GOALI: Neural Networks and Adaptive Critic Designs For Energy Security and Sustainability

建议编号：ECCS-0802047建议标题：GOALI：用于能源安全和可持续发展的神经网络和自适应批评设计PI名称：哈雷，RonaldPI机构：GA技术研究公司-GA理工学院GOALI：用于能源安全和可持续发展的神经网络和自适应批评设计目标和方法目标和方法是开发一种智能广域控制器来监控和协调包括风电场、太阳能发电场、抽水蓄能系统和其他传统发电机在内的电网的大范围区域，以便在缓慢变化的条件下(如移动的云层、风速变化)以及在暂态扰动期间优化利用所有这些资源。该方法将使用自适应批评设计来开发最优广域控制器，该控制器不需要对电力系统进行任何数学描述。这包括首先进行仿真，然后在实时数字仿真器上进行验证，最后在科罗拉多州的一个小型2000kW电力系统上进行演示。智力价值学习自适应批评设计的理论与细胞神经网络和对象网络的比例有多好。风电场和太阳能发电场的最佳利用，将提高可靠性，在电网干扰时提供减振，并协助控制室操作员。这些发现将提供对多时基系统识别器和神经控制器的更好理解。广泛影响该项目的结果将使太阳能和风电场能够像其他更传统的能源一样被工程师识别和使用，并与控制措施完全集成，以帮助维护电网稳定、能源安全和可持续性。包括少数族裔在内的熟练劳动力将被生产出来，作为未来追求“清洁”能源和减少二氧化碳和其他温室气体排放的冠军。

AbstractProposal Number: ECCS-0802047Proposal Title: GOALI: Neural Networks and Adaptive Critic Designs For Energy Security and SustainabilityPI Name: Harley, RonaldPI Institution: GA Tech Research Corporation - GA Institute of TechnologyGOALI: NEURAL NETWORKS AND ADAPTIVE CRITIC DESIGNS FOR ENERGY SECURITY AND SUSTAINABILITYObjectives and ApproachThe objective is to develop an intelligent wide area controller to monitor and coordinate wide areas of a power network that includes a windfarm, solarfarm, pumped storage system and other traditional generators, in order to optimally use all of these resources both during slow changing conditions such as moving cloud cover, and variations in wind speed, as well as during transient disturbances. The approach will use adaptive critic designs to develop an optimal wide area controller, which does not need any mathematical descriptions of the power system. This involves first simulation, then validation on a real time digital simulator and finally, a demonstration on a small scale 2000kW power system in Colorado.Intellectual MeritLearning how well the theory of adaptive critic designs scales up with cellular neural networks and ObjectNets. Optimum utilization of the wind and solar farms, will improve reliability, damping during disturbances in the power network, and assist control room operators. The findings should provide a better understanding of multiple-time base system identifiers and neurocontrollers.Broader ImpactThe outcome of this project will allow solar and wind farms to be recognized and used by engineers in the same way as other more traditional energy sources and to be fully integrated with controls to assist in maintaining grid stability, and energy security and sustainability. Skilled manpower including minorities will be produced to serve as future champions in the quest for generating "clean" energy and reducing CO2 and other greenhouse gases emissions.