Energy saving in a home environment is an emerging concept in the field of the smart home, that opens up a diverse set of downstream applications, such as dynamic pricing and demand replication techniques, whose goals are to reduce power consumption from the main utility and power cost. while achieving demand-side management (DSM) in a residential household. 1 2. Before, to control residential peak load, two techniques have been utilized by utilities: Direct load control(DLC) and time differing price programs. In DLC, the utility controls some substantial energy engrossing appliances like air conditioners, refrigerators, water heaters etc… Since Direct load control does not require progressed metering and interest in direct communication and it can work with the current metering framework, it has been fruitful in decreasing demands at peak pricing circumstances. However, DLC is poor in tackling in a certain area where customers who do not have large energy consuming appliances like air conditioners are not eligible to use program benefits 3 4. Whereas, the time-differing price works with a pricing signal. Since cost is higher at crest times, it urges clients to move their power utilization from crest to off-crest hours. For this situation, an inward financial decision-making process drives client reaction and the load adjustments are totally intentional. 5
Home energy management system (HEMS) foundation in late year, imagining the possibilities of smart home concept to achieve demand-side management (DSM) in the residential environment 6. This involves signaling to clients to indicate that when is the usage of electricity is desirable or when it is not 7. A combination of different hardware devices and software programs like smart metering and information and communication technology (ICT) facilitate bidirectional communication 8. Further, DSM solution involves the integration of renewable energy resources (RES) like photovoltaic (PV) power or wind power in the smart home 9. Nevertheless, Utilization of generated power is not completely utilized by the smart home due to improper nature of RES. A promising answer to overcome this problem is to incorporate a certain set of batteries as energy storage system ESS to HEMS 10. ESS is capable to accomplish shaving from peak demand and power cost reduction by accumulating the power from main grids during off-crest price periods or when there is unreasonable RES generation and later, providing accumulated energy to end client during peak price periods or at the time of insufficient power generation from RES. The integration of RES and ESS facilitate HEMS to a new approach called Demand Response (DR). Demand response means changes in usage of utility power by end clients from their normal consumption patterns in response to changes in the cost of utility power over time” 11. Cost based demand response DR deem in flattening demand variations as the main objective. DR benefits both customers and utility If facilitate customers to reduce crest demand in response to incentives 12. Whereas in utility side, by maintaining in the reduction of high peak demand, DR program is successful in safeguarding grid from the risk of outages. During peak load periods, control the supply-demand ratio, and better the grid reliability 13 14.
In contrast to DR approach, integration of RES and ESS provides reliable and efficient in saving of cost and energy along with grid stability in the smart residential environment, but uncertainty
in renewable energy production, high discharge rate in ESS which reduces battery life and uncertain power demand make HEMS system very hard to deal with, To overcome this uncertain situation, we deployed self-learning Fuzzy control (SLFC) algorithm. Where Fuzzy logic control is utilized in 15 to resolve the load shifting of appliances so that both power cost and comfort level violation of consumers were reduced. The control rule of fuzzy logic in 16 optimizes the power distribution among micro-grid and improves the life cycle of the batteries. The activity of cost reduction in a smart grid against predicted heavy load is achieved by utilizing RES as the main power source and ESS as the buffer in 17. A flexible FLC with optimized rules using membership function as in 18 to reduce grid power profile variation while maintaining a state of charge (SOC) within the secure limit.
Fuzzy controller inputs are dynamic price, SOC of ESS and load demand, controller inputs is not constant. Hence, parameters need to be self-updated. Self-learning of the parameter in the proposed FLC is tuned on-line by genetic algorithm (GA) 19 20. For on-line tuning, a close-fisted parameterization scheme for an FLC called orthogonal modulated membership functions (MMF) is utilized 20 21.
A fuzzy controller as a feature of universal approximation capability 22. It is widely utilized in partially or completely unknown systems. Because of their homogeneous nature, fuzzy controller often works parallel to the adaptive controller to amend system performance 23. In this paper focus on self-learning of parameters using a genetic algorithm (GA) to update weightings to increase performance. Genetic algorithms have been widely used as an effective search technique to perform searches ranging from general to specific and from simple to complex 24 and boost the computing efficiency of covariance matching of the system 25. Basically this method is used for optimization 26. Genetic algorithms are implemented by generating a population and creating a new population by performing the following procedures: reproduction, crossover, and mutation 27. The control framework proposed in this paper provides use of GA to adjust input and output parameters in the antecedent and consequent parts of the fuzzy controller with modulated membership function 28. This paper discusses the creation of a fitness function based on Lyapunov stability analysis to enable the controller to perform online tuning of its parameters as in 29.
HEMS being best application oriented system in field of home energy management. Parallel to it, many challenges were identified to attain effective output like saving, cost saving and also in satisfying load demands. First, the control strategies of I-HEMS designed using SLFC approach relied on the fuzzy and its update of parameters that were finely tuned with respect to a given smart home environment, hence their applicability in other household conditions is varied. Second, majority of the computational intelligence based HEMS strategies neglected the dynamic electricity price factor, therefore their electricity cost minimization capability were not guaranteed. Finally, proposed I-HEMS is comparably successful in minimizing power cost, leading to their effective performances in varied real environments.
In this paper, the equipment exhibit of an intelligent home power management system, considering day-ahead power tax, equipped with a PV module and an ESS. A HEMS structure is then created utilizing a Self-learning fuzzy controller to acquire the power drawn from the main grid by considering the power produced by solar, power stored in ESS, household power demand and power costs. By utilizing the power generated from local PV module and power from ESS, the proposed HEMS intends to limit the power drawn from lattice required for fulfilling the heap requests. Note that the ideal parameter settings of the fuzzy controller in HEMS may vary according to various household conditions Physically tuning of fuzzy parameters by experimentation is dull and not so effective, to make system comparably effective A number of delicate mechanisms are introduced in this paper to facilitate the self-learning capability of fuzzy controller based on historical data and to improve the efficiency of parameter learning. The specialized curiosity and principle commitments of the proposed work in this paper are summarized as follows:
1) The proposed self-learning fuzzy controller (SLFC) to HEMS aims to minimize the electricity cost and also satisfying household load demands by prioritizing the use of power generated local PV cells and stored energy in ESS.
2) Implemented self-learning scheme uses genetic algorithm is to enable the automatic parameter tuning of SLFC based on historical data, aiming to enhance the robustness of SLFC under different household environments.
3) An efficient parameterization scheme is used to represent each membership function in the antecedent and consequent parts for the entire fuzzy rule base for SLFC with fewer parameters. The GA is then applied to learn as few as nine parameters and yet successfully learns the entire fuzzy rule base of SLFC within shorter learning time.