May 26, 2020
By: Lyat Avidor Peleg
Case study: Titan Cement Group saves money and increases productivity through predictive analytics
It can be challenging to demonstrate the value of a predictive analytics system when it is typically measured by how well it prevented failures and incidents from occurring. Our client, the Titan Cement Group, experimented with different methods of implementing SAM GUARD’s solution and has seen real value from various methods of implementation.
Titan is a cement company that was founded at the beginning of the 20th century and is still run by the same family. It operates 14 cement plants, quarries, grinding plants, and other cement infrastructure across over 10 countries in 5 continents and employs over 5,400 workers. Titan’s annual sales came to €1,610 million in 2019.
Digitization and automation are part of Titan’s company ethos. The entire Titan value chain, from supplier to customer, is connected through a single integrated real-time data platform, and from that vantage point, Titan is able to apply advanced AI-powered tools to optimize performance in different areas.
The goal: lowering costs and making the entire business more reliable
Like in most plants, the control rooms at Titan’s cement plants were always busy, 24/7. Control room employees could be overwhelmed with data, alerts, and beeping alarms, and were constantly chasing fires.
Titan wanted to enable its employees to see the bigger picture so they could deal proactively with the important items instead of being consumed by the most urgent ones. It wished to replace the current norm of struggling with information overload and strengthen its employees with more trouble-shooting capabilities, in order to lower costs and make the entire business more reliable.
Failure is a process, not an instant, explained Fokion Tasoulas, Executive Director, Group Engineering and Technology at the Titan Group. There are several progressive levels of failure, building up to a stage when human senses can see, hear, or smell its effects. By that point, it’s almost always too late to save the item. It’s close to total failure, and is very expensive to repair.
The Titan Group wanted to dial back the point in the timeline at which they could pick up impending failures of parts, process, and quality, so they could plan and schedule repairs effectively, and reduce loss of productivity.
The means: predictive analytics
Implementing SAM GUARD in the control room
The Titan Group decided to implement SAM GUARD’s predictive analytics solution through a “test and learn” approach, beginning in one plant and then learning from that experience and rolling it out to other plants. They selected a representative cement plant with an average level of digital maturity, and looked for projects with the highest projected impact to try out the solution.
Titan tried to identify the right roles to take responsibility for the solution, and began by implementing SAM GUARD in the control room. They encouraged control room operators to engage with SAM GUARD, follow up on its alerts, and troubleshoot the issues it flags.
Over the course of several months, the team learned the value of the system. For example, SAM GUARD picked up a slight change in one of the coolers. The team, still unfamiliar with SAM GUARD, dismissed the alert, as well as the next one that SAM GUARD generated 3 days later. 10 days from the first alert, the cooler motor failed. This convinced the team that SAM GUARD’s alerts should be listened to.
Improving processes, additional roll outs
In order to get the most out of the solution, Titan’s management decided that rather than leaving it to the control room staff, there should be a dedicated person at a high level of technical management who is responsible for ensuring that alerts are reviewed and dealt with. It needed to be someone who is technical and experienced enough to know what other team members to bring in for each alert, and someone senior enough to be able to make it happen. This person, together with the others that are mobilized in response to alerts, became responsible for preventing the next failure.
The team would have a short meeting at the beginning of each day to discuss alerts that the system had raised and consider how best to troubleshoot them. They would then work on finding and fixing the root cause of the problems flagged by SAM GUARD.
This approach worked well, and Titan then rolled out SAM GUARD to two more plants in this way, and were able to do it much more quickly, since they had already built the blueprint of how to do it through their earlier trials. Modelling the first plant took 1.5 years, a process Titan managed completely on its own (Precognize typically does this in just a couple of weeks but Titan wanted to gain the experience in-house). After that, Titan modelled a second plant in just 2 months, and then a third in only 1 month.
Once the third plant was set up, they then decided that it didn’t make sense to have separate troubleshooting efforts at each plant, each benefiting from local knowledge only. So they decided to set up a dedicated Remote Analytical Monitoring Center, or RAMC, that would monitor the five plants together.
Creating a Remote Analytical Monitoring Center
Titan set up a space outside of its plants being monitored to house its Remote Analytical Monitoring Center or RAMC; each of the plants being monitored had implemented SAM GUARD software. The RAMC is staffed by employees who have a deep domain knowledge of maintenance, process, and troubleshooting in Titan plants. From the RAMC, they can measure and monitor performance across multiple sites.
With the basic process of the RAMC fully operational, Titan moved on to establish a strong communication and governance system between plants. The RAMC provides regular reports on performance KPIs with local plant experts. High level experts monitor events daily within the system, perform additional evaluation on each alert by correlating other process variables and operating parameters, and carry out a first root cause analysis. Then, RAMC operators have an open dialogue with the experts on the ground to deliver their insights and hear the domain experts’ opinion. Often, the RAMC can troubleshoot cases faster and with more accuracy than employees who are only familiar with a single plant.
It was at this point that the coronavirus pandemic struck, proving even further the value of remote analytical monitoring. Titan quickly rolled out SAM GUARD in another 2 plants and connected them to the RAMC. Throughout the pandemic, Titan’s RAMC has been working to process simple alerts which enable it to fix problems early and quickly, using the skeleton teams that are currently on site. Thanks to its RAMC, Titan has not needed to keep large numbers of employees in its plants to monitor performance and productivity, thereby reducing the
danger of infection among its workers.
Bottom line results: Increased savings and productivity
This approach was extremely effective. In addition to generating early alerts about impending equipment failure, SAM GUARD helped raise the reliability factor of the kiln in one of its plants, which is the heart of the plant. Reliability increased steadily to unprecedented levels, reaching 98% reliability, considered World Class, within six months of its initial implementation of SAM GUARD, and remaining above that level for almost all of 2020.
With SAM GUARD in place, Titan increased the annual capacity of this plant by 5% without any extra costs; during a period of 10 months, they avoided potential costs of $900,000 due to averted breakdowns, and prevented around 500 hours of downtime in its most important assets, which would have cost even more.
On a Remote Analytical Monitoring Center (RAMC) level, the benefits include:
- Developing a blueprint to rollout predictive analytics implementation at every factory.
- Creating a team of people focused on predictive analytics who built new capabilities and mastered troubleshooting.
- Producing a pool of knowledge that collects experience from multiple different plants, enabling RAMC operators to identify the repeating patterns and generate deeper insights into why issues, events, and failures may occur.
- Overcoming turnover of people who store the collective wisdom in each plant. Instead of losing that knowledge when an employee leaves or retires, the wisdom can be collected, preserved, and used to guide future plant decisions.
- Connecting plants with each other so they can communicate better and learn useful strategies and when to implement, building a global team spirit.
For the Titan Group as a whole, the benefits include:
- A new ability to build and monitor performance and KPIs across the entire organization.
- Establishing best practices in all plants.
- Acting to set benchmarks and compare performance across plants
SAM GUARD improves Titan’s bottom line across multiple plants
By experimenting carefully, the Titan Group implemented SAM GUARD predictive analytics software in its plants in a way that enabled the organization to reap maximum value. With SAM GUARD operating in remote analytical monitoring centers, the Titan Group reduced repairs and maintenance costs, increased performance, and was able to take proactive data-driven decisions to fine tune productivity across multiple plants, with the result that its revenue measurably improved. With such a system already in place, the Titan Group was well positioned to manage any crisis that came along, enabling it to continue to operate efficiently even during the corona virus pandemic.
Hear Fokion Tasoulas of Titan describe the benefits himself in this webinar.