Scholarlink

<p>One of the alternative air conditioning technologies that is emerging as an energy-efficient and eco-friendly solution is the liquid desiccant air conditioning. It generally consists of two major sub-systems, the dehumidifier and the regenerator. The current study utilizes machine learning to study the behavior of the liquid desiccant regenerator under varying conditions. The data is collected from various sources, and input, output features are selected, followed by dividing the data into training and testing sets. The Artificial Neural Network (ANN) model is then developed using the training dataset and validated with the test dataset. As compared to the previous studies where the ANN model was limited to specific regenerator dimensions and operating conditions, the present work aims to develop a more generalized ANN model independent of regenerator size and operational parameters. The model can predict the performance of the system with an R2 of 0.9466. Sensitivity analysis of the model’s input parameters indicated that variations in solution concentration had a comparatively greater effect on the model output. Thereafter, the performance of the liquid desiccant regenerator (LDR) is analyzed by varying the process parameters and is compared with the experimental observations. The process parameters are then optimized using a Genetic Algorithm (GA) for the optimal solutions. For the given range of process parameters, the maximum attainable Moisture Removal Rate (MRR) and Regenerator Effectiveness (RE) are approximately 0.1511&#xa0;g/s, 0.4357, respectively.</p>

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

Investigation of a Liquid Desiccant Regenerator using Machine Learning

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • Mrinal Pradhan,
  • Koushik Das,
  • Rajat Subhra Das

摘要

One of the alternative air conditioning technologies that is emerging as an energy-efficient and eco-friendly solution is the liquid desiccant air conditioning. It generally consists of two major sub-systems, the dehumidifier and the regenerator. The current study utilizes machine learning to study the behavior of the liquid desiccant regenerator under varying conditions. The data is collected from various sources, and input, output features are selected, followed by dividing the data into training and testing sets. The Artificial Neural Network (ANN) model is then developed using the training dataset and validated with the test dataset. As compared to the previous studies where the ANN model was limited to specific regenerator dimensions and operating conditions, the present work aims to develop a more generalized ANN model independent of regenerator size and operational parameters. The model can predict the performance of the system with an R2 of 0.9466. Sensitivity analysis of the model’s input parameters indicated that variations in solution concentration had a comparatively greater effect on the model output. Thereafter, the performance of the liquid desiccant regenerator (LDR) is analyzed by varying the process parameters and is compared with the experimental observations. The process parameters are then optimized using a Genetic Algorithm (GA) for the optimal solutions. For the given range of process parameters, the maximum attainable Moisture Removal Rate (MRR) and Regenerator Effectiveness (RE) are approximately 0.1511 g/s, 0.4357, respectively.