CN115018376A - Load regulation and control optimization method considering novel power system characteristics - Google Patents

Abstract

The invention relates to load regulation and control optimization considering novel power system characteristics. According to the method, based on the load regulation strategy of new energy output and user-side conventional load prediction, the charging plan of the electric vehicle is combined with the new energy power data peak value and the user-side conventional load valley value, so that the new energy consumption is improved, the system load peak-valley difference value is reduced, and the stability of a novel power system is promoted. The invention can reduce the space-time characteristic of transferring the output of new energy by deploying energy storage equipment in a large scale, reduce the cost of system construction and maintenance, improve the utilization rate of the new energy, reduce the power demand on a superior power grid and promote the realization of a double-carbon target.

Description

A load regulation and optimization method considering the characteristics of a new type of power system

technical field

The invention relates to a load regulation and optimization method characterized by a novel power system, and belongs to the technical field of power system regulation and optimization.

Background technique

New energy sources such as photovoltaics and wind power have the characteristics of volatility and intermittence. The openness and uncertainty of equipment access such as clustered/distributed energy storage and large-scale electric vehicles all affect the planning, operation and regulation of new power systems. and analysis present new challenges.

In order to reduce the impact of the volatility of new energy power generation, the system is often equipped with a considerable scale of energy storage equipment to transfer the time and space uncertainty of new energy, which increases the cost of system construction and maintenance, and reduces the utilization rate of energy. In recent years, with the development of intelligent forecasting methods, the mastery of new energy power generation laws has become more mature, which can provide reasonable support for guiding user-side electricity consumption behavior. Therefore, the power generation capacity of new energy and traditional energy, as well as the user-side load and the energy consumption habits of electric vehicles, can be considered as a whole, and the user-side load regulation and optimization strategy can be reasonably designed to improve the "consumption" level of the new power system.

The Chinese invention patent application with the publication number CN112134272A proposes a load regulation method for electric vehicles in a distribution network. By monitoring the charging behavior of electric vehicles in the system in real time, and frequently requiring users to submit charging prices and charging requirements to confirm the charging demand, it lacks consideration of the system. The ability to provide internal power and the impact of large-scale charging behavior on power peaks and valleys. The Chinese invention patent application with publication number CN110807598A proposes a user load regulation value evaluation method for participating in orderly electricity consumption, which takes the peak and valley value of the load on the user side into consideration and involves in the value evaluation system, but lacks new energy and other Considering the consideration of non-traditional energy supply methods and electric vehicles and other equipment that can have the characteristics of energy consumption in time and space, the objective function of which is based on the profit of the power supply company and the cost of electricity consumption of users, cannot provide support for improving the "consumption" capacity of the new power system.

SUMMARY OF THE INVENTION

The purpose of the present invention is to consider the power generation capacity and energy consumption habits in the system as a whole, and to design the user-side load regulation and optimization in view of the "double high" characteristics of high-penetration renewable energy and high-proportion power electronic equipment connection in the new power system. strategies to promote the "consumption" level of the new power system.

In order to achieve the above object, the technical solution of the present invention provides a load regulation optimization considering the characteristics of a novel power system, which is characterized in that it includes the following steps:

Step 1: Obtain the historical meteorological sample information, use the BP neural network to predict the new energy power generation a few days ago, and obtain the new energy power generation data _Pre-power (t);

Step 2: Obtain historical daily type sample information, use time series neural network to predict the conventional power load in the station area, and obtain the conventional power load prediction data P _consumption (t) on the user side before the day before;

Step 3: Obtain the historical data and real-time data of electric vehicles in the Taiwan area, and estimate the daily charging demand of electric vehicles in the Taiwan area through the calculation model of the electric vehicle charging demand in the Taiwan area, including the following steps:

Step 301: N electric vehicles in the grid charging state at 0:00 of the day, and K electric vehicles in the off-grid state; the capacity of the _i -th electric vehicle in the grid charging state is Qi, The charging power of the i-th electric vehicle in the on-grid charging state at time 0 of the day is P _i,t=0 , the SOC state is SOC _i,t=0 , i=1,2,...,N; it is off-grid The capacity of the jth electric vehicle in the state is Q _j , j=1,2,...,K;

Through formula (1), it is estimated that the time required for each electric vehicle to be fully charged at 0:00 on the day is _Test,i,t=0 :

Step 302: When the SOC state SOC _j of the j-th electric vehicle in the off-grid state at 0:00 of the day is lower than 20%, it will be charged, and the charging power of its recharging is P _j , then the j-th electric vehicle in the off-grid state will be charged. The charging demand time _Test,j of the electric vehicle on the day is calculated according to formula (2):

In formula (2), the recharging charging power P _j of the j-th electric vehicle that is off-grid at 0:00 on the day is estimated from the sample historical charging behavior data using formula (3):

分别为样本历史充电行为数据中第j台电动汽车第m次充电的平均值及中位数，M为样本历史充电行为数据的总充电次数；In formula (3),

are the average and median of the mth charging of the jth electric vehicle in the sample historical charging behavior data, respectively, and M is the total charging times of the sample historical charging behavior data;

Step 4: Carry out the optimization of load regulation on the user side, and calculate the plan including the charging power and charging time of the electric vehicle in the station area, which includes the following steps:

The new energy power generation data P _re-power (t) obtained in step 1 and the conventional power load forecast data P _consumption (t) on the user side obtained in step 2 are compared by time dimension, and the overall charging of each time interval is calculated. The power margin, where the overall charging power margin of the t-th time interval is expressed as P _total,t , there is the following formula (4):

In formula (4), P _{t=peak,peak-1,...,1} represents the new energy power generation data predicted in step 1 from the corresponding power generation at the peak time to the corresponding power generation at the sub-peak time, until the corresponding power generation at the trough time;

After sorting the recharge charging power P _j of the K electric vehicles that are off-grid at 0:00 of the day from large to small, the principle of arranging the high charging power demand at the time of high charging margin, the following formula (5) is for recharging The charging behavior of electric vehicles is arranged:

In formula (5), t _j is the planned charging time of the jth electric vehicle, t _{peak, peak-1,...,1} is the corresponding time of the charging margin from high to low, p _j is the planned charging time of the jth electric vehicle power.

The following formula (6) constrains the charging behavior of rechargeable electric vehicles, and the charging power sum ∑P _{j,t=peak,peak-1,...,1} within the same time interval does not exceed the charging power margin P _total,t , charging The time is arranged at more than 20% of the new energy power generation _Pre-power (t) capacity to ensure that the new energy power generation is in a continuous and stable state:

In formula (6), P _{j,t=peak,peak-1,...,1} represents the electric vehicle charging power corresponding to the predicted new energy power generation data peak time in step 1 decreasing one by one, T _j represents the electric vehicle recharging time, P _{re-power-peak} represents the peak value of the new energy power generation data predicted in step 1;

Step 5: Calculate the charging power P _j and charging time t _j of the electric vehicle according to Step 4 and issue a charging plan.

Preferably, the step 1 specifically includes the following steps:

Step 101: Select the light intensity information data and the temperature information data as the feature input of the BP neural network, take the photovoltaic output as the feature output of the BP neural network, and inject the historical data of a certain time interval into the BP neural network to train it, Thereby, a photovoltaic output prediction model based on BP neural network is established;

Step 102: Select the wind speed information data as the feature input of the BP neural network, take the fan output as the feature output of the BP neural network, and inject the historical data of a certain time interval into the BP neural network for training, so as to establish a BP neural network based on BP neural network. Fan output prediction model of the network;

Step 103: Use the weather and meteorological information of the previous day as the input of the photovoltaic output prediction model and the fan output prediction model, and obtain the photovoltaic output prediction result P _solar (t) and the fan output prediction result P _wind ( t), superimpose the photovoltaic output prediction result P _solar (t) and the wind turbine output prediction result P _wind (t) according to the time dimension to obtain the new energy power generation data _Pre-power (t).

Preferably, the step 2 specifically includes the following steps:

Step 201: Select the highest temperature, the lowest temperature, the average temperature, the average relative humidity, and the rainfall as the feature input of the time-series neural network, and take the conventional power load in the station area as the feature output of the time-series neural network, and send it to the time-series neural network. Fill in the historical data of a certain time interval for training, so as to establish a conventional power load forecasting model based on a time series neural network;

Step 202 : Using the day type information as the input of the conventional power load prediction model, obtain the conventional power load prediction data P _consumption (t) on the user side before the day output by the conventional power load prediction model.

Compared with the prior art, the present invention has the following beneficial effects:

(1) The load regulation strategy based on new energy output and conventional load forecasting on the user side combines the electric vehicle charging plan with the peak value of the new energy power data and the valley value of the conventional load on the user side. It is also beneficial to reduce the difference between the peak and valley of the system load and promote the stability of the new power system;

(2) Reduce the spatiotemporal characteristics of large-scale deployment of energy storage equipment to transfer new energy output, while reducing system construction and maintenance costs, it can also improve the utilization rate of new energy, reduce the power demand for the upper-level power grid, and promote "double carbon" goals.

Description of drawings

Fig. 1 is the load regulation strategy frame diagram that considers the characteristic of the novel power system designed by the present invention;

FIG. 2 is a comparison diagram of the predicted value and the actual value of the photovoltaic output prediction based on the BP neural network of the present invention;

FIG. 3 is a comparison diagram of the predicted value and the actual value of the fan output prediction based on the BP neural network according to the present invention;

Fig. 4 is the comparison diagram of the predicted value and the actual value of load forecasting based on the time series neural network of the present invention;

FIG. 5 is a comparison diagram of new energy power generation and user-side load prediction before the present invention;

FIG. 6 is a response diagram of an electric vehicle charging demand in the present invention.

Detailed ways

The present invention will be further described below in conjunction with specific embodiments. It should be understood that these examples are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of the present application.

The structure diagram of the load regulation strategy designed by the present invention considering the characteristics of the new power system is shown in Figure 1. The corresponding input is obtained through the new energy power generation prediction module, the user-side load prediction module and the electric vehicle charging demand estimation module in the station area. The load regulation and optimization module designed for the power peak-valley difference in the time dimension delivers the calculated electric vehicle charging power and charging time to complete the "consumption" of new energy power generation in the system and reduce the power peak-to-valley difference in the Taiwan area .

In order to facilitate the understanding of the control scheme of the present invention, the following describes the strategy scheme of the present invention with reference to FIG. 2 , FIG. 3 , FIG. 4 , FIG. 5 and FIG. 6 .

First, according to the historical meteorological information, the BP neural network is built for the photovoltaic output and the fan output respectively for training, and the light intensity, temperature and wind speed are selected as the corresponding model feature input, and the network parameters are as follows.

name type number of hidden layers number of nodes training times learning rate target minimum error Photovoltaic BP neural network 1 8 1000 0.01 0.000001 fan BP neural network 1 4 3000 0.05 0.000001

The training results are shown in Figure 2 and Figure 3, respectively. The average absolute errors are 0.00139 and 0.025, respectively, and the prediction results are relatively accurate. Some historical data are as follows:

date time temperature light intensity Photovoltaic output wind speed Fan output 2020/3/1 11:30 25 531 677.03 8.72 693.26 2020/3/1 12:00 15 543 738.48 8.97 718.89 2020/3/1 12:30 26 537 680.11 9.49 761.25

According to the historical day type information, build a time series neural network for the conventional power load on the user side for training, and select the highest temperature, the lowest temperature, the average temperature, the relative humidity (average), and the rainfall as the characteristic input. The network parameters are as follows:

name type number of hidden layers number of nodes training times learning rate target minimum error load time series neural network 2 [6,6] 3000 0.001 0.000001

The training results are shown in Figure 4. The average absolute error is 0.00271, and the prediction results are relatively accurate. Part of the historical load data is as follows:

YMD T0700 T0730 T0800 T0830 T0900 T0930 20141201 4340.233 4833.116 6159.911 7581.279 8009.952 8253.632 20141202 5296.319 5713.094 6935.537 8158.838 8491.981 8688.653 20141203 5447.405 5866.201 6971.504 8273.659 8567.503 8745.38

The data of some historical day types are as follows:

Maximum temperature °C Minimum temperature °C Average temperature °C Relative humidity (average) Rainfall (mm) 20141201 25.6 16.5 19.9 70 0.3 20141202 16.5 12.4 13.9 82 3.6 20141203 17.9 12.8 15.1 92 6

From the meteorological station and other platforms to obtain the weather information and daily type information feature quantity, and give it to the corresponding model for prediction, the new energy power generation data and the conventional load data on the user side can be obtained, as shown in Figure 5.

Then, set a total of 100 electric vehicles in the Taiwan area, and the capacity and number are as follows:

Capacity (kWh) Number of units 16 5 twenty four 8 30 17 32 7 57 20 60 11 100 10 111.5 7 150 15

It is set that 56 units are charged on the network and 44 units are off-grid at 0:00 of the day. Some parameters are as follows:

car number Capacity (kWh) 0:00 Status SOC when accessing 2 16 1 0.61 7 111.5 1 0.44 15 twenty four 1 0.82 16 twenty four 1 0.69 twenty three 100 1 0.37

According to step 3, set the SOC of the electric vehicle to be between 10% and 20% for recharging, and perform calculation to obtain the _Test,i,t=0 , _Test,j and _Pj of each electric vehicle in the station area. Some of the results are shown in the table below. :

car number Capacity (kWh) 0:00 Status SOC when accessing Power when connected (kW) Full time (mins) Reconnection power (kW) SOC at re-access Time-consuming (mins) to be fully charged after re-connection 5 16 1 0.20 49 15.73 0 0.00 0.00 11 111.5 0 0.00 0 0 37 0.10 162.00 17 twenty four 1 0.54 46 14.52 0 0.00 0.00 twenty three 100 1 0.37 12 315.83 0 0.00 0.00 35 150 0 0.00 0 0 30 0.12 264.53 50 32 1 0.70 twenty four 23.74 0 0.00 0.00 56 60 0 0.00 0 0 38 0.15 80.48 66 30 1 0.34 40 29.66 0 0.00 0.00 82 57 1 0.55 46 33.41 0 0.00 0.00

Then, according to the constraints designed by the present invention in step 4, the optimal regulation and control arrangement of the charging behavior is carried out with the new energy power generation peak as the center point. Part of the optimization results are as follows.

Claims (3)

1. A load regulation optimization considering the characteristics of a novel power system, characterized in that it comprises the following steps: Step 1: Obtain the historical meteorological sample information, use the BP neural network to predict the new energy power generation a few days ago, and obtain the new energy power generation data _Pre-power (t); Step 2: Obtain historical daily type sample information, use time series neural network to predict the conventional power load in the station area, and obtain the conventional power load prediction data P _consumption (t) on the user side before the day before; Step 3: Obtain the historical data and real-time data of electric vehicles in the Taiwan area, and estimate the daily charging demand of electric vehicles in the Taiwan area through the calculation model of the electric vehicle charging demand in the Taiwan area, including the following steps: Step 301: N electric vehicles in the grid charging state at 0:00 of the day, and K electric vehicles in the off-grid state; the capacity of the _i -th electric vehicle in the grid charging state is Qi, The charging power of the i-th electric vehicle in the on-grid charging state at time 0 of the day is P _i,t=0 , the SOC state is SOC _i,t=0 , i=1,2,...,N; it is off-grid The capacity of the jth electric vehicle in the state is Q _j , j=1,2,...,K; Through formula (1), it is estimated that the time required for each electric vehicle to be fully charged at 0:00 on the day is _Test,i,t=0 :

Step 302: When the SOC state SOC _j of the j-th electric vehicle in the off-grid state at 0:00 of the day is lower than 20%, it will be charged, and the charging power of its recharging is P _j , then the j-th electric vehicle in the off-grid state will be charged. The charging demand time _Test,j of the electric vehicle on the day is calculated according to formula (2):

In formula (2), the recharging charging power P _j of the j-th electric vehicle that is off-grid at 0:00 on the day is estimated from the sample historical charging behavior data using formula (3):

In formula (3),

are the average and median of the mth charging of the jth electric vehicle in the sample historical charging behavior data, respectively, and M is the total charging times of the sample historical charging behavior data; Step 4: Carry out the optimization of load regulation on the user side, and calculate the plan including the charging power and charging time of the electric vehicle in the station area, which includes the following steps: The new energy power generation data P _re-power (t) obtained in step 1 and the conventional power load forecast data P _consumption (t) on the user side obtained in step 2 are compared by time dimension, and the overall charging of each time interval is calculated. The power margin, where the overall charging power margin of the t-th time interval is expressed as P _total,t , there is the following formula (4):

In formula (4), P _{t=peak,peak-1,...,1} represents the new energy power generation data predicted in step 1 from the corresponding power generation at the peak time to the corresponding power generation at the sub-peak time, until the corresponding power generation at the trough time; After sorting the recharge charging power P _j of the K electric vehicles that are off-grid at 0:00 of the day from large to small, the principle of arranging the high charging power demand at the time of high charging margin, the following formula (5) is for recharging The charging behavior of electric vehicles is arranged:

In formula (5), t _j is the planned charging time of the jth electric vehicle, t _{peak, peak-1,...,1} is the corresponding time of the charging margin from high to low, p _j is the planned charging time of the jth electric vehicle power. The following formula (6) constrains the charging behavior of rechargeable electric vehicles, and the charging power sum ∑P _{j,t=peak,peak-1,...,1} within the same time interval does not exceed the charging power margin P _total,t , charging The time is arranged at more than 20% of the new energy power generation _Pre-power (t) capacity to ensure that the new energy power generation is in a continuous and stable state:

In formula (6), P _{j,t=peak,peak-1,...,1} represents the electric vehicle charging power corresponding to the predicted new energy power generation data peak time in step 1 decreasing one by one, T _j represents the electric vehicle recharging time, P _{re-power-peak} represents the peak value of the new energy power generation data predicted in step 1; Step 5: Calculate the charging power P _j and charging time t _j of the electric vehicle according to Step 4 and issue a charging plan. 2. The load regulation optimization considering the characteristics of the novel power system according to claim 1, wherein the step 1 specifically comprises the following steps: Step 101: Select the light intensity information data and the temperature information data as the feature input of the BP neural network, take the photovoltaic output as the feature output of the BP neural network, and inject the historical data of a certain time interval into the BP neural network to train it, Thereby, a photovoltaic output prediction model based on BP neural network is established; Step 102: Select the wind speed information data as the feature input of the BP neural network, take the fan output as the feature output of the BP neural network, and inject the historical data of a certain time interval into the BP neural network for training, thereby establishing a BP neural network based on the BP neural network. Fan output prediction model of the network; Step 103: Use the weather and meteorological information of the previous day as the input of the photovoltaic output prediction model and the fan output prediction model, and obtain the photovoltaic output prediction result P _solar (t) and the fan output prediction result P _wind ( t), superimpose the photovoltaic output prediction result P _solar (t) and the wind turbine output prediction result P _wind (t) according to the time dimension to obtain the new energy power generation data _Pre-power (t). 3. The load regulation optimization considering the characteristics of the novel power system according to claim 1, wherein the step 2 specifically comprises the following steps: Step 201: Select the highest temperature, the lowest temperature, the average temperature, the average relative humidity, and the rainfall as the feature input of the time series neural network, take the conventional power load in the station area as the feature output of the time series neural network, and send the data to the time series neural network. Fill in the historical data of a certain time interval for training, so as to establish a conventional power load forecasting model based on a time series neural network; Step 202 : Using the day type information as the input of the conventional power load prediction model, obtain the conventional power load prediction data P _consumption (t) on the user side before the day output by the conventional power load prediction model.

Images