Wen-Chiao Lin

A diagnoser algorithm for anomaly detection in DEDS under partial and unreliable observations: characterization and inclusion in sensor configuration optimization

Complex engineering systems have to be carefully monitored to meet demanding performance requirements, including detecting anomalies in their operations. There are two major monitoring challenges for these systems. The first challenge is that information collected from the monitored system is often partial and/or unreliable, in the sense that some occurred events may not be reported and/or may be reported incorrectly (e.g., reported as another event). The second is that anomalies often consist of sequences of event patterns separated in space and time. This paper introduces and analyzes a diagnoser algorithm that meets these challenges for detecting and counting occurrences of anomalies in engineering systems. The proposed diagnoser algorithm assumes that models are available for characterizing plant operations (via stochastic automata) and sensors (via probabilistic mappings) used for reporting partial and unreliable information. Methods for analyzing the effects of model uncertainties on the diagnoser performance are also discussed. In order to select configurations that reduce sensor costs, while satisfying diagnoser performance requirements, a sensor configuration selection algorithm developed in previous work is then extended for the proposed diagnoser algorithm. The proposed algorithms and methods are then applied to a multi-unit-operation system, which is derived from an actual facility application. Results show that the proposed diagnoser algorithm is able to detect and count occurrences of anomalies accurately and that its performance is robust to model uncertainties. Furthermore, the sensor configuration selection algorithm is able to suggest optimal sensor configurations with significantly reduced costs, while still yielding acceptable performance for counting the occurrences of anomalies.

Selecting observation platforms for optimized anomaly detectability under unreliable partial observations

Diagnosers for keeping track on the occurrences of special events in the framework of unreliable partially-observed discrete-event dynamical systems were developed in previous work. This paper considers observation platforms consisting of sensors that provide partial and unreliable observations and of diagnosers that analyze them. Diagnosers in observation platforms typically perform better as sensors providing the observations become more costly or increase in number. This paper proposes a methodology for finding an observation platform that achieves an optimal balance between cost and performance, while satisfying given observability requirements and constraints. Since this problem is generally computational hard in the framework considered, an observation platform optimization algorithm is utilized that uses two greedy heuristics, one myopic and another based on projected performances. These heuristics are sequentially executed in order to find best observation platforms. The developed algorithm is then applied to an observation platform optimization problem for a multi-unit-operation system. Results show that improved observation platforms can be found that may significantly reduce the observation platform cost but still yield acceptable performance for correctly inferring the occurrences of special events.

Sensor configuration selection for discrete-event systems under unreliable observations

Algorithms for counting the occurrences of special events in the framework of partially-observed discrete-event dynamical systems (DEDS) were developed in previous work. Their performances typically become better as the sensors providing the observations become more costly or increase in number. This paper addresses the problem of finding a sensor configuration that achieves an optimal balance between cost and the performance of the special event counting algorithm, while satisfying given observability requirements and constraints. Since this problem is generally computational hard in the framework considered, a sensor optimization algorithm is developed using two greedy heuristics, one myopic and the other based on projected performances of candidate sensors. The two heuristics are sequentially executed in order to find best sensor configurations. The developed algorithm is then applied to a sensor optimization problem for a multi-unit-operation system. Results show that improved sensor configurations can be found that may significantly reduce the sensor configuration cost but still yield acceptable performance for counting the occurrences of special events.

Sequential window diagnoser for discrete-event systems under unreliable observations

This paper addresses the issue of counting the occurrence of special events in the framework of partially-observed discrete-event dynamical systems (DEDS). Developed diagnosers referred to as sequential window diagnosers (SWDs) utilize the stochastic diagnoser probability transition matrices developed in [9] along with a resetting mechanism that allows on-line monitoring of special event occurrences. To illustrate their performance, the SWDs are applied to detect and count the occurrence of special events in a particular DEDS. Results show that SWDs are able to accurately track the number of times special events occur.

Synthesis and optimization of a Bayesian belief network based observation platform for anomaly detection under partial and unreliable observations

Complex engineering systems, such as nuclear processing systems, need to be closely monitored to meet given operational requirements. Previous work has developed diagnosers for detecting and counting occurrences of anomaly patterns (e.g., physical faults, facility misuse) in such systems within discrete event dynamic system (DEDS) framework. This work illustrates the application of this general methodology for the design and optimization of a diagnoser based on Bayesian belief networks (BBNs). Two advantages of this approach are as follows. The first is that current monitoring implementations using BBNs, which is popular in the industry, can be easily expanded and optimized based on the BBN-based diagnosers developed here. The second is that BBN-based diagnosers for tracking anomaly patterns do not require as much computer memory and computation effort as DEDS-based diagnosers. For the BBN-based diagnosers designed here, an optimization problem for finding a sensor configuration that balances sensor cost and diagnoser performance is formulated and solved. Simulation results show that a BBN-based diagnoser performs well in detecting and counting the occurrences of anomalies, while sensor configuration optimization results indicate that improved sensor configurations can be found such that sensor cost is significantly reduced while maintaining acceptable monitoring performance.

Inclusion of game-theoretic formulations for resilient condition assessment monitoring

Monitoring systems collect information from sensors distributed around a monitored plant to assess its health condition. These sensors are prone to be compromised by an attacker. Consequently, two clearly distinct agents exist, namely, a monitoring system and an attacker, both having opposite objectives regarding the accuracy of plant condition assessments. Under this context, a game between these two players arises. This paper considers the inclusion of game-theoretic formulations into resilient condition assessment monitoring (ReCAM) systems. In particular, proposed game calculations periodically identify best sensor networks to be used by the ReCAM system for sensor adaptation based on estimated attacks that an attacker may use. The resulting ReCAM system is then applied to a simplified power plant model and its performance is evaluated via simulations.