CN107122845A - A kind of modified neutral net overhead transmission line wind speed forecasting method based on division gentle breeze area - Google Patents

Abstract

The invention discloses an improved neural network wind speed prediction method for overhead lines based on the division of breeze zones. The steps are as follows: S1. Preprocessing the original data; Step S3, if not entered, go to step S4; S3, extract the wind speed data and corresponding time data in the breeze area, and filter, clean and standardize the extracted data; S4, substitute the data into the RBF neural network for prediction ; S5, using the wind speed prediction data as the wind speed prediction value of the overhead line. This method first defines the concept of the breeze area, because the wind speed fluctuation in the breeze area is small, which can make the neural network have a better fit, and then use the RBF neural network method to predict. When the external weather conditions are bad, the prediction error of the breeze is reduced, and finally the accuracy of the prediction results is improved, so that the safety of the prediction can be improved by the neural network.

Description

An Improved Neural Network Wind Speed Prediction for Overhead Lines Based on Dividing Breeze Zones method

technical field

The invention relates to the technical field of data discrimination and RBF (radial basis function) neural network prediction, in particular to an improved neural network wind speed prediction method for overhead lines based on division of breeze zones.

Background technique

Overhead transmission lines are an important part of power grid transmission and distribution. In recent years, under the background of my country's rapid economic development and the improvement of people's quality of life, the society's demand for electricity has increased rapidly, and the requirements for power supply reliability have also become higher. In the early stage of line planning and design, it is formulated according to the needs of users for a certain period of time. Therefore, the planning and development of power grids seriously lag behind the requirements of users for electric energy. In developed areas, power shortage often becomes a bottleneck restricting economic development. In 2005, the State Grid Corporation clearly proposed to actively tap the transmission capacity of the existing power grid. On the issue of improving the existing power transmission capacity, Europe and the United States and other countries put forward the concept of smart grid in the early years. There is the transmission capacity of the line and the operating efficiency of the line.

How to make full use of the capacity of existing power transmission equipment and improve the utilization efficiency of power grid equipment is an effective measure to solve the bottleneck of power transmission capacity of the existing power grid, and it is also one of the visions of smart grid construction. One method to realize capacity increase of existing transmission lines is dynamic capacity increase technology. However, the dynamic capacity increase technology needs to monitor the environmental conditions and the status of the wires. Although the existing sensing technology is developing rapidly, it is still difficult to monitor accurately, especially in the real environment where the wind speed and wind direction fluctuate greatly, especially at low wind speed conditions. Inaccurate, but the heat dissipation of the conductor is most obviously affected by convection heat dissipation, which is one of the dangerous conditions of local overheating of the conductor. Therefore, accurate forecasting of meteorological information in the future is of great significance for correctly evaluating the current-carrying capacity of overhead transmission lines and improving the current-carrying capacity to a certain extent.

Wind speed is one of the most important meteorological factors affecting the current-carrying capacity of overhead transmission lines. In summer with no wind or light wind, on the one hand, the power load will increase greatly, which will greatly increase the current of transmission lines; on the other hand, the weather conditions will be severe. This reduces the current-carrying capacity of the overhead transmission line, which poses a severe challenge to the power transmission and distribution of the power grid, so a method that can predict the wind speed more accurately is needed. However, due to the relatively large fluctuation of wind speed, the general neural network method will have large errors.

Contents of the invention

The purpose of the present invention is in order to solve the above-mentioned defect in the prior art, provide a kind of improved neural network wind speed forecasting method based on dividing the breeze zone, the method first proposes the breeze zone (wind speed＜0.5m/s, duration exceeds 45min) concept, through the method of data discrimination, to determine whether to enter the breeze area; if not into the breeze area, directly into the RBF (radial basis function) neural network method for prediction; if entering the micro-zone, use the data of the breeze area, Because the wind speed fluctuation in the breeze area is small, the neural network can have a better fit, and then use the RBF (radial basis function) neural network method to predict. When the external weather conditions are bad, this reduces the forecast error for the breeze, and finally improves the accuracy of the forecast results, so that the safety of the forecast can be improved.

The purpose of the present invention can be achieved by taking the following technical solutions:

An improved neural network overhead line wind speed prediction method based on division of breeze zones, said method comprising the following steps:

S1. Preprocessing the original data;

S2. According to the mathematical discrimination method, judge whether to enter the breeze zone, if enter the breeze zone, enter the next step S3, if not, go to step S4;

S3. Extract wind speed data and corresponding time data in the breeze area, and perform screening, cleaning and normalization processing on the extracted data;

S4. Substituting the data into the RBF neural network for prediction;

S5. The wind speed prediction data is obtained as the wind speed prediction value of the overhead line.

Further, the process of step S1 is as follows:

Filter, clean and normalize the original data in turn, where the normalization formula is as formula (1):

v' is the wind speed after normalization, v is the wind speed of the original data, v _min is the minimum value of the original data, and v _max is the maximum value of the original data.

Further, the mathematical discrimination method is specifically as follows:

The conditions for defining the breeze zone are: wind speed v<0.5m/s, duration t≥45min, and whether to enter the breeze zone can be judged by judging whether the conditions of the breeze zone are met.

Further, the extracted data in the breeze area or the data in the non-breeze area are brought into the RBF neural network, and the wind speed prediction data is obtained through the output layer f(x), where the expression of f(x) is as the formula:

Among them, x ₁ , x ₂ ,…,x _n are the input layer, the wind speed data is substituted into the input layer, w ₀ ,w ₁ ,…,w _n are the weights, is the radial basis function, and its expression is:

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

An improved neural network wind speed prediction method for overhead lines based on the division of the breeze area proposed by the present invention, by dividing the breeze area, including data pre-processing, in the case of bad external weather conditions, the prediction error of the breeze is reduced, and at the same time When forecasting wind speeds, the established breeze zone allows for a fast approximation of the forecast. In terms of specific implementation, the toolbox is invoked through MATLAB programming, which is simple and convenient to implement. The invention can effectively predict the wind speed of the overhead power transmission line under the condition of low wind speed, and can be used as one of the big data prediction methods of the dynamic carrying capacity of the overhead power transmission line.

Description of drawings

Fig. 1 is a flow chart of an improved neural network overhead line wind speed prediction method based on division of breeze zones disclosed by the present invention;

Fig. 2 is a structural diagram of the RBF neural network used in the present invention.

detailed description

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

Embodiment one

When studying the changes in the carrying capacity of overhead lines in the study area, the carrying capacity is greatly affected by meteorological factors, such as temperature, wind speed, humidity, etc., and the wind speed is one of the most important meteorological factors affecting the carrying capacity of overhead transmission lines. Accurate prediction Wind speed is of great significance for correctly evaluating the current-carrying capacity of overhead transmission lines and improving the current-carrying capacity in a certain range. In the windless or breezy summer, on the one hand, the power load increases greatly, which greatly increases the current of the transmission line; on the other hand, the weather conditions are relatively bad, which reduces the current-carrying capacity of the overhead transmission line. It brings severe challenges, so a method that can predict wind speed more accurately is needed. However, due to the relatively large fluctuation of wind speed, the general neural network method will have large errors. And the method that the present invention proposes, at first proposes the concept of breeze zone (wind speed＜0.5m/s, duration surpasses 45min), by the method for data discrimination, judge whether to enter breeze zone. If you do not enter the breeze area, directly bring it into the RBF (Radial Basis Function) neural network method for prediction; if you enter the micro-zone, use the data of the breeze area, because the wind speed fluctuation in the breeze area is small, the neural network can have a better performance Fitting, and then use the RBF (Radial Basis Function) neural network method to predict. Ultimately improve the accuracy of the prediction results.

As shown in Figure 1, an improved neural network wind speed prediction method for overhead lines based on the division of breeze zones disclosed in this embodiment specifically includes the following steps:

S1. Preprocessing the original data;

In the specific implementation manner, the preprocessing of data mainly involves the screening, cleaning and normalization of data, wherein the normalization formula is such as formula (1)

v' is the wind speed after normalization, v is the wind speed of the original data, v _min is the minimum value of the original data, and v _max is the maximum value of the original data.

S2. According to the mathematical discrimination method, judge whether to enter the breeze zone, if enter the breeze zone, enter the next step S3, if not, go to step S4;

In a specific embodiment, the mathematical discrimination method is specifically as follows:

The conditions for defining the breeze zone are: wind speed v<0.5m/s, duration t≥45min, and whether to enter the breeze zone can be judged by judging whether the conditions of the breeze zone are met.

This step mainly involves judging whether the wind speed has entered the breeze zone. Here, a region with a wind speed v<0.5m/s and a duration of t≥45min is defined as the breeze zone. If it is judged to be a breeze zone, go to step S3, and if it is judged not to be a micro zone, go to step S4.

S3. Extract wind speed data and corresponding time data in the breeze area, and perform screening, cleaning and normalization processing on the extracted data;

Extract the data of the breeze area, extract the wind speed data of this period and the corresponding time data, and filter, clean and standardize the data. Wherein, the normalization formula is as the above-mentioned formula (1).

S4. Substituting the data into the RBF neural network for prediction;

Bring the extracted data in the breeze area or the data in the non-breeze area into the RBF neural network (radial basis function) in MATLAB. The structure diagram of the neural network is shown in Figure 2.

In the figure, x ₁ , x ₂ ,…,x _n are the input layer, and the wind speed data is substituted into the input layer, w ₀ , w ₁ ,…,w _n are the weights,

In the hidden layer, is the radial basis function, whose expression is formula (2)

f(x) is the output layer, its expression is as formula (3)

S5. The wind speed prediction data is obtained as the wind speed prediction value of the overhead line.

In this step, the obtained wind speed prediction data is exported, and finally the wind speed prediction value of the overhead line is obtained.

To sum up, this method first defines the concept of the breeze area, and defines the wind speed <0.5m/s as the breeze area for a duration of more than 45 minutes. Because the wind speed fluctuation in the breeze area is small, the neural network can have a better fit, and then use the RBF neural network method to predict. When the external weather conditions are bad, the prediction error of the breeze is reduced, and finally the accuracy of the prediction results is improved, so that the safety of the prediction can be improved by the neural network.

The above-mentioned embodiment is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited by the above-mentioned embodiment, and any other changes, modifications, substitutions, combinations, Simplifications should be equivalent replacement methods, and all are included in the protection scope of the present invention.

Claims (4)

1. it is a kind of based on the modified neutral net overhead transmission line wind speed forecasting method for dividing gentle breeze area, it is characterised in that described Method comprises the following steps：

S1, former data are pre-processed；

S2, according to Math judgment method, discriminate whether enter gentle breeze area, if into gentle breeze area, into next step S3, if not entering Enter, go to step S4；

S3, the air speed data for extracting gentle breeze area and corresponding time data, and to extract data screened, clear up and Standardization processing；

S4, by data substitution RBF neural be predicted；

S5, wind speed value of the forecasting wind speed data as overhead transmission line will be drawn.

2. it is according to claim 1 a kind of based on the modified neutral net overhead transmission line forecasting wind speed side for dividing gentle breeze area Method, it is characterised in that the process of the step S1 is as follows：

Former data are screened, cleared up and standardized successively, wherein, normalizing such as formula (1)：

<mrow> <msup> <mi>v</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mfrac> <mrow> <mi>v</mi> <mo>-</mo> <msub> <mi>v</mi> <mrow> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> </msub> </mrow> <mrow> <msub> <mi>v</mi> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> </msub> <mo>-</mo> <msub> <mi>v</mi> <mi>min</mi> </msub> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>

V ' is the wind speed after standardization, and v is former data wind speed, v_minFor the minimum value of former data, v_maxFor the maximum of former data.

3. it is according to claim 1 a kind of based on the modified neutral net overhead transmission line forecasting wind speed side for dividing gentle breeze area Method, it is characterised in that the Math judgment method is specific as follows：

The condition for defining gentle breeze area is：Wind speed v<0.5m/s, duration t >=45min, by judging whether to meet gentle breeze area Whether condition distinguishing enters gentle breeze area.

4. it is according to claim 1 a kind of based on the modified neutral net overhead transmission line forecasting wind speed side for dividing gentle breeze area Method, it is characterised in that

The data in gentle breeze area will be extracted or the data in non-gentle breeze area are brought into RBF neural, pass through output layer f (x) Forecasting wind speed data are obtained, wherein, f (x) expression formula such as formula：

Wherein, x₁,x₂,…,x_nFor input layer, air speed data is updated to input layer, w₀,w₁,…,w_nFor weight,It is radial direction base Function, its expression formula is：

<mrow> <mi>&amp;phi;</mi> <mrow> <mo>(</mo> <mi>r</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>exp</mi> <mo>(</mo> <mrow> <mo>-</mo> <mfrac> <msup> <mi>r</mi> <mn>2</mn> </msup> <mrow> <mn>2</mn> <msup> <mi>&amp;sigma;</mi> <mn>2</mn> </msup> </mrow> </mfrac> </mrow> <mo>)</mo> <mo>.</mo> </mrow> 1