Prognostics and Health Management of Choke Valves Subject to Erosion

Oil and gas industries are constantly aiming at improving the efficiency of their operations. In this respect, focus is on the development of technology, methods, and work processes related to equipment condition and performance monitoring in order to achieve the highest standards in terms of safety and productivity. To this aim, a key issue is represented by maintenance optimization of critical structures, systems, and components. A way towards this goal is offered by Condition-Based Maintenance (CBM) strategies. CBM aims at regulating maintenance scheduling based on data analyses and system condition monitoring and bears the potential advantage of obtaining relevant cost savings and improved operational safety and availability. A critical aspect of CBM is its integration with condition monitoring technologies for handling a wide range of information sources and eventually making optimal decisions on when and what to repair. In this chapter, a CBM case study concerning choke valves utilized in Norwegian offshore oil and gas platforms is proposed and investigated. The objective is to define a procedure for optimizing maintenance of choke valves by on-line monitoring their condition and determining their Remaining Useful Life (RUL). Choke valves undergo erosion caused by sand grains transported by the oil-water-gas mixture extracted from the well. Erosion is a critical problem which can affect the correct valve functioning, resulting in revenue losses and cause environmental hazards.

A Particle Filtering Based Approach for Gear Prognostics

In this chapter, a particle filtering based gear prognostics method using a one-dimensional health index for spiral bevel gear subject to pitting failure mode is presented. The presented method effectively addresses the issues in applying particle filtering to mechanical component remaining useful life (RUL) prognostics by integrating a couple of new components into particle filtering: (1) data mining based techniques to effectively define the degradation state transition and measurement functions using a one-dimensional health index obtained by a whitening transform; and (2) an unbiased l-step ahead RUL estimator updated with measurement errors. The presented prognostics method is validated using data from a spiral bevel gear case study.

Modeling Multi-State Equipment Degradation with Non-Homogeneous Continuous-Time Hidden Semi-Markov Process

The multi-state reliability analysis has received great attention recently in the domain of reliability and maintenance, specifically for mechanical equipment operating under stress, load, and fatigue conditions. The overall performance of this type of mechanical equipment deteriorates over time, which may result in multi-state health conditions. This deterioration can be represented by a continuous-time degradation process with multiple discrete states. In reality, due to technical problems, directly observing the actual health condition of the equipment may not be possible. In such cases, condition monitoring information may be useful to estimate the actual health condition of the equipment. In this chapter, the authors describe the application of a general stochastic process to multi-state equipment modeling. Also, an unsupervised learning method is presented to estimate the parameters of this stochastic model from condition monitoring data.

Principles of Classification

This chapter discusses the main principles of the creation and use of a classifier in order to predict the interpretation of an unknown data sample. Classification offers the possibility to learn and use learned information received from previous occurrences of various normal and fault modes. This process is continuous and can be generalized to cover the diagnostics of all objects that are substantially of the same type. The effective use of a classifier includes initial training with known data samples, anomaly detection, retraining, and fault detection. With these elements an automated, a continuous learning machine diagnostics system can be developed. The main objective of such a system is to automate various time intensive tasks and allow more time for an expert to interpret unknown anomalies. A secondary objective is to utilize the data collected from previous fault modes to predict the re-occurrence of these faults in a substantially similar machine. It is important to understand the behaviour and functioning of a classifier in the development of software solutions for automated diagnostic methods. Several proven methods that can be used, for instance in software development, are disclosed in this chapter.

Combining Health Monitoring and Control

This chapter proposes the combination of system health monitoring with control and prognosis creating a new paradigm, the health-aware control (HAC) of systems. In this paradigm, the information provided by the prognosis module about the component system health should allow the modification of the controller such that the control objectives will consider the system’s health. In this way, the control actions will be generated to fulfill the control objectives, and, at the same time, to extend the life of the system components. HAC control, contrarily to fault-tolerant control (FTC), adjusts the controller even when the system is still in a non-faulty situation. The prognosis module, with the main feature system characteristics provided by condition monitoring, will estimate on-line the component aging for the specific operating conditions. In the non-faulty situation, the control efforts are distributed to the system based on the proposed health indicator. An example is used throughout the chapter to illustrate the ideas and concepts introduced.

A Temporal Probabilistic Approach for Continuous Tool Condition Monitoring

In this chapter, a temporal probabilistic approach based on hidden semi-Markov model is proposed for continuous (real-valued) tool condition monitoring in machinery systems. As an illustrative example, tool wear prediction in CNC-milling machine is conducted using the proposed approach. Results indicate that the additional flexibility provided in the new approach compared to the existing hidden Markov model-based approach improves the performance. 482 features are extracted from 7 signals (three force signals, three vibration signals and acoustic emission) that are acquired for each experiment. After the feature extraction phase, Fisher’s discriminant ratio is applied to find the most discriminant features to construct the prediction model. The prediction results are provided for three different cases, i.e. cross-validation, diagnostics, and prognostics. The possibility of incorporating an asymmetric loss function in the proposed approach in order to reflect and consider the cost differences between an under- and over-estimation in tool condition monitoring is also explored and the simulation results are provided.

Degradation Based Condition Classification and Prediction in Rotating Machinery Prognostics

A transition stage exists during the equipment degradation, which is between the normal condition and the failure condition. The transition stage presents small changes and may not cause significant function loss. However, the transition stage contains the degradation information of the equipment, which is beneficial for the condition classification and prediction in prognostics. The degradation based condition classification and prediction of rotating machinery are studied in this chapter. The normal, abnormal, and failure conditions are defined through anomaly determination of the transition stage. The condition classification methods are analyzed with the degradation conditions. Then the probability of failure occurrence is discussed in the transition stage. Finally, considering the degradation processes in rotating machinery, the condition classification and prediction are carried out with the field data.

Data Driven Prognostics for Rotating Machinery

A prognostic is an estimate of the remaining useful life of a monitored part. While diagnostics alone can support condition based maintenance practices, prognostics facilitates changes to logistics which can greatly reduce cost or increase readiness and availability. A successful prognostic requires four processes: 1) feature extraction of measured data to estimate damage; 2) a threshold for the feature, which, when exceeded, indicates that it is appropriate to perform maintenance; 3) given a future load profile, a model that can estimate the remaining useful life of the component based on the current damage state; and 4) an estimate of the confidence in the prognostic. This chapter outlines a process for data-driven prognostics by: describing appropriate condition indicators (CIs) for gear fault detection; threshold setting for those CIs through fusion into a component health indicator (HI); using a state space process to estimate the remaining useful life given the current component health; and a state estimate to quantify the confidence in the estimate of the remaining useful life.

Fatigue Damage Prognostics and Life Prediction with Dynamic Response Reconstruction Using Indirect Sensor Measurements

This study presents a general methodology for fatigue damage prognostics and life prediction integrating the structural health monitoring system. A new method for structure response reconstruction of critical locations using measurements from remote sensors is developed. The method is based on the empirical mode decomposition with intermittency criteria and transformation equations derived from finite element modeling. Dynamic responses measured from usage monitoring system or sensors at available locations are decomposed into modal responses directly in time domain. Transformation equations based on finite element modeling are used to extrapolate the modal responses from the measured locations to critical locations where direct sensor measurements are not available. The mode superposition method is employed to obtain dynamic responses at critical locations for fatigue crack propagation analysis. Fatigue analysis and life prediction can be performed given reconstructed responses at the critical location. The method is demonstrated using a multi degree-of-freedom cantilever beam problem.

Structure Reliability and Response Prognostics under Uncertainty Using Bayesian Analysis and Analytical Approximations

This study presents an efficient method for system reliability and response prognostics based on Bayesian analysis and analytical approximations. Uncertainties are explicitly included using probabilistic modeling. Usage and health monitoring information is used to perform the Bayesian updating. To improve the computational efficiency, an analytical computation procedure is proposed and formulated to avoid time-consuming simulations in classical methods. Two realistic problems are presented for demonstrations. One is a composite beam reliability analysis, and the other is the structural frame dynamic property estimation with sensor measurement data. The overall efficiency and accuracy of the proposed method is compared with the traditional simulation-based method.