Planning method and system for optimally matching wind and light storage capacity with hydrogen production ammonia alcohol scale

Abstract

The invention relates to the technical field of renewable energy system planning, and discloses a planning method and system for optimally matching wind and light storage capacity with a hydrogen production ammonia alcohol scale. According to the method, wind and light resource time series data are generated through the long-short-term memory network, wind and light output and chemical load demands of unit capacity are predicted by using the artificial neural network, and the optimization basis of the system is expanded from historical static data to a scene set predicted in the future, so that the perspectiveness of a planning scheme is remarkably improved; the actual guidance value of a planning scheme can be greatly improved, and the problem that a system configuration method in the prior art depends on static data and ignores an actual scheduling rule is solved.

Description

Technical Field

[0001] This invention relates to the field of renewable energy system planning technology, and in particular to a planning method and system for optimizing and matching wind, solar and energy storage capacity with the scale of hydrogen production, ammonia and methanol production. Background Technology

[0002] Currently, in green chemical systems, the capacity configuration of renewable energy (wind power, photovoltaic) and energy storage systems is the primary factor that directly determines whether a project can be successfully implemented and the economic efficiency of its subsequent operation. Its importance runs through the entire life cycle of the project.

[0003] Currently, mainstream wind, solar, and energy storage capacity configuration methods in the industry often focus solely on optimizing the system's own economics (e.g., reducing the levelized cost of electricity (LCOE)) or improving system operational reliability. They fail to deeply integrate with the specific electricity consumption characteristics (e.g., load intensity, continuous power supply requirements) and hard indicators of green energy utilization rates of downstream industries such as hydrogen production and ammonia / methanol synthesis. Wind and solar energy are significantly intermittent and volatile due to natural conditions. These differences place higher demands on the rational configuration of energy storage systems and the refined scheduling of the entire system. Furthermore, existing configuration methods fail to fully incorporate various constraints in actual engineering projects during the early planning stages, leading to a disconnect between planning schemes and project implementation needs. Meanwhile, traditional optimization algorithms (such as linear programming and dynamic programming) often face the "curse of dimensionality" (i.e., the computational complexity increases exponentially with the increase in variable dimensionality) when dealing with complex optimization problems like wind, solar, energy storage, and chemical engineering system capacity configuration. They may struggle to overcome the limitations of local optima and fail to efficiently obtain globally optimal configuration schemes, further restricting the scientific and economical nature of the planning schemes.

[0004] It is evident that the current system configuration method suffers from the problem of relying on static data and ignoring actual scheduling rules. Summary of the Invention

[0005] This invention provides a planning method and system for optimizing the matching of wind, solar and energy storage capacity with the scale of hydrogen production, ammonia and methanol production, in order to solve the problem that existing system configuration methods rely on static data and ignore actual scheduling rules.

[0006] Firstly, this application provides a planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol), including: Constructing prediction models for wind and solar energy resources; Construct a wind and solar power output prediction model; Construct a power load prediction model for the chemical industry; Based on the wind and solar energy resource prediction model, the wind and solar power output prediction model, and the chemical power load prediction model, a multi-constraint adaptive particle swarm optimization algorithm model is constructed. The final optimization variables are obtained based on the multi-constraint adaptive particle swarm optimization algorithm model, and planning is performed based on the optimization variables.

[0007] Secondly, this application provides a planning system for optimizing and matching wind, solar and energy storage capacity with the scale of hydrogen production, ammonia and methanol production, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method described in the first aspect above.

[0008] The present invention has the following beneficial effects: This application presents a planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production, ammonia, and methanol production. This method constructs wind and solar energy resource prediction models, wind and solar power output prediction models, and chemical power load prediction models. Based on these models, a multi-constraint adaptive particle swarm optimization (PSO) algorithm model is built. Then, the final optimization variables are obtained from the PSO algorithm model, and planning is performed based on these variables. This approach utilizes multi-dimensional, long-term scenarios of wind and solar resources generated by long short-term memory networks and wind and solar power output per unit capacity and chemical load demand generated by artificial neural networks. This expands the optimization basis from historical static data to a set of future predicted scenarios, significantly improving the foresight of the planning scheme and enhancing its practical guiding value. It also solves the problem of existing system configuration methods relying on static data and ignoring actual scheduling rules. Furthermore, the PSO algorithm effectively handles complex optimization problems involving high dimensions, nonlinearity, and multiple constraints, avoiding local optima and finding globally approximate optimal capacity configuration schemes.

[0009] Further technical solutions aim to minimize the overall cost per kilowatt-hour, directly reducing the production costs of green hydrogen and its derivatives (ammonia, alcohols) and enhancing the project's economic competitiveness.

[0010] Further technical solutions incorporate constraints such as energy storage discharge sequence based on power source type, grid-connected electricity ratio, and differentiated grid-connected electricity price, making the optimization results more feasible, in line with grid dispatching practices, and more aligned with actual engineering rules. Furthermore, by imposing strict constraints on the green electricity rate, it ensures that chemical products such as hydrogen ammonia and alcohols meet national or international green certification standards.

[0011] In addition to the objectives, features and advantages described above, the present invention has other objectives, features and advantages.

[0012] The present invention will now be described in further detail with reference to the figures. Attached Figure Description

[0013] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is one of the flowcharts of a preferred embodiment of the present invention for a planning method that optimizes the matching of wind and solar storage capacity with the scale of hydrogen production (ammonia and methanol). Figure 2 This is a second flowchart of a planning method for optimizing and matching wind, solar and energy storage capacity with the scale of hydrogen production (ammonia and methanol) according to a preferred embodiment of the present invention. Figure 3 This is a schematic diagram of the neural network of the wind and solar resource prediction model according to a preferred embodiment of the present invention; Figure 4 This is a schematic diagram of the neural network for a preferred embodiment of the wind and solar power output prediction model of the present invention; Figure 5 This is a schematic diagram of the neural network of the chemical load prediction model according to a preferred embodiment of the present invention; Figure 6 This is a schematic diagram of the power supply logic of the method according to a preferred embodiment of the present invention. Detailed Implementation

[0014] The technical solution of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms "an" or "a," and similar terms, do not indicate a quantity limitation, but rather indicate the presence of at least one.

[0016] It should be understood that the planning method provided in this application for optimizing the matching of wind, solar and energy storage capacity with the scale of hydrogen production (ammonia and methanol) can be used in the planning stage of chemical projects (such as green electricity to hydrogen production (ammonia and methanol)) to carry out the collaborative optimization design of energy system and chemical system, so as to achieve the optimal balance of economic efficiency and stability throughout the system's life cycle. This is only an example and is not intended to limit the scope.

[0017] Please see Figure 1 This application provides a planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol), including: Constructing prediction models for wind and solar energy resources; Construct a wind and solar power output prediction model; Construct a power load prediction model for the chemical industry; Based on the wind and solar energy resource prediction model, the wind and solar power output prediction model, and the chemical power load prediction model, a multi-constraint adaptive particle swarm optimization algorithm model is constructed. The final optimization variables are obtained based on the multi-constraint adaptive particle swarm optimization algorithm model, and planning is performed based on the optimization variables.

[0018] The aforementioned planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol) utilizes wind and solar resources generated by long short-term memory networks and multi-dimensional, long-term scenarios of wind and solar unit capacity output and chemical load demand generated by artificial neural networks. This expands the optimization basis from historical static data to a set of future predicted scenarios, significantly improving the foresight of the planning scheme and greatly enhancing its practical guiding value. It also addresses the problem in existing system configuration methods that rely on static data and ignore actual scheduling rules. Furthermore, the particle swarm optimization algorithm effectively handles complex optimization problems involving high dimensions, nonlinearity, and multiple constraints, avoiding getting trapped in local optima and finding a globally approximate optimal capacity configuration scheme.

[0019] It is worth noting that the planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production, ammonia, and methanol production in this application comprises two parts: a multi-dimensional data prediction layer driven by intelligent algorithms and a multi-constraint adaptive particle swarm optimization solution layer. The multi-dimensional data prediction layer driven by intelligent algorithms includes a wind and solar resource and power generation prediction module and a chemical industry electricity load prediction model. The wind and solar resource and power generation prediction module includes wind and solar energy resource prediction models and wind power and photovoltaic power conversion models. The multi-constraint adaptive particle swarm optimization solution layer utilizes the multi-dimensional data generated in the first layer, considering constraints, to optimize the matching of wind, solar, and energy storage installed capacity.

[0020] Below, as Figure 2 As shown, the steps of the above-described planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol) are described using a complete example: Step S1: Collect long-series data of meteorological parameters such as wind speed and direction, temperature, humidity, air pressure, solar radiation, and solar altitude angle at different altitudes to establish a wind and solar energy resource prediction model. This model uses an LSTM (Long Short-Term Memory) network. The input is the collected long-series data, and the output is time-series data of wind speed, wind direction, air density, and total solar radiation at horizontal levels. A schematic diagram of the network is shown below. Figure 3 As shown. The LSTM model function is as follows:

[0021] in, Here, H represents the LSTM nonlinear mapping function, and H is the measured height. For wind turbine hub height, Wind speed at height H The air density at height H. Wind direction at height H Temperature at height H H represents high humidity. H represents atmospheric pressure at altitude, G represents total solar radiation, D represents scattered solar radiation, and B represents direct solar radiation in the normal direction. For terrain slope, Due to the slope of the terrain, Let t be the solar altitude angle, t be the current time, and m be a future time.

[0022] It is worth noting that meteorological data is a decisive factor affecting the power generation capacity of wind and solar power. Specifically, the output power of wind turbines depends on parameters such as wind speed, wind direction, and air density at the hub height; while the output power of solar modules is directly affected by parameters such as solar irradiance. In this application, the meteorological forecast parameters for the project site are derived through a black-box model based on measured data. This is mainly based on the following two reasons: First, in terms of wind energy resource assessment, wind measurement towers are usually only used for observation at a specific height. For example, a 140-meter-high wind measurement tower can only obtain measured data at that height and below. However, the designed hub height of the wind turbine may exceed the measured range of the wind measurement tower (e.g., reaching 180 meters), or be at an unmeasured intermediate height. In this case, it is necessary to calculate the corresponding parameters at the designed hub height using existing measured data. This method employs a black-box model that, through training and learning, can calculate meteorological parameters in the vertical direction. Secondly, in solar energy resource assessment, the deployment of solar measurement stations is not comprehensive, often requiring the use of measured data from station A to extrapolate solar resources at station B. Therefore, this method uses geographical features such as slope and aspect, along with measured data, as input parameters to predict solar radiation at the project site using a black-box model.

[0023] In this application, the LSTM model function described above can be used to obtain predicted data for future times. The LSTM network model used in this application can process time-series data. It takes time-series data as input and outputs not only predicted data for a specific time t, but also data for subsequent times. For example, based on historical data from multiple years, it can predict data for future times (e.g., 8760 hours) through rolling prediction or multi-step prediction. This predicted data can provide long-series data for subsequent optimization models (rather than optimizing only for a single moment), making the optimization more practically meaningful.

[0024] Step S2: Referring to the following physical model of wind power and photovoltaic power conversion, establish a wind power and photovoltaic power conversion model. This model uses an ANN artificial neural network, and the network diagram is shown below. Figure 4 As shown. The time series data output by the LSTM obtained in step S1, the collected geographic information data, and the equipment information are imported into the wind power and photovoltaic power conversion model to obtain the predicted wind and solar power output data for 8760 hours or longer.

[0025] Physical model for wind power output conversion:

[0026] This refers to the output power of wind power. Where C is the swept area of ​​the wind turbine rotor, and C is the wind energy utilization coefficient, a dimensionless coefficient characterizing the efficiency of the wind turbine in capturing energy from the wind, typically between 0.4 and 0.6. The air density at the height of the wind turbine hub. The wind speed is measured at the height of the wind turbine hub.

[0027] Photovoltaic power conversion physical model:

[0028] in, For photovoltaic power output, The total area of ​​the photovoltaic array The overall efficiency of the photovoltaic system.

[0029] ANN model function for predicting wind and solar power output: h fg =σ fg ( *x fg + ) P fg =f fg ( *h fg + ) x fg The input feature vector consists of: time-series data output by the LSTM (wind speed, wind direction, air density, and total solar radiation at the horizontal plane at hub height); geographic information data (slope, aspect, and altitude); and equipment and layout information (wind turbine hub height, turbine power, module power, module tilt angle, module spacing, and wind energy utilization coefficient, etc.), σ. fg It is the activation function of the hidden layer, f fg It is the activation function of the output layer. , These are the weight matrix and bias vector of the hidden layer, respectively. , These are the weight matrix and bias vector of the output layer, h, respectively. fg P is the output vector of the hidden layer. fg The predicted values ​​for wind and solar power output per unit capacity are given (the wind and solar power output prediction model outputs two terms; since the above formula is a neural network schematic, P will be omitted in subsequent calculations for ease of expression). fg Split into P wind and P pv ).

[0030] In this application, the input parameters of the wind and solar power output prediction model will be integrated with the parameters in the physical model, and data such as resource time series data, geography, equipment, and layout information will be added, which can make the model more accurate than the existing physical models.

[0031] Step S3: Establish a predictive model for electricity consumption in the chemical industry. This model uses an artificial neural network (ANN), and the network diagram is shown below. Figure 5 As shown. The model inputs include time characteristics, product type (e.g., hydrogen, ammonia, alcohol), planned output (e.g., hydrogen, ammonia, alcohol output), production calendar, equipment commissioning plan, historical operating data of similar plants, and meteorological conditions. The output is the electricity load for the chemical production of hydrogen, ammonia, and alcohol.

[0032] The ANN model function for predicting chemical loads is as follows: h hg =σ hg ( *x hg + ) P hg =f hg ( *h hg + ) In the formula, x hg The input feature vector (such as time features, product type, planned output, production calendar, equipment commissioning plan, historical operating data of similar factories, meteorological conditions, etc.), σ hg It is the activation function of the hidden layer, f hg It is the activation function of the output layer. , These are the weight matrix and bias vector of the hidden layer, respectively. , These are the weight matrix and bias vector of the output layer, h, respectively. hg P is the output vector of the hidden layer. hg This is a predicted value for the chemical load.

[0033] Step S4: Construct a multi-constraint adaptive particle swarm optimization algorithm model based on the wind and solar energy resource prediction model, the wind and solar power output prediction model, and the chemical power load prediction model, defining the objective function, optimization variables, and constraints. It is worth noting that in this application, the objective function is the LCOE of the wind-solar-storage system; the optimization variables are wind power installed capacity, photovoltaic installed capacity, wind power energy storage installed capacity, and photovoltaic energy storage installed capacity; the constraints include green electricity rate constraints, power feedback constraints, installed capacity constraints, power supply balance constraints, and power supply logic constraints, etc.

[0034] It is worth explaining that the wind and solar energy resource prediction model, the wind and solar power output prediction model, and the chemical load prediction model are independent of the optimization model. Among them, the wind and solar energy resource prediction model is the basis of the wind and solar power output prediction model, and the prediction results of the wind and solar energy resource prediction model are used as the input parameters of the wind and solar power output prediction model; the chemical load prediction model is parallel to the wind and solar power output prediction model. The advantage of establishing the wind and solar energy resource prediction model is that it makes the wind and solar power output prediction more accurate and obtains the prediction data for future time periods. The wind and solar power output prediction model and the chemical load prediction model are established for use in the optimization model. In the optimization algorithm model, the "cost, discount rate, and number of operating years" in the objective function are set by the project and do not require the use of prediction data. The "power generation" is predicted by the wind and solar power output prediction model. Among the constraints of the optimization algorithm model, constraints (5) to (8) are related to the wind and solar power output prediction model and the chemical load prediction model. The optimization algorithm model and the established prediction models exist in a fusion, and multiple objectives can be optimized through particle swarm optimization.

[0035] In this application, the multi-constraint adaptive particle swarm optimization solution layer takes the multi-dimensional, long-time-series data parameters (wind and solar power output and load demand), economic parameters and constraint parameters generated by the first layer as algorithm input, and the optimal capacity configuration of wind power, photovoltaic and their corresponding energy storage as output, with the goal of minimizing the levelized cost of electricity (LCOE) of the wind-solar-storage system throughout its entire life cycle.

[0036] The economic parameters of the constrained particle swarm optimization solution layer include system equipment unit price, operation and maintenance cost, discount rate, grid connection price, and off-grid price. The constraint parameters of the constrained particle swarm optimization solution layer include the green electricity rate of chemical load, reverse power transmission ratio, upper and lower limits for wind power configuration, upper and lower limits for photovoltaic configuration, upper and lower limits for wind-solar-storage configuration, and wind-solar-storage power supply logic.

[0037] Specifically, the objective function is optimized as follows:

[0038] Where q represents the project's operation and maintenance years. This refers to all costs incurred in year n, including initial investment costs for wind, solar, and energy storage, operation and maintenance investment costs, grid power purchase investment costs, grid connection revenue, and depreciation residual value. For the wind and solar power generation in year n, Here, represents the discount rate, and n represents the number of years of operation. The grid purchase cost and grid connection revenue can be referenced from State Grid documents or local industrial and commercial electricity prices and the actual situation of operating power stations.

[0039] The particle position vector of the optimization variable: Y i =[Z wind Z pv Z esswind Z esspv ] Among them, Z wind For wind power installed capacity, Z pv For photovoltaic installed capacity, Z esswind For wind power energy storage installed capacity, Z esspv This refers to the installed capacity of photovoltaic energy storage.

[0040] Constraints: Wind power installed capacity constraints: Z wind (1) Photovoltaic installed capacity constraints: Z pv (2) Wind power energy storage installed capacity constraints: Z esswind (3) Photovoltaic energy storage installed capacity constraints: Z esspv (4) Constraints on the proportion of green electricity generated and consumed by the user: (5) Green electricity direct supply ratio constraints: (6) Power feedback ratio constraint: (7) Power balance constraints: (8) in, For the developable capacity of the wind power project area, For the developable capacity of the photovoltaic project area, The required proportion of self-consumption of new energy electricity. To meet the required green electricity direct supply ratio, To meet the required proportion of new energy sources connected to the grid, This refers to the amount of green electricity directly supplied by wind power (excluding energy storage). This refers to the amount of green electricity directly supplied by photovoltaic power (excluding energy storage). Directly supplying green electricity to wind power storage Directly supplying green electricity to photovoltaic energy storage For wind power generation, For photovoltaic power generation, This refers to the amount of wind power fed into the grid (excluding energy storage). This refers to the amount of electricity generated by photovoltaic power grid connection (excluding energy storage). Wind power storage grid-connected electricity, Photovoltaic energy storage grid-connected electricity, In order to purchase electricity from the grid, Electricity consumption for the chemical production of hydroammonium alcohol.

[0041] It is worth explaining that when conducting system planning for a specific project, in addition to pursuing the lowest LCOE, it is also necessary to fully consider the objective conditions and actual constraints faced by the project. First, the constraints (1) and (2) set the wind power and photovoltaic installed capacity to be greater than or equal to zero and not exceed the regional development limit. This is to avoid the optimization results from deviating from the reality of resource endowment. For example, if the regional wind power development capacity is only 200MW, without this constraint, the optimization scheme may propose an installed capacity that exceeds the actual development capacity, resulting in the result losing its engineering value. Second, the constraints (3) and (4) require the supporting energy storage installed capacity to be a non-negative value, ensuring the physical rationality of the configuration scheme. Third, the constraints (5) to (7) are the setting of constraints such as green electricity ratio and reverse power transmission, which are in response to the green certification policy and the requirements for grid safety operation, ensuring that the hydrogen ammonia products generated by the project meet green standards, while maintaining grid stability and economy. Finally, constraint (8) is a real-time power balance constraint during system operation. It is a necessary condition to ensure that the sum of wind, solar, and energy storage output and grid-purchased electricity always matches the demand of chemical loads. If this condition is not met, it will lead to a power supply gap and make it impossible to achieve the predetermined production target. Therefore, by incorporating the above-mentioned multiple constraints, the optimization model can generate a capacity configuration scheme with practical guiding significance under the premise of technical feasibility and policy compliance.

[0042] Furthermore, the reasoning process for the above formula is detailed below:

[0043] in, Let m be the current moment, and m be a future moment. The predicted wind and solar power generation over a time period of m hours (m≤n*8760). For the period from time t to time t+m (i.e., within the time interval m), the corresponding and The total output of wind and solar installed capacity, i.e., the predicted wind and solar power generation under that installed capacity. and By P fg It is obtained by breaking it down, specifically, and These are the predicted values ​​of wind power output per unit capacity and photovoltaic output per unit capacity at a certain moment, respectively.

[0044]

[0045]

[0046]

[0047] in, , and These are, respectively, wind power generation, wind power direct supply of green electricity (excluding energy storage), wind power energy storage direct supply of green electricity, wind power grid connection (excluding energy storage), wind power energy storage grid connection, photovoltaic power generation, photovoltaic direct supply of green electricity (excluding energy storage), photovoltaic energy storage direct supply of green electricity, photovoltaic grid connection (excluding energy storage), photovoltaic energy storage grid connection, and electricity consumption for the production of hydrogen ammonia and methanol within a time period of m hours. For the period from time t to time t+m (i.e., within the time interval m), the corresponding The total wind power output of the installed wind power capacity. For the period from time t to time t+m (i.e., within the time interval m), the corresponding The total photovoltaic output of the installed photovoltaic capacity This represents the total load of chemical production of hydrogen ammonia from time t to time t+m (i.e., within the time period m).

[0048] In addition, the optimization algorithm is incorporated into the wind-solar-storage power supply logic, such as Figure 6 As shown. Following the principles of "availability priority, economic dispatch, and grid backup," the system first determines whether wind power, solar power, and their associated energy storage are available. If all are unavailable, the entire load is borne by the grid. If available, the real-time grid-connected electricity prices of wind power and solar power are compared, prioritizing the dispatch of the energy source with the lower price. The system first calls upon the power source itself to generate electricity; if the generation is insufficient, its associated energy storage is called upon to discharge. If this still does not meet the load demand, another energy source and its associated energy storage are called upon. If, at this moment, all wind, solar, and energy storage are exhausted and there is still a load gap, the grid will supplement the power supply.

[0049] Step S5: Initialize the algorithm, run the optimization, and obtain the optimization results.

[0050] The number of initial particles and the number of iterations can be adjusted according to the scale of the planned project.

[0051] In the particle initialization phase, for multi-objective cases, Latin hypercube sampling is used for population initialization. Compared with traditional random initialization, the Latin hypercube sampling method can generate a more uniform and representative initial particle swarm in the solution space, ensuring that all regions are covered, thereby accelerating the convergence speed and increasing the probability of finding the global optimum.

[0052] During the fitness evaluation phase, for the wind-solar-storage configuration scheme corresponding to the particle's current position, the simulation of the scheme's operation at every moment in a future time period (e.g., 8760 hours) is conducted. This process strictly adheres to physical laws and grid dispatch rules, executing operations such as charging and discharging, power curtailment, and grid purchase / sales. Ultimately, the scheme's level of economic efficiency (LCOE) and technical indicators (such as green electricity rate and power return rate) are calculated. When updating the individual historical optimum and global optimum, two criteria are employed to ensure the search direction always points towards the feasible region and the region with better performance: 1) a feasible solution is always superior to an infeasible solution; 2) between two feasible solutions, the objective function is compared, and the one with the lower cost is selected.

[0053] When judging convergence and termination conditions, the algorithm terminates when any of the following conditions are met: 1) the preset maximum number of iterations is reached; 2) the improvement of the global optimal solution in multiple consecutive generations is less than a minimum threshold; 3) the variance of all feasible particles in the population is lower than the threshold, indicating that the algorithm has fully converged.

[0054] After convergence or termination, output the optimization results.

[0055] Finally, the final plan is made based on the installed capacity of wind power, photovoltaic power, wind power energy storage, and photovoltaic energy storage.

[0056] This application also provides a planning system for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol production). The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the aforementioned method. This planning system for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol production) can implement various embodiments of the above-described planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol production) and achieves the same beneficial effects; further details are omitted here.

[0057] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen, ammonia, and methanol production, characterized in that, include: Constructing prediction models for wind and solar energy resources; Construct a wind and solar power output prediction model; Construct a power load prediction model for the chemical industry; Based on the wind and solar energy resource prediction model, the wind and solar power output prediction model, and the chemical power load prediction model, a multi-constraint adaptive particle swarm optimization algorithm model is constructed. The final optimization variables are obtained based on the multi-constraint adaptive particle swarm optimization algorithm model, and planning is performed based on the optimization variables.

2. The planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen, ammonia, and methanol production according to claim 1, characterized in that, The construction of wind and solar energy resource prediction models includes: Long-series data of meteorological parameters at different altitudes are collected, including wind speed, wind direction, temperature, humidity, air pressure, solar radiation, and solar altitude angle. A wind energy and solar energy resource prediction model is established based on the long-series data of the meteorological parameters. The input of the wind energy and solar energy resource prediction model is the long-series data of the meteorological parameters, and the output of the wind energy and solar energy resource prediction model is wind speed, wind direction, air density and total solar radiation at horizontal level with time sequence.

3. The planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol) according to claim 1 or 2, characterized in that, The functions of the wind and solar energy resource prediction models satisfy the following relationship: in, Here, H represents the LSTM nonlinear mapping function, and H is the measured height. For wind turbine hub height, Wind speed at height H The air density at height H. Wind direction at height H Temperature at height H H represents high humidity. H represents atmospheric pressure at altitude, G represents total solar radiation, D represents scattered solar radiation, and B represents direct solar radiation in the normal direction. For terrain slope, Due to the slope of the terrain, Let t be the solar altitude angle, t be the current time, and m be a future time.

4. The planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol) according to claim 1, characterized in that, The construction of the wind and solar power output prediction model includes: The physical model for wind power output conversion is established as follows: ； In the formula, This refers to the output power of wind power. Let C be the swept area of ​​the wind turbine rotor, and C be the wind energy utilization coefficient, a dimensionless coefficient characterizing the efficiency of the wind turbine in capturing energy from the wind. The air density at the height of the wind turbine hub. The wind speed is measured at the height of the wind turbine hub. The physical model for photovoltaic power conversion is established as follows: ； in, For photovoltaic power output, The total area of ​​the photovoltaic array For the overall efficiency of photovoltaic systems, Total solar radiation; Based on the aforementioned wind power output conversion physical model, photovoltaic power output conversion physical model, and wind and solar energy resource prediction models, a wind and solar power output prediction model is constructed as follows: h fg =s fg ( *x fg + ) P fg =f fg ( *h fg + ) In the formula, x fg It is the input feature vector, i.e., the output data of the wind and solar energy resource prediction model, σ fg It is the activation function of the hidden layer, f fg It is the activation function of the output layer. , These are the weight matrix and bias vector of the hidden layer, respectively. , These are the weight matrix and bias vector of the output layer, h, respectively. fg P is the output vector of the hidden layer. fg The predicted value for the power output per unit capacity of wind and solar power.

5. The planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol) according to claim 1, characterized in that, The construction of the chemical industry electricity load prediction model includes: A predictive model for electricity load in the chemical industry is established. The inputs of the predictive model are time characteristics, product type, planned output, production calendar, equipment commissioning plan, historical operating data of similar factories, and meteorological conditions. The output is the electricity load for the production of hydrogen ammonia in the chemical industry. The chemical load forecasting model is as follows: h hg =s hg ( *x hg + ) P hg =f hg ( *h hg + ) In the formula, x hg It is the input feature vector, σ hg It is the activation function of the hidden layer, f hg It is the activation function of the output layer. , These are the weight matrix and bias vector of the hidden layer, respectively. , These are the weight matrix and bias vector of the output layer, h, respectively. hg is the output vector of the hidden layer, and Phg is the predicted value of the chemical load.

6. The planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen, ammonia, and methanol production according to claim 1, characterized in that, The objective function of the multi-constraint adaptive particle swarm optimization algorithm model is the LCOE of the wind-solar-storage system; the optimization variables include wind power installed capacity, photovoltaic installed capacity, wind power energy storage installed capacity, and photovoltaic energy storage installed capacity; the constraints include green electricity rate constraints, power return amount constraints, installed capacity constraints, power supply balance constraints, and power supply logic constraints.

7. The planning method for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen production (ammonia and methanol) according to claim 1, characterized in that, The objective function of the multi-constraint adaptive particle swarm optimization algorithm model is as follows: Where q represents the project's operation and maintenance lifespan. For all costs incurred in year n, For the wind and solar power generation in year n, The discount rate is n, and the number of years the system operates. The particle position vector of the optimization variable in the multi-constraint adaptive particle swarm optimization algorithm model is as follows: Y i =[Z wind ,Z pv ,Z esswind ,Z esspv ] Among them, Z wind For wind power installed capacity, Z pv For photovoltaic installed capacity, Z esswind For wind power energy storage installed capacity, Z esspv For photovoltaic energy storage installed capacity; The constraints of the multi-constraint adaptive particle swarm optimization algorithm model are as follows: Wind power installed capacity constraints: Z wind Photovoltaic installed capacity constraints: Z pv Wind power energy storage installed capacity constraints: Z esswind Photovoltaic energy storage installed capacity constraints: Z esspv Constraints on the proportion of green electricity generated and consumed by the user: Green electricity direct supply ratio constraints: Power feedback ratio constraint: Power balance constraints: in, For the developable capacity of the wind power project area, For the developable capacity of the photovoltaic project area, The required proportion of self-consumption of new energy electricity. To meet the required green electricity direct supply ratio, To meet the required proportion of new energy sources connected to the grid, For direct supply of green electricity from wind power, For direct photovoltaic green electricity supply, Directly supplying green electricity to wind power storage Directly supplying green electricity to photovoltaic energy storage For wind power generation, For photovoltaic power generation, This refers to the amount of wind power generated and fed into the grid. For photovoltaic power generation, Wind power storage grid-connected electricity, Photovoltaic energy storage grid-connected electricity, In order to purchase electricity from the grid, Electricity consumption for the chemical production of hydroammonium alcohol.

8. A planning system for optimizing the matching of wind, solar, and energy storage capacity with the scale of hydrogen, ammonia, and methanol production, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of any of the methods described in claims 1-7.

Images