Taking Asset Performance & Reliability to the next level

Context

Setting up any manufacturing or Industrial unit needs a high amount of Capital expenditure. For a manufacturing unit, it includes setting up plant, purchase, and setup of machinery, establishing production process and training workforce for production and quality processes. This effort and expenditure can be successfully leveraged only if output is maximized (close to 100% of designed output). However, each machine or equipment is subjected to mechanical wear and tear over a period, and hence needs repair, maintenance as well as change of worn parts. The manufacturers plan the shutdown based on the recommendation of the equipment supplier and actual wear & tear to enhance the life of equipment. However, if there is an unplanned downtime despite regular maintenance, then there is substantial loss of production and erosion of margins in addition to the regular maintenance cost incurred.

While preventive maintenance scheduled on regular basis helps to maintain the health of equipment. However as per various industry reports unplanned downtime of equipment result in loss of billions of dollars from productive hours. Asset intensive industries like Metals, Mining, Minerals, Fertilizers, Chemicals, Oil & Gas and similar such industries were already under stress owing to their cost intensive operations vis-a-vis lackluster demand even during the pre-pandemic period. The asset heavy nature of their operations is highly maintenance hungry, and any unplanned downtime has substantial impact on the operating margins. Globally, margin erosion resulting from sub-optimal asset utilization is estimated to be almost to the tune of 3-4 Billion USD. The situation has been further aggravated by the Covid-19 Pandemic. The impact of the pandemic has not only created a supply chain disruption but has also caused a huge depletion of workforce. Post pandemic geo-political situations have also not helped the scenario. The organizations are now grappling with the challenge of catering to the pent-up demand with a leaner workforce. This has made the cost of downtime and in turn lower asset utilization even more unaffordable for the organizations.

“An average oil and gas company goes through at least 27 days of unplanned downtime each year, costing US$38 million. Even if the downtime lasts for just 3.65 days, the resulting losses can be as high as US$5 million….?Costs of unplanned downtime (Kimberlite, 2016)

Asset Management is not a new area

Regular maintenance & overhaul of equipment has been a critical activity ever since the evolution of industrialization. Maintenance activities have continuously undergone modernization along with every phase of industrial revolution. The key objective has been to ensure highest levels of reliability with preventive maintenance at regular predefined intervals. The approach of preventive maintenance has been modernized through Computerized Maintenance Management Systems (CMMS).

CMMS are based on Time-Based Maintenance (TBM) approach that cater to the need of Preventive Maintenance. The frequency of TBM is sometimes suggested by OEM and are also finetuned for specific context based on the workload scenario, uptime needs etc. The frequency of TBM is defined in the CMMS and maintenance engineers are alerted with the trigger of work orders to carry out maintenance tasks.

CMMS has evolved over time as per need of modern-day maintenance teams; added the practice of condition-based maintenance (CBM) with vibration monitoring. CBM activities are carried out by a team of inspectors who do a periodic inspection at pre-defined intervals for the set of critical equipment and take need-based action. However, since the CBM is triggered as per timelines defined for TBM, sometimes there are unpleasant surprises of unplanned downtime in between the timelines causing substantial loss in production in addition to the maintenance cost already incurred.

Some of the industry experts have realized that TBM along with CBM at predefined intervals is still not sufficient to avoid unplanned downtime for reasons explained above. They have implemented online CBM systems. There are multiple online vibration systems which have evolved. Certainly, online vibration monitoring systems have helped reduce the unplanned downtime to an extent but not to the desired levels.

Where do traditional approaches fall short?

Even with the advent of online vibration monitoring along with the CMMS based TBM practice there are still surprises and downtimes occur without a warning from the vibration system or even after frequent inspections done by the maintenance team. Such incidents not only cause substantial production loss, but also incur high maintenance costs due to frequent maintenance activities, consumables & spares.

While TBM typically happens with gaps of months between two schedules, vibration monitoring can be more frequent (period of weeks) as per the fixed thresholds prescribed by the OEMs. But the failures which are incipient in nature occur even within the fixed thresholds. One would need to spot them at an early stage before they become catastrophic. But they are very difficult to detect even with the most modern automation systems like SCADA, DCS, online vibration & CMMS. All these systems are configured on fixed thresholds and are not able to detect any anomaly within the threshold limits defined. This is where an intervention of data driven intelligence is required by maximizing the past data of the equipment.

The Advent of Digital Twins

Advent of new digital technology such as Cloud Computing, Machine Learning & Artificial Intelligence has enabled storage, and processing of large amount of data at real-time without significant investment or running costs. This helps in unleashing the power of data by democratizing AI and bring real-time actionable insights from data.

AI driven Digital Twins as a definition are digital replica of a physical object such as Machine or process. The digital replica is built by using machine learning models analyzing past operational data of the equipment/ process. The digital model when fed with real time data provides unique actionable insights which could not have been possible by first principles models alone. In other words Digital twins are the fusion of physical and virtual world which can generate data driven actionable insights which can in turn drive predictability & adaptability across industrial operations.. The core enablers of Digital Twin are IT-OT integration, Cloud data storage and processing, Machine Learning and AI.

With Edge and Cloud topology getting more matured, Digital Twins can be deployed closer to machines and are able to get real-time data and provide insights and alerts to stakeholders.

Digital Twin help overcome shortcoming of Traditional Approach

The need of the hour is to deploy SMART Decision support systems enabled by AI powered digital twins, which can analyze the plant data in the background 24X7X365 and come up with actionable insights at an early stage with enough time for preventing failure .. Such systems would complement the existing CMMS or CBM system with foresight by triggering alerts to the maintenance engineers at an early stage. While the existing systems help gather further insight by doing a detailed Root Cause Analysis (RCA) at that stage to mitigate the risk completely.

Such solutions should be less data hungry and easy to deploy requiring minimum infrastructure and data. To expect all the failure signatures and huge set of data from the customer to start the solution deployment would be unrealistic. It would also be a highly time-consuming exercise requiring substantial investment. The solutions should ensure shorter payback period with minimum investment and high IRR. The post pandemic world has accelerated the need of such AI powered predictive maintenance systems for organizations to deploy & prevent unplanned outages especially with a work force which is much leaner than before.

Conclusion

Digital Twins based Predictive Maintenance should be the preferred approach for critical assets for organizations across industry verticals. The critical assets are the ones where the cost of failure is very high. Those who have adopted CMMS or vibration monitoring through standard industry products, can further augment their approach with Digital Twins for critical assets which shall in turn compliment their existing landscape as mentioned above.

We see many organizations experimenting the Digital Twins for improvement in Asset performance and reliability. Many are in the POC stage and trying to understand the benefits. However, some PoCs fail, discouraging further investment in Digital Twins. The difference between failure and success of POCs/pilots are primarily based on whether right solution approach has been applied and whether the right use case has been identified. There are organizations who have successfully implemented Predictive maintenance and are deriving continuous benefit YOY. We hope that such solutions are further democratized and successfully adopted by the industry which can in turn revolutionize the entire maintenance strategy to drive performance and throughput.

Authors

1. Sayantan Roy, Head, Digital Industrial Consulting and Solutions

Tata Consultancy Services, Bangalore, India

Email ID:?[email protected]

2. Jaishankar GL, Senior Consultant, Digital Industrial Consulting and Solutions

Tata Consultancy Services, Bangalore, India

Email ID:?[email protected]

3. Gaurav Joshi , Head Technology Advisory Services

Tata Consultancy Services , Pune India

Email id: [email protected]