There are 2 repositories under electricity-demand-forecasting topic.
Electricity demand forecasting for Austin, TX, using a combination of timeseries methods and regression models
Multi-User-Personality-Electricity-Load-Forecasting
This project will present an applied and game-like approach to simulating the load growth, investment decisions by two types of generation technologies, demand-price responsiveness, and reliability, of a test-case power system. The simulation begins as a 9-bus system with existing generation (3 generators) and transmission lines (8 lines). System topology can be viewed in a figure throughout the game with the yearly generation and load at each bus. In addition, dynamic color-coding is used to highlight transmission lines that exceed MVA ratings and highlight bus voltages that violate any limits. The winning objective of the player company (you) is to maximize his profit. Reliability can be tracked by viewing the N-1 generator and line contingencies every year, but this does not influence profits. There are two generation technologies used: coal and gas turbine. Each technology will have a similar competitor in the simulation. The competitor can bring down the market price and reduce the player’s profits significantly. The clock starts at T=0 in the investment game with a historical record of past prices and projected prices based on lack of investment. As time moves forward in yearly increments, the load, prices, investment costs, and other variables are adjusted to that of the player’s performance. The player has the opportunity to study various profitable and unprofitable investment alternatives each year of the simulation. If he invests at the right location, and in the right planning year, his company can make windfall profits. Competitors randomly participate in adding extra generation in random areas of the system based on the competition level settings. The challenge for the user is to study the effects of his investment decisions on market prices, reliability, and his profitability.
Forecasting time series data with MLP by Google TensorFlow.
Multi-User-Personality-Electricity-Load-Forecasting
A study on energy demand forecasting based on smart meters data. The report and the presentation of the study are also provided in this repository.
Robust regression for electricity demand forecasting against cyberattacks
Statistical evaluation of renewable and non-renewable electricity generation in the EU.
A Fuzzy system to predict the hourly electricity demand. Used triangular membership funcitons with 13 real world rules.