Energy modeling with meteorological data and multiobjective optimization of a confectionery stove

Abstract Confectionery stoving is the process used to produce soft-jelled candies. The stoving process has many steps, but the most energy-intensive step is heating and drying the candies, which can consume up to 84% of the energy at a factory. Drying is achieved by heating and ventilating the candy stove with large volumes of outside air, so the impact of the local climate on confectionery facilities can be significant. However, there is a lack of literature investigating local climate impact on the stoving process. This work demonstrates how Multiobjective Optimization (MOO) can be used to find a balanced solution between first cost and energy usage. Furthermore, the work investigates local climate impact to study whether strategic facility siting in a specific climate zone can significantly reduce energy needs and installation costs. The paper proposes a generic energy model that uses actual meteorological data and a product-driven recipe for calculating the energy requirement of a confectionery stove. Four distinct climate zones were analyzed, and it was found that operating a stove in a dry, hot climate, such as Las Vegas, requires 15% less energy than a hot-humid climate, such as Houston. The energy model incorporates a Multiobjective Evolutionary Algorithm (MOEA) procedure to minimize energy usage and initial cost. The algorithm varies the airflow rate and percentage of outdoor air used in the stove. A Pareto Front search procedure is defined, which compares all solutions against a “base case” solution. The base case can be the minimum of any objective, but for this work, it is the lowest energy consuming solution. The search procedure found a preferred solution where a small increase in energy consumption, 0.2%, reduces first cost by 12%–45%. The methods outlined in this paper apply to any energy-consuming manufacturing process that is influenced by climate or any such process that has a broad set of functional operating parameters.