Motor Oil Tests SAP Predictive Maintenance to Reduce Refinery Downtime

Motor Oil Tests SAP Predictive Maintenance to Reduce Refinery Downtime

Reading time: 6 mins

By Fred Donovan, Senior Editor, SAPinsider

A combination of stagnating oil prices, supply chain disruptions, and the push to reduce carbon emissions by developed countries has forced oil and gas organizations to rely on digital technologies, such as predictive maintenance, to improve their bottom line.

An EY survey of oil and gas executives found that respondents view digital technology as the best option to improve efficiency and productivity in their operations and boost profits.

A full 80% of respondents are investing at least a moderate amount in digital technologies. Around 43% of respondents identified big data analytics as one of the top trends that will possibly affect their business in the coming years.

The survey found that the average oil and gas company experiences nearly 27 days of unplanned downtime annually, resulting in losses ranging between 38 million and 88 million US dollars every year.

Motor Oil Group, an oil refiner and retailer of oil and gas products based in Greece, wanted to invest in digital technology to significantly lower unplanned downtime and cut costs.

The company needed a more accurate way to predict and address potential equipment problems before they occurred at its Corinth refinery and use sensor data to better forecast abnormal events. In addition, it wanted to ensure that the right people have access to data with real-time updates and automatic notifications to alert maintenance workers and technicians in case of potential abnormal events.

The company processes 185,000 barrels of crude oil per stream day at the refinery, which is the largest privately owned industrial complex in Greece and one of the top refineries in Europe. Even a brief disruption at the refinery has a significant impact on production.

Predictive Maintenance Benefits

Dimitrios Michalopoulos, Industrial Applications Head of the IT Division, Motor Oil, tells SAPinsider that the company held discussions with machine learning and artificial intelligence experts to determine if predictive maintenance was a promising investment area.

After speaking with the experts, the company identified several predictive maintenance benefits: cost savings, downtime and equipment failure avoidance, increased safety, greater quality and effectiveness of maintenance planning, and expanded technical knowledge about correlations and interactions of past abnormal events concerning equipment.

Motor Oil, which has been using an SAP ERP since 1999, naturally looked to SAP for its predictive maintenance technology. It set up virtual workshops with SAP about the predictive maintenance capabilities of and customer case studies associated with SAP Business Technology Platform.

“The main goal of these workshops was to help us identify challenges and opportunities, discover and prioritize use cases, and assess the value of the technology for Motor Oil,” Michalopoulos says.

Once Motor Oil decided to move ahead with the predictive maintenance project, it worked with SAP to develop a roadmap for implementation. It then ran a proof of concept with SAP for three and a half weeks to gather data from existing equipment.

“We decided to run a proof of concept to identify potential benefits for us,” he explains.

The results were “encouraging,” and in February 2021 the company launched a pilot project involving critical processes and unique equipment to identify potential benefits of the technology.

“During the pilot project, we analyzed four years of historical sensor data on five compressors to learn from them and develop time-series forecasting and the root-cause-analysis mechanisms,” Michalopoulos says.

However, the company determined that it did not have the technical skills in-house to implement the project by itself.

“We first contacted SAP to identify what options we have and what was available in the market. We tried to find our implementation partner who was experienced in such projects. So, we decided to run the project with Accenture. From the results, I can say that it was the right choice,” he says.

Working with Accenture, Motor Oil was able to increase the accuracy of predictions for abnormal events at the refinery by more than 77%, providing Michalopoulos and his team with up to 120 hours lead time to respond to upcoming technical issues.

Motor Oil included in the pilot project representatives from its distributed control system (DCS) and maintenance departments. They analyzed the data retrieved from the Historian database and identified areas of improvement in the current processes at the refinery.

During the pilot project, the company trained its staff on how to use machine learning and artificial intelligence so they would have a better understanding of how to use the technology in the future. “We organized some training sessions, and we have built some training resources that will help us in the future,” he explains.

SAP HANA Cloud and SAP Analytics Cloud

In addition to the SAP Business Technology Platform, Motor Oil used SAP HANA Cloud to store the data used in the analysis to produce predictive models about abnormal events at the refinery. Additional tools provided by SAP HANA Cloud enabled the company to develop programs, analyze data, build models, and simulate results, Michalopoulos relates.

SAP HANA Cloud provides users with a framework to develop programs in Python and have access to the data stored in HANA. Programs that analyze sensor measurements, identify abnormal behaviors, and produce forecasts are written in Python.

SAP HANA Cloud enabled Motor Oil to use root-cause analysis algorithms from the SAP HANA predictive analysis library on abnormal events to explain them with greater accuracy.

Motor Oil used SAP Analytics Cloud in conjunction with SAP HANA Cloud, giving its employees access to intuitive dashboards where sensors and equipment were monitored in real time. If abnormal events were predicted, the right users received email notifications in advance, allowing technicians to solve the problem before a shutdown occurred.

“We developed a user interface with multiple windows through which maintenance people could access the predictions that the models produced,” Michalopoulos says.

The company analyzed the refinery’s operations and revealed patterns about how different sensors affect each other. The platform also continuously improved, enabling predictions to become more accurate.

Technicians could repair sensors instead of replacing them entirely, which saved on maintenance costs. This was in addition to the cost savings of avoiding production shutdowns, and it created a safer working environment at the Corinth refinery.

“For the majority of sensors, the predictions are extraordinary. I couldn’t imagine having a forecast 24 hours before predicting how our equipment would behave for the next day. It is fantastic,” he enthuses.

Michalopoulos recommends that organizations analyze the data before undertaking a major IT project to determine if it is even possible.

“Before deciding the full scope of work, explore the data that you have available with subject matter experts. Let the data instruct you if it can answer your questions or not. The data will show you which questions to ask to get the right answers. Maybe you want the answer to a business question for which you do not have the right data. It is worthless to invest money in something that cannot be answered with the data,” he advises.

Motor Oil is currently in the evaluation phase of the pilot project’s deliverables. The company plans to expand the number and type of equipment that the system monitors as its next step in implementing predictive maintenance.

What does this mean for SAPinsiders?

  • Determine the type of data you will need to answer business questions about a proposed project. Motor Oil’s Michalopoulos advises companies to let the data instruct you about whether your project is worth pursuing. “It is worthless to spend money on a project for which you do not have the data to answer the business questions,” he notes.
  • Conduct a pilot project to identify the benefits and challenges of a significant task. A pilot project can assess whether a product or process will effectively operate in an environment as promised. It considers the feasibility of the undertaking and whether it will provide a return on investment without disrupting equipment and processes.
  • Determine whether you have the technical skills in-house to implement a project. If you don’t have the skills, find an implementation partner with experience in your particular project area. Motor Oil consulted with SAP and reached out to Accenture after determining that it did not have the expertise to implement the project as envisioned. “From the results, I can say that this was the right choice,” concludes Michalopoulos.


Motor Oil

  • Headquarters: Maroussi, Greece
  • Industry: Oil and Gas
  • Employees: 2,500
  • Revenue: 7.3 billion US dollars
  • Company details: Motor Oil Group is one of the largest companies in Greece and the Greek energy sector.
  • SAP solutions: SAP Business Suite powered by SAP HANA, SAP R/3 Enterprise (transitioning to SAP HANA by July 2022), SAP Business Technology Platform, SAP HANA Cloud, SAP Analytics Cloud, SAP SuccessFactors, SAP Business Warehouse, SAP Business Planning and Consolidation

More Resources

See All Related Content