CN110867851B - Switch station structure design method - Google Patents

Abstract

The invention discloses a switching station structure design method, which belongs to the technical field of power distribution network planning and design, and particularly relates to a switching station network architecture design method for power grid or power supply unit planning and design, wherein each switching station in a medium-voltage power distribution network of a power supply unit is divided into a plurality of switching station groups according to predicted loads; in the switch station group, each outbound bus of each switch station is connected with one section of outbound bus of the adjacent switch station through a connecting wire; the outbound bus is connected with the interconnecting link through an interconnecting link. According to the invention, on one hand, the power distribution network in the independent power supply unit is more robust under the condition that the number of short-term substations is fixed, and on the other hand, when the switch station group is divided into a plurality of groups, the line from the switch station in the power supply unit to the power consumer is less changed when the switch station group is excessively planned for a long term; on the other hand, the output robustness of the adjacent switch stations is improved, the resource utilization rate is improved, and the design redundancy is more flexible.

Description

Switch station structure design method

Technical Field

The invention belongs to the technical field of planning and designing of power distribution networks, and particularly relates to a switching station network architecture design method for planning and designing a power supply grid or a power supply unit.

Background

The switching station is also called an opening and closing station, and plays an important role in grid management. In gridding planning construction, the load density of a main city is steadily improved, the land cost is continuously increased, and a high-capacity main transformer and a 110kV voltage level become the development trend of a high-voltage distribution network, but also bring new problems of shortage of interval resources, difficulty in releasing power supply capacity and the like. The medium-voltage switch station is used as a bus extension of a transformer substation, has the characteristics of small occupied area and flexible distribution point, and has remarkable effect in urban distribution networks. The power distribution network planning of a plurality of large cities in China takes a 10kV switching station as a main power supply of a future power distribution network. Therefore, the power supply reliability of the switching station is guaranteed to become a foundation for improving the overall safety level of the power distribution network. Typical switchyard wiring designs are based on the "N-1" safety guidelines, and are difficult to adapt to special areas with high power reliability requirements without regard to "N-2" or failure conditions in service conditions.

Disclosure of Invention

The invention provides a switching station structure design method, and a switching station network architecture designed based on the method can provide higher reliability and more convenient automatic configuration.

The technical scheme provided by the invention is as follows: a switching station structure design method divides each switching station in a medium-voltage distribution network of a power supply unit into a plurality of switching station groups according to predicted loads; in the switch station group, each outbound bus of each switch station is connected with one section of outbound bus of the adjacent switch station through a connecting wire; the outbound bus is connected with the interconnecting link through an interconnecting link.

Further, the load prediction method comprises a ground simulation type prediction method, a load density index method, a multiple variable prediction method or a trend type prediction method.

Further, when the predicted load covered by the switchyard group is larger than the design load of the first switchyard in the switchyard group, planning a second switchyard is additionally arranged around the switchyard with the minimum design load, and the switchyard group crossing or adjacent to the coverage area of the switchyard group is re-divided according to the design load of the second switchyard.

Further, in a power grid including the power supply unit, the ratio of the number of switching stations to the number of substations is not lower than 3:1.

Further, the first switchyard is a switchyard with the minimum design load in the switchyard group.

Further, the total number of all upper-level power transformation stations of one switch station group is not smaller than the total number of switch stations of the group.

Furthermore, the power supply incoming line and the connecting line of the switch station are provided with optical fiber longitudinal differential protection devices.

Furthermore, the optical fiber longitudinal differential protection device uses independent optical fiber for communication.

Further, the independent optical fibers are laid synchronously with the power supply incoming lines and the connecting lines.

Further, the optical fiber longitudinal differential protection device is configured according to the following scheme: once the fault occurs, the non-fault power-losing section can be switched to the standby power supply through the standby power automatic switching device and the extended standby power automatic switching device. The longitudinal differential protection device, the spare power automatic switching device and the expansion spare power automatic switching device are used for carrying out communication between stations through the Ethernet switch, and the inter-station communication optical cable is laid synchronously with the connecting line.

The beneficial effects brought by the invention are as follows: on one hand, the switch station group is divided in the power supply unit, each switch station in the switch station group forms a standby relation of N-2, so that the power distribution network in the independent power supply unit is more robust under the condition that the number of short-term substations of the power supply unit is fixed, and on the other hand, when the switch station group is divided to be convenient for planning excessively for a long term, the line from the switch station in the power supply unit to a power user is less changed; on the other hand, the design redundancy of the existing switchyard is fully utilized to increase the output robustness of the adjacent switchyard, the resource utilization rate is improved, and the design redundancy is more flexible.

Drawings

FIG. 1 is a network architecture of a switchyard provided by a prior art design approach;

fig. 2 is a network structure of a switching station according to an embodiment of the present invention;

FIG. 3 is a network architecture of a switchyard designed according to the method provided by the present invention;

fig. 4 is a network structure of switching stations that is reformed after planning a second switching station in the embodiment of fig. 3.

The bus comprises a bus body, a bus switch body, a bus bar, a bus-bar switch, a circuit breaker, a connecting line and a connecting switch. 6. A power supply unit 7, a switchyard 71, a first switchyard 72, a second switchyard.

Detailed Description

It should be noted that the switching station is the main power source of the urban distribution network in the future; the reliability of the switching station is enhanced, and the switching station is an important means for improving the overall power supply reliability of the power distribution network. The switch station is generally built in a load-intensive area, is provided with dual power supplies and is respectively taken from different substations or different buses of the same substation; 2 power supply incoming lines, 6-12 outgoing lines and single bus sectionalized wiring are generally adopted, and an outgoing line breaker is protected. The switch station is connected with the distribution transformer by adopting a circuit breaker, the distribution transformer can be configured according to the requirement, the distribution transformer connected with the switch station can be arranged inside and outside a transformer substation or can be arranged at one position with a sectional bus of the switch station according to the arrangement of a distribution network, wherein the switch station provided with the distribution transformer is called a KT station, and the switch station not provided with the distribution transformer is called a KF station. As shown in fig. 1, three switching stations are arranged in a power supply unit of a meshed power distribution network, the three switching stations are respectively provided with double power supplies, the three switching stations are respectively taken from different substations or different buses 1 of the same substation, and under the condition that one power supply is lost, the lower part of each load line can be recovered through a bus-bar switch 2 used for segmentation between the upper and lower part of the buses 1, and as the bus-bar switch 2, a circuit breaker 3 and a connecting switch 5 in the technical scheme of the invention are common switching equipment, the technical scheme of the invention is not designed for replacing the existing equipment, but is a network design method, the bus-bar switch 2 and the circuit breaker 3 are not distinguished in a graph, all are rectangular symbols, a solid rectangle represents that the switching equipment is in a closed and conducting state, and a hollow rectangle represents that the switching equipment is in a breaking and open-circuit state. In fig. 1, each switching station only meets the safety criterion of "N-1", and when the switching station is in the overhauling state of "N-1", the theoretical power supply reliability of the grid structure will be reduced, and the overall reliability of the power distribution network will be lower at this time, when any switching station in fig. 1 is in the overhauling state of "N-2", the switching station will lose power from the whole station, the reliability is 0, and the safety criterion of "N-2" is not met, and the high reliability cannot be achieved in the overhauling state of the upper power supply or the incoming line.

As shown in fig. 2. The technical idea of the first aspect of the present invention is to utilize a network structure with high reliability. The network structure is characterized in that each switching station in a medium-voltage distribution network of a power supply unit is divided into a plurality of switching station groups according to a basis; within these groups of switchyard, each outgoing bus bar 1 of each switchyard is connected to a section of outgoing bus bars 1 of the adjacent switchyard by a tie line 4; the outgoing bus in the switching station is connected with the connecting wire 4 led out by the connecting switch 5. The basis for such a division of the switching station groups is preferably based on the predicted load of the power supply unit which has been acquired during the gridding planning process. The prediction load can be understood as a space vector field, and the space vector field can be divided into a plurality of subspaces by using clustering, cell dividing and the like, and all switch stations corresponding to each subspace form a switch station group. Since the low-voltage distribution part of the power supply unit is generally directly connected to the power consumer via the low-voltage distribution transformer, no switching stations are provided, which are mainly referred to as individual switching stations in the medium-voltage distribution network of the power supply unit. The load prediction method comprises a ground simulation type prediction method, a load density index method, a multiple variable prediction method or a trend type prediction method.

In the power distribution network planning of the technical scheme of the invention, the switching stations are regarded as main planning power supplies of medium-voltage lines, and the ratio of the number of the switching stations to the number of substations is recommended to be not lower than 3:1. As shown in fig. 3 and 4, the embodiment of the invention is a network structure design method of a 10kV medium-voltage switch station, which aims at a power supply grid with a load density of more than 35mW/km and denser special line users. The 10kV medium-voltage switch station is an indoor power distribution station provided with medium-voltage power distribution incoming and outgoing lines and used for redistributing power. The 10kV medium-voltage switch station can be used as expansion of transformer substation buses, the number of 10kV buses is greatly increased, allocation of interval resources is optimized, the number of cables in the same path is reduced, and the equipment utilization rate is improved. The dashed lines in fig. 3 represent the coverage area of a substation, where it can be seen that two distribution substations cover 6 power supply units 6, each power supply unit 6 comprising switching stations 7 of unequal 3 to 6, and each switching station 7 in each power supply unit 6 being divided into groups of switching stations 7 connected by tie lines 4, each switching station 7 in each group forming a backup loop by at least two adjacent switching stations. As shown in fig. 4, in a power supply unit 6 having two switchyard groups, when the design load of the first switchyard 71 is smaller than the predicted load covered by the switchyard group where it is located, the second switchyard 72 is planned and designed around it for sharing its output load, and is divided into one switchyard group again, and the corresponding tie line 4 is added. In this embodiment, the second switchyard 72 is not divided into the first switchyard 71, but in some other embodiments, the switchyard group where the first switchyard 71 is located may be divided into the switchyard groups where the first switchyard 71 and the second switchyard 72 are located.

Preferably, in a power grid comprising said power supply units, the ratio of the number of switchyard to the number of substations is not lower than 3:1.

Preferably, a switching station is a switching station with the smallest design load in the switching station group. Preferably, the total number of all upper-level power transformers of one of the switchyard groups is not smaller than the total number of switchyards of the group.

Preferably, the power supply incoming line and the connecting line of the switching station are provided with optical fiber longitudinal differential protection devices. In some embodiments, the fiber optic differential protection device communicates using a separate optical fiber. In some embodiments, the independent optical fibers are laid in synchronization with the power supply line and tie. The optical fiber longitudinal differential protection devices are configured according to the following scheme: once the fault occurs, the non-fault power-losing section can be switched to the standby power supply through the standby power automatic switching device and the extended standby power automatic switching device. The longitudinal differential protection device, the spare power automatic switching device and the expansion spare power automatic switching device are used for carrying out communication between stations through the Ethernet switch, and the inter-station communication optical cable is laid synchronously with the connecting line.

The first prediction method, mathematical model prediction method, is applicable to the present invention.

According to the modeling modes of the mathematical model, the mathematical model prediction method is divided into a time sequence model, a gray system model and a correlation analysis model.

A time series model including a regression model, an exponential smoothing model, a moving average model, and the like: wherein (1) the regression model, the unified linear regression, is the existing conventional regression model, since the data obtained in the material collection stage is a function of time (or other variables such as temperature), if the load is represented by X, it is a linear function of time (or other variables) to which unified linear regression based on the number of observed data and the time axis is performed; nonlinear regression, among the fit curves that may be employed, most fit curves require the application of nonlinear regression. Some non-linear problems can be addressed by changing to linear regression. The curve model is transformed into a linear model, and then the estimated value of unknown parameters is calculated by using the least square method principle, so that the calculated simulation curve is obtained. For some of the most typical nonlinear regression equations, the analysis can be performed by transforming the variables into linear regression problems. (2) The fundamental principle of the exponential smoothing model on which the exponential smoothing method depends is that the closer the history time is, the greater the influence on the future is, and the farther the history time is, the less the influence on the future is. Another principle on which the exponential smoothing method depends is to continuously correct the new predicted value by using the summarized difference of the predicted misload prediction knowledge points, i.e. to correct by using the "error feedback" principle. Thus, its basic concept is: it is assumed that the time series has a certain characteristic, i.e. there is a certain basic data pattern, and that these observations reflect both this basic data pattern and the random variations. The goal of the exponential smoothing method is to use "smoothing" historical data to distinguish between fundamental data patterns and random variations. This corresponds to eliminating maxima or minima in the historical data, thereby obtaining a "smoothed value" of the time series, and taking it as a predicted value for the future period; linear exponential smoothing is a non-stationary process when there is a tendency for the time series to increase or decrease over time. Such time series are prevalent in nature and socioeconomic activities. Such as the continuous increase in the incomes of residents in recent years, and the gradual increase in sales of various home appliances. Linear exponential smoothing is an effective method of predicting such time series by which trend is continually adjusted taking into account the average of the increments per period. A trending time series is predicted which can be broken down into two parts, one part being the current level condition and the other part being the delta. Based on this principle, linear exponential smoothing can be used for prediction; the conic index is smooth, and some time sequences, although having a tendency to increase or decrease, are not necessarily linear and may increase or decrease in the shape of the conic. For such non-stationary time series, it may be more efficient to employ a conic exponential smoothing method. It features that not only the linear growth factor but also the growth factor of quadratic parabola are considered. Although the method of quadratic curve exponential smoothing is somewhat complex, it is quite effective for predicting non-stationary time series, and it can adjust the predicted value as the time series increases in a parabolic fashion. Generally, both the quadratic curve exponential smoothing method and the linear exponential smoothing method have advantages, and the two models can be used simultaneously in practical applications, so that one model with higher prediction precision can be selected. Note that: the smoothing coefficient alpha (the selection of the smoothing coefficient alpha has great influence on the prediction result, and for a stable number sequence, alpha is preferably 0.1-0.2). (3) Moving average model, which is the simplest adaptive model, is also the oldest time series prediction method. It can be used not only independently to make simple predictions, but it is also an important component of decomposition predictions. In mathematical statistics, if a set of data is to be processed, the arithmetic mean value thereof is often taken to represent the overall load taking prediction knowledge point summary value level, and the method is simple and effective. However, the arithmetic mean value can only be used for explaining the general situation, and high points and low points in the data cannot be seen, and development processes and trends cannot be reflected. If a set of data is segmented averaged, then this is an improvement, however, the segmented averaging approach reduces the data too much and the graph is inaccurate. The effect is significantly improved if the segment averaging method is not fixed on a segment, but is successively shifted one bit back to average with the spacing of each segment remaining unchanged. For example, every segment is 5, then the first segment 1,2,3,4,5 data points are averaged, the second segment 2,3,4,5,6 are averaged, and so on. This method of segment averaging is called the moving average method. The key of the moving average method is that only the last few data points are taken to obtain the average value, the weights of the points participating in calculating the moving average value are equal, namely, the weights of all the points are equal to 1/N, and the weights of the previous data points are zero. When the moving average method is used for time series prediction, data points with abnormal large and small are smoothed, and abnormal data has little influence on the moving average. The speed of reflection of the moving average on the data change and the ability to smooth the disturbance depend on the number of data points N for which the moving average is calculated. As N decreases, the prediction system is sensitive to the reflection of data changes, but the smoothing ability decreases, i.e., the prediction accuracy of the estimated value decreases. Prediction accuracy and reflection of data changes are contradictory and cannot be used at the same time. The value of N should therefore generally be determined by a compromise, as the case may be. In the present algorithm, the moving average model is classified into a moving average growth value method and a moving average growth rate method. Moving average growth value: and taking the annual growth value of the load as a target quantity, carrying out moving average calculation on the annual growth value of the load, thus obtaining the growth value of the load of the year to be predicted relative to the previous year, and finally obtaining the load prediction knowledge point summary load of the year to be predicted. Moving average growth rate: and taking the annual growth rate of the load as a target quantity, carrying out moving average calculation on the annual growth rate, so as to obtain the growth rate of the load of the year to be predicted relative to the load of the year before the annual growth rate, and finally obtaining the load of the year to be predicted.

The gray theory model of the load prediction is that a differential equation is established by adopting a historical data column as a model of the power load prediction. Because the gray prediction requires less original load data, does not consider distribution rules and change trends, and is simple and convenient to operate, the electric power load prediction model based on the gray theory is more and more favored and paid attention to electric power scientific research and planning personnel. Since the gray theory regards the random quantity as a gray quantity varying within a certain range, the random process as a gray process varying within a certain range and a certain time zone, and the load variation itself of the power system is affected by multiple factors, it is a random variation quantity, and it is a gray process within a certain time and range. The factors influencing the power load change cannot be completely determined, the factors are irregular, the gray theory generates the history data and the series which are irregular, the history data and the series are accumulated, the history data and the series are obviously exponentially increased compared with the original values, and the form of differential equation solution is exponentially increased, so that the data series which are exponentially increased after being generated can be naturally fitted by using a differential equation model, and the load can be easily predicted by using the differential equation, and then the load can be reversely generated and restored to an actual load predicted value. Wherein (1) a normal gray model, a GM model, is a gray model (GreyModel), GM (1, 1) is the most commonly used and simplest gray model, which is a model composed of a differential equation containing only univariates, and is a special case of GM (1, n) model. The essence of grey model prediction is 'one-time accumulation generation', namely, after the original number series are accumulated once, obvious exponential rules are formed, then a curve is used for fitting and accumulation generation, and then the accumulation reduction is carried out, so that a predicted value can be obtained. According to the prediction result, performing subtraction reduction to obtain a gray prediction model of the original data sequence; the common gray GM (1, 1) model has certain limitations when applied to load prediction. It is only suitable for short-medium term load prediction (e.g. G-type curve around 5 years) with constant growth rate of the predicted amount or exponentially growing. And for E-type and S-type curves without obvious change rules and long-term load prediction, the prediction error is larger. (2) The gray model method of the equivalent-dimensional innovation has the important reason that when long-term data and data with large fluctuation are predicted by the gray GM (1, 1) model, prediction deviation is large, wherein parameters a and b are regarded as constants, new information is not brought into the model, and old information is not removed. Based on this, an isovitamin recurrence gray GM (1, 1) model was developed. The model regards parameters a and b as a function of time t, predicts the two parameters first and then predicts the original data by a grey prediction method. The basic method comprises the following steps of tracking parameters a (k) and b (k), establishing a parameter estimation value sequence { a (k), b (k) }, analyzing a finding rule of a (k) and b (k), establishing a corresponding mathematical model, and predicting parameters a (k+1) and b (k+1) of the next step. The model applies metabolic techniques thereto. As the information value of the old data is gradually reduced along with the development of the system, the old data is removed in time, and the modeling sequence can more practically reflect the current characteristics of the system. Especially when the system undergoes qualitative leaps or mutations with the accumulation of the amount, it is clearly reasonable to remove old data that do not reflect the current characteristics of the system. Over time, new random disturbances or drivers will also enter the system, affecting the development of the system, in which case new information will be continually replenished. In the equal-dimensional innovation recurrence model, a GM (1, 1) model is firstly utilized to predict a value (refer to a first common gray model), then the predicted value is supplemented to known data, the oldest data is removed at the same time, the equal dimensions of a series are maintained, the GM (1, 1) model is built again, and the next data is predicted, and the cycle is repeated. The dual functions of parameter prediction and metabolism greatly improve the prediction precision of the equal-dimensional innovation recurrence gray GM (1, 1) model. (3) The envelope gray prediction model can predict a model which is difficult to predict gray by a number sequence by constructing an upper boundary GM (1, 1) and a lower boundary GM (1, 1) of the envelope model. According to the description defined above, the upper envelope GM (1, 1) of the disarranged sequence x is constructed, the general requirement is that the peak points of x are often not equally spaced, for which reason it is necessary to specify interpolation points by observation or experience to become equally spaced, and similarly, the lower envelope GM (1, 1) can be constructed. The envelope model is in fact a GM (1, 1) model with human empirical intervention, a model of contour coverage. Envelope ash prediction is a prediction of an interval, and specifically has the following characteristics: a. extending the upper envelope line and the lower envelope line through prediction to obtain a range of future change of the behavior variable; b. the predicted value of the original disarray sequence x (if x can be modeled as GM (1, 1)) is constrained. Or the grant falling within the envelope region, and discarding the predicted value outside the envelope region; c. GM (1, 1) modeling is performed on the envelope center point to obtain an x-predicted whitening value. Modeling: the first step: the upper envelope curve and the lower envelope curve of the sequence x are drawn to outline an upper envelope curve xU and a lower envelope curve xL which are separated from the disordered sequence x; summarizing load prediction knowledge points; and a second step of: the data are selected from the upper envelope line and the lower envelope line at equal intervals, wherein the data must comprise the peak points (not necessarily all the peak points) of x, and the upper envelope line is provided with a plurality of peak points; or the valley of x (not necessarily all valleys), for the lower envelope. Step three: GM (1, 1) modeling and predicting the upper envelope sequence and the lower envelope sequence; step four: modeling and predicting an original sequence x; step five: modeling and predicting the center sequence of the envelope region; step six: and obtaining a prediction result.

The correlation analysis model, i.e. correlation method, is to study the correlation between the predicted object and the factors and extrapolate this relationship to the future, and calculate the predicted value of the predicted object from the future changes of other factors. Theoretically, the core of the mathematical theory of load prediction is how to obtain the historical change law of the predicted object. The predictive model is in fact a mathematical function that expresses this law of variation. Specifically, the method comprises (1) unit regression analysis, wherein the basic idea of unit linear regression for prediction is as follows: based on the observations of X and Y, a regression equation is determined by looking for reasonable values of regression coefficients a and b, taking X, Y as the known number. Using the regression equation, a and b are taken as known numbers to estimate future changes in the X and Y values, or to predict the value of Y from the value of X. Unit linear regression model: y=a+bx+epsilon, X is the independent variable, Y is the dependent variable, a and b are regression coefficients, epsilon is the random error. And calculating according to the relation between the output value and the electric quantity load, the historical increase of the output value and the future development trend, and recording the output value from the historical year to the target year. (2) Multiple regression analysis, or multiple regression model, is a linear regression model that has two or more independent variables in the equation. The multiple linear regression prediction is a prediction made by using multiple influencing factors for the load electric quantity problem by using a multiple linear regression model.

The second prediction method, the load density index method, is applicable to the present invention. The load density index method not only can predict the change rule of future load quantity, but also can make corresponding prediction for the future load geographic distribution condition. By way of example, the steps include: A. zoning: dividing the regional planning into non-communication regional cells according to regional planning by referring to regional properties, wherein each regional cell only comprises one regional property (non-overlapping and non-crossing); according to the characteristics of the terrain, the middle areas (non-overlapping and non-crossing) are divided by taking railways, rivers and the like as boundaries. B. Type of load density index: there are generally two indices for different use, namely, a building area load density index and a floor area load density index. Different indexes are adopted for different land properties, such as a road square land generally adopts an occupied area index, and industrial land, resident living land and the like generally adopts a building area index. C. Load density index selection: the current power utilization indexes of different land properties of different areas are researched by adopting research indexes, load composition is selected to be similar according to the land properties of the planned areas, the planned distant load, the near-future electric quantity, the load forecast historical electric quantity and the near-future planned load forecast knowledge points of the near-future load distribution are selected to be more accurate than the planned regional distant load density index, and the power utilization indexes of the areas with similar years are summarized due to the regional differences and the uncertainty of development. (classified summarization of load density indexes of different areas is recommended) D. Determination of the inter-cell concurrence rate of the prediction results: the load of the cell is directly obtained by calculation according to the land property diagram and the load density index, so that the maximum load appears at different moments due to different load characteristics of different properties, and the time rate among the cells needs to be considered when the load of the cell is summarized. The inter-cell timing rate should be selected with emphasis on grasping the overall land-use characteristics of the region macroscopically while taking into account the differences between the main electrical loads. The current situation mainly adopts empirical data, but the credibility is to be demonstrated.

The third prediction method, the correlation coefficient method, is applicable to the present invention. The correlation coefficient prediction method is to find the predicted value of the power load by searching the relation between the power load history data and the history value of a certain predictable factor and the predicted value of the factor, assuming that the increase of the power load is similar to the change rule of the factor. This class of methods includes: an electric power elastic coefficient method, a GDP comprehensive electricity consumption and a yield unit consumption method. This type of method is similar, and is described by taking the power elastic coefficient as an example, the ratio of the power demand increase rate to the national economy increase rate is called the power elastic coefficient (power demand elastic coefficient). According to the calculated range or the comparison object, the electric power elastic coefficient can be divided into an industrial electric power elastic coefficient, an agricultural electric power elastic coefficient, an industrial and agricultural total output value electric power elastic coefficient and a national economic electric power elastic coefficient. Therefore, the development speed and the development level of the power demand can be calculated as long as the magnitude of the power elasticity coefficient value and the development level and the development speed of national economy or industry and agriculture in the prediction period are determined. Under the market economic condition, the electric power elasticity coefficient has become unstable, and with the rapid development of scientific technology, the electricity-saving technology and the management of the electric power demand side, the continuous generation and development of new economy (such as knowledge economy and information economy), the continuous expansion of the range of replacing other non-electric energy sources with electric energy, the rapid change of the relation between electric power and economy, the serious imbalance of the change steps of the electric power demand and the economic development, the difficulty in capturing the elasticity coefficient, the difficulty in obtaining satisfactory effect of the electric power demand prediction by using the elasticity coefficient method, and the gradual desalination.

The load prediction method applicable to the present invention is specifically described above so that those skilled in the art can understand the concept of predicting load according to the present invention, and other load prediction methods in the prior art may also be applied to the present invention, which will not be described herein.

In this embodiment, after obtaining the coincidence distribution of one power supply unit by the above or existing method

In a specific embodiment, the total utilization rate of each interval of 110kV and 35kV substations in the original power supply grid reaches 90% -95%, the interval utilization rate is larger than 15%, and the interval resources are relatively tense. According to the network scheme provided by the design method provided by the application, 57.9% of special line users in the area can be supplied with power by the switch station, about 35% of 10kV lines can be cut over to the switch station through optimization transformation, the interval utilization rate of transformer substations in the area can be reduced to 80% -85%, and the interval utilization rate can be reduced to 0%.

Claims (10)

1. A structural design method of a switching station is characterized in that: dividing each switchyard in a medium-voltage distribution network of a power supply unit into a plurality of switchyard groups according to the predicted load of the switchyard; in the switch station group, each outbound bus of each switch station is connected with one section of outbound bus of other adjacent switch stations in the group through a connecting wire; the outbound bus is connected with the interconnecting link through an interconnecting link.

2. A switchyard structure design method according to claim 1, characterized in that: the load prediction method comprises a ground simulation type prediction method, a load density index method, a multiple variable prediction method or a trend type prediction method.

3. A switchyard structure design method according to claim 1, characterized in that: when the predicted load covered by the switchyard group is larger than the design load of the first switchyard in the switchyard group, planning a second switchyard to share the output load of the first switchyard around the switchyard with the minimum design load of the switchyard group, and re-dividing the switchyard group crossing or adjacent to the coverage area of the switchyard group according to the design load of the second switchyard.

4. A switchyard structure design method according to claim 1, characterized in that: in a power grid comprising said power supply units, the ratio of the number of switchyard to substation is not lower than 3:1.

5. A switchyard structure design method according to claim 3, characterized in that: the first switchyard is a switchyard with the minimum design load in the switchyard group.

6. A switchyard structure design method according to claim 1, characterized in that: the total number of all upper-level transformer stations of one switch station group is not smaller than the total number of switch stations of the group.

7. A switchyard structure design method according to claim 1, characterized in that: the power supply incoming line and the connecting line of the switch station are provided with optical fiber longitudinal differential protection devices.

8. The switchyard structure design method of claim 7, wherein: the optical fiber longitudinal differential protection device uses independent optical fiber for communication.

9. The switchyard structure design method of claim 8, wherein: and the independent optical fibers are laid synchronously with the power supply incoming lines and the connecting lines.

10. The switchyard structure design method of claim 7, wherein: the optical fiber longitudinal differential protection device is configured according to the following scheme: once a fault occurs, the non-fault power-losing section can be switched to a standby power supply through the standby power automatic switching device and the extended standby power automatic switching device; the longitudinal differential protection device, the spare power automatic switching device and the expansion spare power automatic switching device are used for carrying out communication between stations through the Ethernet switch, and the inter-station communication optical cable is laid synchronously with the connecting line.

Images