An informatics design of demand response system for the future smart grids

Abstract

Extreme weather events, natural disasters and global warming have already been recognized as consequences of the overuse of fossil fuels. Over the years, the generation of electricity has been dependent on the burning of fossil fuels. Clearly, a solution to address the environmental problems would be to reduce the demand for electricity and enable the wider use of renewable energy sources. Subsequently, the concept of smart grids has been proposed in order to control the demand for electricity by supporting real time demand response. Demand response is matched to real time energy demand and supply in order to reduce energy wastage due to peak reserves. A full functioned demand response system includes energy demand forecasting, demand monitoring, load management and end user site energy management functions. To facilitate the above functions, an energy management method, load prediction algorithm and advance metering infrastructure are urgently required. Firstly, the end user site energy management solution is investigated. A proactive classified fuzzy intelligence, namely an Energy Consumption Optimization algorithm (ECO algorithm), implemented in energy management systems, is proposed to optimize the energy consumption for residential apartments and offices. The exceptional intelligence is a fusion of a classification algorithm, Interrupted ON-OFF model prediction and fuzzy logic designed for the tradeoff between comfort and energy consumption. Moreover, being at the heart of the system, the intelligence is adopted by the coordinator of the system and controls the peripherals by means of a wireless sensor network in the home automation system. With appropriate system configuration, the ECO-algorithm could cater for consumption control of a range of different end user sites such as hospitals, schools, offices. However, dedicated consumption control at the end user site alone cannot complete the mission of demand response. The next challenge is to enable interactive communication between the end user site and the utility. A Multi-Interface ZigBee Building Area Network (MIZBAN) for High Rise Advance Metering Infrastructure (HRAMI) has been developed. A MIZBAN aids meter management such as Demand Response (DR), Meter Reading Order (MRO) and Load Profile (LP) for smart grid applications. To cater for high traffic communications in high rise buildings, the Multi-interface management framework (MIMF) is defined and designed to coordinate the operations between multiple interfaces based on a newly defined tree-based mesh ZigBee (T-Mesh) which supports both mesh and tree routing in a single network. In a trial undertaken in a 23-storey building, measurements revealed that MIZBAN improved the backbone and floor network round trip time by 75% and 67% respectively. In order to meet the US government's minimum DR requirement (latency < 0.25s), this dissertation forms six recommendations for MIZBAN design. To complete the demand response system design, a load prediction algorithm is essential in analyzing the data from AMI and at the same time provides important insight into energy generation. An hourly energy demand forecast model for a small zone, such as a residential area, is also discussed in this dissertation. Benefiting from the mature of cloud computing technology, instant energy consumption data is collected from the end user site via the AMI and so the predicted results are more precise. However, the prediction model must be able to generate the predicted result in a short time and therefore the least amount of input data should be used. This dissertation asserts that 7.5% of end user sites are sufficient to maintain the error percentage of predicted results below 8%. The proposed model can be applied across a wide range of conditions regardless of building type, region or climate. Finally, after taking the whole demand response system into consideration, attention is drawn to how the system can help the utilities control energy demand and reshape the load profile in residential areas.

Date of Award	15 Jul 2013
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Kim Fung TSANG (Supervisor)