Deciphering the point names of various equipment within commercial buildings remains a vexing challenge. Both the lack of standards, and the usual idiosyncrasies in naming conventions, add to the complexity faced by engineers as they attempt to analyse and optimise a building’s systems. For example, T, Tmp, Temp, C, and Celsius could all be used to denote a temperature sensor. Likewise, the delimiters (e.g., ., -, or _) that apportion parts of a point name varies considerably. This results in the possibility of both f1-AHU_tmp and Floor02.AHU.Temperature appearing in the same dataset (and often even within the same building). The initial work here explores foundational deep-learning techniques to help a human expert understand previously unseen building point names. We’ve applied a simplified yet effective neural-network architecture called AutoEncoders (AE) to the task, with tests done against an expansive dataset comprising of point names gathered from publicly available Brick and Mortar datasets, as well as a private corpus. This novel implementation is augmented by a unique sequential tokenising method that applies tri-grams in a sliding window over the point names, which preserves both the linearity and delimiters within point names. This is diametrically opposed to prior approaches that have either ignored the character sequences and/or removed the presence of delimiting characters. Early results show promising 91% accuracy levels achieved against previously unseen point names.

Forecasting the power generation from intermittent renewable energy sources, such as Photovoltaic (PV) systems, is crucial for the reliable operations of power systems. In this paper, we consider the task of spatially aggregated PV power generation from large-scale, grid-connected and geographically dispersed PV sites. PV power generation data is highly uncertain, non-linear and non-stationary, making accurate forecasting very challenging. We present a new approach, Wavelet Convolutional Neural Networks (WCNNs), by combining Wavelet Transformation (WT) with Convolutional Neural Networks (CNNs). The WCNNs approach first applies time-invariant WT to decompose the highly fluctuating PV power time series into multiple components. It then predicts the approximation (i.e., low frequency smoothed time series) and details (i.e., high frequency random noise) using CNNs and linear regression, respectively. Extensive evaluation using a real dataset from the Australian Energy Market Operator (AEMO) shows that WCNNs is an effective approach and outperforms the state-of-the-art machine learning models both with and without WT.

The variable nature of solar power output from PhotoVoltaic (PV) systems is the main obstacle for penetration of such power into the electricity grid. Thus, numerous methods have been proposed in the literature to construct forecasting models. In this paper, we present a comprehensive comparison of a set of prominent methods that utilize weather prediction for future. Firstly, we evaluate the prediction accuracy of widely used Neural Network (NN), Support Vector Regression (SVR), k-Nearest Neighbours (kNN), Multiple Linear Regression (MLR), and two persistent methods using four data sets for 2 years. We then analyze the sensitivities of their prediction accuracy to 10–25% possible error in the future weather prediction obtained from the Bureau of Meteorology (BoM). Results demonstrate that ensemble of NNs is the most promising method and achieves substantial improvement in accuracy over other prediction methods.