Innovative Approach to Build a Scalable Hybrid Model for Prescriptive Maintenance

Objective and Scope Analysis of operational plant data needs experts in order to interpret detected anomalies which are defined as unusual operation points. The next step on the digital transformation journey is to provide actionable insights into the data. Prescriptive Maintenance defines in advance which kind of detailed maintenance and spare parts will be required. This paper details requirements to improve these predictions for rotating equipment and show potential to integrate the outcome into an operational workflow. Methods, Procedures, Process First principle or physics-based modelling provides additional insights into the data, since the results are directly interpretable. However, such approaches are typically assumed to be expensive to build and not scalable. Identification of and focus on the relevant equipment to be modeled in a hybrid model using a combination of first principle physics and machine learning is a successful strategy. The model is trained using a machine learning approach with historic or current real plant data, to predict conditions which have not occurred before. The better the Artificial Intelligence is trained, the better the prediction will be. Results, Observations, Conclusions The general aim when operating a plant is the actual usage of operational data for process and maintenance optimization by advanced analytics. Typically a data-driven central oversight function supports operations and maintenance staff. A major lesson-learned is that the results of a rather simple statistical approach to detect anomalies fall behind the expectations and are too labor intensive. It is a widely spread misinterpretation that being able to deal with big data is sufficient to come up with good prediction quality for Prescriptive Maintenance. What big data companies are normally missing is domain knowledge, especially on plant critical rotating equipment. Without having domain knowledge the relevant input into the model will have shortcomings and hence the same will apply to its predictions. This paper gives an example of a refinery where the described hybrid model has been used. Novel and Additive Information First principle models are typically expensive to build and not scalable. This hybrid model approach, combining first principle physics based models with artificial intelligence and integration into an operational workflow shows a new way forward.

Optimizing Rotating Equipment Maintenance Through Machine Learning Algorithm

This work highlights the development and results of a Rotating equipment predictive maintenance tool that allows to monitor the status of rotating machines through a synthetic "health index" and early detection of anomalies. The data-driven proposed solution is of great help to maintenance engineers, who, alongside the existing methodologies, can apply an effective tool based on artificial intelligence for early prevention of failures. Taking advantage of the high availability of remote sensors data, an anomaly detection machine learning model, which relies on Principal Component Analysis (PCA) and Kernel Density Estimation (KDE), has been built. This model is capable of estimating, in real time, the health status of the machine, by matching the sensors actual values with the reference ones based on the Normal Operating Conditions (NOC) periods, that have been previously identified. If an anomalous behavior is detected, the Fault Isolation step of the model allows to evaluate which are the most contributing sensors for the investigated anomaly. These outcomes, combined with a failure mode matrix, which links the sensors deviations with the possible malfunctions, allows to highlight the most likely failure modes to be associated to the investigated anomaly. The developed predictive tool has been implemented on operating sites and it has demonstrated the capability to generate accurate warnings and detect anomalies to be processed by the maintenance engineers. These alerts may be aggregated into events in order to be monitored and analyzed by remote and on site specialists. The availability of alerts gives to the users the possibility to predict any deterioration of the machines or process fluctuations, that could lead to unplanned events with consequent mechanical breakdowns, production losses and flaring events. As a consequence, tailored operative adjustment to prevent critical events can be taken. Thanks to the tool, it is also possible to monitor over time the equipment behavior in order to provide suggestions for maintenance plans optimization and other useful statistics concerning the most recurrent failure. The tool's innovative feature is the ability to utilize the giant amount of data and to reproduce complex field phenomena by means of artificial intelligence. The proposed tool represents an innovative predictive approach for rotating equipment maintenance optimization.

Building a Low-Manned Production Installation: Considerations for Rotating Equipment, Electrical & Automation Systems, and Digitalization

Nearly all oil and gas operators and engineering companies in the offshore sector today are engaged in programs to advance concepts for low-manned and/or normally unattended production installations (NUIs). When it comes to the design of these facilities, topsides rotating equipment and electrical, instrumentation, control, and telecommunications (EICT) packages represent key areas of interest for decision-makers, owing to the significant impact they can have on required manning levels. Over the past decade, the author's company has worked closely with major Operators in the U.S. and the North Sea to look at how existing technologies can be applied in these areas to safely facilitate de-manning of both brownfields and greenfields. This paper provides insight into these efforts. It also presents projected manpower and cost savings from de-manning, using data derived from both studies and real-world projects.

DWT-LSTM-Based Fault Diagnosis of Rolling Bearings with Multi-Sensors

Bearings are widely used in many steam turbine generator sets and other large rotating equipment. With the rapid development of contemporary industry, there is a great number of rotating equipment in various large factories, such as nuclear power plants. As the core component of rotating machinery, the failure of rolling bearings may lead to serious accidents during the industrial production operation. In order to accurately diagnose the fault status of rolling bearings, a novel long short-term memory (LSTM) model with discrete wavelet transformation (DWT) for multi-sensor fault diagnosis is proposed in this paper. The main purpose of this paper is to use the DWT-LSTM model to diagnose the health of rolling bearings. Firstly, the DWT is used to obtain detailed fault information in both different frequency and time scales. Then, the LSTM network is employed to characterize the long-term dependencies hidden in the time series of the fault information. The proposed DWT-LSTM method makes full use of the advantages of feature extraction based on expert experience and deep network learning to discover complex patterns from a large amount of data. Finally, the feasibility and efficiency of the proposed method are illustrated by comparison with the existing methods.