Ian Bennett

Multidimensional prognostics for rotating machinery: A review

Determining prognosis for rotating machinery could potentially reduce maintenance costs and improve safety and availability. Complex rotating machines are usually equipped with multiple sensors, which enable the development of multidimensional prognostic models. By considering the possible synergy among different sensor signals, multivariate models may provide more accurate prognosis than those using single-source information. Consequently, numerous research papers focusing on the theoretical considerations and practical implementations of multivariate prognostic models have been published in the last decade. However, only a limited number of review papers have been written on the subject. This article focuses on multidimensional prognostic models that have been applied to predict the failures of rotating machinery with multiple sensors. The theory and basic functioning of these techniques, their relative merits and drawbacks and how these models have been used to predict the remnant life of a machine are discussed in detail. Furthermore, this article summarizes the rotating machines to which these models have been applied and discusses future research challenges. The authors also provide seven evaluation criteria that can be used to compare the reviewed techniques. By reviewing the models reported in the literature, this article provides a guide for researchers considering prognosis options for multi-sensor rotating equipment.

An Assessment of the Emissions and Global Warming Potential of Gas Turbines for LNG Applications

This paper concentrates on the emissions module of what is part of a wider project dealing with various aspects of gas turbine usage as drivers for Liquefied Natural Gas (LNG) production. The framework is known as TERA, a Techno-Economic and Environmental Risk Analysis, developed at Cranfield University with the core of the study being the performance module whilst the risk, economics and environmental modules are built around the performance. Whilst TERA exists for aviation and power production no such system is available for assessment of LNG production. With environmental issues high on the public agenda new legislation on emissions can be expected, especially in Europe. This will mean Oil & Gas companies will have to look for ways to reduce their emissions. One way to reduce turbo machinery losses is to replace out dated and/or obsolete machinery having less overall energy efficiency. The selection of turbomachinery involves assessments of risk, both economic and technical, as well as environmental impacts of the new technology. The core to all of this is the performance assessment, the primary basis on which selection is made. An aviation emissions model, developed at Cranfield University, is adapted for industrial applications. Technical performance calculations are made using the inhouse software called Turbomatch. Performance results for three typical days of the year (summer, winter and spring/autumn) are fed into the emissions model to get the levels of NOx, CO2, H2O, CO and unburnt hydrocarbon emissions. Later, NOx, CO2 and H2O emissions levels are fed into the environmental module to estimate the damage the engine causes to the environment over 100 years with respect to global warming. Two hypothetical engine configurations are investigated based on engine data available in the public domain. The first one, an 85MW single spool industrial machine (SSI-85), is used as the baseline to compare against a 100MW triple spool, intercooled aeroderivative (ITSA-100). The results suggest that the ITSA-100 produces more NOx but has less carbon emissions and consequently less global warming effects. This has varied economic impacts depending on which emission is a priority for reduction. CO2 and H2O emissions are more important than NOx for LNG gas turbine applications. The paper shows how this simple but effective system can be utilised to give a viable comparison between one or more proposed solutions for turbomachinery selection and replacement. The scope of the system is expanded as other modules come together to give a total assessment in terms of technical, economic, environmental and risk perspectives for LNG production.

Canonical Variable Analysis for Fault Detection, System Identification and Performance Estimation

Condition monitoring of industrial processes can minimize downtime and maintenance costs while enhancing the safety of operation of plants and increasing the quality of products. Multivariate statistical methods are widely used for condition monitoring in industrial plants due to the rapid growth and advancement in data acquisition technology. However, the effectiveness of these methodologies in real industrial processes has not been fully investigated. This paper proposes a CVA-based approach for process fault identification, system modeling and performance estimation. The effectiveness of the proposed method was tested using data acquired from an operational industrial centrifugal compressor. The results indicate that CVA can be effectively used to identify abnormal operating conditions and predict performance degradation after the appearance of faults.

Rotating Machine Prognostics Using System-Level Models

The prognostics of rotating machines is crucial for the reliable and safe operation as well as maximizing usage time. Many reliability studies focus on component-level prognostics. However, in many cases, the desired information is the residual life of the system, rather than the lifetimes of its constituent components. This review paper focuses on system-level prognostic techniques that can be applied to rotating machinery. These approaches use multi-dimensional condition monitoring data collected from different parts of the system of interest to predict the remaining useful life at the system level. The working principles, merits and drawbacks as well as field of applications of these techniques are summarized.

Addressing Missing Data for Diagnostic and Prognostic Purposes

One of the major targets in industry is minimising the downtime of a machine while maximising its availability, with maintenance considered as a key aspect towards achieving this objective. Condition based maintenance and prognostics and health management, which relies on the concepts of diagnostics and prognostics, is a policy that has been gaining ground over several years. The successful implementation of this methodology is heavily dependent on the quality of data used which can be undermined in scenarios where there is missing data. This issue may compromise the information contained within a data set, thus having a significant effect on the conclusions that can be drawn, hence it is important to find suitable techniques to address this matter. To date a number of methods to recover such data, called imputation techniques, have been proposed. This paper reviews the most widely used methodologies and presents a case study using actual industrial centrifugal compressor data, in order to identify the most suitable technique.

Dealing with missing data as it pertains of e-maintenance

Purpose Centrifugal compressors are integral components in oil industry, thus effective maintenance is required. Condition-based maintenance and prognostics and health management (CBM/PHM) have been gaining popularity. CBM/PHM can also be performed remotely leading to e-maintenance. Its success depends on the quality of the data used for analysis and decision making. A major issue associated with it is the missing data. Their presence may compromise the information within a set, causing bias or misleading results. Addressing this matter is crucial. The purpose of this paper is to review and compare the most widely used imputation techniques in a case study using condition monitoring measurements from an operational industrial centrifugal compressor. Design/methodology/approach Brief overview and comparison of most widely used imputation techniques using a complete set with artificial missing values. They were tested regarding the effects of the amount, the location within the set and the variable containing the missing values. Findings Univariate and multivariate imputation techniques were compared, with the latter offering the smallest error levels. They seemed unaffected by the amount or location of the missing data although they were affected by the variable containing them. Research limitations/implications During the analysis, it was assumed that at any time only one variable contained missing data. Further research is still required to address this point. Originality/value This study can serve as a guide for selecting the appropriate imputation method for missing values in centrifugal compressor condition monitoring data.

Reciprocating Compressor Prognostics

Reciprocating compressors are vital components in oil and gas industry though their maintenance cost can be high. The valves are considered the most frequent failing part accounting for almost half the maintenance cost. Condition Based Maintenance and Prognostics and Health Management which is based on diagnostics and prognostics principles can assist towards reducing cost and downtime while increasing safety and availability by offering a proactive means for scheduling maintenance. Although diagnostics is an established field for reciprocating compressors, there is limited information regarding prognostics, particularly given the nature of failures can be instantaneous. This paper compares several prognostics methods (multiple liner regression, polynomial regression, K-Nearest Neighbours Regression (KNNR)) using valve failure data from an operating industrial compressor. Moreover, a variation about Remaining Useful Life (RUL) estimation process based on KNNR is proposed along with an ensemble method combining the output of all aforementioned algorithms. In conclusion it is showed that even when RUL is relatively short given the instantaneous failure mode, good estimates are feasible using the proposed methods.

Dealing with missing data for prognostic purposes

Centrifugal compressors are considered one of the most critical components in oil industry, making the minimization of their downtime and the maximization of their availability a major target. Maintenance is thought to be a key aspect towards achieving this goal, leading to various maintenance schemes being proposed over the years. Condition based maintenance and prognostics and health management (CBM/PHM), which is relying on the concepts of diagnostics and prognostics, has been gaining ground over the last years due to its ability of being able to plan the maintenance schedule in advance. The successful application of this policy is heavily dependent on the quality of data used and a major issue affecting it, is that of missing data. Missing datas presence may compromise the information contained within a set, thus having a significant effect on the conclusions that can be drawn from the data, as there might be bias or misleading results. Consequently, it is important to address this matter. A number of methodologies to recover the data, called imputation techniques, have been proposed. This paper reviews the most widely used techniques and presents a case study with the use of actual industrial centrifugal compressor data, in order to identify the most suitable ones.