Design And Implementation Of Disaggregating Residential Energy Consumption System With Nonintrusive Sensing

Power and water shortages have become one of the main leading factors which could interferethe development of all countries in the world. Especially in recent years the population increaseand environmental deterioration lead to the cities facing increasing household electrical and waterdemand. At present, the main way of electric power production mode in China is still in the formsof high investment, high energy consumption, high emission and low efficiency which cause greatdamages to the environment. Thus to improve residents’ energy saving consciousness and theefficiency of power usage, will be a great promoting of the energy conservation and emissionreduction throughout the whole society. The more solutions to addressing water shortages are toexploit new sources of supply and build advanced water supply facilities.Homes currently have no explicit means to evaluate whether their household energyconsumption is more reasonable. Unfortunately their monthly utility bills don’t provide insight tounderstanding where the energy is being used. Aim to disaggregate household energyconsumption, a system of identification and classification of electric appliances and water fixturesis designed applying nonintrusive sensing method in this paper. This system can not only providecustomers with the energy consumption in details at the individual device-level but also morereasonable suggestions on how to use household energy. It promises a variety of significantapplications, including energy conservation, home healthcare, energy monitoring, government andsocial organization policy making.The main research content of this paper is divided as followed:In the first place, this paper is to study the components of electromagnetic interference (EMI)noises in residential power line, and to discuss the basic theory of switched mode power supplies(SMPS). The EMI noises generated during SMPS devices operation continuously propagates intothe power line of a home with different characteristics in frequency domain, making it suitable asa unique identifier for individual device. The water pressure in residential piping holds at a stablestate, so when the valve of a water fixture is opened or closed, it will be effect changes of the pressure wave in pipe. According to different transmission path and valve types, each water fixtureCorresponds to a specific pressure wave, and it can be used as an identifier for classifying waterfixtures.Secondly, a computing processing system is designed using FPGA as a microprocessor. Itsperipheral equipments consist of an AD converter, a preprocessing circuit, a LCD display interfaceand a network interface. Applying nonintrusive single point sensing technology with existinginfrastructures in home, the EMI noises in residential power-line can be acquired by plugging thissystem into any electrical outlet in the home. In the same way, this system can be installed in anypipe outlet to obtain changes of the pressure wave. The results of data processing transmit throughnetwork to the PC, presenting a more profound understanding to users.Thirdly, the device identification and classification method is to propose. For EMI noises inresidential power-line, to reduce the interference of base-noise, sliding window is used to averageincoming raw data. FFT and power spectrum vector chasing algorithm are applied to identifyelectrical events. The supervised pattern recognition algorithm is adopted to classify electricalappliances in accordance with spectral characteristics of SMPS. As the pressure wave is uniquewhen water fixtures open and close, it can be identified and classified by using sliding windowtechnique and pattern recognition algorithm based on inner product of vector.Finally, to estimate the feasibility and sensitivity of the system, the prototype system isdeployed in real homes. By compared with the actual electric and water consumption, it showsthat this system has a higher accuracy in identifying and classifying individual device, and realizesthe purpose of disaggregating residential energy consumption.