefficient, q The calorific value of coal per kilogram. N 1 This represents the coal consumption at the end of the previous unit of time. N 2 This represents the initial coal consumption for the next unit of time. f ( N ) is a function of coal consumption, where is the amount of coal consumed per unit time.

[0213] Specifically, the calculation method for the power generation per unit time of a single wind turbine generator is as follows:

[0214] .

[0215] In the formula, E w This refers to the power generation per unit time of a single wind turbine generator. v 1 represents the cut-in speed of the wind turbine generator set; v 2 represents the cut-out speed of the wind turbine generator set; p ( v For wind turbine generator sets at wind speeds v Power generation at that time; This represents the probability distribution of wind speed at the hub height of a wind turbine generator.

[0216] Specifically, the calculation method for the power generation per unit time of a single photovoltaic panel is as follows:

[0217] .

[0218] In the formula, E s This refers to the power generation per unit time of a single photovoltaic panel. H A This represents the total solar radiation on a horizontal surface. P AZ The installation capacity of a single photovoltaic panel; E A 1 represents the irradiance under standard conditions; K is the overall efficiency coefficient, typically taken as 0.65~0.85.

[0219] Specifically, the calculation method for the power generation per unit time of a single fossil fuel generator unit is as follows:

[0220] .

[0221] In the formula, E f The amount of electricity generated per unit time by a single fossil fuel generator unit; q The calorific value per kilogram of coal is generally taken as 20929~29300 kcal / kg. KJ / Kg ; N This represents the amount of coal consumed per unit of time.

[0222] S212. Specifically, the equivalent emission reduction... EQE for:

[0223] .

[0224] In the formula, EQE represents the equivalent emission reduction. C This refers to the carbon dioxide emissions generated per unit of electricity produced by a single fossil fuel generator unit. for T The number of fossil fuel generators that generate all of the grid's demand at any given time.

[0225] Specifically, the carbon dioxide emissions generated per unit time by a single fossil fuel power generation unit are calculated as follows:

[0226] .

[0227] In the formula, C The carbon dioxide produced per unit time of electricity generated by a single fossil fuel generator unit. Car The carbon content, a basic element of coal; OF The carbon oxidation rate of coal.

[0228] S22. Specifically, the constraints of the wind-solar-electricity coordinated energy supply multi-objective optimization model are the optimization quantity constraints, equivalent quantity constraints, demand constraints, and integer constraints of the power generation equipment in the target oilfield well site.

[0229] S221. Specifically, the optimal quantity constraints for wind turbine generators, photovoltaic panels, and fossil fuel generators within the target oilfield well site are as follows:

[0230] .

[0231] In the formula N wmax It refers to the number of existing wind turbine generators at the target oilfield well site. N smax It refers to the number of existing photovoltaic panels at the target oilfield well site. N fmax It refers to the number of existing fossil fuel generator units at the target oilfield well site.

[0232] S222. Specifically, within the target oilfield well site T The constraint for fossil fuel generators that generate all the electricity demand of the grid at any given time is:

[0233] .

[0234] S223. Specifically, the demand constraints within the target oilfield well site are as follows:

[0235] .

[0236] S224. Specifically, the integer constraint is:

[0237] .

[0238] In the formula, It is a positive integer.

[0239] S23. Specifically, construct a multi-objective optimization model for wind-solar-electric coordinated energy supply:

[0240] .

[0241] Step 3: Based on the wind-solar-electricity coordinated energy supply optimization model constructed in Step 2, the optimization model is solved using the dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode, such as... Figure 3 .

[0242] S31. Set population size N and number of iterations Iteration .

[0243] S32. Dung beetle populations are initialized using chaotic mapping. Dung beetle populations initialized using chaotic mapping have a more uniform distribution of individuals and richer population diversity.

[0244] Chaotic mapping includes, but is not limited to Chebyshev , Circle , 4]Logistic , Tent , Piecewise and Fuch Methods. These mapping methods are all based on corresponding functional expressions: Chebyshev mapping is based on Chebyshev polynomials, Circle mapping is based on nonlinear functions of circular motion, Logistic mapping is based on quadratic polynomials, Tent mapping is based on piecewise linear functions divided into two segments by periodic factors, Piecewise mapping is based on piecewise linear functions divided into four segments by piecewise control factors, and Fuch mapping is based on one-dimensional discrete functions of cosine functions.

[0245] S33. Calculate the fitness value of individuals in the dung beetle population based on S211 and S212. .

[0246] S34. Calculate the non-dominated solutions in the population based on the initial population and store them in the Pareto front. A non-dominated solution is one where no other solution is better than the current solution on all objective functions. The Pareto front is the set of solutions formed by the non-dominated solutions.

[0247] S35. Calculate the crowding distance for each nondominated solution in the Pareto front. The distance between each nondominated solution and other nondominated solutions is the crowding distance.

[0248] S36. Update the individual positions of the dung beetle population initialized in S32 using dancing, rolling, foraging, and stealing strategies. Specifically:

[0249] S361. The dung beetle population contains *Gnaphalium affine*. Based on S32, initialize the individual positions of the dung beetle population, determine if any obstacles exist, and update the positions using a dancing strategy if present, otherwise using a rolling strategy. The specific method for determining if obstacles exist is to check if the individual *Gnaphalium affine* is located on the boundary. The specific methods for updating individual positions using the dancing and rolling strategies are as follows:

[0250] .

[0251] .

[0252] .

[0253] In the formula, x i ( t+1) is the [number]th species in the dung beetle population. i Only rolling dung beetle t+ The position of the first iteration; x i ( t ) is the [number]th [number] species in the dung beetle population. i Only rolling dung beetle t The position of the next iteration; x i ( t- 1) is the [number]th species in the dung beetle population. i Only rolling dung beetle t- The position of the first iteration; Due to the influence of natural conditions, the value is -1 or 1; Influenced by light source conditions; for t The position of the worst global position in the next iteration; From the perspective of the dung beetle's dancing strategy, Values ; k and b It is a fixed value.

[0254] S362. An egg-laying dung beetle exists within the dung beetle population. The individual positions of the dung beetle population are initialized according to S32. After the egg-laying dung beetles lay their eggs, the hatched dung beetles forage for food. The specific foraging strategy for updating individual positions is as follows:

[0255] Calculate the oviposition area of ​​the dung beetle population and record the oviposition.

[0256] .

[0257] .

[0258] In the formula, in the formula, For the current number t The local optimum position of the dung beetle population in the next iteration; This marks the lower boundary of the spawning area. This is the upper boundary of the spawning area; R The boundary convergence factor; Lb This serves as the lower bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. Ub This serves as the upper bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. f b Fitness value of an individual in a dung beetle population The maximum value, f w Fitness value of an individual in a dung beetle population The minimum value. Specifically, the formula for calculating the boundary convergence factor in the original dung beetle optimization algorithm is: This application replaces the adaptive population fitness convergence factor with a different one. The amplitude of the replaced convergence factor is exchanged, resulting in a more stable search space. The formula for calculating the boundary convergence factor in the dung beetle optimization algorithm after the replacement is as follows:

[0259] The position of dung beetle egg balls is updated using a dynamic sine and cosine update strategy.

[0260] .

[0261] .

[0262] In the formula, B i ( t+1 ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t+1 The location of spawning in the next iteration; B i ( t ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t The location of spawning in the next iteration; b 1 and b 2 represents two sizes, 1× D Independent random vectors; D Dimensions for optimizing the wind-solar-power coordinated energy supply model; l A random number in the range [-1, 1]. r For dynamic sine and cosine shape parameters; m A random number in the range [-1, 1]. n The constants that determine the shape of the dynamic sine and cosine.

[0263] Dung beetle eggs hatch into baby dung beetles, and the optimal foraging area for the baby dung beetles is calculated; the baby dung beetles are the offspring, and the dung beetle population is the parent generation, which is a renewal process.

[0264] .

[0265] .

[0266] In the formula, X b For the current number t The position of the global optimum in the next iteration; Lb b This is the lower bound of the optimal foraging area (the optimal area for offspring); Ub b This is the upper limit of the optimal foraging area.

[0267] The position of the dung beetle is updated using an iterative strategy based on the golden section line.

[0268] .

[0269] In the formula, x i ( t+ 1) is the [number]th species in the dung beetle population. i The little dung beetle t+ The position of the first iteration; x i ( t ) is the [number]th [number] species in the dung beetle population. i The little dung beetle t The position of the next iteration; R 1 is Random numbers; R 2 is Random numbers; C 1 and C 2 represents the coefficient introduced by the golden ratio. , .

[0270] S363. A dung beetle population contains thieving cockroaches. Based on the initialization of individual dung beetle positions in S32, the thieving cockroaches prey on the food of smaller dung beetles, updating their individual positions through a theft strategy. The specific method for updating individual positions using the theft strategy is as follows:

[0271] .

[0272] In the formula, x i ( t+ 1) is the [number]th species in the dung beetle population. i Only stealing dung beetles t+ The position of the first iteration; x i ( t ) is the [number]th [number] species in the dung beetle population. i Only stealing dung beetles t The position of the next iteration; X b The best location for food; g 1× D Independent random vectors; D Dimensions of optimization variables for the wind-solar-electricity coordinated energy supply optimization model; S It is a constant.

[0273] S37. Calculate the new fitness value of individuals in the dung beetle population based on S211 and S212. The optimal position of the dung beetle population is updated based on the fitness value to obtain a new dung beetle population.

[0274] S38. Based on the new dung beetle population, calculate the new non-dominated solutions in the new dung beetle population and save them in the Pareto front. Recalculate the crowding distance of each non-dominated solution in the Pareto front and prioritize the retention of non-dominated solutions with larger crowding distances.

[0275] S39.Judgment t Is it greater than or equal to? Iteration If so, output the Pareto front result, such as... Figure 6 If not, return to S36 and set t = t + 1.

[0276] Step 4: Based on the solution results of the optimization model, adjust the number of wind turbine generators, photovoltaic panels and fossil fuel generators operating in the target oilfield well site.

[0277] Step 5: Record T Real demand of power grid at any time Update the historical demand sequence of the target oilfield well site and re-execute steps 1-4 to achieve dynamic optimization.

[0278] The following section uses the historical demand sequence of the target oilfield well site shown in Table 1 as an example to illustrate a dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields as shown in this embodiment.

[0279] Step 1: Obtain the historical demand sequence of the target oilfield well site and establish... T A real-time power grid demand forecasting model takes historical demand sequences as input and forecasts the target oilfield well sites. T The grid demand at any given time. The historical demand sequence for the target oilfield well site is shown in Table 1. A stacked bar chart showing the historical power generation of the target oilfield well site, the power generation per unit time of fossil fuel generators, wind turbines, and photovoltaic panels is shown below. Figure 4 As shown, all of them meet the historical demand of the power grid.

[0280] Table 1 Historical Demand Sequence of Target Oilfield Well Sites

[0281]

[0282] Establish T The real-time grid demand prediction model, in this embodiment, is achieved through... STL Decomposition to determine seasonal difference step size S =24 and obtain the seasonal series and residual series; determine the Akaike information criterion. SARIMA The order of non-seasonal autoregression in the model p =2. Period-by-period difference order d =1. Order of non-seasonal moving average q =1. Seasonal autoregressive order P =1. Order of seasonal differenceD =0 and seasonal moving average order Q =1. The SARIMA model was used to predict the seasonal sequence, and LSTM-Attention was used to predict the residual sequence. The model weights were calculated, and the final prediction results are shown in Table 2. The historical demand and predicted values ​​of the target oilfield well site are as follows: Figure 5 As shown.

[0283] Table 2. Forecast Results of Historical Demand Series for Target Oilfield Well Sites

[0284]

[0285] Based on the above model prediction T Grid demand at time 25 hours =6507.58 kWh

[0286] Step 2: Based on the predictions obtained in Step 1 T Real-time grid demand We set the objective function and constraints, and constructed a multi-objective optimization model for wind-solar-electricity coordinated energy supply.

[0287] The specific parameters of the target oilfield well site in this embodiment of the invention are shown in Table 3, which are used to construct a multi-objective optimization model for wind-solar-electricity coordinated energy supply.

[0288] Table 3. Data on various parameters of the target oilfield well site

[0289]

[0290] Based on the above parameters, a multi-objective optimization model for wind-solar-electricity coordinated energy supply is constructed:

[0291] .

[0292] Step 3: Based on the wind-solar-electricity coordinated energy supply optimization model constructed in Step 2, the dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode is used to solve the optimization model.

[0293] Set population size N=50 and number of iterations Iteration = 20 The population is initialized using the Tent chaotic mapping, and the Pareto front graph, obtained using steps S33-S39 and based on the dung beetle optimization algorithm of a high-dimensional swarm intelligence search model, is shown below. Figure 6 As shown

[0294] Step 4: Based on the solution results of the optimization model, select the solution result with the maximum equivalent emission reduction, and adjust the number of wind turbine generators, photovoltaic panels and fossil fuel generators operating in the target oilfield well site to 5, 24 and 0 respectively.

[0295] Step 5: Record T Real demand of power grid at any time =6412.42, update the historical demand sequence of the target oilfield well site, and repeat step 1 to obtain... =6281.46, and then steps 2 and 3 are executed again to obtain the number of wind turbine generators, photovoltaic panels and fossil fuel generators operating at the next time T: 6, 21 and 0 respectively. Steps 1-5 are repeated in this way to achieve dynamic optimization.

[0296] In one exemplary embodiment, a dynamic optimization system for wind-solar-electricity coordinated energy supply in oilfields is provided, comprising:

[0297] Data Prediction Module 101: Obtain historical demand sequences for the target oilfield well site. Seq Construct a model to predict the power grid demand at time T in the target oilfield well site. T Real-time grid demand .

[0298] Optimization Model Construction Module 202: Constructs a multi-objective optimization model for wind-solar-electricity coordinated energy supply, based on the data prediction module 101. Based on data such as the number and basic parameters of wind turbines, photovoltaic panels, and fossil fuel generators at the target oilfield well site, a multi-objective optimization model for wind-solar-electricity coordinated energy supply is constructed.

[0299] Intelligent optimization module 303: Utilizes the dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode to solve the multi-objective optimization model of wind-solar-electricity coordinated energy supply constructed by optimization model construction module 202, and obtains the optimization results.

[0300] Adjustment feedback module 404: Adjusts the operating quantity of wind turbine generators, photovoltaic panels, and fossil fuel generators based on the optimization results of intelligent optimization module 303, and... T Real demand of power grid at any time The data is fed back to the data prediction module 101 for prediction at the next time step.

[0301] In one exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described dynamic optimization method for wind-solar-electricity coordinated power supply in oilfields.

Claims

1. A dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields, characterized in that, include: Acquire target oilfield well site T Historical demand sequence of the power grid at any given time; Based on the target oilfield well site T Historical demand sequences of the power grid at all times are used to construct target oilfield well sites. T A time-based power grid demand forecasting model; Target oilfield well site T The historical demand sequence of the power grid at any given time is input into the demand prediction model to obtain... T Forecasted demand for the power grid at any given time; Based on the predicted demand, a multi-objective optimization model for wind-solar-electric coordinated energy supply is constructed. The dung beetle optimization algorithm is used to solve the multi-objective optimization model of wind-solar-electricity coordinated energy supply to obtain the optimal strategy of the power grid; the dung beetle optimization algorithm is a dung beetle optimization algorithm based on a high-dimensional swarm intelligence search mode; The number of wind turbines, photovoltaic panels, and fossil fuel generators operating in the target oilfield well site is adjusted based on the aforementioned optimal power grid strategy. Based on the target oilfield well site T Historical demand sequences of the power grid at all times are used to construct target oilfield well sites. T The demand forecasting model for the power grid at any given time includes: Will STL Decomposition model, seasonal autoregressive integral moving average model, LSTM-Attention Model and error weight allocation: Model initialization, all model parameters are assigned a value of 0; STL The parameters of the decomposition model include the seasonal difference step size. S The parameters of the seasonal autoregressive integral moving average model include the order of the non-seasonal autoregression. p Period-by-period difference order d non-seasonal moving average order q Seasonal autoregression order P Seasonal difference order D and seasonal moving average order Q , LSTM- Attention The model parameters include window length. h The parameters of the error weighting model include the weights of the seasonal autoregressive integral moving average model and the weights of the residual correction model. The residual correction model is composed of... LSTM-Attention Obtained through model training and optimization; Input the historical demand sequence into the initialized data. STL Decompose the model to obtain the seasonal difference step size. S Seasonal series and residual series; The seasonal difference step size S The seasonal autoregressive integral moving average model initialized with the seasonal sequence input is used to determine the parameters of the seasonal autoregressive integral moving average model using the Akaike information criterion, thus obtaining the parameter-determined seasonal autoregressive integral moving average model. Sure LSTM-Attention Model window length h Slide the window sequentially along the residual sequence, with the values ​​within the window... h The training takes one set of data as input and the first data outside the window as output. LSTM-Attention The model utilizes the stochastic gradient descent algorithm to... LSTM-Attention The model hyperparameters were optimized to obtain the residual correction model; The seasonal autoregressive integral moving average model, with its seasonal series input parameters determined, yields the target oilfield well site. T Initial forecast of grid demand at any given time; Inputting the residual sequence into the residual correction model yields the target oilfield well site. T The residual forecast value of the power grid demand at any given time; The initial demand forecast, the residual demand forecast, the seasonal series, and the residual series are input into the error weight allocation model to obtain the weights of the seasonal autoregressive integral moving average model with determined parameters, as well as the weights of the residual correction model. The target oilfield well site is obtained by multiplying the weights of the seasonal autoregressive integral moving average model determined by the parameters by the initial demand forecast. T The initial weighted forecast value of the power grid at any given time is used to multiply the weights of the residual correction model by the demand residual forecast value to obtain the target oilfield well site. T The residual weighted prediction value of the power grid at any given time is obtained by adding the initial weighted prediction value and the residual weighted prediction value. T Forecasted demand for the power grid at any given time; Sure STL The decomposition model, the seasonal autoregressive integral moving average model with defined parameters, the residual correction model, and the error weight allocation model are used for the target oilfield well site. T A time-based power grid demand forecasting model.

2. The dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields according to claim 1, characterized in that, After adjusting the number of operating wind turbines, photovoltaic panels, and fossil fuel generators at the target oilfield well site based on the aforementioned optimal grid strategy, the following is also included: Acquire target oilfield well site T Measured values ​​of power grid demand at any given time; The measured demand of the power grid at time T in the target oilfield well site is added as the last element to the historical demand sequence to obtain the target oilfield well site. T+1 Historical demand sequence of the power grid at any given time; make T=T+1 And return to the step "Target oilfield well site" T The historical demand sequence of the power grid at any given time is input into the demand prediction model to obtain... T "The demand forecast of the power grid at any given time." 3. The dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields according to claim 1, characterized in that, Target oilfield well site T The historical demand sequence of the power grid at any given time is input into the demand prediction model to obtain... T The demand forecast for the power grid at any given time includes: Target oilfield well site T Historical demand sequences of the power grid at any given time are input into... STL Decompose the model to obtain the seasonal series and residual series; The seasonal data series is input into a seasonal autoregressive integral moving average model with defined parameters to obtain the target oilfield well site. T Initial forecast of grid demand at any given time; Inputting the residual sequence into the residual correction model yields the target oilfield well site. T The residual forecast value of the power grid demand at any given time; The initial demand forecast, the residual demand forecast, the seasonal series, and the residual series are input into the error weight allocation model to obtain the weights of the seasonal autoregressive integral moving average model with determined parameters, as well as the weights of the residual correction model. The target oilfield well site is obtained by multiplying the weights of the seasonal autoregressive integral moving average model determined by the parameters by the initial demand forecast. T The initial weighted forecast value of the power grid at any given time is used to multiply the weights of the residual correction model by the demand residual forecast value to obtain the target oilfield well site. T The residual weighted prediction value of the power grid at any given time is obtained by adding the initial weighted prediction value and the residual weighted prediction value. T The forecast value of the power grid demand at any given time.

4. The dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields according to claim 1, characterized in that, The multi-objective optimization model for wind-solar-electric coordinated energy supply includes an economic benefit objective function, an equivalent emission reduction objective function, optimization quantity constraints, equivalent quantity constraints, demand constraints, and integer constraints.

5. The dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields according to claim 4, characterized in that, The objective function for economic benefits is: ； ； ； ； In the formula, EB For economic benefits, c o At the current oil price, C o This represents the cumulative oil production of the target oilfield well site within a unit of time. E w This refers to the power generation per unit time of a single wind turbine generator. N w The number of wind turbine generators. E s This refers to the power generation per unit time of a single photovoltaic panel. N s The number of photovoltaic panels. E f This refers to the power generation per unit time of a single fossil fuel generator unit. N f The number of fossil fuel generator sets. P t Time-of-use electricity pricing for the target oilfield well site. C w Let α be the installation cost of a single wind turbine generator set, and α be the depreciation period of the wind turbine generator set. OP w The unit time operation and maintenance cost of a single wind turbine generator. C s β represents the installation cost of a single photovoltaic panel, and β represents the depreciation period of the photovoltaic generator set. OP s The unit time operation and maintenance cost of a single photovoltaic panel. C f γ represents the installation cost of a single fossil fuel generator set, and γ represents the depreciation period of the fossil fuel generator set. OP f The unit time operation and maintenance cost of a single fossil fuel generator unit. v 1 represents the cut-in speed of the wind turbine generator per unit time. v 2 represents the cut-out speed of the wind turbine generator per unit time. p ( v For wind turbine generator sets at wind speeds v Power generation at that time f ( v ( ) represents the probability distribution of wind speed at the hub height of a wind turbine generator. H A This represents the total solar radiation on a horizontal surface. P AZ The installation capacity of a single photovoltaic panel. E A1 The irradiance under standard conditions at the end of the previous unit of time. E A2 The irradiance under the initial standard conditions for the next unit time. f ( E A () is a function of irradiance variation under standard conditions. K The overall efficiency coefficient, q The calorific value per kilogram of coal. N 1 This represents the coal consumption at the end of the previous unit of time. N 2 This represents the initial coal consumption for the next unit of time. f ( N ) is a coal consumption function used to describe the amount of coal consumed per unit time; The objective function for the equivalent emission reduction is: ； ； In the formula, EQE For equivalent emission reductions, C This refers to the carbon dioxide emissions produced per unit time by a single fossil fuel power generation unit. for T The number of fossil fuel generators that generate all of the grid's demand at any given time. Car The carbon content, a basic element of coal. OF The carbon oxidation rate of coal; The optimized quantity constraints include optimizing the number of wind turbine generators, photovoltaic generators, and fossil fuel generators. The optimization quantity constraint is: ； In the formula, N wmax It refers to the number of existing wind turbine generators at the target oilfield well site. N smax It refers to the number of existing photovoltaic panels at the target oilfield well site. N fmax It refers to the number of existing fossil fuel generator units at the target oilfield well site; The equivalent quantity constraint for the fossil fuel generator sets is: ； The power generation demand constraint at time T is: ； In the formula, This represents the predicted demand of the power grid at time T. The integer constraint is: ； In the formula, It is a positive integer.

6. The dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields according to claim 4, characterized in that, The dung beetle optimization algorithm is used to solve the multi-objective optimization model of wind-solar-electricity coordinated energy supply to obtain the optimal power grid strategy, including: Set the population size N and the maximum number of iterations (Iteration); The objective functions for economic benefits and equivalent emission reductions are determined as fitness functions. Let the number of iterations t=1; Based on the population size N, the dung beetle population is initialized using chaotic mapping, and the initialized population is used as the dung beetle population at the t-th iteration. Calculate the fitness value of individuals in the dung beetle population at the t-th iteration; Calculate the non-dominated solutions in the dung beetle population at the t-th iteration and store them in the Pareto front; Calculate the crowding distance for each nondominated solution in the Pareto front to obtain multiple crowding distances; The dung beetle population at iteration t+1 is updated using dancing, rolling, foraging, and stealing strategies. Calculate the fitness value of individuals in the undetermined dung beetle population at the (t+1)th iteration; The dung beetle population at the (t+1)th iteration is determined based on the fitness values ​​of individuals in the undetermined dung beetle population at the (t+1)th iteration and the fitness values ​​of individuals in the dung beetle population at the (t)th iteration. Calculate the non-dominated solutions in the dung beetle population at the (t+1)th iteration and store them in the Pareto front; Calculate the crowding distance for each nondominated solution in the Pareto front, and prioritize retaining nondominated solutions with larger crowding distances; Determine whether the iteration number t is greater than or equal to the maximum iteration number Iteration, and obtain the determination result; If the judgment result is negative, then let t = t + 1, and return to the step "update the dung beetle population at the (t+1)th iteration using the dancing strategy, rolling strategy, foraging strategy, and stealing strategy"; If the judgment result is yes, then the Pareto front is output as the optimal strategy for the power grid.

7. The dynamic optimization method for wind-solar-electricity coordinated energy supply in oilfields according to claim 6, characterized in that, The foraging strategy is as follows: Calculate the oviposition area and oviposition location of a dung beetle population: ； ； ； In the formula, For the current number t The local optimum position of the dung beetle population in the next iteration; This marks the lower boundary of the spawning area. This is the upper boundary of the spawning area; R The boundary convergence factor; Lb This serves as the lower bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. Ub This serves as the upper bound for the optimization variables in the wind-solar-electricity coordinated energy supply optimization model. f b Fitness value of an individual in a dung beetle population The maximum value, f w Fitness value of an individual in a dung beetle population The minimum value; The position of dung beetle egg balls is updated using a dynamic sine and cosine update strategy: ； ； In the formula, B i ( t+1 ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t+1 The location of spawning in the next iteration; B i ( t ) is the [number]th [number] species in the dung beetle population. i Dung beetles that only lay eggs t The location of spawning in the next iteration; b 1 and b 2 represents two sizes, 1× D Independent random vectors; D Dimensions for optimizing the wind-solar-power coordinated energy supply model; l A random number in the range [-1, 1]. r For dynamic sine and cosine shape parameters; m A random number in the range [-1, 1]. n The constants that determine the shape of the dynamic sine and cosine waves; Dung beetle eggs hatch into baby dung beetles; the optimal foraging area for the baby dung beetles is calculated. ； ； In the formula, X b for t The position of the global optimum in the next iteration; Lb b This is the lower boundary of the optimal foraging area; Ub b This is the upper limit of the optimal foraging area; The position of the dung beetle is updated using an iterative strategy based on the golden section line. ； In the formula, x i ( t+ 1) is the [number]th species in the dung beetle population. i The little dung beetle t+ The position of the first iteration; x i ( t ) is the [number]th [number] species in the dung beetle population. i The little dung beetle t The position of the next iteration; R 1 is Random numbers; R 2 is Random numbers; C 1 and C 2 represents the coefficient introduced by the golden ratio; ； 。 8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the dynamic optimization method for wind-solar-electric coordinated energy supply in oilfields as described in any one of claims 1-7.

Images