Intelligence, data, and oil: PETRONAS does it all

The company won three trophies at the Malaysia Technology Excellence Awards for integrating its innovations into action. 

In a data-driven age of doing business, using innovations in technology and artificial intelligence helps companies predict market trends, improve operational productivity and efficiency, or make better business decisions. Malaysia’s Petroliam Nasional Berhad, better known as PETRONAS, introduced many breakthroughs in Machine Learning (ML) and Analytics Intelligence (AI) based solutions. In doing so, the company managed to win the Analytics - Engineering Award, Analytics - Oil & Gas Award, and the AI - Oil & Gas Award at the Malaysia Technology Excellence Awards 2022.  

The awards programme recognises breakthroughs, key innovations, and significant players in the tech industry of Malaysia, and PETRONAS solidified their position as a top company with their PETRONAS Intelligent Substations (IS), PETRONAS Digital Twin Enterprise Optimisation Centre (EOC), and PETRONAS Machinery Monitoring and Prescriptive Diagnostics (P-MMPD).

PETRONAS Intelligent Substations (IS)

PETRONAS IS works as an integrated digitally enabled electrical technology solution, made up of crucial applications such as Alternator Health Indicator (AHI), Electrical System Digital Simulator (ESDS) and Programme Analysis of Transformer Health (PATH). These features were established as a cloud-based solution to address plant reliability issues from electrical equipment failure. Such failures usually lead to high operating expenditure (OPEX) from unplanned plant shutdowns and network disturbances. 

IS aims to address pain points and electrical equipment challenges through an in-house algorithm utilising real-time and offline data for prognosis and trending. Thanks to its big-data analytics combined with artificial intelligence, the solution enables real-time data monitoring of electrical systems and equipment and predictive and prescriptive action before failure happens. 

With the implementation of PETRONAS IS, electrical equipment reliability and Overall Equipment Effectiveness (OEE) of plants significantly improved. The predictive analytics allowed for early identification of issues and risks in over 18 plants in the PETRONAS Upstream and Downstream Malaysia Assets. The company aims to look further by rolling out the IS to all PETRONAS Operating Plant Units (OPU) groupwide and globally. 

The AHI has helped reduce generator maintenance costs in 17 PETRONAS assets, providing significant cost savings and plant reliability improvement.

Since implementing the IS, PETRONAS has generated significant cost avoidance and value creation, proving to be a key achievement for the company in pursuit of techno-digital excellence, further reinforced by their Analytics - Engineering Award at the MTEA 2022.

PETRONAS Digital Twin: Connecting the Dots

Digital Twin’s Enterprise Optimisation Centre (EOC) provides a virtual and connected “digital twin” of all PETRONAS operations. It offers end-to-end integrated views and insights of the entire company‘s operations, all condensed into 17 transformational yet interactive views, which took approximately five months to develop. These views provide instant insight into molecular flow across the entire complex enterprise, including more than 40 processing plants, 20 feeds, and 60 products. 

The first phase of the EOC, named “Unified Experience”, aimed to address one main business pain point: “How might we provide end-to-end visibility of key enterprise data and insights to enable better value-driven decision making?” 

The team collaborates with 25 SMEs and more than 30 engineering and digital experts, including a partner working from five different countries (the UK, US, India, Singapore, and Malaysia). They were able to identify 650 critical data points across the integrated value chain, representing all current and planned molecular movements. Numerous business-centric data sources are connected through EOC, allowing visibility across the entire enterprise operations. It simplifies PETRONAS operations into 17 uniquely curated interface visuals supported by interactive graphical views, all through four cloud-based servers that enable remote implementation. 

Now, EOC is working towards incorporating Greenhouse Emissions (GHG) flows with financial flows that allow optimisation at a corporate level to deliver “organisational agility”. As the dynamics of business continue to change, with high volatility while ESG and energy transition remain pervasive, organisations need to be flexible enough to reshape themselves continuously and be able to move with pace for both opportunities and threats. With EOC in place, PETRONAS now can have near real-time immersion at an enterprise level, allowing for much more rapid decision-making, rightfully earning their spot with the Analytics - Oil & Gas Award. 

PETRONAS Machinery Monitoring and Prescriptive Diagnostics (P-MMPD)

The P-MMPD is an analytic tool that uses artificial intelligence (AI) through machine learning (ML). It enables early detection of anomalies and failures of machinery through its ability to highlight the slightest deviation from the standard operational profiles. Through further analysis, identifying underperforming assets and the initiation of remedial action to improve equipment efficiency ahead of time becomes possible. 

Traditionally, equipment failures took a considerable amount of time and resources to diagnose and resolve correctly. Operations and maintenance teams worked separately by leveraging different systems to track and analyse equipment performance and health. P-MMPD was established in 2020 to address this pain point of increasing turbomachinery reliability by early detection of mechanical failure, including the impact of process parameter changes by correlating collected data from multiple sensors on the equipment and its adjacent. These features on the P-MMPD allow it to pinpoint the slightest deviations from the standard expected results or “known good behaviour” before catastrophic failure is computed by the AI. 

With the integration of the Failure Mode and Effects Analysis (FMEA) library, the AI capability was further enhanced and optimised with years of machine operation experience onto the digital platform. The P-MMPD will then prescribe corrective actions for each combination of triggered anomalies in an ML model, eliminating time-consuming manual investigations and speeding up the time needed to make decisions, boosting productivity. 

In its first year of deployment, P-MMPD accurately predicted failures in advance, allowing the PETRONAS teams to fix issues before they translate into actual losses. Over 200 ML models were deployed, and 51 prominently early warnings, of which 12 were high-impact warnings were identified, resulting in a total value of 14x ROI and reduced unscheduled downtime -- saving millions of dollars for PETRONAS. 

Cloud-based technology allows the deployment of 459 ML models for 317 pieces of equipment in 21 business locations across PETRONAS as of December 2021. It will enable all levels to immerse themselves with new ways of working through AI technology and slowly be implemented out of the intended use of turbomachinery equipment such as boilers, generators, and production lines. Recognising the value of prevention with early detection and the utility of artificial intelligence, PETRONAS raised its productivity and saved millions in avoided failures, earning its spot at the AI - Oil & Gas Award.