Development of a novel intelligent approach for estimation of electrical power consumption in subway stations

一種嶄新的評估地鐵站耗電量智慧分析方法建模

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

  • Chan Man Philip LEUNG

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date15 Jul 2015

Abstract

With the growing demand for energy due to rapid increase in population and expansion of economies, energy saving has become a global issue of concern. To lower the stress we put on the environment, energy saving is one of the quickest and cheapest ways to provide a sustainable environment for future generations. For the analysis of energy usage of buildings, many researchers conduct extensively studies covering many aspects. With the aid of software, whole energy analyses were carried out for different types of buildings including institutional, office, commercial, industrial and residential. Studies also focus on the design of building envelop and HVAC systems in order to find ways of improving energy performance. However, not many studies have been conducted on the energy consumption analysis of subway station buildings. The extensive use of subway system initiates the engineers and scientists’ research on the energy consumptions of the subway stations. Pioneering works show that the prediction of energy consumption of a subway station is non-linear in nature. Different mathematical models were applied to simulate the energy consumption. With traditional energy analysis tools such as calculation methods or simulation software, the analysis process is timeconsuming and tedious. These approaches can only perform energy analysis each time for only one scenario based on detailed design parameters including architecture, construction, MVAC system design, lighting, equipment, occupant numbers and behavior, and their relationships to energy usage. These parameters imply the need for large amounts of data, which have to be carefully collected, handled and input into models established by traditional analysis tools. Any mistakes in these procedures may cause great errors. With the advantage of handling non-linear problems which are inherently noisy and capturing fine relationships among large amount data, ANN provides a rapid approach for architects and engineers to estimate the electrical consumption of a new station so that they can optimize the designs of the mechanical and electrical systems to achieve maximum efficiency. This study develops an intelligent approach using ANN model to predict the energy consumption of railway stations. Multi-layered Perceptron (MLP) model is adopted to mimic the non-linear correlation between energy consumption, the spatial design of the station, meteorological factors and also the usage of the 19 stations selected. A coefficient of correlation is obtained between the MLP predicted results and the actual collected data to evaluate the performance of the prediction. A statistical approach is applied to assess the performance of the developed MLP model. It shows that minimum a coefficient of correlation is 0.96 with a 95% confidence level which is considered sufficient for engineering application. This approach is also adopted to predict the profile of the weekly electrical consumption of a selected station. The predicted profile reasonably agrees with that of the actual consumption. This study develops a useful tool to estimate the electrical power consumption of new MTR stations With the established ANN model, this thesis demonstrates one of its applications on spatial design of subway station in respect of energy saving. By learning from historical data of outdoor temperature (dry-bulb), outdoor relative humidity, area of concourse, area of platform, area of shops, plant room area and staff accommodation area, the ANN model can be trained to capture fine relationships between these data sets, and in turn make predictions. In the thesis, computational simulation using EnergyPlus is also used to verify the ANN prediction. Energy consumption analyses based on randomly selected cases and the optimal case predicted by ANN are also conducted to verify the ANN result. The result shows that the optimal case consumes less energy than the randomly selected cases and the reliability of the ANN result. Hence, the study concludes that the established ANN model can help architects and engineers to make decision on the configuration design of subway station in respect of energy saving.

    Research areas

  • Forecasting, Railroad stations, Energy conservation, Electric power consumption