Selection of building thermal insulation materials using robust optimization

Abstract

Life cycle assessment (LCA) is commonly used to analyze the environmental profile of a material and product. For example, the selection of building insulation is crucial to ensure that energy and financial conservation goals can be met. Most of the current methods used in material selection are deterministic in nature, which ignores or neglects the effects caused by uncertainties present in various factors, including how the thermal insulation properties of insulation may change with time. In addition, during the use phase, uncertainty in operating conditions may also affect the performance of the selected material, affecting the overall environmental and economic performance. In practice, when material usage is in large quantities, uncertainty may have significant impacts. This article proposes a novel optimization model-based approach in selecting suitable building insulations using LCA results and economic consideration together with its uncertainties. This work presents an optimization-based approach to select building thermal insulation materials in the presence of data uncertainties. Firstly, we developed a deterministic model, incorporating life cycle assessment and costing to evaluate various environmental impacts and costs of the different materials. Next, by considering the different uncertainties in the data used for the environmental and cost assessments, a robust optimization approach is proposed to derive the second model. A novel solution algorithm is then developed to obtain model solutions efficiently. Computational studies based on a high-rise apartment in Shanghai were performed to test the applicability and performance of the proposed solution. The results demonstrate that the proposed robust optimization model is effective in mitigating the data uncertainties in the material selection optimization and outperforms the deterministic model significantly in terms of probability of achieving cost and environmental impact requirements under uncertainty. In addition, from our case studies, it was identified that parametric uncertainties and payback period exert great influence in decision-making. The work which had expanded the basic deterministic model to include multiple aspects rather than a single objective, and taking into account of parametric uncertainties to mitigate the risks of uncertainties in real applications allows building engineers to select appropriate building material(s) based on both environmental and economic consideration. Furthermore, this work also proposes a novel solution approach to address a wide range of intractable nonlinear optimization problems.