Effective and Efficient Smart Meter Data Analytics

Sprypoint总部设在PEI的夏洛特敦，是一家为电力、天然气、水、废水和电信公用事业行业开发创新解决方案的软件公司。其软件产品已帮助北美各地的许多公用事业公司节约能源，节省运营成本，并提供更好的客户服务。随着公用事业行业的计算需求随着时间的推移而变化，这些产品也会随着时间的推移而发展。近年来，智能电网开始在全球范围内取代传统电网，许多公用事业公司已经将智能电表部署为智能电网的基本组件。这些智能电表通常在每个时间间隔结束时收集消耗数据，时间间隔可能短至5分钟。因此，在一年或更长的时间内从一个城市的所有消费者那里收集的数据量是巨大的。Sprypoint已经开发了从这些数据中提取信息并将其可视化的软件产品，例如，构建消费者每小时或每天消费的直方图。然而，为了帮助公用事业公司更好地设计动态定价策略(以降低峰值需求)，并促进适合不同消费群体的节能计划，需要开发更多的分析方法，以自动发现海量智能电表数据中的消费者模式和趋势。**根据最近与客户的讨论，Sprypoint确定了以下两项重要的分析任务：1)了解家庭和建筑物的热敏感度，特别是识别那些在室外温度变化时电力消耗迅速变化的人；2)了解消费者的典型日常电力消耗习惯并对其进行分类。因此，这个拟议的项目将开发执行这两项任务的方法，以构建有效的智能电表数据分析系统。此外，由于智能电表数据量巨大，我们还将调查这些方法的效率，以确保这些任务能够快速执行。*

Based in Charlottetown, PEI, Sprypoint is a software company that develops innovative solutions to the electric, gas, water, wastewater and telecom utility industries. Its software products have helped many utilities across North America to conserve energy, to save operational costs and to provide better customer service. These products also evolve over time, as the computational needs of the utility industry change over the years. In recent years, smart grids have started to replace conventional power grids worldwide, and many utility companies have deployed smart meters as fundamental components of smart grids. These smart meters typically collect consumption data at the end of each time interval which could be as short as 5 minutes. Thus, the amount of data collected from all the consumers in a city over a period of one year or more is large. Sprypoint has built software products that extract information from these data and visualize it, e.g., building histograms of a consumer's hourly or daily consumption. However, to help utilities better design dynamic pricing strategies (to reduce peak demand) and promote energy saving programs suitable for different consumer groups, more analytics methods need to be developed to automatically discover consumer patterns and trends over the massive amount of smart meter data. **Based on recent discussions with their clients, Sprypoint has identified that the following two analytics tasks are important: 1) understanding the thermal sensitivity of households and buildings, especially to identify those whose power consumption changes rapidly when the temperature outside changes, and 2) understanding the typical daily power consumption habits of consumers and categorizing them. This proposed project will thus develop methods to perform these two tasks, to build an effective smart meter data analytics system. Furthermore, since the amount of smart meter data is massive, we will also investigate the efficiency of these methods, to ensure that these tasks can be performed quickly.******