A Digital Twin (DT) refers to a digital replica of physical assets, processes and systems. DTs integrate artificial intelligence, machine learning and data analytics to create living digital simulation models that are able to learn and update from multiple sources, and to represent and predict the current and future conditions of physical counterparts. However, the current activities related to DTs are still at an early stage with respect to buildings and other infrastructure assets from an architectural and engineering/construction point of view. Less attention has been paid to the operation & maintenance (O&M) phase, which is the longest time span in the asset life cycle. A systematic and clear architecture verified with practical use cases for constructing a DT would be the foremost step for effective operation and maintenance of buildings and cities. To this end, this paper presents a system architecture for DTs, which is specifically designed at both the building and city levels. Based on current research about multi-tier architectures, this proposed DT architecture enables integration of heterogeneous data sources, supports effective data querying and analysing, supports decision-making processes in O&M management, and further bridges the gap between human relationships with buildings/cities. Based on this architecture, a DT demonstrator of the West Cambridge site of the University of Cambridge was developed. This paper aims at going through the whole

No abstract

Data-driven prognostics solutions underperform under the conditions of limited failure data availability since the number of failure data samples is insufficient for training prognostics models effectively. In order to address this problem, we present a novel methodology for generating real-valued failure data which allows training datasets to be augmented so that the number of failure data samples is increased. In contrast to existing data generation techniques which duplicate or randomly generate data, the proposed methodology is capable of generating new and realistic failure data samples. To this end, we utilised the conditional generative adversarial network and auxiliary information pertaining to the failure modes. The proposed methodology is evaluated in a real-world case study involving the prediction of air purge valve failures in heavy trucks. Two prognostics models are developed using gradient boosting machine and random forest classifiers. It is shown that when these models are trained on the augmented training dataset, they outperform the best prognostics solution previously proposed in the literature for the case study by a large margin. More specifically, costs due to breakdowns and false alarms are reduced by 44%.

When failure data are limited, data-driven prognostics solutions underperform since the number of failure data samples is insufficient for training prognostics models effectively. In order to address this problem, we present a novel methodology for generating failure data which allows training datasets to be augmented so that the number of failure data samples is increased. In contrast to existing data generation techniques which duplicate or randomly generate data, the proposed methodology is capable of generating new and realistic failure data samples. The methodology utilises the conditional generative adversarial network and auxiliary information pertaining to failure modes to control and direct the failure data generation process. The theoretical foundation of the methodology in a non-parametric setting is presented and we show that it holds in practice using empirical results. The methodology is evaluated in a real-world case study involving the prediction of air purge valve failures in heavy-trucks. Two prognostics models are developed using the gradient boosting machine and random forest classifiers. When these models are trained on the augmented training dataset, they outperformed the best solution previously proposed in the literature for the case study by a large margin. More specifically, costs due to breakdowns and false alarms are reduced by 44%. INDEX TERMS Equipment prognostics, expert knowledge, generative modeling, limited failure data, physics of failure.

No abstract