Appliance Disaggregation and Optimal Scheduling

伦敦水电是一家当地分销公司，为安大略省伦敦市超过15万消费者提供服务。LG目前**正在实施一项实时能源管理试点，为每个参与的家庭提供一个智能能源集线器、智能插头和一个可远程驱动多达五台电器的应用程序。通过这项试点，LH将评估**关键峰谷电价(CPP)对用电模式的影响，通过移动应用平台提供的上下文信息**与消费者互动的能力，以及连接到智能插头的本地设备**运行时间表的自动化过程。这一努力对激励和促进消费者经济和可持续地使用电力具有重大影响。**拟议的项目由两个主要部分组成。首先，机器学习将被应用于分解智能电表和插头产生的电力**使用数据，以便在一天的不同**间隔内识别消费者家庭中的活动电器。拟议的分类器将使用在路易斯安那最先进的研究**设施中产生的高度细粒度的数据进行培训。消费者将能够通过他们的移动**应用程序提供关于本地分类准确性的反馈，然后使用该应用程序来定制和微调每个家庭的分类器。在第二阶段，利用家电分解的优化**调度将集成到智能能源中心，以便根据经济可行性、可持续性和消费者偏好自动执行本地家电的运营**调度。

London Hydro (LH) is a Local Distribution Company serving more than 150 000 consumers in the region of London, Ontario. LH is**currently implementing a real-time energy management pilot that makes available to each participating household a smart energy**hub, smart plugs and an application to remotely actuate up to five appliances. Through this pilot, LH will assess the impact of**critical peak pricing (CPP) on power consumption patterns, the ability to engage with consumers through contextual information**provided through the mobile application platform and the automation process of the operation schedules of local appliances**connected to smart plugs. This effort has significant implications in motivating and enabling economical and sustainable usage of**electricity by consumers.**The proposed project is composed of two main components. First, machine learning will be applied to disaggregate the power**usage data generated by smart meters and plugs in order to identify active appliances in a consumer's household during various**intervals over a day. The proposed classifier will be trained using highly granular data generated in the state-of-the-art research**facility at LH. Consumers will be able to provide feedback regarding the accuracy of local disaggregation through their mobile**application which is then used to customize and fine-tune the classifier for each household. In the second phase, optimal**scheduling that leverages on appliance disaggregation will be integrated into smart energy hubs for automating the operation**schedule of local appliances based on economical viability, sustainability and consumer preferences.