Altair > Case Studies > Data Analytics for Heavy Equipment: Reducing Gas Turbine Generator Downtime and Failures

Data Analytics for Heavy Equipment: Reducing Gas Turbine Generator Downtime and Failures

Technology Category

Analytics & Modeling - Machine Learning
Application Infrastructure & Middleware - Data Visualization

Applicable Industries

Electrical Grids
Oil & Gas

Applicable Functions

Maintenance

Use Cases

Asset Health Management (AHM)
Predictive Maintenance

Services

Data Science Services
System Integration

About The Customer

Serba Dinamik is a multi-billion-dollar supplier of services and power generation systems to the oil and gas industry. Its gas turbine generators are employed by major exploration and drilling firms such as Petronas, Shell, and ExxonMobil to provide power on offshore oil rigs. These turbines power all critical systems on the rigs, including safety systems and pumps. Their reliability is therefore essential to smooth operations on the rigs. Any failure could result in millions of dollars of losses per day, making it crucial for the equipment to be consistently reliable with no unplanned downtime.

The Challenge

Serba Dinamik’s management was seeking to implement an advanced predictive maintenance system that would do more than simply limiting downtime. They needed to provide clients with on-demand visibility into the turbine performance and potential future variances from normal operation. The system needed to detect anomalies and outliers in sensor data that may indicate impending failure of a subcomponent, and provide clear guidance on optimal maintenance scheduling based on planned rig operations and equipment performance. The team set a goal of boosting output for their microturbine power generators by an average of 25%, reducing downtime by 70%, and cutting maintenance expenses by 25%.

The Solution

Serba Dinamik utilized Altair’s design and simulation software to develop, debug, and manufacture turbines. Altair suggested that the artificial intelligence (AI) and machine learning (ML) capabilities of Altair® Knowledge Studio® and the real-time data visualization capabilities of Altair® Panopticon™ would improve the reliability of turbines mounted on offshore rigs. Serba had access to a large set of performance data for their turbines that enabled them to use the unsupervised ML capabilities of Knowledge Studio to identify patterns, trends, and outliers in the historical performance data and then build AI models that would flag emerging patterns in operational data. Working with engineers from three key hardware vendors plus Altair, the Serba team built and tested an AI model to predict when maintenance tasks are required plus a set of analytical dashboards that enable shore-based personnel to monitor the performance of offshore turbines.

Operational Impact

Serba Dinamik’s engineering team worked with Altair and ORS Technologies Sdn. Bhd., a Malaysian engineering consulting firm, to develop a new Smart Predictive Maintenance Data System (SPMDS) utilizing Knowledge Studio and Panopticon. After a six-month testing period, Serba’s crews began using SPMDS to monitor the performance and maintenance requirements for turbines in the field. The maintenance crews use Panopticon-powered dashboards built into SPMDS to monitor every sensor mounted on operating turbines in real time. AI models built with Knowledge Studio identify potential failures or issues that require engineering attention, and, based on that understanding, take turbines offline only when necessary.

Quantitative Benefit