Integrated packing and sequence-optimization problems appear in many industrial applications. As an example of this type of problem, we consider the production of glued laminated timber (glulam) in sawmills: Wood beams must be packed into a sequence of pressing steps subject to packing constraints of the press and subject to sequencing constraints. In this paper, we present a three-stage approach for solving this hard optimization problem: Firstly, we identify alternative packings for small parts of an instance. Secondly, we choose an optimal subset of these packings by solving a set cover problem. Finally, we apply a sequencing algorithm in order to find an optimal order of the selected subsequences. For every level of the hierarchy, we present tailored algorithms, analyze their performance and illustrate the efficiency of the overall approach by a comprehensive numerical study.
This research project presents both innovative multi-scalar modelling methods and production processes aimed at facilitating the design and fabrication of free-form glue-laminated timber structures. The paper reports on a research effort that aims to elucidate and formalize the connection between material performance, multi-scalar modelling (Weinan 2011), and early-stage architectural design, in the context of free-form glue-laminated timber structures. This paper will examine how the concept of multi-scalar modelling as found in other disciplines can also be used to embed low-level material performance of glue-laminated timber into early-stage architectural design processes, thus creating opportunities for feedback across the design chain and an increased flexibility in effecting changes. The research uses physical prototypes as a means to explore and evaluate the methods presented.
Mass timber products have shown promise as an innovative alternative to conventional framing systems for use in tall wood buildings, but this new trend in design and construction poses concerns for the long-term durability of the products. A major challenge that classically faces timber products is the threat of moisture-induced mold and decay fungi, which are a heightened concern in mass timber buildings exposed to the environment for extended duration during construction. Consequently, it is important to understand the hygric and thermal (hygrothermal) conditions that mass timber products can experience in multi-story constructions and to be able to quantify the behavior of the products for their suitable design and implementation. An eight-story mass timber building located in Portland, Oregon was chosen for this study and was instrumented for moisture content monitoring through its production, construction, and in-situ use. Record breaking precipitation levels occurred during the building’s construction and while dimension lumber and glulam products subsequently dried to acceptable levels, cross laminated timber products (CLT) dried more slowly. These measurements have an observed bias and the decay risk for the products is inconclusive. Samples of CLT used in the building were characterized for hygrothermal properties and integrated into WUFI, a simulation software, for analysis of the building. The software showed limitations for correctly simulating the behavior of CLT in isolated lab experiments and therefore a re-calibration was performed for accurate simulation. Preliminary on-site simulation results provide a decent approximation of observed data despite its high variance, but drying rate predicted by the program is lower than what was measured.