Scholarlink

Research shows that heuristic optimization algorithms can find solutions for complex problems in the physical world. Pairing these algorithms with machine learning models for predictive building control applications has been widely explored. This study investigates the efficacy of determining optimal ventilation rates using a particle swarm optimization (PSO) algorithm, a genetic algorithm (GA), and a hybridized genetic particle swarm optimization (GPSO) algorithm developed using MATLAB within an EnergyPlus building energy simulation model. Weather data from Vancouver is used to exemplify a marine climate, where free-cooling opportunities are relatively abundant. The algorithm performance results are collected for both the heating and cooling seasons and are compared against each other for run times, energy savings, and indoor air quality performance. The results are compared against simulation results using a conventional demand control ventilation (DCV) system. Results indicate that when compared to the DCV controller with an economizer mode, heuristic optimization control methods are capable of reducing HVAC energy consumption during a cooling season with free-cooling opportunities by up to 14.1%. Additionally, all three optimization algorithms are capable of minimizing HVAC energy consumption with zero unmet hours for the indoor carbon dioxide concentration setpoint.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluation of Heuristic Optimization Algorithms for Model Predictive Ventilation Control and Reducing Energy Consumption

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ryan Bielenda,
  • Fitsum Tariku

摘要

Research shows that heuristic optimization algorithms can find solutions for complex problems in the physical world. Pairing these algorithms with machine learning models for predictive building control applications has been widely explored. This study investigates the efficacy of determining optimal ventilation rates using a particle swarm optimization (PSO) algorithm, a genetic algorithm (GA), and a hybridized genetic particle swarm optimization (GPSO) algorithm developed using MATLAB within an EnergyPlus building energy simulation model. Weather data from Vancouver is used to exemplify a marine climate, where free-cooling opportunities are relatively abundant. The algorithm performance results are collected for both the heating and cooling seasons and are compared against each other for run times, energy savings, and indoor air quality performance. The results are compared against simulation results using a conventional demand control ventilation (DCV) system. Results indicate that when compared to the DCV controller with an economizer mode, heuristic optimization control methods are capable of reducing HVAC energy consumption during a cooling season with free-cooling opportunities by up to 14.1%. Additionally, all three optimization algorithms are capable of minimizing HVAC energy consumption with zero unmet hours for the indoor carbon dioxide concentration setpoint.