The core of many typical large-scale industrial infrastructures consists of hundreds or thousands of systems that are similar in their basic design and purpose. For instance, District Heating (DH) utilities rely on a large network of substations to deliver heat to their customers. Similarly, a factory may require a large fleet of specialized robots for manufacturing a certain product. Monitoring these systems is important for maintaining the overall efficiency of industrial operations by detecting various problems due to faults and misconfiguration. However, this can be challenging since a well-understood prior model for each system is rarely available.

In most cases, each system in a fleet or network is fitted with a set of sensors to measure its state at different time intervals. Typically, a data-driven model for each system can be used for their monitoring. However, not all factors that can influence the operation of each system in a fleet have an associated sensor. Moreover, sufficient data instances of normal, atypical, and faulty behavior are rarely available to train such a model. These issues can impede the effectiveness of a system-level data-driven model. Alternatively, it can be assumed that since all the systems in a fleet are working on a similar task, they should all behave in a homogeneous manner. Any system that behaves differently from the majority is then considered an outlier. It is referred to as a global or fleet-level model. While the approach is simple, it is less effective in the presence of non-stationary working conditions. Hence, both system-level and fleet-level modeling approaches have their limitations.

This thesis investigates system-level and fleet-level models for large-scale monitoring of systems. It proposes to rely on an alternative way, referred to as a reference-group based approach. Herein, the operational monitoring of a target system is delegated to a reference-group, which consists of systems experiencing a comparable operating regime along with the target system. Thus, the definition of a normal, atypical, or faulty operational behavior in a target system is described relative to its reference-group. This definition depends on the choice of the selected anomaly detection model. In this sense, if the target system is not behaving operationally in consort with the systems in its reference-group, then it can be inferred that this is either due to a fault or because of some atypical operation arising at the target system due to its local peculiarities. The application area for these investigations is the large-scale operational monitoring of thermal energy systems: network of DH substations and fleet of heat-pumps.

The current findings indicate three advantages of a reference-group based approach. The first is that the reference operational behavior of a target system in the fleet does not need to be predefined. The second is that it provides a basis for what a target system’s operational behavior should have been and what it is. In this respect, each system in the reference-group provides evidence about a particular behavior during a particular period. It can be very useful when the description of a normal, atypical, and faulty operational behavior is not available. The third is that it can detect atypical and faulty operational behavior quickly compared to fleet-level models of anomaly detection.

Increasing technical complexity, design variations, and customization options of IoT units create difficulties for the construction of monitoring infrastructure. These units can be associated with different domains, such as a fleet of vehicles in the mobility domain and a fleet of heat-pumps in the heating domain. The lack of labeled datasets and well-understood prior unit and fleet behavior models exacerbates the problem. Moreover, the time-series nature of the data makes it difficult to strike a reasonable balance between precision and detection delay. The thesis aims to develop a framework for scalable and cost-efficient monitoring of industrial fleets. The investigations were conducted on real-world operational data obtained from District Heating (DH) substations to detect anomalous behavior and faults. A foundational hypothesis of the thesis is that fleet-level models can mitigate the lack of labeled datasets, improve anomaly detection performance, and achieve a scalable monitoring alternative.

Our preliminary investigations found that operational heterogeneity among the substations in a DH network can cause fleet-level models to be inefficient in detecting anomalous behavior at the target units. An alternative is to rely on subfleet-level models to act as a proxy for the behavior of target units. However, the main difficulty in constructing a subfleet-level model is the selection of its members such that their behavior is stable over time and representative of the target unit. Therefore, we investigated various ways of constructing the subfleets and estimating their stability. To mitigate the lack of well-understood prior unit and fleet behavior models, we proposed constructing Unit-Level and Subfleet-Level Ensemble Models, i.e., ULEM and SLEM. Herein, each member of the respective ensemble consists of a Conformal Anomaly Detector (CAD). Each ensemble yields a nonconformity score matrix that provides information about the behavior of a target unit relative to its historical data and its subfleet, respectively. However, these ensemble models can give different information about the nature of an anomaly that may not always agree with each other. Therefore, we further synthesized this information by proposing a Combined Ensemble Model (CEM). We investigated the advantages and limitations of decisions that rely on the information obtained from ULEM, SLEM, and CEM using precision and detection delay. We observed the decisions that relied on the information obtained through CEM showed a reduction in overall false alarms compared to those obtained through ULEM or SLEM, albeit at the cost of some detection delay. Finally, we combined the components of ULEM, SLEM, and CEM into what we refer to as TRANTOR: a conformal anomaly detection based indusTRiAl fleet moNiTORing framework. The proposed framework is expected to enable fleet operators in various domains to improve their monitoring infrastructure by efficiently detecting anomalous behavior and controlling false alarms at the target units.