Load peak regulation method considering user demand response

Abstract

The invention discloses a load peak regulation method considering user demand response. The method comprises the following steps: acquiring power grid operation data; establishing a load prediction model, and identifying a peak load; setting a target function based on the power consumption scene of the user; and carrying out model solving to obtain optimized peak regulation parameters, and providing economic excitation or non-economic excitation for users participating in peak regulation. The invention belongs to the technical field of power loads, and particularly provides a load peak regulation method considering user demand response, which introduces response of a user side, and improves the peak regulation flexibility of a power grid by using the load regulation capability of a user.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric loads, and specifically refers to a load peak regulation method taking user demand response into consideration. Background Art

[0002] With the large-scale access of renewable energy, the difference between peak and valley loads in the power grid has increased significantly, posing a challenge to the stable operation of the power system. Traditional load peaking methods mainly rely on power companies to adjust the output of generators. However, this method has limited flexibility and high economic costs.

[0003] Therefore, a new method is needed to utilize user-side resources to participate in peak load regulation. Summary of the invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a load peak-shaving method that takes user demand response into consideration, introduces the response of the user end, utilizes the user's load regulation capability, and improves the peak-shaving flexibility of the power grid.

[0005] The technical solutions adopted in this application are as follows:

[0006] This solution provides a load peak regulation method considering user demand response, including the following steps:

[0007] S1 obtains power grid operation data;

[0008] S2 builds a load forecasting model to identify peak loads;

[0009] S3 sets the objective function based on the user's electricity usage scenario;

[0010] S4 solves the model to obtain optimized peak-shaving parameters and provides economic incentives or non-economic incentives for users participating in peak-shaving.

[0011] This solution provides a load peak regulation system that takes into account user demand response, including:

[0012] A data collection module, which collects user-side electricity consumption data, environmental data, and device status information through smart meters and IoT devices;

[0013] A computational modeling module, which constructs a user preference model based on the data collected by the data collection module, and establishes a load forecasting model based on the user preference model to predict the peak load and regulation demand in a specific period of time;

[0014] Dynamic pricing and incentive module, the dynamic pricing and incentive module includes a price signal generation unit and an incentive mechanism unit. The price signal generation unit sends price signals to users through a dynamic pricing mechanism based on real-time load conditions, wherein the price is higher during peak hours and lower during off-peak hours. The incentive mechanism unit provides economic incentives (electricity discounts) or non-economic incentives (points redemption rewards) to users participating in peak load regulation;

[0015] The optimization decision-making and scheduling module uses multi-objective optimization algorithms (such as genetic algorithms and particle swarm algorithms) to optimize the operating sequence of equipment while meeting user comfort, and realizes automatic start and stop of equipment (such as adjusting the operating time of air conditioners and water heaters, and the production period of industrial equipment) through smart home devices or industrial controllers, thereby achieving the effect of power optimization and distributed control.

[0016] As a preferred solution, the smart meter is used to record the user's power consumption, voltage, current, power factor and other data in real time; the IoT device includes an IoT sensor for collecting environmental data such as temperature and humidity, light intensity, and device status, as well as an intelligent control terminal for connecting to the user's electrical equipment to achieve two-way communication.

[0017] As a preferred solution, the user preference model includes comfort requirements (such as temperature, lighting) and flexibility range (such as air conditioning start-up time, industrial equipment working hours).

[0018] The beneficial effects achieved by the present invention using the above technical solution are as follows:

[0019] 1. Through adjustments on the user side, the peak output demand on the power generation side is reduced, the start-up and shutdown costs of standby units are reduced, and at the same time, the load adjustment capabilities of users are utilized to improve the peak-shaving flexibility of the power grid.

[0020] 2. Use multi-objective optimization algorithms to optimize the operating sequence of equipment while meeting user comfort, thereby achieving power optimization and distributed control effects.

[0021] 3. Through flexible incentive mechanisms, users can be transformed from passive electricity users to active participants.

[0022] 4. Use blockchain technology to record regulatory behaviors and ensure the fairness of the platform. At the same time, use smart devices and optimization algorithms to achieve seamless scheduling and reduce the burden of user participation. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of the composition of a load peak regulation method considering user demand response provided in this scheme. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0025] Example 1

[0026] This solution provides a load peak regulation method considering user demand response, including the following steps:

[0027] S1 obtains power grid operation data;

[0028] S2 builds a load forecasting model to identify peak loads;

[0029] S3 sets the objective function based on the user's electricity usage scenario;

[0030] S4 solves the model to obtain optimized peak-shaving parameters and provides economic incentives or non-economic incentives for users participating in peak-shaving.

[0031] Please refer to Figure 1 As shown, this embodiment provides a load peak regulation system considering user demand response, which is used to implement the above-mentioned peak regulation method, and specifically includes:

[0032] Data acquisition module: The data acquisition module collects user-side electricity consumption data, environmental data, and device status information through smart meters and IoT devices.

[0033] The smart meter is used to record the user's power consumption, voltage, current, power factor and other data in real time; the IoT device includes an IoT sensor for collecting environmental data such as temperature and humidity, light intensity, and device status, as well as an intelligent control terminal for connecting to the user's electrical equipment to achieve two-way communication.

[0034] Computational modeling module: constructs a user preference model based on the data collected by the data collection module, and establishes a load forecasting model based on the user preference model to predict the peak load and regulation demand in a specific period of time.

[0035] First, the user's electricity consumption behavior is modeled: after collecting user-related data, the user's electricity consumption pattern is identified through machine learning algorithms including but not limited to cluster analysis and time series prediction, thereby building a load flexibility assessment model.

[0036] Secondly, build a load forecasting model: based on historical load data and weather forecasts, use deep learning methods (such as LSTM or Transformer) to predict the load curve in future periods and identify peak loads in advance.

[0037] The user preference model includes the construction of user portraits, starting from the two aspects of comfort parameters and flexibility analysis. Among them, the comfort parameters include: the temperature range of air conditioner settings and the start time of water heaters for residential users, and the operating time of user production equipment for industrial users. In terms of flexibility analysis, the adjustability of equipment is evaluated based on historical data, such as the charging time window of electric vehicles, and the start and stop constraints of equipment are defined.

[0038] Dynamic pricing and incentive module: The dynamic pricing and incentive module includes a price signal generation unit and an incentive mechanism unit. The price signal generation unit sends price signals to users through a dynamic pricing mechanism based on real-time load conditions, wherein prices are higher during peak hours and lower during off-peak hours. The incentive mechanism unit provides economic incentives (electricity discounts) or non-economic incentives (points redemption rewards) to users participating in peak load regulation.

[0039] In the composition of this module, the design is based on dynamic pricing strategy, incentive mechanism implementation, and user participation method. In the dynamic pricing strategy, time-of-use electricity prices are implemented, and different electricity prices are set for peak, normal, and valley hours. In addition, real-time electricity prices are introduced to dynamically adjust price signals according to the grid load forecast. In addition to the economic incentives and non-economic incentives introduced above, the incentive mechanism implementation includes direct economic rewards, which can also provide electricity fee reductions or cash subsidies for users participating in load regulation. At the same time, users can obtain points based on their participation, and then use the points to redeem gifts or value-added services. Non-economic incentives provide energy-saving and emission reduction labels and environmental protection contribution rankings to enhance the social honor of user participation.

[0040] There are two ways for users to participate: active participation: through the mobile application, users can choose to adjust the time period and equipment; passive participation: authorizing the intelligent control terminal to automatically optimize the operation of the equipment according to the electricity price signal and adjustment needs.

[0041] Optimization decision and scheduling module: Use multi-objective optimization algorithms (such as genetic algorithms, particle swarm algorithms) to optimize the operating sequence of equipment while meeting user comfort, and use smart home devices or industrial controllers to realize automatic start and stop of equipment (such as adjusting the operating time of air conditioners and water heaters, and the production period of industrial equipment), thereby achieving the effect of power optimization and distributed control.

[0042] The design of the optimization decision and scheduling module includes three aspects:

[0043] 1. Optimize model design:

[0044] The core of the optimization algorithm is to reflect the peak-shaving demand of the system through the objective function, while taking into account the user experience and cost effect. The elements of the optimization algorithm include the objective function and constraints, as follows:

[0045] The objective functions include: minimizing the peak load of the power grid; minimizing user discomfort (such as temperature deviation from the preset value); and maximizing user incentive benefits.

[0046] 1.1 Minimize the peak load on the power grid

[0047] Minimize:P peak =max(P t )

[0048] P t is the total load in each time period.

[0049] 1.2 Minimize User Inappropriateness

[0050]

[0051] △C i is the comfort deviation of user i (e.g., temperature deviation from the preset value); i is the weight of each user's comfort.

[0052] 1.3 Maximize user incentive benefits

[0053]

[0054] G i is the revenue that the platform obtains through user participation in demand response; C i is the cost of user response adjustment.

[0055] 1.4 Comprehensive objectives

[0056] After normalizing the above objectives, the weighted combined trip comprehensive objective function is:

[0057] Minimize:w1P peak -w2R+w3Σλ i ·ΔC i

[0058] w1, w2, and w3 are weight factors that can be flexibly adjusted according to system requirements.

[0059] Constraints include: start-up time and operating time of electrical equipment; user comfort and safety requirements; and grid operation safety constraints (such as voltage stability and frequency fluctuation limits).

[0060] 2. Optimization algorithm design: Three algorithms are provided for optimization, including:

[0061] 2.1 Genetic algorithm: suitable for multi-objective optimization, optimizing equipment start-up and shutdown plans through selection, crossover and mutation operations;

[0062] The optimization process of the genetic algorithm is as follows:

[0063] 2.1.1 Initialization: Randomly generate the initial population of user equipment start and stop plans.

[0064] 2.1.2. Fitness function: Evaluate the performance of each plan (based on the objective function).

[0065] 2.1.3 Selection, crossover, and mutation:

[0066] Choose a plan with high adaptability;

[0067] Generate new plans by crossing (e.g., crossing time periods);

[0068] Make random changes to some plans (such as adjusting the start and stop times of equipment).

[0069] 2.1.4 Iteration: Repeatedly generate new populations until convergence or the maximum number of iterations is reached.

[0070] Advantages: strong global optimization capability, suitable for complex problems. Disadvantages: slow convergence speed, easy to fall into local optimum.

[0071] 2.2 Particle Swarm Optimization Algorithm: It simulates the rapid convergence of group intelligent behavior and is suitable for large-scale user optimization; it is also suitable for scenarios with a single objective function and a small number of device types.

[0072] Among them, the optimization process of the particle swarm optimization algorithm is as follows:

[0073] 2.2.1 Initialize particle swarm: Each particle represents a load regulation scheme.

[0074] 2.2.2 Update speed and position: Update according to the current particle position, speed and historical optimal solution:

[0075] V i+1 =W·V i +c1·r1·(P best -X i )+c2·r2·(g best -X i )

[0076] X i+1 =X i +V i+1

[0077] W is the inertia weight, c1 and c2 are learning factors; P best , g best are the local and global optimal positions of the particle.

[0078] 2.2.3 Iteration: Update the particle swarm state until the target or maximum number of iterations is reached.

[0079] Advantages: high computational efficiency, suitable for real-time optimization scenarios. Disadvantages: high requirements on the smoothness of the objective function, easy to fall into local optimum.

[0080] 2.3 Reinforcement learning algorithm: By training intelligent agents, the optimal dispatching strategy is learned in a dynamic environment. It is suitable for scenarios where the power grid status is complex and changes in real time.

[0081] Among them, the optimization process of the reinforcement learning algorithm is as follows:

[0082] 2.3.1. Status: grid load, user equipment status, real-time electricity price, etc.;

[0083] 2.3.2. Action: Adjust the start and stop status of the equipment;

[0084] 2.3.3. Reward function: designed by combining load regulation effect and user satisfaction.

[0085] Reinforcement learning algorithms include: Deep Q learning (DQN): suitable for discrete action space; Policy gradient method (such as PPO): suitable for continuous action space.

[0086] Advantages: Ability to adjust adaptively in dynamic environments, stable results. Disadvantages: Long training time, requiring a lot of computing resources.

[0087] The implementation case of the optimization algorithm is as follows:

[0088] Case 1: Residential user equipment regulation

[0089] 1. Scenario: Optimization of operating timing of air conditioners, water heaters, and electric vehicles;

[0090] 2. Methods:

[0091] Using a genetic algorithm, the start and stop times of the air conditioner are adjusted alternately with the charging time of the electric vehicle, with the goal of reducing the load during peak hours;

[0092] The fitness function is dynamically evaluated by combining user comfort and electricity price signals.

[0093] Case 2: Load Optimization for Industrial Users

[0094] 1. Scenario: Scheduling of multiple production equipment;

[0095] 2. Methods:

[0096] Use particle swarm optimization algorithm to quickly find the best equipment operation plan during off-peak hours;

[0097] Reduce peak load while meeting production cycle constraints.

[0098] Case 3: Load peak regulation under dynamic grid conditions

[0099] 1. Scenario: complex power grid with wind power and photovoltaic access;

[0100] 2. Methods:

[0101] Adopt reinforcement learning algorithm to adjust load response in real time;

[0102] Through dynamic electricity price signals and reward mechanisms, user-side equipment is encouraged to actively cooperate.

[0103] 3. Distributed scheduling design:

[0104] Using blockchain technology, we build a distributed energy trading and dispatching platform to ensure that the regulation process is transparent and credible. We also deploy edge computing nodes on the user side to reduce the reliance of real-time control on centralized servers and improve response speed.

[0105] As an example to further illustrate:

[0106] The system also includes a real-time feedback and adjustment module, which includes a feedback mechanism unit and a dynamic adjustment unit.

[0107] The feedback mechanism unit consists of two functions: data feedback and load monitoring. During the data feedback process, the operating status of the user's equipment is transmitted back in real time through the intelligent terminal. The load monitoring work uses the power grid monitoring system to analyze the actual load changes and compare them with the expected regulation effect, so as to achieve the effect of the evaluation plan.

[0108] The dynamic adjustment unit makes dynamic adjustments based on the data from the feedback mechanism unit. If the actual load does not reach the target, the electricity price signal or incentive policy will be automatically corrected. If it is based on user preference feedback (such as inappropriate complaints), the regulation strategy will be adjusted or the participation of some equipment will be cancelled.

[0109] Example 2

[0110] Embodiments of residential user scenarios, commercial user scenarios, and industrial user scenarios are provided.

[0111] Residential electricity usage scenarios:

[0112] 1) Electrical equipment:

[0113] Air conditioners, water heaters, electric cars, smart washing machines;

[0114] 2) Implementation process:

[0115] 2.1 During peak hours, electricity prices increase, and users authorize the platform to delay the operation of the water heater to off-peak hours;

[0116] 2.2 Users set the charging window for electric vehicles (e.g. midnight to early morning) through the mobile app, and the platform automatically optimizes the charging time to avoid peak hours;

[0117] 2.3 The platform provides points rewards, which users can redeem for electricity discounts or environmental protection certificates.

[0118] Commercial user scenarios:

[0119] 1) Electrical equipment:

[0120] Central air conditioning, lighting system, elevator;

[0121] 2) Implementation process:

[0122] 2.1 During the peak hours at noon, slightly increase the air conditioning temperature by 1-2°C and optimize the lighting system to reduce unnecessary energy consumption;

[0123] 2.2 Start the air conditioner to cool down in advance at night to avoid the peak load during the day;

[0124] 2.3 The platform provides real-time electricity price signals, and users adjust equipment operation in real time through the building management system.

[0125] Industrial user scenarios:

[0126] 1) Electrical equipment:

[0127] Large-scale production equipment, energy storage systems;

[0128] 2) Implementation process:

[0129] 2.1 Non-critical production tasks (such as auxiliary production lines) are adjusted to low-peak periods;

[0130] 2.2 The energy storage system charges during off-peak hours and discharges during peak hours, reducing the burden on the power grid;

[0131] 2.3 The platform records user adjustment behaviors through smart contracts to ensure the fairness of reward settlement.

[0132] Example 3

[0133] In addition, in this embodiment, the load peak regulation method between the residential side, the commercial side, and the industrial side is as follows:

[0134] Residents: During peak hours, the charging task of electric vehicles is allocated to off-peak hours;

[0135] Commercial side: adjust the operating hours of central air conditioners in office buildings to avoid peak loads;

[0136] Industrial side: Optimize the operation plan of production equipment and transfer some non-critical tasks to low-operation periods.

[0137] It should be noted that although the embodiments of the present invention have been shown and described, for ordinary technicians in this field, without departing from the purpose of the invention, structural methods and embodiments similar to the technical solution are designed without creativity and should all fall within the scope of protection of the present invention.

Claims

1. A load peak regulation method considering user demand response, characterized in that: The following steps are involved: S1 obtains power grid operation data; S2 builds a load forecasting model to identify peak loads; S3 sets the objective function based on the user's electricity usage scenario; S4 solves the model to obtain optimized peak-shaving parameters and provides economic incentives or non-economic incentives for users participating in peak-shaving.

2. A load peak shaving method considering user demand response according to claim 1, characterized in that: Multi-objective optimization algorithms are used to optimize the operating sequence of equipment while meeting user comfort, achieve power optimization and distributed control, and reflect the system's peak-shaving needs through objective functions, while taking into account user experience and cost effects.

3. A load peak regulation system according to the load peak regulation method considering user demand response according to any one of claims 1-2, characterized in that: The system comprises: Data collection module: collects user-side electricity consumption data, environmental data, and equipment status information through smart meters and IoT devices; Computational modeling module: constructing a user preference model based on the data collected by the data collection module, and establishing a load forecasting model based on the user preference model to predict the peak load and regulation demand in a specific period of time; Dynamic pricing and incentive module: includes a price signal generation unit and an incentive mechanism unit. The price signal generation unit sends price signals to users through a dynamic pricing mechanism based on real-time load conditions. The incentive mechanism unit provides economic incentives or non-economic incentives to users participating in peak load regulation. Optimization decision and scheduling module: Use multi-objective optimization algorithms to optimize the operating sequence of equipment while meeting user comfort, and realize automatic start and stop of equipment through smart home devices or industrial controllers, thereby achieving power optimization and distributed control effects.

4. A load peak shaving system considering user demand response according to claim 3, characterized in that: It also includes a real-time feedback and adjustment module, which includes a feedback mechanism unit and a dynamic adjustment unit; The feedback mechanism unit consists of data feedback and load monitoring. During the data feedback process, the operating status of the user equipment is transmitted in real time through the intelligent terminal. The load monitoring work uses the power grid monitoring system to analyze the actual load changes and compare them with the expected regulation effect, so as to achieve the effect of the evaluation plan; The dynamic adjustment unit is used to make dynamic adjustments based on the data from the feedback mechanism unit. If the actual load does not reach the target, the electricity price signal or incentive policy will be automatically corrected. If it is based on user preference feedback, the regulation strategy will be adjusted or the participation of some equipment will be cancelled.

5. A load peak regulation system considering user demand response according to claim 4, characterized in that: The multi-objective optimization algorithms in the optimization decision and scheduling module include genetic algorithm, particle swarm optimization algorithm and reinforcement learning algorithm.

6. A load peak shaving system considering user demand response according to claim 5, characterized in that: The optimization algorithm elements include objective function and constraints; The objective function includes: minimizing the peak load of the power grid; minimizing user discomfort; maximizing user incentive benefits; the constraints include: the start-up time and operating time of electrical equipment; user comfort and safety requirements; and power grid operation safety constraints.

7. A load peak shaving system considering user demand response according to claim 6, characterized in that: The specific modeling steps of the computational modeling module are as follows: First, the user's electricity consumption behavior is modeled: after collecting relevant user data, the user's electricity consumption pattern is identified through machine learning algorithms including but not limited to cluster analysis and time series prediction, so as to build a load flexibility assessment model; Secondly, a load forecasting model is constructed: based on historical load data and weather forecasts, a deep learning method is used to predict the load curve for future periods and identify peak loads in advance; the adjustability of the equipment is evaluated based on historical data, and the start and stop constraints of the equipment are defined.

Images