Intelligent centralized control method and system of electric energy router based on B / S architecture

Abstract

A kind of B / S architecture-based electric energy router intelligent centralized control method and system, method includes: according to the physical system of electric energy router, the day-ahead optimization scheduling model with minimum system cost as target is established;The day-ahead optimization scheduling model is input into the intelligent centralized control system established based on B / S architecture, and the load data in future T time is predicted through long short-term memory neural network, to determine the constraint condition of day-ahead optimization scheduling model;The day-ahead optimization scheduling model is solved by NSGA-III algorithm, and the optimal power flow control scheme of electric energy router is output.The present application can provide the day-ahead optimization scheduling scheme of electric energy router system, the compatibility of the intelligent centralized control system established based on B / S architecture is good, can be integrated into larger integrated control platform, can realize the accurate prediction of system load power and the efficient solution of multi-objective function optimization model, still has better performance optimization effect when target function dimension is higher.

Description

Technical Field

[0001] This invention belongs to the field of power system automation dispatching technology, specifically relating to an intelligent centralized control method and system for power routers based on a B / S architecture. Background Technology

[0002] As the installed capacity of renewable energy increases, microgrid control is becoming increasingly complex and diverse, and the amount of data generated by grid operation is growing, making optimized scheduling increasingly difficult. Power routers, which integrate information technology and power electronic conversion technology, can largely solve this problem because they enable the efficient utilization and transmission of distributed energy.

[0003] A power router mainly consists of two parts: a physical system and a control system. The physical system, based on power electronic conversion technology, provides the necessary power interface for various types of distributed power sources, energy storage devices, and new loads. Furthermore, due to the high controllability of power electronic equipment, the direction and magnitude of energy flow at each node within the distribution network can be precisely controlled according to user needs, achieving multi-directional energy flow and active control of power flow. The information technology-based control system enables the power router to have communication and intelligent decision-making capabilities. It can monitor the power router's operating status, provide optimized scheduling schemes, and distribute them to the physical system, thus realizing the intelligent operation of the power router.

[0004] The physical system technology of power routers is relatively mature, and various power electronics equipment manufacturers have developed DC / DC and DC / AC converters for different demonstration scenarios. However, because the control of current power router systems is based on converter-level control, the control principle is simple and the system data volume is relatively small, resulting in the immaturity of the power router control system. This is mainly reflected in the following two points: First, there is a lack of a control system architecture with intelligent decision-making capabilities; second, with the increase in system complexity and data volume, traditional optimization scheduling algorithms are no longer able to meet the power flow control requirements of power router systems.

[0005] Currently, the main optimization scheduling method used in power routers is day-ahead optimization scheduling. After the power router system modeling method is determined, the quality of the day-ahead optimization scheduling method depends mainly on two aspects: first, the accuracy of load power prediction; and second, how to obtain the optimal solution of the model.

[0006] Load data required for load power forecasting exhibits strong temporal and periodic characteristics, thus time-series forecasting methods are commonly used. Deep learning neural networks, such as LSTM (Long Short-Term Memory), perform excellently in time-series forecasting, but a single power router system cannot provide sufficient training data. Load data from different power routers show strong correlations and can be used as training data to jointly train an LSTM; however, the varying sizes of training data result in poor training performance for the neural network. Therefore, finding a deep learning prediction algorithm capable of learning and building a unified training model using load data of different sizes would significantly improve the accuracy of power router load power forecasting.

[0007] The goal of model solving is to obtain the optimal power flow control scheme for a power router. However, due to the increasing system complexity of power routers and the refinement of system modeling methods, the number of objective functions in power router models is constantly increasing, making model solving increasingly difficult. Currently, commonly used algorithms for multi-objective optimization problems include NSGA-2 and MOEA / D. Among them, the NSGA-2 algorithm exhibits an exponential increase in non-dominated individuals as the number of objective functions increases, making it difficult for Pareto dominance relationships to distinguish between good and bad individuals; MOEA / D, on the other hand, cannot guarantee the distribution of solutions in high-dimensional multi-objective optimization problems. Therefore, finding an optimization algorithm that can maintain good performance even with high-dimensional objective functions will be beneficial for obtaining the optimal power flow control scheme for power routers. Summary of the Invention

[0008] The purpose of this invention is to address the problems in the prior art by providing an intelligent centralized control method and system for power routers based on a B / S architecture. This method can provide a day-ahead optimization scheduling scheme for power router systems and achieve good optimization performance even when the objective function dimension is high during the process of solving the optimal power flow control scheme for power routers.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A method for intelligent centralized control of a power router based on a B / S architecture includes:

[0011] A day-ahead optimization scheduling model with the goal of minimizing system cost is established based on the physical system of the power router.

[0012] The day-ahead optimization scheduling model is input into an intelligent centralized control system based on a B / S architecture, and the load data within the next T time period is predicted by a long short-term memory neural network to determine the constraints of the day-ahead optimization scheduling model.

[0013] The day-ahead optimization scheduling model is solved using the NSGA-III algorithm, and the optimal power flow control scheme for the power router is output.

[0014] As a preferred embodiment, the physical system of the power router includes a power grid, a wind power generation system, a photovoltaic power generation system, a water electrolysis hydrogen production-hydrogen storage tank-fuel cell system, user loads, converters, and AC and DC buses;

[0015] The AC bus is connected to the power grid through an isolation transformer, the wind power generation system is connected to the AC bus through AC / DC and DC / AC converters, and the user load is connected to the AC bus.

[0016] The photovoltaic power generation system and the hydrogen fuel cell power generation system and the water electrolysis hydrogen production-hydrogen storage tank-fuel cell system are respectively connected to the DC bus via DC / DC converters. The hydrogen storage tank in the water electrolysis hydrogen production-hydrogen storage tank-fuel cell system is simultaneously connected to the fuel cell of the hydrogen fuel cell power generation system and the electrolyzer of the water electrolysis hydrogen production system.

[0017] The AC and DC buses are connected via a DC / AC converter.

[0018] As a preferred option, the day-ahead optimization scheduling model with the goal of minimizing system cost includes a wind power generation system model, a photovoltaic power generation system model, a hydrogen fuel cell power generation system model, a water electrolysis hydrogen production system model, and a hydrogen storage tank model.

[0019] The model expression for a wind power generation system is as follows:

[0020]

[0021] in, Let be the maximum power output of the wind turbine at time t. The rated power of the wind turbine unit. For real-time wind speed, and These are the cut-in wind speed, cut-out wind speed, and rated wind speed of the wind turbine.

[0022] The model expression for a photovoltaic power generation system is as follows:

[0023]

[0024] in, Let be the maximum output power of the photovoltaic generator set at time t. This refers to the rated power of the photovoltaic generator set; and These are the solar irradiance and photovoltaic module surface temperature under standard conditions, respectively; k is the power temperature coefficient by which the photovoltaic module temperature affects the maximum output power. The operating temperature of the photovoltaic module is calculated using the following formula:

[0025]

[0026] In the formula, Ambient temperature; It is the real-time solar irradiance intensity; The rated operating temperature of the photovoltaic module;

[0027] The model expression for a hydrogen fuel cell power generation system is as follows:

[0028]

[0029] In the formula, Let be the maximum output power of the fuel cell at time t. The rated output power of the fuel cell, Let δ be the remaining energy storage capacity of the hydrogen storage tank at time t. FC The conversion efficiency of hydrogen fuel cells; The lower limit of the hydrogen storage capacity of the hydrogen storage tank; Δt is the length of the control cycle;

[0030] The model expression for the water electrolysis hydrogen production system is as follows:

[0031]

[0032] in, Let be the maximum power consumption of the electrolytic cell at time t. The rated power of the electrolytic cell, Let t be the remaining hydrogen storage capacity of the hydrogen storage tank at time t. This represents the upper limit of the hydrogen storage tank's energy storage capacity; δ E The hydrogen production efficiency of the electrolyzer;

[0033] The model expression for the hydrogen storage tank is as follows:

[0034]

[0035] In the formula, It is the actual power consumption of the electrolytic cell at time t; It is the actual electrical power output of the hydrogen fuel cell at time t.

[0036] As a preferred embodiment, the day-ahead optimization scheduling model, which aims to minimize system cost, further includes the following objective function:

[0037] minC=C system +C offset

[0038] In the formula, C represents the total system cost. systemFor system operating costs, C offset For power deviation costs;

[0039] System operating cost C system The calculation expression is as follows:

[0040]

[0041] In the formula, These are the total cost of system-grid interaction during time period t, and the maintenance costs of photovoltaic panels, wind turbines, electrolyzers, fuel cells, and hydrogen storage tanks, respectively.

[0042] Total cost of system-grid interaction during time period t The costs include both electricity price and environmental protection costs, calculated using the following formula:

[0043]

[0044]

[0045] In the formula, The electricity price paid by the power router to the grid during time period t; The environmental costs of a large power grid, namely, the costs of pollutant treatment; γ k It refers to the emissions of Class K pollutants generated by power generation from a large power grid; c k The cost coefficient for treating pollutants of type k;

[0046] Operating costs and power generation are directly proportional to the following:

[0047]

[0048] In the formula, ρ E ρ is the loss coefficient during the operation of the electrolytic cell. FC ρ is the loss factor during the operation of the hydrogen fuel cell; W ρ is the loss coefficient of the wind turbine during operation. PV ρ is the loss factor of the optoelectronic unit during operation. tank This is the loss coefficient when hydrogen is introduced into or removed from the hydrogen storage tank. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit;

[0049] Electricity Deviation Cost C offset The calculation expression is as follows:

[0050]

[0051]

[0052] In the formula, ΔP t Let μ1 be the offset power of the power router system at time t; μ2 be the penalty coefficient for the offset power of the power purchased from the grid; μ3 be the penalty coefficient for the offset power of the wind and solar power curtailment. Let t be the maximum output power of the photoelectric generator. Let t be the maximum output power of the wind turbine.

[0053] As a preferred embodiment, the day-ahead optimization scheduling model with the objective of minimizing system cost also includes the following constraints:

[0054] System power balance constraints:

[0055]

[0056] In the formula, These are, respectively, the system load power at time t, the power purchased by the system from the grid, the actual output power of the photovoltaic generator, and the actual output power of the wind turbine;

[0057] Actual output power constraints for each device:

[0058]

[0059] In the formula, This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This represents the upper limit of the power that the system can purchase from the grid.

[0060] As a preferred embodiment, the intelligent centralized control system based on the B / S architecture includes physical devices at the bottom layer, servers at the second layer, and browsers at the third layer.

[0061] The underlying physical devices establish a communication connection with the second-layer server according to the transmission protocol, and are responsible for uploading physical device data to the server and receiving control commands from the server to control the physical devices.

[0062] The server uses a database to store data. It can receive data uploaded by physical devices and instructions sent by the browser, save the data in the database and respond accordingly. It can also upload the data in the database to the browser for display and send control instructions to the underlying physical devices based on pre-written decision logic.

[0063] Browsers can read data from the server's database for display, and can also send control commands to the server for execution.

[0064] As a preferred embodiment, the step of predicting load data within the next time period T using a long short-term memory neural network includes:

[0065] The training set is based on the historical load data of a set number of power router systems, the prediction set is based on the historical load data of this power router system, and the prediction target is the load data of this power router system within a future time T.

[0066] Before inputting the training set payload data into the Long Short-Term Memory neural network for training, each time series is grouped and standardized by dividing each time series by a standardization factor v. i Standardization factor v i The value is taken as the average load data. To avoid the average being 0, the average load data is incremented by 1. The calculation expression is as follows:

[0067]

[0068] Among them, v i is the standardization factor for each load data point; l is the sequence length of the load data. This is the data for the i-th power load sequence;

[0069] After training, the prediction set data is first standardized and then input into the trained long short-term memory neural network for prediction. The model's output is multiplied by the corresponding standardization factor to obtain the load data for the next time period T.

[0070] As a preferred embodiment, the steps of solving the day-ahead optimal scheduling model using the NSGA-III algorithm and outputting the optimal power flow control scheme for the power router include:

[0071] N power flow control schemes are randomly generated in the feasible solution space formed by the constraints.

[0072] Entering the iterative process, let P... t Let P be the initial power flow control scheme for the t-th iteration, with a quantity of N, and the initial power flow control scheme for the t-th iteration be P. t In the simulation of two-point crossover and polynomial mutation, a new set of control schemes of number N, Q, is generated by randomly selecting a control scheme. t By merging the initial scheme and the new scheme, a control scheme set R of size 2N is obtained. t =P t ∪Q t ;

[0073] The control scheme set R is sorted using the fast non-dominated sorting algorithm.t It is divided into multiple dominance layers (F1, F2, ...);

[0074] Based on the reference point selection mechanism, N power flow control schemes are selected from each level of non-dominated layer to enter the next cycle as the initial solution for the next cycle;

[0075] Return and continue iterating until the required number of iterations is reached;

[0076] Optimal power flow control scheme for output power router system.

[0077] A centralized intelligent control system for power routers based on a B / S architecture includes:

[0078] The day-ahead optimization scheduling model establishment module is used to establish a day-ahead optimization scheduling model with the goal of minimizing system cost based on the physical system of the power router;

[0079] The load data prediction module is used to input the day-ahead optimization scheduling model into the intelligent centralized control system based on the B / S architecture, and predict the load data in the next T time period through the long short-term memory neural network to determine the constraints of the day-ahead optimization scheduling model.

[0080] The day-ahead optimization scheduling model solution module is used to solve the day-ahead optimization scheduling model using the NSGA-III algorithm and output the optimal power flow control scheme for the power router.

[0081] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the intelligent centralized control method for a power router based on a B / S architecture.

[0082] Compared with the prior art, the present invention has at least the following beneficial effects:

[0083] The intelligent centralized control system based on the B / S architecture boasts good compatibility and can be integrated into larger comprehensive control platforms; it offers strong sharing capabilities, requires zero client-side maintenance, and offers simple and convenient functional expansion. By predicting load data within the next T time period using a Long Short-Term Memory (LSTM) neural network, the constraints of the day-ahead optimal scheduling model are determined. The NSGA-III algorithm is then used to solve the day-ahead optimal scheduling model, enabling accurate prediction of system load power and efficient solution of multi-objective function optimization models, while maintaining good optimization performance even with high objective function dimensions. This invention provides a day-ahead optimal scheduling scheme for power router systems, and its integration with the power router physical system enhances the adaptive control capability of the power router system's power flow. This invention truly integrates information technology into the power router physical system, enabling the power router to possess advanced intelligent decision-making capabilities. Attached Figure Description

[0084] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0085] Figure 1 Physical structure topology diagram of the power router according to an embodiment of the present invention;

[0086] Figure 2 A schematic diagram of an intelligent centralized control system based on a B / S architecture according to an embodiment of the present invention;

[0087] Figure 3 The flowchart of the intelligent centralized control method for power routers based on B / S architecture in this invention is shown in the embodiment. Detailed Implementation

[0088] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, those skilled in the art can obtain other embodiments without creative effort.

[0089] This invention proposes an intelligent centralized control method for power routers based on a B / S architecture, applicable to MW-level power routers, such as... Figure 3 As shown, the specific implementation steps include:

[0090] Step 1: Establish the physical model of the power router.

[0091] like Figure 1 As shown, the power router consists of a physical system and a control system. The physical system comprises a power grid, a wind power generation system, a photovoltaic power generation system, a water electrolysis hydrogen production-hydrogen storage tank-fuel cell system, user loads, converters, and AC and DC buses.

[0092] The power router system operates under grid-connected conditions, with the grid ensuring the stability of the system bus voltage. Therefore, the system's AC bus is connected to the grid via an isolation transformer. The wind power generation system is connected to the AC bus via an AC / DC + DC / AC converter, and the system load is connected to the AC bus.

[0093] The photovoltaic power generation system, hydrogen fuel cell power generation system, and water electrolysis hydrogen production system are connected to the DC bus via a DC / DC converter, and the hydrogen storage tank is connected to both the fuel cell and the electrolyzer.

[0094] The AC and DC buses are connected via a DC / AC converter.

[0095] Step 2: Establish a mathematical model of the power router system.

[0096] Establishing a mathematical model of the system is a prerequisite for optimizing the scheduling of complex systems. Step 2 establishes a day-ahead optimization scheduling model with the goal of minimizing system cost, based on the physical system of the power router described in Step 1.

[0097] Step 2.1, Modeling the wind power generation system.

[0098] The maximum power output of the wind turbine at time t Subject to its own rated power and real-time wind speed The wind turbine's power generation is limited. When the wind speed is less than the cut-in wind speed or greater than the cut-out wind speed, the wind turbine's power generation is 0; when the wind speed is greater than the cut-in wind speed but less than the rated wind speed, the wind turbine's power generation increases linearly with increasing wind speed; when the wind speed is greater than the rated wind speed but less than the cut-out wind speed, the wind turbine's power generation remains constant at the rated power. In summary, It can be represented as:

[0099]

[0100] in, and These are the cut-in wind speed, cut-out wind speed, and rated wind speed of the wind turbine.

[0101] Step 2.2, Photovoltaic power generation system modeling.

[0102] The maximum power output of a photovoltaic (PV) generator at time t is limited by solar irradiance, ambient temperature, and the rated power of the PV generator. According to the NMOT test conditions published in IEC 61215, the operating temperature of the PV module... This can be approximated by measuring ambient temperature and solar irradiance:

[0103]

[0104] in, Ambient temperature; It is the real-time solar irradiance intensity; Let t be the rated operating temperature of the photovoltaic module. Then, the maximum output power of the photovoltaic generator set at time t is... It can be represented as:

[0105]

[0106] in, This is the rated power of the photovoltaic generator set; and These are the solar irradiance and photovoltaic module surface temperature under standard conditions, respectively; k is the power temperature coefficient by which the photovoltaic module temperature affects the maximum output power.

[0107] Step 2.3: Establishment of the model for the water electrolysis hydrogen production-hydrogen storage tank-fuel cell system.

[0108] Step 2.3.1, Establishment of hydrogen fuel cell model.

[0109] The maximum output power of the fuel cell at time t Subject to the rated output power of the fuel cell itself and the remaining energy storage capacity of the hydrogen storage tank at time t limit, It can be represented as:

[0110]

[0111] Where, δ FC It refers to the conversion efficiency of hydrogen fuel cells; Δt is the lower limit of the hydrogen storage capacity of the hydrogen storage tank; Δt is the length of the control cycle.

[0112] Step 2.3.2, Establishment of the model for the water electrolysis hydrogen production system.

[0113] The water electrolysis hydrogen production equipment consists of an electrolyzer system, a hydrogen purification system, and a control system, with the electrolyzer system being the core component. The maximum power consumption of the electrolyzer at time t is... Subject to its own rated power and the remaining hydrogen storage capacity of the hydrogen storage tank at time t limit, It can be represented as:

[0114]

[0115] in, This is the upper limit of the energy storage capacity of the hydrogen storage tank; δ E It refers to the hydrogen production efficiency of the electrolyzer.

[0116] Step 2.3.3, Establishment of hydrogen storage tank model.

[0117] Hydrogen storage cylinders use compressed gas to store hydrogen. The real-time hydrogen storage capacity of the storage tank is equivalent to stored energy, which can be expressed as:

[0118]

[0119] in, It is the actual power consumption of the electrolytic cell at time t; It is the actual electrical power output of the hydrogen fuel cell at time t.

[0120] Step 2.4, Establishment of the mathematical model of the objective function.

[0121] The system is already built and operational, therefore installation costs are not considered. The total system cost consists of the system operating cost C. system And electricity deviation cost C offset The composition and optimization of the system scheduling objective are equivalent to finding the minimum cost:

[0122] min C = C system +C offset (7)

[0123] Step 2.4.1, Establish the system operating cost model.

[0124] The system operating costs mainly include: the total cost of the interaction between the system and the power grid, and the maintenance costs of photovoltaic panels, wind turbines, electrolyzers, fuel cells, and hydrogen storage tanks.

[0125]

[0126] in, These figures represent the total cost of system-grid interaction during time period t, and the maintenance costs of photovoltaic panels, wind turbines, electrolyzers, fuel cells, and hydrogen storage tanks. The total cost of system-grid interaction mainly includes electricity costs and environmental costs.

[0127]

[0128]

[0129] in, The electricity price paid by the power router to the grid during time period t; The environmental costs of a large power grid, namely, the costs of pollutant treatment; γ k It refers to the emissions of Class K pollutants generated by power generation from a large power grid; c k This represents the cost coefficient for treating pollutants of type k.

[0130] The operating cost of wind, solar and energy storage modules is approximately proportional to the amount of electricity generated; here, the proportionality coefficient is set as ρ.

[0131]

[0132] Where, ρ E ρ is the loss coefficient during the operation of the electrolytic cell. FC ρ is the loss factor during the operation of the hydrogen fuel cell; W ρ is the loss coefficient of the wind turbine during operation. PV ρ is the loss factor of the optoelectronic unit during operation. tank This is the loss coefficient when hydrogen is introduced into or removed from the hydrogen storage tank. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit;

[0133] Step 2.4.2, Establish the system power deviation cost model.

[0134] There are two main types of power deviations in the power router system. The first is that the total power generation of hydrogen fuel cells, wind power, and solar power is insufficient to support the total system load, and power needs to be purchased from the grid. The second is that the total power generation of wind and solar power is greater than the actual wind and solar power used in the dispatch, resulting in wind and solar curtailment.

[0135] Deviation power is detrimental to system co-optimization and is an undesirable quantity in the system's optimization results. Therefore, in the context of solving for the minimum real-time operating cost, this portion of power is multiplied by a higher cost coefficient μ.

[0136]

[0137]

[0138] Where: ΔP t Let μ1 be the offset power of the power router system at time t; μ2 be the penalty coefficient for the offset power of the power purchased from the grid; μ3 be the penalty coefficient for the offset power of the wind and solar power curtailment. Let t be the maximum output power of the photoelectric generator. Let t be the maximum output power of the wind turbine.

[0139] Step 2.5: Establish system constraints.

[0140] Step 2.5.1: Establish system power balance constraints.

[0141]

[0142] in, These are the system load power at time t, the power purchased by the system from the grid, the actual output power of the photovoltaic generator, and the actual output power of the wind turbine.

[0143] Step 2.5.2: Establish constraints on the operating range of the equipment.

[0144] The output power constraints of each device are shown in the following formula:

[0145]

[0146] In the formula, This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This is the upper limit of the power that the system can purchase from the grid.

[0147] At this point, the mathematical modeling of the power router topology in step 1 is complete. The model input is... The output of the model is The objective function of the model is given by equations (7)(8)(9); the constraints of the model are given by equations (10)(11).

[0148] in, This represents the remaining hydrogen storage capacity in the hydrogen storage tank at the initial moment.

[0149] Step 3: The intelligent centralized control system is implemented.

[0150] Step 3.1: Establishment of the intelligent centralized control system architecture.

[0151] The main architecture of a centralized intelligent control system is as follows: Figure 2 As shown, the control system is mainly divided into three layers.

[0152] The lowest layer is the device layer, which is the foundation for the implementation of the intelligent control system. Each device at the lowest level establishes a communication connection with the server at the second layer based on transmission protocols such as Modbus, and is responsible for uploading device data to the server and receiving control commands from the server to control the physical devices.

[0153] The second layer is the server, the core of the control system's functionality. The server uses a database for data storage, receiving data uploaded from lower-level devices and commands issued by the web browser. It stores the data in the database and responds accordingly. It can also upload data from the database to the browser for display and issue control commands to lower-level devices based on pre-written decision logic. The server embeds an intelligent optimization scheduling algorithm that calculates and stores the day-ahead optimal scheduling scheme for the power router in the database, which can then be invoked by its internal intelligent decision logic. The server communicates with the web client via protocols such as IEC 104.

[0154] The third layer is the browser. The browser can read data from the server's database for display and can also send control commands to the server for execution. All client machines with a browser installed can interact with the server after gaining access.

[0155] Step 3.2, implementation of intelligent optimization scheduling algorithm.

[0156] In step 2, a power router model was established. The model requires the following input variables:

[0157] These are light intensity, wind speed, and temperature, respectively. Local meteorological data published by the meteorological bureau for the next T time period can be directly used as forecast data. Combined with specific physical equipment data, the output constraints of wind turbines can be established according to equation (1); the output constraints of photovoltaic generators can be established according to equations (2) and (3). Equal to the last moment of the previous optimized scheduling cycle It can be calculated based on the optimized scheduling scheme of the previous cycle; combined with specific physical equipment data and equations (4), (5), and (6), the output constraints of hydrogen fuel cells and the power consumption constraints of electrolyzers can be established. This is the current electricity price data, which can be set in advance based on the latest data released locally.

[0158] In summary, a complete set of output power constraint equations (11) for each device can be established. To achieve day-ahead optimal scheduling of the power router system, two problems still need to be solved: First, to achieve load power... First, it enables accurate prediction and improves system power balance constraints; second, it enables efficient solution of multi-objective function optimization models.

[0159] Step 3.2.1, System load power prediction is achieved.

[0160] In this embodiment of the invention, the prediction of system load data uses LSTM (Long Short-Term Memory Neural Network) as the main algorithm, a large amount of historical load data of the power router system as the training set, the historical load data of the current power router system as the prediction set, and the load data of the current power router in the future within T time period as the prediction target, so as to achieve the prediction of system load data.

[0161] To address the issue of inconsistent training set data sizes mentioned in the background, this embodiment of the invention performs grouping and standardization on each time series before inputting the training set payload data into the LSTM for training. Each time series is divided by a standardization factor v. i Standardization factor v i The value is taken as the average load data. In order to avoid the average value being 0, the average load data is processed by adding 1 (Equation 12) to avoid the problem of dividing by 0 during the calculation process.

[0162]

[0163] Among them, v i is the standardization factor for each load data point; l is the sequence length of the load data. This is the i-th power load sequence data.

[0164] After training, the prediction set data is first standardized according to equation (12), and then input into the trained neural network for prediction. The output of the model is multiplied by the corresponding standardization factor to obtain the load data for the future time period T, which is the system model in step 2. Thus, the power balance constraint equation (10) of the power router model can be established.

[0165] The model constraints have now been established.

[0166] Step 3.2.2, implementation of the model solution method.

[0167] The optimization objective established in step 2.5, equation (7), represents the total system cost. The total cost can be broken down into two optimization objectives: equation (8) represents the system operating cost, and equation (9) represents the system power deviation cost. The magnitude of the system operating cost depends on… The power setting and the magnitude of the system power deviation cost depend on The power settings. Finding the optimal power flow control scheme for the power router is equivalent to finding the optimal solution for the two objective functions (8) and (9).

[0168] The model solving method used in this embodiment of the invention is based on the NSGA-III algorithm. The algorithm modifies the mechanism of the NSGA-II algorithm, which selects the best control scheme based on the crowding of the objective function, and instead adopts a reference point-based selection mechanism after non-dominated sorting. This still achieves good optimization results when the objective function has a high dimension.

[0169] The main steps of the algorithm in this embodiment of the invention are as follows:

[0170] a) Randomly generate N power flow control schemes in the feasible solution space formed by the constraints.

[0171] b) Enter the iterative process, let P t Let N be the initial power flow control schemes for the t-th iteration, with a quantity of N, in P t In the simulation of two-point crossover and polynomial mutation, a new set of control schemes of number N, Q, is generated by randomly selecting a control scheme. t By merging the initial scheme and the new scheme, a control scheme set R of size 2N is obtained. t =P t ∪Q t .

[0172] c) The control scheme set R is sorted using a fast non-dominated sorting algorithm. t It is divided into multiple dominance layers (F1, F2, ...).

[0173] d) Select N power flow control schemes from each level of non-dominated layer based on the reference point selection mechanism to enter the next cycle as the initial solution for the next cycle.

[0174] e) Return to step b and continue iterating until the required number of iterations is reached.

[0175] f) Optimal power flow control scheme for the output power router system.

[0176] Another embodiment of the present invention also proposes an intelligent centralized control system for power routers based on a B / S architecture, comprising:

[0177] The day-ahead optimization scheduling model establishment module is used to establish a day-ahead optimization scheduling model with the goal of minimizing system cost based on the physical system of the power router;

[0178] The load data prediction module is used to input the day-ahead optimization scheduling model into the intelligent centralized control system based on the B / S architecture, and predict the load data in the next T time period through the long short-term memory neural network to determine the constraints of the day-ahead optimization scheduling model.

[0179] The day-ahead optimization scheduling model solution module is used to solve the day-ahead optimization scheduling model using the NSGA-III algorithm and output the optimal power flow control scheme for the power router.

[0180] Another embodiment of the present invention also proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the intelligent centralized control method for power routers based on a B / S architecture.

[0181] For example, the instructions stored in the memory can be divided into one or more modules / units. These modules / units are stored in a computer-readable storage medium and executed by the processor to complete the intelligent centralized control method for power routers based on a B / S architecture according to the present invention. The one or more modules / units can be a series of computer-readable instruction segments capable of performing specific functions, and these instruction segments describe the execution process of the computer program on the server.

[0182] The electronic device may be a smartphone, laptop, PDA, or cloud server, among other computing devices. It may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the electronic device may also include more or fewer components, or combinations of certain components, or different components; for example, it may also include input / output devices, network access devices, buses, etc.

[0183] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0184] The memory can be an internal storage unit of the server, such as a hard drive or RAM. Alternatively, it can be an external storage device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory can include both internal and external storage units. The memory is used to store computer-readable instructions and other programs and data required by the server. It can also be used to temporarily store data that has been output or will be output.

[0185] It should be noted that the information interaction and execution process between the above-mentioned module units are based on the same concept as the method embodiment. For details on their specific functions and technical effects, please refer to the method embodiment section. They will not be repeated here.

[0186] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0187] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0188] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0189] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for intelligent centralized control of a power router based on a B / S architecture, characterized in that, include: A day-ahead optimization scheduling model with the goal of minimizing system cost is established based on the physical system of the power router. The day-ahead optimization scheduling model is input into an intelligent centralized control system based on a B / S architecture, and the load data within the next T time period is predicted by a long short-term memory neural network to determine the constraints of the day-ahead optimization scheduling model. The day-ahead optimization scheduling model is solved using the NSGA-III algorithm, and the optimal power flow control scheme for the power router is output. The day-ahead optimization scheduling model with the goal of minimizing system cost includes the following objective function: In the formula, The total system cost, For system operating costs, For power deviation costs; System operating costs The calculation expression is as follows: In the formula, , , , , , They are respectively t Total cost of interaction between the time-of-use system and the power grid, maintenance costs of photovoltaic panels, wind turbines, electrolyzers, fuel cells, and hydrogen storage tanks; t Total cost of interaction between time-of-use system and power grid The costs include both electricity price and environmental protection costs, calculated using the following formula: In the formula, for t The electricity price that the time-of-use power router purchases from the grid; This refers to the environmental costs of a large power grid, namely, the costs of pollutant treatment. It is generated by the power generation of the large power grid. k Emissions of pollutants of this type; To process k Cost coefficients for pollutants of this type; Operating costs and power generation are directly proportional to the following: In the formula, This is the loss coefficient during the operation of the electrolytic cell; This represents the loss factor during the operation of the hydrogen fuel cell. This is the loss factor of the wind turbine during operation. This is the loss factor when the optoelectronic unit is working; This is the loss coefficient when hydrogen is introduced into or removed from the hydrogen storage tank. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit; Electricity Deviation Cost The calculation expression is as follows: In the formula, for Offset power of the instantaneous power router system; The penalty coefficient for the deviation in power purchased from the power grid; This is the penalty coefficient for the deviation in power output due to wind and solar curtailment. for The maximum output power of the photoelectric generator at any given time. for The maximum output power of the wind turbine at any given time; express The power that the time system purchases from the power grid.

2. The intelligent centralized control method for power routers based on B / S architecture according to claim 1, characterized in that, The physical system of the power router includes a power grid, a wind power generation system, a photovoltaic power generation system, a water electrolysis hydrogen production-hydrogen storage tank-fuel cell system, user loads, converters, and AC and DC buses; The AC bus is connected to the power grid through an isolation transformer, the wind power generation system is connected to the AC bus through AC / DC and DC / AC converters, and the user load is connected to the AC bus. The photovoltaic power generation system and the hydrogen fuel cell power generation system and the water electrolysis hydrogen production-hydrogen storage tank-fuel cell system are respectively connected to the DC bus via DC / DC converters. The hydrogen storage tank in the water electrolysis hydrogen production-hydrogen storage tank-fuel cell system is simultaneously connected to the fuel cell of the hydrogen fuel cell power generation system and the electrolyzer of the water electrolysis hydrogen production system. The AC and DC buses are connected via a DC / AC converter.

3. The intelligent centralized control method for power routers based on B / S architecture according to claim 2, characterized in that, The day-ahead optimization scheduling model with the goal of minimizing system cost includes a wind power generation system model, a photovoltaic power generation system model, a hydrogen fuel cell power generation system model, a water electrolysis hydrogen production system model, and a hydrogen storage tank model. The model expression for a wind power generation system is as follows: in, For wind turbine units t Maximum power output at any given time The rated power of the wind turbine unit. For real-time wind speed, , and These are the cut-in wind speed, cut-out wind speed, and rated wind speed of the wind turbine. The model expression for a photovoltaic power generation system is as follows: in, For photovoltaic generator sets in t Maximum output power at any given time This refers to the rated power of the photovoltaic generator set; and These are solar irradiance and photovoltaic module surface temperature under standard conditions, respectively. The power temperature coefficient is the effect of photovoltaic module temperature on maximum output power. The operating temperature of the photovoltaic module is calculated using the following formula: In the formula, The ambient temperature; It is the real-time solar irradiance intensity; The rated operating temperature of the photovoltaic module; The model expression for a hydrogen fuel cell power generation system is as follows: In the formula, For fuel cells in t Maximum output power at any given time The rated output power of the fuel cell, Hydrogen storage tank Remaining energy storage capacity at all times The conversion efficiency of hydrogen fuel cells; This is the lower limit of the hydrogen storage capacity of the hydrogen storage tank; To control the cycle time length; The model expression for the water electrolysis hydrogen production system is as follows: in, For the electrolytic cell in t Maximum power consumption at any time The rated power of the electrolytic cell, Hydrogen storage tank t Remaining hydrogen storage capacity at all times This represents the upper limit of the energy storage capacity of the hydrogen storage tank; The hydrogen production efficiency of the electrolyzer; The model expression for the hydrogen storage tank is as follows: In the formula, yes The actual power consumption of the electrolytic cell at any given time; yes The actual electrical power output of the hydrogen fuel cell at any given time.

4. The intelligent centralized control method for power routers based on B / S architecture according to claim 3, characterized in that, The day-ahead optimization scheduling model with the goal of minimizing system cost also includes the following constraints: System power balance constraints: In the formula, , , , They are Real-time system load power, power purchased from the grid, actual output power of photovoltaic units, and actual output power of wind turbine units; Actual output power constraints for each device: In the formula, This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This is the upper limit of the power that the system can purchase from the grid.

5. The intelligent centralized control method for power routers based on B / S architecture according to claim 1, characterized in that, The intelligent centralized control system based on the B / S architecture includes physical devices at the bottom layer, servers at the second layer, and browsers at the third layer. The underlying physical devices establish a communication connection with the second-layer server according to the transmission protocol, and are responsible for uploading physical device data to the server and receiving control commands from the server to control the physical devices. The server uses a database to store data. It can receive data uploaded by physical devices and instructions sent by the browser, save the data in the database and respond accordingly. It can also upload the data in the database to the browser for display and send control instructions to the underlying physical devices based on pre-written decision logic. Browsers can read data from the server's database for display, and can also send control commands to the server for execution.

6. The intelligent centralized control method for power routers based on B / S architecture according to claim 1, characterized in that, The step of predicting load data within the next T time period using a long short-term memory neural network includes: The training set is based on the historical load data of a set number of power router systems, the prediction set is based on the historical load data of this power router system, and the prediction target is the load data of this power router system within a future time T. Before inputting the training set payload data into the Long Short-Term Memory neural network for training, each time series is grouped and standardized by dividing each time series by a standardization factor. Standardization factor The value is taken as the average load data. To avoid the average being 0, the average load data is incremented by 1. The calculation expression is as follows: in, The standardization factor for each load data point; The sequence length of the load data; It is the first i Power load sequence data; After training, the prediction set data is first standardized and then input into the trained long short-term memory neural network for prediction. The model's output is multiplied by the corresponding standardization factor to obtain the load data for the next time period T.

7. The intelligent centralized control method for power routers based on B / S architecture according to claim 1, characterized in that, The steps of solving the day-ahead optimal scheduling model using the NSGA-III algorithm and outputting the optimal power flow control scheme for the power router include: Randomly generated in the feasible solution space constituted by the constraints. One power flow control scheme; Entering the iterative process, let... For the first The initial power flow control scheme for the next iteration has a number of [number]. In the The initial power flow control scheme for the next iteration The random selection of control schemes simulates two-point crossover and polynomial mutation, generating a number of... New control scheme set The initial and new schemes are combined to obtain a quantity of 2. control scheme set ; The control scheme set is sorted using the fast non-dominated sorting algorithm. Divided into multiple dominance layers ; From the reference point-based selection mechanism in each level of non-dominated layer, select... Each power flow control scheme enters the next cycle as the initial solution for the next cycle; Return and continue iterating until the required number of iterations is reached; Optimal power flow control scheme for output power router system.

8. A centralized intelligent control system for a power router based on a B / S architecture, characterized in that, include: The day-ahead optimization scheduling model establishment module is used to establish a day-ahead optimization scheduling model with the goal of minimizing system cost based on the physical system of the power router; The load data prediction module is used to input the day-ahead optimization scheduling model into the intelligent centralized control system based on the B / S architecture, and predict the load data in the next T time period through the long short-term memory neural network to determine the constraints of the day-ahead optimization scheduling model. The day-ahead optimal scheduling model solution module is used to solve the day-ahead optimal scheduling model using the NSGA-III algorithm and output the optimal power flow control scheme for the power router. The day-ahead optimization scheduling model with the goal of minimizing system cost includes the following objective function: In the formula, The total system cost, For system operating costs, For power deviation costs; System operating costs The calculation expression is as follows: In the formula, , , , , , They are respectively t Total cost of interaction between the time-of-use system and the power grid, maintenance costs of photovoltaic panels, wind turbines, electrolyzers, fuel cells, and hydrogen storage tanks; t Total cost of interaction between time-of-use system and power grid The costs include both electricity price and environmental protection costs, calculated using the following formula: In the formula, for t The electricity price that the time-of-use power router purchases from the grid; This refers to the environmental costs of a large power grid, namely, the costs of pollutant treatment. It is generated by the power generation of the large power grid. k Emissions of pollutants of this type; To process k Cost coefficients for pollutants of this type; Operating costs and power generation are directly proportional to the following: In the formula, This is the loss coefficient during the operation of the electrolytic cell; This represents the loss factor during the operation of the hydrogen fuel cell. This is the loss factor of the wind turbine during operation. This is the loss factor when the optoelectronic unit is working; This is the loss coefficient when hydrogen is introduced into or removed from the hydrogen storage tank. This represents the actual power consumption of the electrolytic cell. This represents the actual output power of the hydrogen fuel cell. This represents the actual output power of the wind turbine. This represents the actual output power of the optoelectronic unit; Electricity Deviation Cost The calculation expression is as follows: In the formula, for Offset power of the instantaneous power router system; The penalty coefficient for the deviation in power purchased from the power grid; This is the penalty coefficient for the deviation in power output due to wind and solar curtailment. for The maximum output power of the photoelectric generator at any given time. for The maximum output power of the wind turbine at any given time; express The power that the time system purchases from the power grid.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent centralized control method for power routers based on B / S architecture as described in any one of claims 1 to 7.

Images