Voltagemanagementis aeffective wayto reduceenergyconsumption anddemand which some utilities have usedfor many years. Studies in the context ofthis issue idoneeither byetworkmodelling that need accurate information of the network loads and the percentage of different load composition in daily load profile and or information obtained by Advanced metering infrastructure[AMI] that are installedfor each customer. However, it is difficult toobtaiaccurate instantaneous information of each customer load in the distribution network and also using too many expensive AMI for voltage management program with the related telecommunication network, is very costly. So, amixed method using a simple network model and also a few number ofmart metering devices, could be effective. In this thesis, whilst introducing the concept of voltage management and its methods for demand and energy reduction, a new method based on artificial neural networks for implementation of voltage management program on a distribution network is presented. The proposed method without any information of instantaneous customer loads and only using a few smart meters in the network, estimates the voltage profile of the distribution network and the voltage management program will be implemented using this estimation. Also a new algorithm for determining the benefits of the voltage management program using total load curves will be offered. Also, ithis study DG penetration at thedistributioetwork,as one of the influentialfactoron voltage profile andetwork model is considered in voltage management studies. This thesis uses distributed generatioaa solution totrengtheweakointin thegrid againstvoltage drop,a way toincreasethe maximumreachable voltage reduction amount in the voltage management program. 33 bus distribution test grid taking into account the modern electrical appliances is used for case study andthe simulation resultothis network denote that theimplementation of the voltage management plans have its meritimoderdistribution networks. Key Words Voltage management, energy conservation, peak load reduction, artificial neural network, distributed generation.
