Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article


Doctor of Philosophy


Electrical and Computer Engineering


Srikantha, Pirathayini

2nd Supervisor

McIssac, Ken

Joint Supervisor


If a smart grid is to be described in one word, that word would be ’connectivity’. While electricity production and consumption still depend on a limited number of physical connections, exchanging data is growing enormously. Customers, utilities, sensors, and markets are all different sources of data that are exchanged in a ubiquitous digital setup. To deal with data complexity, many researchers recently focused on machine learning (ML) applications in smart grids. Much of the success in ML is attributed to discriminative learning where models define boundaries to categorize data. Generative learning, however, reveals how data is generated by learning the underlying distribution functions. In the past few years, generative models brought new dimensions to various domains. Computers became painters and composers.

This thesis identifies three applications in the smart grid where generative learning has great potential. On the demand side, residential loads such as dishwashers and clothes driers are simulated using generative models. In specific, the latest developments in generative adversarial networks and kernel density estimators are levered to learn the underlying distributions of individual loads for both power consumption patterns and usage habits. Being data-driven, the learning process eliminates any biases introduced by rule-based models where predetermined fixed formulas describing each load are considered. The study demonstrates the flexibility, viability, and remarkable accuracy of the proposed framework. The resulting synthetic power consumption patterns and usage habits for individual loads are valuable sources for researchers to build or improve their data-driven models for demand-side studies.

Non-intrusive load monitoring (NILM) is the second topic researched on the demand side. The goal in NILM is to identify the status of individual loads in a household by merely relying on a smart meter’s measurements without any hardware installations. The research focuses on identifying the operational condition of individual loads by developing a novel hybrid algorithm that combines the widely used generative technique, namely, hidden Markov model, with k-means clustering. The hybrid model is demonstrated to accurately identify the operation conditions of individual loads based on the ingested aggregate signal.

Finally, for power transmission, a combination of generative models is proposed to estimate power states from a set of redundant measurements. Power state estimation is a fundamental technique in shedding light on the operational condition of the grid. A traditional state estimator is typically executed online and, in its non-linear formulation, involves a high level of computational complexity. Generative models shift that burden to the offline learning process. On the other hand, bad data detection and identification is a central feature in traditional estimators. As such, the developed framework integrated that feature in the data-driven state estimator by incorporating forward and backward generative adversarial networks. Simple domain knowledge is further incorporated in the model to improve its accuracy against the benchmark data-driven model. The proposed framework remarkably detected tampered measurements including false data injection.

Summary for Lay Audience

As per Hegel, ontological categorizations are concepts that a simple person uses to recognize the world [1]. Indeed, classification is a fundamental rational activity that scientists brought to machines in the era of ”machine learning”. Artificial neural networks made this possible through discriminative learning, e.g. machines learn to discriminate dogs from cats. But this is not the end of the story. People’s mental power extends to more than a mere classification of objects. People can extract common features among a set of objects and generate new ideas. While architects can discriminate between good and bad designs, they can learn from these historical records to come up with novel designs. This is why it is widely accepted among researchers that brains learn generative models [2]. To mimic human intelligence, generative learning is considered an essential part of artificial intelligence.

This thesis applies recent developments in generative modelling by tackling three problems in the modern electrical grid. First, a model is developed to simulate the operation of electrical loads in the residential sector. The generative model eventually learns how to behave like a dishwasher, a cloth dryer, and so on. If the model is instructed to simulate a dishwasher, it will generate a signal that looks similar to that generated by a real dishwasher in terms of both signal’s shape and its occurrence in time. The generated synthetic data is a valuable tool that can be used by researchers in further studies such as non-intrusive load monitoring (NILM). This is the second area where generative modelling is applied. In simple words, NILM is the process of identifying the operational status of individual loads inside a house by just reading the measurements recorded by a modern electric meter. The recorded aggregate power consumption for a household is passed to the developed generative model that will, in turn, identify which individual appliances were operating during that time. One of the advantages of NILM is to help consumers to have a better planning of their power consumption and, hence, reduce electricity bills.

Finally, generative techniques are leveraged to assist in quickly monitoring the status of the power grid. Typically, in an electrical grid, measurements (e.g. power values) are taken at various locations along the grid and sent back to a control centre. These measurements are fed to the pre-trained generative model which will instantly provide the operator with a snapshot of the voltage complex values at different buses, i.e. common connections. These values can be used to determine the status of the whole grid. In addition, the generative model can identify any bad measurements that do not seem to be within the acceptable range of error that is usually associated with the measuring process. The developed framework reflects the great benefit that generative models bring to the industry

Available for download on Tuesday, August 01, 2023