Sustainability of natural gas transmission infrastructure is highly related to the system’s ability to decrease emissions due to ruptures or leaks. Although traditionally such detection relies in alarm management system and operator’s expertise, given the system’s nature as large-scale, complex, and with vast amount of information available, such alarm generation is better suited for a fault detection system based on data-driven techniques. This would allow operators and engineers to have a better framework to address the online data being gathered. This paper presents an assessment on multiple fault-case scenarios in critical infrastructure using two different data-driven based fault detection algorithms: Principal component analysis (PCA) and its dynamic variation (DPCA). Both strategies are assessed under fault scenarios related to natural gas transmission systems including pipeline leakage due to structural failure and flow interruption due to emergency valve shut down. Performance evaluation of fault detection algorithms is carried out based on false alarm rate, detection time and misdetection rate. The development of modern alarm management frameworks would have a significant contribution in natural gas transmission systems’ safety, reliability and sustainability.

No abstract

The implementation of fault detection techniques in industrial systems for process monitoring has proven to be a useful tool to process operators supervising the plant’s operation conditions. As plants become more instrumented, more data is available for fault detection applications, if they are capable of demonstrate anticipation and low false alarm rates. A regional Natural Gas transportation system deals with these types of drawbacks. While the improvements are carried out, some effort should be done in order to improve the safety in operations. In this paper a data-driven technique was used to detect fault conditions along the pipeline, sectioning it in five partitions to increase the detection sensibility. To overcome the lack of quality in data, simulation software intended to gas controllers training and pipeline operation was used to simulate leaks scenarios. Some historic data with high quality is also used to create normal operation condition models by means of Principal Component Analysis. All simulated faults were detected in a reduced time gap and recent events related to third-party actions showed the tool proficiency to detecting faults in real time. In addition, it considers a fault normalized index per section indicating the fault persistence and aggressiveness in a single plot.

Fault detection and diagnosis schemes based on data-driven statistical modelling are highly dependent on an accurate and exhaustive feature extraction procedure to deliver a superior performance as a monitoring strategy. Otherwise conducted, a deficient feature extraction procedure leads to a monitoring structure widely deviated from normal operating conditions. If an operating state is not identified as it, an increment in false alarm rate would be evidenced whenever the process shifts towards that condition and the monitoring scheme triggers the abnormal condition warning. On the other hand, if two similar operating conditions could not be individualized i.e. to be identified as a single operating state, a lack of sensitivity for minor — yet typical — deviations would render a monitoring strategy with prominent misdetection rates. Although Multimode Operational Mapping requires the proper identification of a finite set of normal process states, it is a challenging task especially for large-scale systems. Its complexity derives from a broad universe of monitoring variables, highly interactuating process units integrated over very dynamic network systems, among others. This is the case of natural gas transmission infrastructure, as it deals with variable upstream production rates, diverse consumption trends from customers, internal processes constrains, merged in a stringent operating scheme. This paper proposes a novel strategy to address the identification and feature extraction of normal conditions on multimode operation systems. The proposed framework uses a segmentation approach based on operator’s knowledge, the Takagi-Sugeno-Kang fuzzy engine and k-means algorithm to characterize the normal operation states of the system. The results show an improvement in the performance of Principal Component Analysis during abnormal conditions detection, in addition an increase on the sensitivity of Hotelling and Q statistics.

No abstract