Comparison and Evaluation of Model-Based Optimal Sensor Placement Implementation: A Case Study of a Hybrid Cross-Laminated Timber and Steel Building

The construction industry has reshaped timber building practice into the integration of timber structures with other materials such as steel, termed hybrid timber-based structures, as an attractive, sustainable, and aesthetic solution. However, novel hybrid timber-based structures may require critical validation of design assumptions and structural dynamic performance. This paper conducts a study of model-based optimal sensor placement (OSP) implementation in which two OSP methods and six evaluation criteria are compared and evaluated to ensure the most significant dynamic information content by analyzing two comparison metrics for a resulting sensor configuration in the dynamic identification of a hybrid cross-laminated timber (CLT)-steel building with structural irregularities. The Engineering Design and Innovation building at The Pennsylvania State University is a rectangular 4-story hybrid building with moment-resisting steel frames and composite floors of reinforced concrete layer and CLT panels. For this case study, the OSP problem is a sensitive factor, motivated by the potential torsional effects of structural irregularities from the lateral force-resisting system during serviceability, in the capture of the most significant information content and identification of dynamic properties by a cost-effective structural health monitoring (SHM) strategy. The proposed investigation compares and examines the backward sequential sensor placement (BSSP) and genetic algorithm (GA) methods using different evaluation criteria to capture and maximize the most relevant dynamic information, evaluated by a Pareto analysis constructed based on the modal assurance criterion (MAC) trace and Fisher information matrix (FIM) determinant metrics simultaneously, for the numerical mode shape identification. The model-based OSP implementation plays a critical role in SHM applications, simplifying data collection and improving dynamic information quality for dynamic identification and damage detection in order to better comprehend the structural health condition and performance of existing civil infrastructure such as the hybrid mass timber-steel building presented.