Electronic Thesis and Dissertation Repository

Thesis Format



Master of Science


Computer Science

Collaborative Specialization

Artificial Intelligence


Lutfiyya, Hanan L.

2nd Supervisor

Haque, Anwar



In the Smart Grid environment, the advent of intelligent measuring devices facilitates monitoring appliance electricity consumption. This data can be used in applying Demand Response (DR) in residential houses through data analytics, and developing data mining techniques. In this research, we introduce a smart system approach that is applied to user's disaggregated power consumption data. This system encourages the users to apply DR by changing their behaviour of using heavier operation modes to lighter modes, and by encouraging users to shift their usages to off-peak hours. First, we apply Cross Correlation to detect times of the occurrences when an appliance is being used. We then use two approaches to recognize the operation mode used: The Dynamic Time Warping (DTW), and Machine Learning using K-Means and K-Nearest Neighbors (KNN).

Summary for Lay Audience

Technology is rapidly evolving. The electricity network is changing. This network is getting smarter with the use of smarter measuring devices which are capable of seamlessly collecting various types of information on consumer electricity consumption. This information is collected frequently by devices that sense through the wired connections then store the electricity consumption data in a storage device that is accessible by the hydro company. Consequently, the hydro company performs analysis on the stored information and then gets back to the consumer with tips and recommendations on reducing electricity bill by urging the consumer to change certain habits in operating house appliances.

In this research, we introduce a smart system approach that is applied to a household's electricity consumption data for each appliance. The main objective of this system is to encourage the consumers to change their behavior with certain appliances by switching using heavier operation modes to lighter operation modes. Also, the system aims to urge users to shift their appliances usage to the times of the day where the demand is lower and electricity is cheaper.

To achieve the goals of our study, we analyzed publicly available electricity consumption data for a group of appliances and came up with a representation of this data. We then built an algorithm to mimic this data with the same form, so that we can generate as much data as we need. Thereafter, we developed an algorithm that can search this data and find the times when an appliance is turned on using a searching technique (cross-correlation) that relies on prior knowledge about each appliance electricity consumption pattern. Using the detected time, we applied two other algorithms to determine the operation mode (light, medium, heavy) for each detected instance of the house appliances. We used an algorithm based on comparison (Dynamic Time Warping), and the other algorithm is based on artificial intelligence (K Nearest Neighbors).

Creative Commons License

Creative Commons Attribution-Noncommercial 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 License.