Back to search

University of British Columbia

Control of an integrated solar thermal system based on intelligent iterative learning for hot water demand prediction

Abstract

dc:description

In this thesis, an Iterative Learning (IL) approach to disturbance prediction that uses intelligent iteration grouping is proposed for Economic Model Predictive Control (EMPC), and applied to an Integrated Solar Thermal System (ISTS) in order to improve controller performance. An ISTS consists of a Solar Thermal Collector (STC) which collects energy from the sun, a Thermal Storage Tank (TST) which stores this energy for later use, and an auxiliary Heat Pump (HP) which acts as the actuator for the system, providing additional energy as required. The disturbance in the system is then the user hot water demand. In order to optimize the control performance of an ISTS with EMPC, it is important to be able to accurately predict this hot water demand before it happens. To solve this problem, a novel IL-based approach to disturbance prediction for EMPC is presented. This approach involves separating long-term disturbance data, which in this case is user hot water demand, into a number of 24 hour iterations. These iterations are then further divided into groups using unsupervised learning based on the individual iteration profiles. Following the grouping of iterations, each iteration is given features such as the day of the week it occurs on, and a supervised learning classi fier is trained to map from features to groups in order to predict the group of future iterations. Finally, IL is applied to learn patterns within each group iteratively and predict the actual hot water demand trajectory for future iterations. A simulation of an ISTS using real world hot water demand data then demonstrates the effectiveness of the proposed approach to disturbance prediction, achieving higher performance EMPC than can be attained with existing disturbance prediction methods. Specifically, the EMPC implementation using the IL-based disturbance prediction algorithm is shown to prevent constraint violations within the ISTS more effectively than all other EMPC implementations while decreasing the average daily system cost by over 6%.

Degree

thesis:*
Name thesis:degree_name
Master of Applied Science - MASc
Level thesis:degree_level
master's
Discipline thesis:degree_discipline
Mechanical Engineering
Grantor dc:publisher
University of British Columbia
Year dc:date
2020

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Morrison, Jacob

Rights

dc:rights
Statement dc:rights
  • Attribution-NonCommercial-NoDerivatives 4.0 International
Language dc:language
eng

Identifiers

dc:identifier.*
Handle dc:identifier
http://hdl.handle.net/2429/76028

Chain of custody

source
Harvested from
University of British Columbia
Base URL
circle.library.ubc.ca/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Morrison, Jacob. Control of an integrated solar thermal system based on intelligent iterative learning for hot water demand prediction. master's thesis, University of British Columbia, 2020. http://hdl.handle.net/2429/76028