Endeavour Energy, together with Optus, Amazon Web Services, and Unleash live, has deployed its first 5G and AI-enabled drones to improve restoration times for unplanned electricity outages, particularly during natural disasters such as storms, floods, and bushfires. As part of the first demonstration, Endeavour Energy flew the drones over physical electricity infrastructure located in Sydney's western suburb of St Marys. During the flyover, footage of damaged assets was streamed in real-time using 5G to Endeavour Energy's training ground in Hoxton Park. With the demonstration a success, according to Optus, Endeavour Energy will now deploy the solution across infrastructure assets in Penrith and Blacktown, which would remove the need to use a large fleet of vehicles, helicopters, and technicians to physically identify and carry out remediation. "We're thrilled to work with Optus, AWS, and Unleash live, with the support of the Australian government to expedite the use of 5G drone technology to make faster decisions and expedite critical maintenance to continue to keep the lights on for our customers," Endeavour Energy chief asset and operating officer Scott Ryan said.
In this paper, the process of forecasting household energy consumption is studied within the framework of the nonparametric Gaussian Process (GP), using multiple short time series data. As we begin to use smart meter data to paint a clearer picture of residential electricity use, it becomes increasingly apparent that we must also construct a detailed picture and understanding of consumer's complex relationship with gas consumption. Both electricity and gas consumption patterns are highly dependent on various factors, and the intricate interplay of these factors is sophisticated. Moreover, since typical gas consumption data is low granularity with very few time points, naive application of conventional time-series forecasting techniques can lead to severe over-fitting. Given these considerations, we construct a stacked GP method where the predictive posteriors of each GP applied to each task are used in the prior and likelihood of the next level GP. We apply our model to a real-world dataset to forecast energy consumption in Australian households across several states. We compare intuitively appealing results against other commonly used machine learning techniques. Overall, the results indicate that the proposed stacked GP model outperforms other forecasting techniques that we tested, especially when we have a multiple short time-series instances.