Ramin Moghaddass

Identifying the location of faults, anomalies, and failures is a long-standing but critical challenge in power system networks. The implementation of smart meters and advanced sensor measurement technologies in recent years has allowed power systems operators to more accurately identify fault locations, quickly resolve problems, and improve the overall reliability of power networks. However, the presence of complex topology changes, penetration of renewable energy resources, stochastic propagation of anomalies in the network, and missing data require the use of new approaches that are sensitive to these issues. This article introduces an innovative method that uses a graph convolutional network (GCN) combined with a modified probability propagation matrix and dual graphs to identify and locate node/line anomalies using network sensors installed on both nodes and lines. Also, an optimization model was developed to find the most likely sources of anomalies across all nodes and edges of the network. The proposed method, which was evaluated on the IEEE 118-bus system and a set of simulated data, demonstrated outstanding performance in handling complex topologies and missing data. Although the proposed model is designed for power networks, its flexible characteristics make it applicable to many sensor-intensive networks or graph structures (e.g., transportation and social networks) where anomaly detection at nodes and/or edges is critical.

This article introduces a deep state‐space modeling (DSSM) framework tailored for monitoring complex and dynamic deteriorating systems operating under varying operating conditions and sensor‐based condition monitoring. By integrating stochastic recurrent neural networks (RNNs) with a generative model in a variational autoencoder (VAE) form, the proposed framework effectively approximates the complex behaviors inherent in degrading systems with latent states and captures long‐term dependencies and latent dynamics without relying on unrealistic distributional and parametric assumptions. The framework leverages RNNs to model temporal dependencies and VAE to model robust probabilistic inference, enabling accurate latent state estimation and time to event (TE) prediction. Moreover, its flexibility extends to accommodating both continuous and discrete latent states, enriching the representation of underlying data dynamics. By performing joint inference and learning, utilizing VAE for system dynamics modeling offers significant advantages over traditional state‐space models (SSMs), which require high computational resources and tuning. We have tested this framework using both simulation and real‐world datasets. Also, a case study on a wind turbine dataset demonstrates the effectiveness of the proposed framework in early fault detection.

To monitor the dynamic behavior of degrading systems over time, a flexible hierarchical discrete-time state-space model (SSM) is introduced that can mathematically characterize the stochastic evolution of the latent states (discrete, continuous, or hybrid) of degrading systems, dynamic measurements collected from condition monitoring sources (e.g., sensors with mixed-type outputs), and the failure process. This flexible SSM is inspired by Bayesian hierarchical modeling and recurrent neural networks without imposing prior knowledge regarding the stochastic structure of the system dynamics and its variables. The temporal behavior of degrading systems and the relationship between variables of the corresponding system dynamics are fully characterized by stochastic neural networks without having to define parametric relationships/distributions between deterministic and stochastic variables. A Bayesian filtering-based learning method is introduced to train the structure of the proposed framework with historical data. Also, the steps to utilize the proposed framework for inference and prediction of the latent states and sensor outputs are discussed. Numerical experiments are provided to demonstrate the application of the proposed framework for degradation system modeling and monitoring.

Many network/graph structures are continuously monitored by various sensors that are placed at a subset of nodes and edges. The multidimensional data collected from these sensors over time create large-scale graph data in which the data points are highly dependent. Monitoring large-scale attributed networks with thousands of nodes and heterogeneous sensor data to detect anomalies and unusual events is a complex and computationally expensive process. This paper introduces a new generic approach inspired by state-space models for network anomaly detection that can utilize the information from the network topology, the node attributes (sensor data), and the anomaly propagation sets in an integrated manner to analyze the entire network all at once. This article presents how heterogeneous network sensor data can be analyzed to locate the sources of anomalies as well as the anomalous regions in a network, which can be impacted by one or multiple anomalies at any time instance. Experimental results demonstrate the superior performance of our proposed framework in detecting anomalies in attributed graphs. Summary of Contribution: With the increasing availability of large-scale network sensors and rapid advances in artificial intelligence methods, fundamentally new analytical tools are needed that can integrate data collected from sensors across the networks for decision making while taking into account the stochastic and topological dependencies between nodes, sensors, and anomalies. This paper develops a framework to intelligently and efficiently analyze complex and highly dependent data collected from disparate sensors across large-scale network/graph structures to detect anomalies and abnormal behavior in real time. Unlike general purpose (often black-box) machine learning models, this paper proposes a unique framework for network/graph structures that incorporates the complexities of networks and interdependencies between network entities and sensors. Because of the multidisciplinary nature of the paper that involves optimization, machine learning, and system monitoring and control, it can help researchers in both operations research and computer science domains to develop new network-specific computing tools and machine learning frameworks to efficiently manage large-scale network data.

Many real systems have a network/graph structure with many connected nodes and many edges representing deterministic or stochastic dependencies and interactions between nodes. Various types of known or unknown anomalies and disturbances may occur across these networks over time. Developing real-time anomaly detection and isolation frameworks is crucial to enable network operators to make more informed and timely decisions and take appropriate maintenance and operations actions. To monitor the health of modern networks in real time, different types of sensors and smart devices are installed across these networks that can track real-time data from a particular node or a section of a network. In this article, we introduce an innovative inference method to calculate the most probable explanation of a set of hidden nodes in heterogeneous attributed networks with a directed acyclic graph structure represented by a Bayesian network, given the values of a set of binary data observed from available sensors, which may be located only at a subset of nodes. The innovative use of Bayesian networks to incorporate parallelization and vectorization makes the proposed framework applicable for large-scale graph structures. The efficiency of the model is shown through a comprehensive set of numerical experiments.

A new data fusion and network analytics framework is proposed that is based on the topology of large-scale networks and the stochastic dependencies between nodes, edges, and sensor data. The framework can transform real-time sensor data collected from disparate sources in a network to detect the location of anomalies and the nodes that are impacted by the detected anomalies. By intelligently fuzing multidimensional sensor data based on the topology of a large-scale network, this article also contributes to big data analytics for network systems. We will show that the proposed framework not only brings computational benefits, but also results in better anomaly estimates leading to lower false alarm rates and higher detection rates.

With the rising complexity of deteriorating systems and availability of advanced sensors, the need for more robust and reliable methods for condition monitoring and dynamic maintenance decision-making has significantly increased. To generate more reliable results for reliability analysis of complex systems, we propose a new generative framework for failure prognosis utilizing a hybrid state-space model (SSM) that represents the evolution of the system's operating condition and its degradation over time. The proposed model can employ a set of real-time sensor measurements to: 1) diagnose the hidden degradation level of the system and 2) predict the likelihood and the uncertainty of failure without imposing unrealistic heavy distributional assumptions. We provide analytical results for the prediction and update steps of the associated particle filter, as well as for the estimation of model parameters. A single-layer feed-forward neural network (Extreme Learning Machine) was used to model the nonparametric relationship between the multi-dimensional observation process and the rest of system's dynamics. We demonstrate the application of our framework through numerical experiments on a set of simulated data and a turbofan engine degradation data set.