Multi-scale regulation and control method of wind, light and heat storage micro-grid system in oil exploitation

Abstract

The invention discloses a multi-scale regulation and control method of a wind, light and heat storage micro-grid system in oil exploitation. The method comprises the following steps: S1, setting a multi-scale data acquisition point for data acquisition; s2, predicting the wind-solar power generation power, the photo-thermal energy supply power and the petroleum mechanical production load change in the future time period; s3, constructing a master-slave game model which takes the wind-light-heat storage and supply system as a master party and takes the oil mining load as a slave party; s4, solving the master-slave game model by adopting a particle swarm optimization algorithm; and S5, generating a regulation and control instruction according to a solving result of S4. According to the multi-scale regulation and control method of the wind, light and heat storage micro-grid system in oil exploitation, through multi-scale potential tapping and intelligent algorithm application, the energy utilization rate can be increased, the cost can be reduced, and stable operation of mechanical exploitation equipment can be guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of energy dispatching, and in particular to a multi-scale control method of a wind, solar, thermal and storage microgrid system in oil production. Background Art

[0002] In today's global pursuit of sustainable energy development, the oil extraction industry faces an urgent need for energy transformation. As a model of clean and renewable energy utilization, wind, solar, thermal and storage smart microgrid systems have been gradually introduced into the field of oil extraction, but they face many challenges in practical applications. Given that oil fields are an important area of ​​energy production and consumption, promoting the use of new energy to replace traditional energy in the process of oil field extraction has become a key measure to achieve the country's carbon neutrality goals, which has far-reaching significance for the sustainable development of the entire energy industry.

[0003] Wind power generation is highly dependent on wind speed. Instantaneous, daily, and seasonal changes in wind speed will cause large fluctuations in power generation. For example, in a coastal oil production area, wind turbines can be fully powered when the sea breeze is strong during the day, but the wind speed drops sharply at night, and the power generation may drop to an extremely low level. Photovoltaic power generation is also restricted by factors such as light intensity, cloud cover, and sun angle. The power generation between sunny and cloudy days, summer and winter varies significantly. Although the solar thermal utilization system is relatively stable, the heat output is limited when the light is insufficient, and the capacity and efficiency of the heat storage device need to be carefully weighed. The energy storage system is the key to balancing supply and demand, but its charging and discharging characteristics, capacity attenuation, cost and other issues cannot be ignored. Different types of energy storage technologies, such as lithium batteries, lead-acid batteries, and thermal storage, are applicable to different scenarios and need to be reasonably configured.

[0004] There are many types of oil extraction equipment, including oil pumps, water injection pumps, compressors, etc., each with different operating power curves and working hours. The power demand of the oil pump is relatively stable during the oil extraction process, but the impact current is large at the moment of startup; the power of the water injection pump changes in stages according to the water injection pressure demand. Moreover, different stages of oil extraction, such as the exploration period, the initial stage of extraction, the high-yield period, and the decay period, have completely different requirements for the total amount, stability, and quality of energy. Traditional control methods are mostly based on experience or simple timing control, and cannot adapt to this complex and changeable energy supply and demand pattern.

[0005] On the one hand, it is difficult to fully tap the potential of each component of wind, solar, heat and storage, and energy waste occurs frequently. For example, during the peak of wind and solar power generation, due to the lack of effective allocation, the excess electricity is not properly stored or used; on the other hand, it is impossible to ensure the stable operation of key equipment for oil extraction, and power outages, undervoltage and other problems occur from time to time, affecting extraction efficiency, increasing equipment maintenance costs, and may even cause safety hazards. In addition, traditional regulation lacks foresight in responding to long-term environmental changes and oil extraction planning adjustments, and cannot plan energy system optimization and upgrades in advance. Therefore, a new multi-scale integrated regulation method is urgently needed to break the deadlock. Summary of the invention

[0006] The purpose of the present invention is to provide a multi-scale control method for a wind, solar, thermal and storage microgrid system in oil production, which solves the problem of difficult balance between supply and demand of oil machine-produced energy and lack of foresight in traditional control. By using intelligent algorithms, it copes with energy fluctuations and ensures stable operation of machine-produced equipment.

[0007] To achieve the above object, the present invention provides a multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production, comprising the following steps:

[0008] S1. Establish multi-scale data collection points for data collection;

[0009] S2. Predict the changes in wind and solar power generation, solar thermal power supply and oil extraction load in the future;

[0010] S3. Construct a master-slave game model with the wind, solar, heat and storage supply system as the main party and the oil drilling load as the slave party;

[0011] S4, using particle swarm optimization algorithm to solve the master-slave game model;

[0012] S5. Generate control instructions based on the solution result of S4.

[0013] Preferably, in S1, a high-precision sensor is used to collect the real-time power P of wind power generation. w (t), real-time photovoltaic power generation power P p (t), real-time energy supply power P of the solar thermal system h (t), energy storage system state of charge SOC(t), charging power P c (t), discharge power P d (t) and the real-time load power of each oil extraction equipment P l (t); where t represents time, and the acquisition frequency is in seconds. s , hour level h and Japanese level f d .

[0014] Preferably, in S2, the autoregressive moving average model ARIMA and the long short-term memory network LSTM are used to combine historical meteorological data, historical energy production data and historical machine-collected load data to predict the wind power generation in the future period. Photovoltaic power generation Solar thermal power and oil drilling load Changes, forecasting future periods are divided into short-term, medium-term and long-term.

[0015] Preferably, in S3, the main party objective function is to maximize the energy supply benefit, then the energy supply benefit of the wind, solar, thermal storage supply system at time t is calculated as:

[0016] R s (t) = P w (t)·p w (t)+P p (t)·p p (t)+P h (t)·p h (t)

[0017] -P c (t)·Ces(t)-P d (t)·Ces(t);

[0018] Among them, p w (t), p p (t) and p h (t) are the unit electricity prices of wind power, photovoltaic power and solar thermal power at time t, respectively; Ces(t) is the unit charging and discharging cost of energy storage at time t;

[0019] The objective function is to minimize the sum of energy procurement cost and production loss cost. The sum of energy procurement cost and production loss cost of oil machine extraction load at time t is calculated as follows:

[0020] C l (t)=Σ i∈{w,p,h} P l,i (t)·p i (t)+Cidle(t)·(L(t)-Σ i∈{w,p,h} P l,i (t));

[0021] Among them, P l,i (t) is the amount of electricity purchased from the i-th energy component at time t, p i (t) is the purchase price of the corresponding energy, C idle (t) is the idle cost of the machine mining equipment at time t, and L(t) is the load demand of the machine mining at time t;

[0022] The energy supply and demand balance constraint means that the total energy generated by the energy supply system at time t is equal to the total energy consumed by the oil extraction load, that is:

[0023] P w (t)+P p (t)+P h (t)+P d (t)-P c (t) = P l (t);

[0024] The operating capacity and power upper and lower limits of each component are as follows:

[0025] P min,i ≤P i (t)≤P max,i ;

[0026] Among them, i represents different energy components, P min,i and P max,i are the minimum and maximum power operating limits of component i at time t, respectively.

[0027] Preferably, in S4, a particle swarm optimization algorithm PSO is used to solve the master-slave game model, and the particle swarm size N, inertia weight ω, learning factors c1, c2, and maximum number of iterations iter are set. max ;

[0028] Particle velocity update formula:

[0029]

[0030] Among them, m represents the particle number, n represents the dimension, and k represents the number of iterations. is the velocity of the mth particle in the nth dimension in the kth iteration, and are two random numbers of the nth dimension in the kth iteration, is the individual optimal position of the mth particle in the nth dimension in the kth iteration, is the optimal position of the group in the nth dimension at the kth iteration;

[0031] Particle position update formula:

[0032]

[0033] in, is the position of the mth particle in the nth dimension at the kth iteration.

[0034] Preferably, in S5, a control instruction is generated according to the solution result of the particle swarm optimization algorithm in S4:

[0035] Setting ΔP w (t), ΔP p (t) are the instantaneous fluctuation values ​​of wind power generation and photovoltaic power generation at time t, respectively. w (t)|or|ΔP p (t)|Exceeds the threshold ΔP th When the wind, solar, heat and storage supply systems are quickly charged and discharged within the response time of seconds to balance the power;

[0036] During hourly regulation, the operating parameters of the solar thermal system are adjusted in advance and the energy storage charging and discharging plan is optimized based on short-term forecast results; during seasonal regulation, the energy storage capacity planning is adjusted in advance and the maintenance cycle of wind and solar equipment is optimized based on long-term forecasts.

[0037] Preferably, after S5, it also includes establishing a multi-dimensional evaluation indicator system, including life cycle cost, carbon emission reduction, energy utilization efficiency and number of oil machine extraction operation interruptions; regularly evaluating the control effect, and adjusting the prediction model parameters, master-slave game model price and cost parameters and particle swarm optimization algorithm parameters based on the evaluation results and using the feedback mechanism.

[0038] Therefore, the present invention adopts the above-mentioned multi-scale control method of the wind-solar-thermal storage microgrid system in oil production, and the beneficial effects are as follows:

[0039] (1) The present invention uses precise multi-scale integrated control, fully considers the characteristics of various energy components of wind, solar, thermal and energy storage, and the load requirements of oil extraction, optimizes energy generation, storage and distribution, avoids energy waste, significantly improves energy utilization efficiency, reduces dependence on traditional fossil energy, and reduces energy costs.

[0040] (2) The present invention is based on the optimization strategy of master-slave game theory and swarm intelligence algorithm. On the premise of meeting the energy demand of oil extraction, it minimizes the system operation cost, including energy production cost, equipment maintenance cost and energy storage cost, and maximizes the efficiency of oil extraction, thereby improving the overall economic benefits and enhancing the competitiveness of enterprises.

[0041] (3) The present invention can effectively cope with the intermittent and volatile nature of wind and solar energy. Through reasonable control of the charge and discharge of energy storage units and the coordinated complementarity of multiple energy sources, it can ensure that oil extraction equipment can obtain a stable and reliable energy supply under different working conditions and time scales, reduce the risk of production interruption due to unstable energy supply, improve production continuity and stability, and ensure the smooth progress of oil extraction operations.

[0042] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is an overall flow chart of an embodiment of a multi-scale control method of a wind-solar-thermal storage microgrid system in oil production according to the present invention.

[0044] Figure 2 It is an overall framework diagram of an embodiment of a multi-scale control method of a wind-solar-thermal storage microgrid system in oil production according to the present invention. DETAILED DESCRIPTION

[0045] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0046] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0047] like Figure 1 , Figure 2 As shown, a multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production includes the following steps:

[0048] S1. Set up multi-scale data collection points for data collection and use high-precision sensors to collect real-time wind power generation power P w (t) is used to reflect the power generation capacity of the wind turbine at that moment, and its value depends on factors such as wind speed and performance parameters of the wind turbine.

[0049] Wind power model:

[0050] The real-time power calculation formula of wind power generation is as follows:

[0051]

[0052] Where ρ is the air density (kg / m 3 ), A is the swept area of ​​the wind wheel (m 2 ), determined by the rotor radius R, A = πR 2 The larger the radius of the wind rotor, the larger the swept area and the stronger the ability to capture wind energy; v(t) represents the wind speed at a given moment (m / s), which is a key factor affecting wind power generation. The size and change of wind speed directly determines how much wind energy the wind turbine can capture.

[0053] C p (λ(t), β(t) is the wind energy utilization coefficient, which is a function of the blade tip speed ratio λ(t) and the pitch angle β(t). Where ω(t) is the angular velocity of the wind rotor at time t (rad / s); the pitch angle β(t) is the angle between the wind rotor blades and the airflow direction. By adjusting the pitch angle, the wind energy utilization coefficient can be changed, thereby controlling the wind power generation. For example, when the wind speed is low, the pitch angle can be appropriately increased to improve the wind energy utilization coefficient; when the wind speed is too high, the pitch angle can be reduced to protect the wind turbine equipment.

[0054] Photovoltaic power generation real-time power P p (t), which is determined by factors such as light intensity, conversion efficiency of photovoltaic panels, and temperature, and reflects the power generation output of the photovoltaic array at a specific moment.

[0055] Photovoltaic model

[0056] The real-time power calculation formula of photovoltaic power generation is:

[0057]

[0058] Where: N s Indicates the number of PV modules connected in series. The number of modules in series will affect the output voltage. The more modules in series, the higher the output voltage. However, the current of the modules in series must be consistent, otherwise the overall output power will be affected due to the short board effect.

[0059] N p Indicates the number of PV panels connected in parallel. Parallel connection can increase the output current, thereby improving the overall power generation. The more parallel connections there are, the greater the output current.

[0060] I ph It indicates the photocurrent (A) at a given moment, which is determined by the light intensity and the characteristics of the photovoltaic cell. The stronger the light intensity, the greater the photocurrent. Factors such as the material and temperature of the photovoltaic cell will also affect it.

[0061] I0 represents the reverse saturation current (A), which is related to the material and temperature of the photovoltaic cell. When the temperature rises, the reverse saturation current will increase. Generally, the reverse saturation current of silicon-based photovoltaic cells is small at room temperature.

[0062] q represents the charge of an electron and is a constant.

[0063] V(t) represents the output voltage of the photovoltaic cell at time t (V), and its magnitude is affected by factors such as light intensity, temperature, and load. When the light intensity changes or the load changes, the output voltage will change accordingly.

[0064] I(t) represents the output current (A) of the photovoltaic cell at time t, which is the actual output current of the photovoltaic cell and is related to factors such as photocurrent, reverse saturation current and battery internal resistance.

[0065] R s It is the series resistance (Ω), which is mainly composed of the resistance of the battery material itself, the contact resistance between the electrode and the battery, etc. The series resistance will consume part of the electrical energy and reduce the output power. Generally, it is hoped that the series resistance is as small as possible.

[0066] A o It represents the diode ideal factor, and its value range is generally between 1 and 2. It reflects the physical properties of photovoltaic cells. Photovoltaic cells with different materials and manufacturing processes have different values.

[0067] k is the Boltzmann constant, which is a physical constant.

[0068] T(t) represents the temperature of the photovoltaic cell at time t (K). The increase in temperature will reduce the open-circuit voltage of the photovoltaic cell, thereby affecting the output power. For example, in the hot summer, the temperature of the photovoltaic cell increases and the power generation power may decrease.

[0069] Rsh It represents the parallel resistance (Ω), which is mainly caused by leakage at the edge of the battery and defects in the battery body. The smaller the parallel resistance, the greater the leakage current and the lower the output power. High-quality photovoltaic cells usually have a higher parallel resistance.

[0070] Real-time power supply of solar thermal system P h (t), which is related to the lighting conditions, the efficiency of the photothermal conversion device, and the heat storage and transmission conditions, represents the rate at which the photothermal system provides thermal energy to the oil recovery operation.

[0071] Photothermal model

[0072] The calculation formula of real-time energy supply power of solar thermal system is:

[0073] P h (t) = A collect G T (t)η collect ;

[0074] Among them, A collect Indicates the collector lighting area (m 2 ), the larger the lighting area, the more solar radiation energy can be absorbed. Generally, the lighting area of ​​the collector is determined according to actual needs and installation space.

[0075] G T (t) represents the total solar irradiance on the collector surface at time t (W / m 2 ), the size of which depends on factors such as solar radiation intensity, weather conditions (sunny, cloudy, etc.), geographical location (latitude, altitude, etc.) and time (season, different times of the day). For example, at noon, the solar altitude angle is large and the total solar irradiance is usually strong.

[0076] η collect It indicates the heat collection efficiency of the collector, which is a value between 0 and 1, reflecting the ability of the collector to convert solar radiation energy into heat energy. The heat collection efficiency is affected by factors such as the collector type (flat plate collector, vacuum tube collector, etc.), material properties, optical properties, heat loss, etc. For example, the heat collection efficiency of a high-efficiency vacuum tube collector may reach 0.6-0.8, while the heat collection efficiency of an ordinary flat plate collector is relatively low.

[0077] The state of charge SOC(t) of the energy storage system has a value range of 0-1.

[0078] Energy storage model

[0079] State of charge (SOC) calculation formula:

[0080]

[0081] Among them, SOC(t) Indicates the state of charge at time t, with a value range of 0-1, reflecting the proportion of the remaining power of the energy storage system at that moment to the total capacity. For example, it means that the remaining power of the energy storage system is half of the total capacity.

[0082] SOC(t-Δt) represents the state of charge at the previous time t-Δt.

[0083] Δt represents the time interval, which is used to describe the time difference between two adjacent calculations of the state of charge. Its size is determined according to the data acquisition frequency and the system control requirements. For example, in second-level control, Δt may be 1 second.

[0084] I(τ) represents the charge and discharge current (A) in the time interval [t-Δt, t]. A positive value represents the charge current, and a negative value represents the discharge current. The magnitude of the charge and discharge current depends on the working state of the energy storage system (charging, discharging or stationary) and the requirements of the external load or power supply.

[0085] Q n Indicates the rated capacity (Ah) of the energy storage system, which is the maximum amount of electricity that the energy storage system can store. For example, a battery with a rated capacity of 100Ah can theoretically store 100 ampere-hours of electricity when fully charged.

[0086] Charging power P c (t) and discharge power P d (t), that is, a positive value indicates charging, and a negative value indicates discharging. The sizes of the two are affected by the performance of the energy storage device, the charging and discharging control strategy, and the energy balance requirements of the system.

[0087] The relationship between charging power and energy is as follows:

[0088] E charge (t) = E charge (t-Δt)+P c (t)Δtη charge ;

[0089] Among them, E charge (t) is the charging energy (Wh or kWh) at time t, which means the energy stored by the energy storage system through charging at that time.

[0090] E charge (t-Δt) is the charging energy at the previous moment t-Δt;

[0091] η charge is the charging efficiency, which is a value between 0 and 1, reflecting the energy loss caused by heat and other reasons during the charging process, such as η charge=0.9 means that there is a 10% energy loss during the charging process. The charging efficiency is affected by factors such as the type of energy storage device (lithium battery, lead-acid battery, etc.), charging method (constant current charging, constant voltage charging, etc.) and ambient temperature.

[0092] The relationship between discharge power and energy is as follows:

[0093]

[0094] Among them, E discharge (t) is the discharge energy (Wh or kWh) at time t, indicating the energy released by the energy storage system through discharge at that time. η discharge It is the discharge efficiency, which is also a value between 0 and 1, reflecting the energy loss during the discharge process. The discharge efficiency is affected by factors such as the internal resistance of the energy storage device, the discharge current, and temperature. For example, when discharging at a large current, the discharge efficiency may decrease.

[0095] Energy storage system cost calculation formula (unit charging and discharging cost):

[0096]

[0097] in,

[0098] Ces(t) is the unit charging and discharging cost of energy storage at time t (yuan / kWh), which is used to calculate the benefits of wind, solar, thermal and storage supply systems and the cost of oil extraction loads in the master-slave game model.

[0099] C inital It represents the initial investment cost (RMB) of the energy storage system, including the one-time investment costs such as the purchase, installation, and commissioning of the energy storage equipment. For example, the cost of purchasing a lithium battery energy storage system and the materials and labor costs required for installation.

[0100] C maintain It represents the maintenance cost (RMB) of the energy storage system during its service life, covering the cost of daily inspection, upkeep, repair, replacement of parts, etc. The maintenance cost will gradually accumulate as the energy storage system is used for longer.

[0101] C repalcement It represents the replacement cost (RMB) of the energy storage system during its service life. When the energy storage equipment reaches the end of its service life or its performance is severely degraded and cannot meet demand, the cost of replacing the equipment needs to be incurred, such as the cost of replacing the battery pack.

[0102] Q total It indicates the total charge and discharge capacity (kWh) of the energy storage system during its service life, reflecting the total work of the energy storage system during its entire life cycle. The larger the total charge and discharge capacity, the lower the unit charge and discharge cost.

[0103] Real-time load power of each equipment of oil drilling machine l (t) is used to reflect the actual power consumption of various equipment (such as pumping units, water injection pumps, compressors, etc.) during the oil production process. Its changes are related to factors such as the operating status of the equipment, production technology, and oil well conditions.

[0104] Where t represents time, P w (t), P p (t), P h (t), P c (t), P d (t) and P l The unit of (t) is watt (W) or kilowatt (kW), and the acquisition frequency is f s 、f h and f d The units are all Hertz (Hz), which indicates the frequency of data collection for various parameters at the corresponding time scale, and is used to capture the changing characteristics of energy systems and machine-generated loads at different time resolutions. s Used to capture instantaneous fluctuations, hourly f h To assist in short-term scheduling and day-level d Can support medium-term planning.

[0105] S2. Using the autoregressive moving average model ARIMA and the long short-term memory network LSTM, combined with historical meteorological data, historical energy production data and historical machine-collected load data, a prediction model is constructed to predict the wind power generation at time t in the future period. Photovoltaic power generation Solar thermal power and oil drilling load Changes are used to plan energy production, distribution and storage strategies in advance to cope with future changes in energy supply and demand.

[0106] The forecast future period is divided into short-term (hourly, 1-24 hours), medium-term (daily, 1-7 days) and long-term (seasonal, 1-4 seasons). The forecast model is regularly updated and trained to adapt to environmental changes.

[0107] The autoregressive moving average model ARIMA includes stationary processing (difference), autoregressive part (AR) and moving average part (MA):

[0108] Stationarity processing (difference):

[0109] The ARIMA model first performs a difference operation on the non-stationary time series. Let the original time series be y t , if the series is non-stationary, by differencing Convert the series into a stationary series.

[0110] Among them, B is the lag operator, B l y t =y t-l , d is the difference order.

[0111] Autoregressive part (AR):

[0112] The stationary series (The differenced series) can be expressed as a linear combination of past values. For the AR(p) part (p is the autoregressive order), the model can be written as:

[0113]

[0114] in, is the autoregressive coefficient, ε t is a white noise sequence, representing a random disturbance at time t.

[0115] Moving Average (MA):

[0116] For the MA(q) part (q is the moving average order), the stationary series It can be expressed as a linear combination of past white noise, that is:

[0117]

[0118] Among them, θ q is the moving average coefficient.

[0119] The complete ARIMA (p, d, q) model integrates the difference, autoregression and moving average parts to predict time series. When predicting the changes in wind and solar power generation, solar thermal power supply and oil machine load in the future period, it will fit the parameters of p, d, q based on historical data, and use these parameters and historical data to predict future values.

[0120] Long Short-Term Memory (LSTM) architecture

[0121] CellState:

[0122] The core of LSTM is the cell state C t , which is like an information conveyor belt that runs through the entire time series. The cell state can selectively forget or add information, and its initial value C0 is usually a given initial state that is continuously updated during the time series processing.

[0123] ForgetGate:

[0124] The forget gate determines which information is discarded from the cell state. It is implemented through a sigmoid function with the formula:

[0125] f t =σ(W f ·[h t-1 ,x t ]+b f );

[0126] Among them, W f is the weight matrix of the forget gate, b f is the bias, h t-1 is the hidden state of the previous moment, x is the input of the current moment, σ is the sigmoid function, and the output f t is a value between 0 and 1, used to control the cell state C at the previous moment t-1 degree of forgetfulness.

[0127] Input Gate:

[0128] The input gate consists of two parts, a sigmoid layer that decides which values ​​to update, and a tanh layer that creates a new vector of candidate values:

[0129] D t =σ(W D ·[h t-1 ,x t ]+b D ), used to control the degree of update;

[0130] Used to create candidate values.

[0131] Then, update the cell state to:

[0132]

[0133] Output Gate:

[0134] The output gate determines the next hidden state h t The value of . First pass it through a sigmoid function:

[0135] o t =σ(W o ·[h t-1 ,x t ]+b o );

[0136] h t =o t *tanh(C t );

[0137] Then, this hidden state h tIt can be used for subsequent operations such as prediction. When predicting wind and solar power generation, the time series of historical data is processed time by time step, and the last hidden state is used for prediction.

[0138] S3. Construct a master-slave game model with the wind, solar, heat and storage supply system as the master and the oil drilling load as the slave. The master's objective function is to maximize the energy supply revenue. The energy supply revenue of the wind, solar, heat and storage supply system at time t, in yuan, is calculated as follows:

[0139] R s (t) = P w (t)·p w (t)+P p (t)·p p (t)+P h (t)·p h (t)

[0140] -P c (t)·Ces(t)-P d (t)·Ces(t);

[0141] Among them, p w (t), p p (t) and p h (t) are the unit electricity selling prices of wind power, photovoltaic power and solar thermal power at time t, respectively, in yuan / kilowatt-hour (yuan / kWh); Ces(t) is the unit charging and discharging cost of energy storage at time t, in yuan / kilowatt-hour (yuan / kWh), taking into account factors such as battery life degradation, charging and discharging efficiency, and operation and maintenance costs. The objective function aims to maximize the economic benefits of the supply system through reasonable energy allocation and pricing strategies.

[0142] The objective function is to minimize the sum of energy procurement cost and production loss cost. The sum of energy procurement cost and production loss cost of oil machine extraction load at time t is calculated as follows:

[0143] C l (t)=Σ i∈{w,p,h} P l,i (t)·p i (t)+C idle (t)·(L(t)-∑ i∈{w,p,h} P l,i (t));

[0144] Among them, P l,i (t) is the amount of electricity purchased by the slave (oil extraction load) from the i-th energy component (wind power, photovoltaic array, solar thermal system and energy storage system, etc.) at time t, in kilowatt-hours (kWh); i(t) is the purchase price of the corresponding energy, in Yuan / kilowatt-hour (Yuan / kWh); C idle (t) is the idle cost of mechanical mining equipment at time t, in yuan / kilowatt-hour (yuan / kWh); L(t) is the load demand of mechanical mining at time t, in kilowatt-hour (kWh); The purpose of this objective function is to minimize the energy procurement cost and the production loss cost caused by insufficient energy supply under the premise of meeting the load demand of oil mechanical mining.

[0145] Constraints:

[0146] Energy supply and demand balance constraint, which represents the total energy E generated by the energy supply system at time t s (t) is equal to the total energy consumed by the oil extraction load E d (t), that is:

[0147] P w (t)+P p (t)+P h (t)+P d (t)-P c (t) = P l (t);

[0148] E s (t) = E d (t);

[0149] This constraint ensures the stable operation of the energy system and avoids energy shortage or surplus.

[0150] The operating capacity and power upper and lower limits of each component are as follows:

[0151] P min,i ≤P i (t)≤P max,i ;

[0152] Among them, i represents different energy components, P min,i and P max,i They are the minimum and maximum power operating limits of component i at time t, in watts (W) or kilowatts (kW), etc. These limits are set based on the equipment nameplate parameters (model, rated power, input voltage, rated current, number of phases, etc.) and safe operating specifications to ensure that each component operates within a safe and stable range to prevent equipment damage and energy waste.

[0153] S4, use the particle swarm optimization algorithm PSO to solve the master-slave game model, set the particle swarm size N, inertia weight ω, learning factors c1, c2, and maximum number of iterations iter max .

[0154] In simple systems, N is 30-50, and in complex systems it is 80-150. The choice is made based on the complexity of the problem to balance the algorithm's search range and computational efficiency.

[0155] Using a linear decreasing strategy, the initial value ω init is 0.9, the final value ω end The inertia weight ω is used to control the influence of the particle's last velocity on the current velocity. A larger inertia weight is conducive to global search, while a smaller inertia weight is conducive to local search. Through a linear decreasing method, the algorithm can widely explore the solution space in the early stage, and can perform fine search in the better area in the later stage to find the global optimal solution.

[0156] Learning factors c1 and c2, where c1 focuses on individual learning and has a value range of 1.5-2.0; c2 focuses on group learning and has a value range of 2.0-2.5. The learning factor is used to adjust the step size of the particle's learning to its own historical optimal position and the group's historical optimal position. By reasonably setting the learning factor, the individual exploration and group collaboration capabilities of the particles can be balanced, the convergence speed of the algorithm can be accelerated, and at the same time, it can avoid falling into the local optimal solution.

[0157] Maximum number of iterations iter max Determined according to the real-time requirements and the difficulty of problem convergence. For short-time scale control (such as seconds and hours), 30-50 iterations are taken to meet the needs of rapid response; for long-time scale control (such as seasonal level and long-term planning), 80-150 iterations are taken to ensure that the algorithm can fully search for a better solution, while taking into account the limitations of computing resources and time costs.

[0158] Particle position vector X i represents a set of possible energy distribution and load response strategies, the speed vector v i Used to update the position and particle velocity update formula:

[0159]

[0160] Among them, m represents the particle number, n represents the dimension (corresponding to different energy distribution or load response variables), and k represents the number of iterations. is the velocity of the mth particle in the nth dimension in the kth iteration, ω is the inertia weight, and are two random numbers in the nth dimension in the kth iteration, ranging from 0 to 1, used to increase the randomness of the search; is the individual optimal position of the mth particle in the nth dimension at the kth iteration, is the optimal position of the group in the nth dimension at the kth iteration;

[0161] The particle velocity update formula updates the particle velocity by comprehensively considering the particle's current velocity, individual experience, and group experience, guiding the particle to move to a better position in the solution space.

[0162] Particle position update formula:

[0163]

[0164] in, is the position of the mth particle in the nth dimension in the kth iteration. By adding the updated velocity to the current position, we get the new position of the particle in the next iteration. This position represents a possible energy allocation and load response strategy.

[0165] S5. Generate control instructions based on the solution of the particle swarm optimization algorithm in S4:

[0166] Setting ΔP w (t), ΔP p (t) are the instantaneous fluctuation values ​​of wind power generation and photovoltaic power generation at time t, in watts (W) or kilowatts (kW); when |ΔP w (t)|or|ΔP p (t)|Exceeds the threshold ΔP th When the wind, solar, heat and storage supply systems are quickly charged and discharged within the response time of seconds to balance the power;

[0167] Among them, the second-level control threshold ΔP th Determined by reverse calculation based on the voltage and frequency fluctuation range allowed by oil drilling equipment. For equipment that is sensitive to voltage fluctuations, ΔP th It is set to 5% of the rated power (or other appropriate proportion determined by the characteristics of the equipment), which means that when the instantaneous fluctuation of wind and solar power generation exceeds this threshold, the energy storage system needs to charge and discharge quickly within a response time of seconds (such as 0.5 seconds) to balance the power and maintain the stable operation of the mining equipment.

[0168] In hourly regulation, based on short-term forecast results, the solar thermal system operating parameters are adjusted and the energy storage charging and discharging plan is optimized hourly response time (e.g. 1 hour) in advance; in seasonal regulation, based on long-term forecasts, the energy storage capacity planning is adjusted and the maintenance cycle of wind and solar equipment is optimized seasonal response time (e.g. 1 month) in advance.

[0169] Finally, a multi-dimensional evaluation indicator system is established, including the full life cycle cost (including investment costs, operation and maintenance costs, and energy purchase costs, etc.), carbon emission reductions, energy utilization efficiency, and the number of interruptions in oil drilling operations. The regulation effect is evaluated regularly, and the prediction model parameters, the master-slave game model price and cost parameters, and the particle swarm optimization algorithm parameters are adjusted based on the evaluation results and the feedback mechanism.

[0170] Example 1: Initial application of small inland oil production sites

[0171] Data collection: Install adapter sensors on key mining equipment such as wind turbines, 500kW photovoltaic arrays, solar thermal collectors, 500kWh lithium battery energy storage systems, and oil pumps and water injection pumps. Industrial Ethernet is used to transmit data, and the second-level acquisition frequency is used to monitor the instantaneous fluctuations of wind power and photovoltaic power. Hourly-level acquisition assists in analyzing the change pattern of mining load. For example, at noon on a sunny day, it was monitored that the photovoltaic power generation reached 400kW, the wind power generation power was 50kW, the oil pump load was 200kW, and the water injection pump load was 150kW.

[0172] Prediction: Collect the meteorological and mechanical load data of the site in the past three years, and use the ARIMA model combined with the LSTM network to predict the energy supply and demand in the next 24 hours. After training, the average error of the wind and solar power generation forecast for the next 6 hours is controlled within 10%, and the mechanical load forecast error is within 8%. It is predicted that the wind speed will increase at night, the photovoltaic power generation power will drop to 0, and the mechanical load will remain at around 300kW.

[0173] Master-slave game model construction: Combined with the local energy market, the wind power selling price is set at 0.45 yuan / kWh, the photovoltaic power price is set at 0.55 yuan / kWh, the energy storage charging and discharging cost is set at 0.2 yuan / kWh, and the idle cost of the mining equipment is estimated to be 1,000 yuan / hour based on the crude oil price and the downtime loss. Consider the constraints of the rated parameters of each component, such as the maximum power of the wind turbine is 100kW, the maximum power of the photovoltaic array is 500kW, and the maximum charging power of the energy storage is 100kW.

[0174] Swarm intelligence algorithm solution: The particle swarm optimization algorithm was selected, and the number of particles was set to 50, the initial inertia weight was 0.9, the learning factor c1=c2=2, and the maximum number of iterations was 100. After 30 iterations, the optimal energy distribution plan during the day was obtained as 60% photovoltaic power supply, 20% wind power, and energy storage to supplement the remaining demand, and wind power and energy storage were mainly used for power supply at night.

[0175] Control and execution: The central control system controls each device in real time according to the solution results. For example, during the day, the photovoltaic panel inclination is adjusted to increase the power generation efficiency by 10%, and the energy storage system is controlled to charge and discharge according to the plan. At night, when the wind speed increases and the wind power exceeds 80kW, the energy storage system is charged; when the machine mining load increases, the energy storage is discharged and replenished in time. On the day of implementation, the energy utilization efficiency was improved by 12% compared with traditional control, and the oil machine mining operation was not interrupted.

[0176] Example 2: Seasonal Control of Medium-sized Coastal Oil Production Platforms

[0177] Data collection: Sensors continuously collect data and found that the average photovoltaic power generation power dropped to 200kW during the day in winter, the wind power generation power fluctuation increased, and the machine-generated load increased slightly due to process adjustments.

[0178] Forecast: Update weather and load data, retrain the forecast model, and forecast energy supply and demand for the next month. It is found that the available period of wind power in winter nights is shortened, and the peak period of mechanical load is extended.

[0179] Adjustment of the master-slave game model: According to the changes in winter energy costs, the electricity sales price is adjusted, with wind power reduced to 0.4 yuan / kWh and photovoltaic power reduced to 0.5 yuan / kWh. Considering the low-temperature performance of the battery, the energy storage charging and discharging cost rises to 0.25 yuan / kWh.

[0180] Swarm intelligence algorithm re-solved: Particle swarm optimization algorithm parameters were adjusted, with the number of particles being 50, the inertia weight being 0.8, the learning factors c1 and c2 being both set to 1.5, and the maximum number of iterations being 200. The winter strategy was re-solved to prioritize key mining equipment during the day, with less energy storage and more charging, and solar thermal energy increasing the proportion of wellhead insulation energy supply.

[0181] Regulation and implementation: According to the new strategy, during the winter, the number of power outages of machine mining equipment decreased by 70% compared with the same period in previous years, the stability of energy supply was greatly improved, carbon emissions were reduced by 25%, and crude oil extraction efficiency remained stable.

[0182] Example 3: Long-term expansion and upgrading planning of large oil production bases

[0183] After three years of operation, the mining scale has expanded by 50%, and the original energy system needs to be expanded.

[0184] Data collection: Review historical operating data and comprehensively analyze energy supply and demand bottlenecks in each season and time period.

[0185] Forecast: Use big data analysis to predict energy demand in the next three years, estimate the peak load growth rate, combine geological exploration planning, mining progress expectations and other factors, and comprehensively consider the long-term change trend of wind and solar resources.

[0186] Therefore, the present invention adopts the above-mentioned multi-scale control method of the wind, solar, thermal storage microgrid system in oil extraction, and relies on precise potential tapping of energy components and intelligent optimization and allocation to effectively improve energy utilization efficiency, stably ensure oil machine extraction operations, and reduce costs.

[0187] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A multi-scale control method for a wind-solar-thermal-storage microgrid system in oil production, characterized in that: The following steps are involved: S1. Establish multi-scale data collection points for data collection; S2. Predict the changes in wind and solar power generation, solar thermal power supply and oil extraction load in the future; S3. Construct a master-slave game model with the wind, solar, heat and storage supply system as the main party and the oil drilling load as the slave party; S4, using particle swarm optimization algorithm to solve the master-slave game model; S5. Generate control instructions based on the solution result of S4.

2. The multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production according to claim 1 is characterized in that: In S1, high-precision sensors are used to collect the real-time power P of wind power generation. w (t), real-time photovoltaic power generation power P p (t), real-time energy supply power P of the solar thermal system h (t), energy storage system state of charge SOC(t), charging power P c (t), discharge power P d (t) and the real-time load power of each oil extraction equipment P l (t); where t represents time, and the acquisition frequency is in seconds. s , hour level h and Japanese level f d .

3. The multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production according to claim 2 is characterized in that: In S2, the autoregressive moving average model ARIMA and the long short-term memory network LSTM are used to combine historical meteorological data, historical energy production data, and historical machine-collected load data to predict wind power generation in the future period. Photovoltaic power generation Solar thermal power and oil drilling load Changes, forecasting future periods are divided into short-term, medium-term and long-term.

4. The multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production according to claim 3 is characterized in that: In S3, the main objective function is to maximize the energy supply benefits. The energy supply benefits of the wind, solar, thermal and storage supply system at time t are calculated as follows: R s (t)=P w (t)·p w (t)+P p (t)·p p (t)+P h (t)·p h (t)-P c (t)·Ces(t)-P d (t)·Ces(t); Among them, p w (t), p p (t) and p h (t) are the unit electricity prices of wind power, photovoltaic power and solar thermal power at time t, respectively; Ces(t) is the unit charging and discharging cost of energy storage at time t; The objective function is to minimize the sum of energy procurement cost and production loss cost. The sum of energy procurement cost and production loss cost of oil machine extraction load at time t is calculated as follows: C l (t)=∑ i∈{w , p,h} P l,i (t)·p i (t)+C idle (t)·(L(t)-∑ i∈{w,p,h} P l,i (t)); Among them, P l,i (t) is the amount of electricity purchased from the i-th energy component at time t, p i (t) is the purchase price of the corresponding energy, C idle (t) is the idle cost of the machine mining equipment at time t, and L(t) is the load demand of the machine mining at time t; The energy supply and demand balance constraint means that the total energy generated by the energy supply system at time t is equal to the total energy consumed by the oil extraction load, that is: P w (t)+P p (t)+P h (t)+P d (t)-P c (t)=P l (t); The operating capacity and power upper and lower limits of each component are as follows: P min,i ≤P i (t)≤P max,i ; Among them, i represents different energy components, P min,i and P max,i are the minimum and maximum power operating limits of component i at time t, respectively.

5. The multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production according to claim 4 is characterized in that: In S4, the particle swarm optimization algorithm PSO is used to solve the master-slave game model, setting the particle swarm size N, inertia weight ω, learning factors c1, c2, and maximum number of iterations iter max ; Particle velocity update formula: Among them, m represents the particle number, n represents the dimension, and k represents the number of iterations. is the velocity of the mth particle in the nth dimension in the kth iteration, and are two random numbers of the nth dimension in the kth iteration, is the individual optimal position of the mth particle in the nth dimension in the kth iteration, is the optimal position of the group in the nth dimension at the kth iteration; Particle position update formula: in, is the position of the mth particle in the nth dimension at the kth iteration.

6. The multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production according to claim 5 is characterized in that: In S5, according to the solution of the particle swarm optimization algorithm in S4, the control instructions are generated: Setting ΔP w (t), ΔP p (t) are the instantaneous fluctuation values ​​of wind power generation and photovoltaic power generation at time t, respectively. w (t)|or|ΔP p (t)|Exceeds the threshold ΔP th When the wind, solar, heat and storage supply systems are quickly charged and discharged within the response time of seconds to balance the power; During hourly regulation, the operating parameters of the solar thermal system are adjusted in advance and the energy storage charging and discharging plan is optimized based on the short-term forecast results. In seasonal regulation, based on long-term forecasts, energy storage capacity planning is adjusted in advance and the maintenance cycle of wind and solar equipment is optimized.

7. The multi-scale control method of a wind-solar-thermal-storage microgrid system in oil production according to claim 6 is characterized in that: S5 also includes the establishment of a multi-dimensional evaluation indicator system, including life cycle costs, carbon emission reductions, energy efficiency and the number of interruptions in oil drilling operations; regular evaluation of the control effect, and adjustment of the prediction model parameters, master-slave game model price and cost parameters and particle swarm optimization algorithm parameters based on the evaluation results and feedback mechanism.

Images