Transportation of prefabricated modules is a critical process in modular construction and can cause additional stresses which may damage individual modules, leading to additional costs and time during the installation and operational phase if damaged parts are not timely restored or replaced. Therefore, utilizing a tracking and condition monitoring system is useful for modular building manufacturers to detect damaged modules upon arrival to the project job site. The tracking system, developed in this study, consists of a acceleration-sensor-based data acquisition (DAQ) module and a storage module, along with automated data analysis module to assess the structural condition of prefabricated building modules during transportation. The system uses the acceleration Root Mean Square (RMS) parameter and unsupervised clustering algorithms for damage detection. After detailed analysis, using performance measurement techniques: confusion matrix and accuracy score, it was found that the DBSCAN clustering algorithm yielded full accuracy score in the case of more than one level of damage compared to k-means, mean shift, and agglomerative clustering algorithms, which yielded accuracy scores of 0.81, 0.79, and 0.78 respectively. The developed system addresses existing challenges of cost and complexities associated with commercially available hardware systems and can provide evidence to support manufacturers' insurance claims on repair and/or replacement costs. However, the developed system should be tested further on more prefabricated building modules instrumented with a larger number of sensors.

预制模块的运输是模块化结构中的一个关键过程，并且可能会产生额外的应力，这可能会损坏单个模块，如果损坏的部件没有及时修复或更换，则会导致安装和运行阶段的额外成本和时间。因此，利用跟踪和状态监测系统有助于模块化建筑制造商在到达项目工作现场时检测损坏的模块。本研究开发的跟踪系统由基于加速度传感器的数据采集 (DAQ) 模块和存储模块以及自动数据分析模块组成，用于评估预制建筑模块在运输过程中的结构状况。该系统使用加速度均方根 (RMS) 参数和无监督聚类算法进行损伤检测。经过详细分析，使用性能测量技术：混淆矩阵和准确度得分，发现与k-means、均值偏移和凝聚聚类算法相比， DBSCAN聚类算法在多于一级损伤情况下的准确度得分为0.81、0.79 , 和 0.78 分别。开发的系统解决了与商用硬件系统相关的成本和复杂性的现有挑战，并且可以提供证据来支持制造商对维修和/或更换成本的保险索赔。但是，应在配备更多传感器的更多预制建筑模块上进一步测试开发的系统。