The goal for this paper and my research is to reduce overall cost associated with electricity use at UC Merced. UC Merced presents itself as a unique opportunity for to model integration and optimization of renewable energy sources. It will be discussed exactly what makes UC Merced unique and how UC Merced has set a path towards higher energy efficiency on a community level.
Furthermore, I will discuss difficulties involved with integrating renewable resources and then proceed to analyze techniques for further optimization as UC Merced continues its path towards zero net energy. One of these optimization techniques, genetic algorithms; I will discuss in some detail as it was the technique chose to verify the results of the optimization.
The main goal of this study is to determine the effect of moving UC Merced‘s Central Plant load closer to or completely during daylight hours when there is inexpensive (solar) energy available or during the night time when energy pricing is minimum. While it seems logical to shift the cooling load, it has yet to be quantitatively shown that such load shifting would be more cost effective.
Genetic algorithm (GA)-based Artificial Neural Network (ANN) models are used for demand and energy production forecasting and then GA based cost optimization is performed to find optimum time window for load shifting. We determined that loading shifting can be beneficial and the associated savings are presented for both summer and winter seasons.
REDUCING ELECTRICITY COSTS BY MITIGATING POWER OUTPUT FLUCTUATIONS
Figure 3 shows the relationship between PV power output fluctuations of a 1-MW plant and the resulting cost of those fluctuations over each of the 12 months in 2010. The total cost, and the cost of fluctuations are defined respectively.
REDUCING ELECTRICITY COSTS VIA CHILLER LOAD SHIFTING
Block diagram that outlines the general energy management procedure presented in this study. Data comes from multiple sources so processing the data is an important step. Once the inputs are defined, they are fed through an ANN/GA hybrid in order to forecast power loads. These power loads, combined with a detailed price structure are then optimized using another GA. The desired output of this optimization is the start time of chillers that provides the minimal cost.
Forecasting for the net campus load should be relatively consistent throughout the year with the chiller loads removed and solar energy generated added in. This is generally depicted in the above figures; however, the months chosen do have aforementioned peculiarities. In addition to the sudden rise in campus population, some of the hottest days of the year occur red, which adds to the uncertainty in the load.
The overview that begun this study suggested many methods for energy control and load management. This field of efficiently integrating RES into the standard portfolio of electricity generation will only grow as policy continues to take effect, as it is advancing and changing rapidly. Much research in this area has been done over the past years, even more so recently, and various successful methods for load management have been implemented in small communities, factories, and buildings alike.
While GAs have long since been used and proven to solve optimization problems of all kinds, only recently have they been thought of a way to manage energy. In the same way, ANNs have been proven to be great tools for pattern recognition and curve fitting. Moreover, they have been implemented as power output forecasters to great avail in the past. In this study, they were shown to be adequate for the task at hand.
This research showed that ANNs and GAs could be a powerful combination manage and optimize a variable and uncertain energy source such as solar. The end result of this research showed the ANN/GA can yield similar results to that of more traditional simulation based methods. However, the method has not been directly implemented and may face many more hurdles. It also shows that the GA, while very flexible, may be excessive in certain respects. Such as using it to optimize inputs to the ANN when perhaps there may be easier ways to increase forecasting accuracy.
Source: University of California
Author: Bron Davis