Abstract

Rising fuel costs boost energy prices, which is a driving force for improving efficiency of operation of any energy generation facility. This paper focuses on enhancing the operation of distributed integrated energy systems (IES), system that bring together all forms of cooling, heating and power (CCHP) technologies. Described methodology can be applied in power generation and district heating companies, as well as in small-scale systems that supply multiple types of utilities to consumers in industrial, commercial, residential and governmental spheres. Dispatching of such system in an optimal way needs to assess large number of production and purchasing schemes in conditions of continually changing market and variable utility demands influenced by many external factors, very often by weather conditions. The paper describes a combination of forecasting and optimization methods that supports effective decisions in IES system management. The forecaster generates the future most probable utility demand several hours or days ahead, derived from the past energy consumer behaviour. The optimizer generates economically most efficient operating schedule for the IES system that matches these forecasted energy demands and respects expected purchased energy prices.

Keywords: Integrated energy system; Energy management; Demand forecasting; Energy resource allocation

Article Outline

1. Introduction

2. Exploratory analysis of historical data

2.1. Influencing factors

2.1.1. Weather and environmental conditions

2.1.2. Calendar-based variables

2.1.3. Seasonal effects

2.1.4. Economic variables

2.1.5. Nominations

2.2. Load profile charts

2.3. Interactive visual data analysis

2.3.1. Correlation plot

2.3.2. Matrix plot

2.3.3. Parallel plot

2.3.4. Trend plot

3. Energy demand forecasting

3.1. Bayesian locally weighted polynomial regression

3.1.1. System definition

3.1.2. Data set

3.1.3. Weights

3.1.4. Model

3.1.5. Bayesian prediction

3.1.6. Locality

3.2. Forecasting workflow

3.2.1. Data collection

3.2.2. Model building

3.2.3. Performance assessment

3.6. Solution architecture

3.7. Implementation issues

4. Optimal resource allocation

4.1. Problem formulation

4.2. Application workflow

4.2.1. Creation of the technology scheme

4.2.2. Entry of model parameters

4.2.3. Connection to the external data sources

4.3. Case study—hospital utility optimization

5. Conclusion

References