An Integrated Prediction and Optimization Model of a Thermal Energy Production System in a Factory Producing Furniture Components

Halil Akbaş,Gültekin Özdemir

doi:10.3390/en13225999

Abstract

Thermal energy is an important input of furniture components production. A thermal energy production system includes complex, non-linear, and changing combustion processes. The main focus of this article is the maximization of thermal energy production considering the inbuilt complexity of the thermal energy production system in a factory producing furniture components. To achieve this target, a data-driven prediction and optimization model to analyze and improve the performance of a thermal energy production system is implemented. The prediction models are constructed with daily data by using supervised machine learning algorithms. Importance analysis is also applied to select a subset of variables for the prediction models. The modeling accuracy of prediction algorithms is measured with statistical indicators. The most accurate prediction result was obtained using an artificial neural network model for thermal energy production. The integrated prediction and optimization model is designed with artificial neural network and particle swarm optimization models. Both controllable and uncontrollable variables were used as the inputs of the maximization model of thermal energy production. Thermal energy production is increased by 4.24% with respect to the optimal values of controllable variables determined by the integrated optimization model.

Highlights

The world’s fossil fuel reserves have gradually been depleted with the growth in industry, transportation, population, and urban areas
The results obtained by this article prove that thermal energy production system (TEPS), which involves complex operations with chemical and physical reactions, can be modeled and optimized with very high prediction accuracy, chemical and physical reactions, can be modeled and optimized with very high prediction accuracy, low complexity, and short computation time by using the integrated artificial neural network (ANN)-particle swarm optimization (PSO) model
The solution the single-objective optimization model has resulted in a 4.24% increase in the thermal energy production (TEP) under optimized of the single-objective optimization model has resulted in a 4.24% increase in the TEP under operational conditions

Summary

Introduction

The world’s fossil fuel reserves have gradually been depleted with the growth in industry, transportation, population, and urban areas. The environmental impacts of high fossil fuel consumption have increased considerably. Biomass is a suitable renewable energy resource that can substitute for fossil fuels to restrain greenhouse gas emissions [1,2,3]. Biomass resources can be converted to a more useful energy form with thermal processes called combustion, gasification, and pyrolysis [4]. Combustion is the burning of biomass in the air to produce hot gases and ash. Stored chemical energy is converted into heat, which is transformed to kinetic energy through heating water to produce vapor used for gas engines [5]. Great diversity in biomass resources creates the need for adaptable combustion technology in order to achieve efficient and clean biomass combustion [6]

Objectives

Methods

Results

Conclusion