Optimized design of wind and photovoltaic farms for electricity consumption
The optimized design is done by analyzing of the following time series types: 1) daily average of solar irradiation, 2) hourly average of wind speed and 3) hourly (or monthly) average of electricity consumption. Neural networks are used to predict the behavior of energy sources and energy demand of the customer. Thus are calculated optimal dimensions of the electricity generation system.
It has optimized an electrolytic hydrogen production plant supplied by wind and photovoltaic generators. The neural network predicts the hourly average power of electricity generation. Electrolysis power consumption is automatically controlled attending to wind and photovoltaic generation.
It's solved the following problems:
1. Determine the optimized size of wind and/or photovoltaic farms. For that, it's simulated the production of electrical energy from wind and/or photovoltaic farms. Moreover it's simulated the energy consumption of a factory or a village.
2. Determine an optimized plan of monthly electricity consumption for a production plant or a village powered by wind and/or photovoltaic energy.
3. Develop the software for automatic control of energy consumption in a production plant powered by wind and/or photovoltaic electricity. There are considered the following cases: stand-alone systems and grid-connected systems.
To solve the above problems, it's used the following data:
1. Longitude and latitude of the wind and/or photovoltaic farms.
2. Time series of measurements, of at least one year, of hourly average of wind velocity at farm.
3. Average of electricity power consumption (annual or monthly).
4. Time series of hourly average of electricity power consumption.
The time series of daily average of solar irradiation and hourly average of wind speed, allow train a perceptron to predict values of indicated magnitudes. The irradiation time series are obtained from the database of NASA and from the longitude and latitude of photovoltaic farm site.
The time series of hourly average of wind speed must be the result of measurements (of at least one year) on the installation site of wind generators. From the latitude and longitude of the place of photovoltaic farm it's determined the optimal inclination angle of solar panels. It's used time series of hourly average of electricity power consumption, to train a perceptron to predict this magnitude.
There have considered the following ways to solve the two enunciated problems:
1. If it knows average consumption (monthly or yearly) of electricity power, define the optimal dimensions of wind and/or photovoltaic farms.
2. If it knows the time series of hourly (or monthly) average of electricity power consumption, train a perceptron to predict such magnitude. On this basis we simulate the power electricity consumption (at a village or a production plant) and determine the optimal dimensions of the wind and/or photovoltaic farms.
In the case of stand-alone installations, the second way make feasible to set an optimized plan of electricity power consumption. In case of grid-connected systems, it's possible to estimate the energy that will be taken from or injected into the grid.
1. The production planning of factories powered by wind and/or photovoltaic electricity.
2. The decreasing of size of wind and/or photovoltaic farms.
3. The estimation of annual electrical energy taken from or injected into the network.
The proposed technology is intended for all kinds of factories and villages using electricity produced by wind and photovoltaic generators.
RENEWABLE ENERGIES AND ATMOSPHERIC POLLUTION
Jose Manuel López López
+34 947 25 8895