Predicting Failures from Oilfield Sensor Data using Time Series Shapelets

Abstract

Increasing instrumentation of the modem digital oilfield produces streams of data from sensors that monitor the functioning of different components in the field. This data should be converted to actionable information rapidly in order to respond to events as they happen or are predicted. The challenge is therefore to develop technologies that can process these large sensor datasets rapidly and with minimal manual supervision to ensure a data processing system that can scale with the increasing instrumentation.We consider as a use-case an oilfield with several Electrical Submersible Pumps (ESPs), each instrumented with sensors that continually measure electrical properties of the pump (the streams of sensor data), which are then relayed to a central location. In this paper, we demonstrate how a time-series analysis approach can be applied to failure detection and failure prediction from the streams of sensor data. The method involves identifying "shapelets" -short instances that are particularly distinct -in the streams of sensor data.The shapelets approach is particularly applicable to large oil and gas enterprise datasets because the algorithm does not need access to the entire historical data. This greatly reduces the amount of data that needs to be stored for data analysis. Moreover, unlike model-based approaches, shapelet-based analysis does not make any assumptions about the underlying nature of the data, making it practical for applications where a detailed physical model of the pump is not available.We validate our proposed method by analysis on a representative set of instrumented ESPs. We describe the preprocessing steps that were applied in our analysis. We report the results of experiments to study the effects of varying the data processing parameters on the accuracy of fault detection and prediction. These results indicate that shapelet-based approaches are promising for analysis of time-series data in the oil and gas industry.

Digital transformation: a review on artificial intelligence techniques in drilling and production applications

“…Liu et al [51] applied decision trees, support vector machines, and Bayesian networks for understanding sucker rod pump production wells. Patri et al [55] used decision trees and time series shapelets to monitor the behavior of electrical submersible pumps (ESPs). Pennel et al [63] used random forest for studying rod pump artificial lift systems.…”

“…Patri et al [55] present an approach using time series analysis for the detection and prediction of failures, in real time, from data acquired from sensors associated with submersible electric pumps, used for artificial elevation in oil fields. The method developed by the authors aims to identify shapelets -short instances of data -representative of normal behavior or failure behavior in the data sequence produced by the sensors of a given electric pump.…”

“…Failure detection and prediction in artificial lift systems Time series classifier algorithm for oil well pumps failure prediction Liu et al [51] Decision tree classifier, associated with time series shapelets, for electrical submerse pumps failure detection and prediction Patri et al [55] Random forest and cross-validation model for prediction failures in artificial lift system (rod pumps and gas lift) in 800 oil wells Pennel et al [63] Predictive maintenance of equipment Support vector machine (SVM) classifier for warning predictive maintenance in compressors, to avoid production deferments…”

Reservoir Classification Using Well-Testing Pressure Derivative Data

“…The name of this technique is Shapelet. It has been successfully applied in the literature (Xing et al, 2011, Lines et al, 2012, Patri et al, 2014, Hameurlain et al, 2017, Ahmadi et al, 2017b.…”

“…Xing et al in 2011 did early classification in time series data using shapelets (Xing et al, 2011). In 2012, Lines et al proposed to create a new data space transforming data by calculating the distances from a time series to each shapelet and then use this to perform the clas- (Patri et al, 2014, Patri et al, 2016. In 2017, Ahmadi et al applied shapelets to perform well-testing model identification from pressure derivative plots (Ahmadi et al, 2017b).…”