Electricity consumption intelligent prediction system and method

Abstract

The invention discloses an electricity consumption intelligent prediction system and method, wherein the system comprises an intelligent electricity consumption terminal (1) as well as an electric energy collecting module (2) used for managing real-time data collection, a local storage management module (3), a predication point prediction value analyzing and calculating module (4) and a GPRS (general packet radio service) communication control module (5) which are respectively connected with the intelligent electricity consumption terminal (1), wherein the GPRS communication control module (5) is in charge of communication transmission control; and a core prediction processing algorithm of the predication point prediction value analyzing and calculating module (4) adopts a BP (back propogation) neural network. According to the invention, users can be informed of the electricity consumption situation in advance, and effective electricity consumption optimization suggests are provided for the users by combining with the electricity consumption optimization algorithm, the electricity consumption habits of users are improved, unnecessary electricity expenditure of electric appliances of the users is avoided, and electricity expense is saved for the users economically; and the power grid fluctuation caused by electricity consumption in peak time for the users can also be lowered, thus a power grid is more stable, more stable electricity power conveying is provided, the electricity consumption quality of residents is improved, and the production of enterprise electricity power users of production, manufacturing and the like operates stably.

Description

technical field [0001] The invention relates to the technical field of a prediction system and method for providing reference basis for power supply and demand balance in the power industry, and providing reasonable power consumption suggestions for power facility construction, large-scale power-consuming enterprises and ordinary resident users. Background technique [0002] Since electric energy is real-time and non-storable, it is necessary to maintain a balance between power generation and power consumption at all times, otherwise it will cause insufficient power supply or waste of power. On the one hand, power operators need to reasonably plan the construction of power facilities according to the electricity consumption of users; Reasonable application of power resources not only improves the stable operation of the inherent power grid but also saves unnecessary electricity expenses. In order to solve these problems, the prediction of electricity consumption is introduc...

AI Technical Summary

Problems solved by technology

The gray model is only ideal for predicting systems that develop exponentially, and the prediction progress is unstable; the mathematical model of the linear regression model is relatively simple, and the prediction effect can only be achieved under special circumstances where the past, present and future development patterns are consistent. It is relatively good, but in the demand forecasting of electricity consumption, this situation is only an ideal situation, so this method is not adaptable and has relatively large limitations; the single consumption method needs to do a lot of tedious statistical work before forecasting, which is not conducive to high efficiency. In addition, the unit consumption method has uncertainty for medium and long-term power forecasting, and the forecasting effect is not ideal; the trend analysis model method is based on the continuity of the development of things for forecasting, and is only used in the medium and long-term forecasting of power demand. The effect is good, but the effect is poor in short-term forecasting, and it needs to find the law in a large amount of historical data, which is a very time-consuming and labor-intensive work, and its realization is more complicated and difficult; the electric elasticity coefficient method has no The law can be followed, it is difficult to apply in practice, and because of its rapid changes, the gap between the predicted value and the real value is too large, and the credibility is not high

[0004] The above prediction methods are different from the perspective of research, the starting point of modeling, the form of data and the applicable conditions. There are more or less deficiencies and defects in the implementation of forecasted power consumption in the power system. The deviation of the true value is too large or it is difficult to implement and apply due to the complexity of the method itself

Images