The objective of this thesis is to study the load forecasting of distribution feeders and substations for Fong-Shan District of Taiwan Power Company. To increase the accuracy of load forecasting, the load characterization of customers served has been investigated. The typical load patterns of different customers classes and derived by performing the statistic of power consumption data retrieved. The daily load profiles and load consumptions data distribution feeders and substations have been solved by considering the typical load patterns and energy consumption of all customers served. To investigate the correlation ship of temperature and energy consumption of customer classes, the temperature sensitivity of customer energy consumption has been used to update the load composition and the contribution of load change by different customer classes. To perform the load forecasting of distribution systems, the linear, nonlinear and hybrid load forecasting modules have been proposed. The historical load data of distribution feeders and substations in Fong-Shan District have been used to derive the load forecasting modules. To analyze the accuracy of load forecasting by considering the temperature effect, the temperature change is included in the load forecasting module. With the load forecasting derived, the proper load transfers among different distribution feeders and different substations have been determined to achieve the load balancing of service areas.

Accurate short term load forecasting plays a very important role in power system management. As electrical load data is highly non-linear in nature, in the proposed approach, we first separate out the linear and the non-linear parts, and then forecast the load using the non-linear part only. The Semiparametric spectral estimation method is used to decompose a load data signal into a harmonic linear signal model and a nonlinear trend. A support vector machine is then used to predict the non-linear trend. The final predicted signal is then found by adding the support vector machine predicted trend and the linear signal part. With careful determination of the linear component, the performance of the proposed method seems to be more robust than using only the raw load data, and in many cases the predicted signal of the proposed method is more accurate when we have only a small training set.

No description available.

An Integrated Chaos Search Genetic Algorithm (CGA) /Fuzzy System (FS), Tabu Search (TS) and Neural Fuzzy Network (NFN) method for load forecasting is presented in this paper. A Fuzzy Hyper-Rectangular Composite Neural Networks (FHRCNNs) was used for the initial load forecasting. Then we used CGAFS and TS to find the optimal solution of the parameters of the FHRCNNs, instead of Back-Propagation (BP). First the CGAFS generates a set of feasible solution parameters and then puts the solution into the TS. The CGAFS has good global optimal search capabilities, but poor local optimal search capabilities. The TS method on the other hand has good local optimal search capabilities. We combined both methods to try and obtain both advantages, and in doing so eliminate the drawback of the traditional ANN training by BP. This thesis presents a hybrid Chaos Search Immune Algorithm (IA)/Genetic Algorithm (GA) and Fuzzy System (FS) method (CIGAFS) for solving short-term thermal generating unit commitment problems (UC). The UC problem involves determining the start-up and shutdown schedules for generating units to meet the forecasted demand at the minimum cost. The commitment schedule must satisfy other constraints such as the generating limits per unit, reserve and individual units. We combined IA and GA, then added chaos search and fuzzy system approach in it. Then we used the hybrid system to solve UC. Numerical simulations were carried out using four cases; ten, twenty and thirty thermal units power systems over a 24-hour period.

Electric power systems are huge real time energy distribution networks where accurate short term load forecasting (STLF) plays an essential role. This thesis is an effort to comprehensively investigate new and advanced neural network (NN) architectures to perform STLF. Two hybrid and two 3-layered NN architectures are introduced. Each network is individually tested to generate weekday and weekend forecasts using data from three jurisdictions of Canada. Overall findings suggest that 3-layered cascaded NN have outperformed almost all others for weekday forecasts. For weekend forecasts 3-layered feed forward NN produced most accurate results. Recurrent and hybrid networks performed well during peak hours but due to occurrence of constant high error spikes were not able to achieve high accuracy.

Indiana University-Purdue University Indianapolis (IUPUI) / Electric demand charge is a large portion (usually 40%) of electric bill in residential, commercial, and manufacturing sectors. This charge is based on the greatest of all demands that have occurred during a month recorded by utility provider for an end-user. During the past several years, electric demand forecasting have been broadly studied by utilities on account of the fact that it has a crucial impact on planning resources to provide consumers reliable power at all time; on the other hand, not many studies have been conducted on consumer side. In this thesis, a novel Maximum Daily Demand (MDD) forecasting method, called Adaptive-Rate-of-Change (ARC), is proposed by analysing real-time demand trend data and incorporating moving average calculations as well as rate of change formularization to develop a forecasting tool which can be applied on either utility or consumer sides. ARC algorithm is implemented on two different real case studies to develop very short-term load forecasting (VSTLF), short-term load forecasting (STLF), and medium-term load forecasting (MTLF). The Chi-square test is used to validate the forecasting results. The results of the test reveal that the ARC algorithm is 84% successful in forecasting maximum daily demands in a period of 72 days with the P-value equals to 0.0301. Demand charge is also estimated to be saved by $8,056 (345.6 kW) for the first year for case study I (a die casting company) by using ARC algorithm. Following that, a new Maximum Demand Management (MDM) method is proposed to provide electric consumers a complete package. The proposed MDM method broadens the electric consumer understanding of how MDD is sensitive to the temperature, production, occupancy, and different sub-systems. The MDM method are applied on two different real case studies to calculate sensitivities by using linear regression models. In all linear regression models, R-squareds calculated as 0.9037, 0.8987, and 0.8197 which indicate very good fits between fitted values and observed values. The results of proposed demand forecasting and management methods can be very helpful and beneficial in decision making for demand management and demand response program.

Power system planning, reliability analysis and economically efficient capacity scheduling all rely heavily on electricity demand forecasting models. In the context of a deregulated wholesale electricity market, using scheduling a region’s bulk electricity generation is inherently linked to future values of demand. Predictive models are used by municipalities and suppliers to bid into the day-ahead market and by utilities in order to arrange contractual interchanges among neighboring utilities. These numerical predictions are therefore pervasive in the energy industry. This research seeks to develop a regression-based forecasting model. Specifically, electricity demand is modeled as a function of calendar effects, lagged demand effects, weather effects, and a stochastic disturbance. Variables such as temperature, wind speed, cloud cover and humidity are known to be among the strongest predictors of electricity demand and as such are used as model inputs. It is well known, however, that the relationship between demand and weather can be highly nonlinear. Rather than assuming a linear functional form, the structural change in these relationships is explored. Those variables that indicate a nonlinear relationship with demand are accommodated with penalized splines in a semiparametric regression framework. The equivalence between penalized splines and the special case of a mixed model formulation allows for model estimation with currently available statistical packages such as R, STATA and SAS. Historical data are available for the entire New England region as well as for the smaller zones that collectively make up the regional grid. As such, a secondary research objective of this thesis is to explore whether or not an aggregation of zonal forecasts might perform better than those produced from a single regional model. Prior to this research, neither the applicability of a semiparametric regression-based approach towards load forecasting nor the potential improvement in forecasting performance resulting from zonal load forecasting has been investigated for the New England wholesale electricity market.

Building energy load forecasting is becoming an increasingly important task with the rapid deployment of smart homes, integration of renewables into the grid and the advent of decentralized energy systems. Residential load forecasting has been a challenging task since the residential load is highly stochastic. Deep learning models have showed tremendous promise in the fields of time-series and sequential data and have been successfully used in the field of short-term load forecasting at the building level. Although, other studies have looked at using deep learning models for building energy forecasting, most of those studies have looked at limited number of homes or an aggregate load of a collection of homes. This study aims to address this gap and serve as an investigation on selecting the better deep learning model architecture for short term load forecasting on 3 communities of residential buildings. The deep learning models CNN and LSTM have been used in the study. For 15-min ahead forecasting for a collection of homes it was found that homes with a higher variance were better predicted by using CNN models and LSTM showed better performance for homes with lower variances. The effect of adding weather variables on 24-hour ahead forecasting was studied and it was observed that adding weather parameters did not show an improvement in forecasting performance. In all the homes, deep learning models are shown to outperform the simple ANN model. / Master of Science / Building energy load forecasting is becoming an increasingly important task with the rapid deployment of smart homes, integration of renewables into the grid and the advent of decentralized energy systems. Residential load forecasting has been a challenging task since residential load is highly stochastic. Deep learning models have showed tremendous promise in the fields of time-series and sequential data and have been successfully used in the field of short-term load forecasting. Although, other studies have looked at using deep learning models for building energy forecasting, most of those studies have looked at only a single home or an aggregate load of a collection of homes. This study aims to address this gap and serve as an analysis on short term load forecasting on 3 communities of residential buildings. Detailed analysis on the model performances across all homes have been studied. Deep learning models have been used in this study and their efficacy is measured compared to a simple ANN model.

A newly-developed technique for short-term load forecasting is generalized. The algorithm combines features from knowledge-based and statistical techniques. The technique is based on a generalized model for the weather-load relationship, which makes it site independent. Weather variables are investigated, and their relative effect on the load is reported. That effect is modeled via a set of parameters and rules that constitute the rule based technique. This technique is very close to the intuitive judgmental approach an operator would use to make his guess of the load. That is why it provides a systematic way for operator intervention if necessary. This property makes the technique especially suitable for application in conjunction with demand side management (DSM) programs. Moreover, the algorithm uses pairwise comparison to quantify the categorical variables, and then utilizes regression to obtain the least-square estimation of the load. Because it uses the pairwise comparison technique, it is fairly robust. Since the forecast does not depend on any preset model, the technique is inherently updatable. A generalized version of the technique has been tested using data from four different sites in Virginia, Massachusetts, Florida and Washington. The average absolute weekday forecast errors range from 1.30% to 3.10% over all four seasons in a year. Error distributions show that the errors are 5% or less around 91 % of the time. / Master of Science

Resumo: A previsão de carga, em sistemas de energia elétrica, constitui-se numa atividade de grande importância, tendo em vista que a maioria dos estudos realizados (fluxo de potência, despacho econômico, planejamento da expansão, compra e venda de energia, etc.) somente poderá ser efetivada se houver a disponibilidade de uma boa estimativa da carga a ser atendida. Deste modo, visando contribuir para que o planejamento e operação dos sistemas de energia elétrica ocorram de forma segura, confiável e econômica, foi desenvolvida uma metodologia para previsão de carga, a previsão multinodal, que pode ser entendida como um sistema inteligente que considera vários pontos da rede elétrica durante a realização da previsão. O sistema desenvolvido conta com o uso de uma rede neural artificial composta por vários módulos, sendo esta do tipo perceptron multicamadas, cujo treinamento é baseado no algoritmo retropropagação. Porém, foi realizada uma modificação na função de ativação da rede, em substituição à função usual, a função sigmoide, foram utilizadas as funções de base radial. Tal metodologia foi aplicada ao problema de previsão de cargas elétricas a curto-prazo (24 horas à frente) / Abstract: Load forecasting in electric power systems is a very important activity due to several studies, e.g. power flow, economic dispatch, expansion planning, purchase and sale of energy that are extremely dependent on a good estimate of the load. Thus, contributing to a safe, reliable, economic and secure operation and planning this work is developed, which is an intelligent system for multinodal electric load forecasting considering several points of the network. The multinodal system is based on an artificial neural network composed of several modules. The neural network is a multilayer perceptron trained by backpropagation where the traditional sigmoide is substituted by radial basis functions. The methodology is applied to forecast loads 24 hours in advance / Orientador: Carlos Roberto. Minussi / Coorientador: Francisco Villarreal Alvarado / Banca: Anna Diva Plasencia Lotufo / Banca: Maria do Carmo Gomes da Silveira / Banca: Gelson da Cruz Junior / Banca: Edmárcio Antonio Belati / Doutor