From highways to parking lots to tennis courts and more, asphalt is ubiquitous in modern life. It can also be highly dangerous under high temperatures such as those used in processing the petroleum-based substance. According to the US Occupational Safety and Health Administration (OSHA), more than a dozen heated storage tanks for asphalt or No. 6 fuel oil have exploded in the past decade.
To improve the safety of its asphalt operations, US-based Owens Corning has put data analytics to work, leveraging low-code tools to develop a digital platform that incorporates multiple data flows and enables previously plant-specific information to be shared and coordinated across the company’s operations.
“This project was driven by a need to create real-time visibility to data with actionable insights to prevent hazards and enhance safety in operating asphalt processing tanks across our manufacturing network,” says Malavika Melkote, director of IT and the Analytics Center of Excellence (COE) at Owens Corning.
The project, which has earned Owens Corning a CIO 100 Award in IT Excellence, leverages digitizing sensors to extract data from asphalt tanks. This data is integrated with a range of other data points, providing easy-to-use visuals for plant operators to analyze, Melkote says. “They can quickly assess potential hazards and risks in real-time and proactively take preventive actions.”
The increased visibility into potential hazards enabled by the loss prevention platform, combined with preventive maintenance, has minimized unplanned production outages and created sizable cost savings via reduced equipment losses. Melkote says the time required to make a decision and take action at its plants has gone from days to minutes, thanks to the platform.
The MVP approach
Monitoring and managing an asphalt tank’s vapor space is critical to safety and compliance with Title V of the federal Clean Air Act. Prior to developing the loss prevention platform, Owens Corning and the rest of the asphalt industry collected vapor space data in offline databases. Those databases were plant-specific and the information they contained was difficult to share and coordinate across the company.
“Through our Loss Prevention discovery process, it became evident that the monitoring and managing of our asphalt tanks’ vapor space was a critical component to safety and regulatory compliance,” says Frank Burg, asphalt manufacturing support leader at Owens Corning. “While there was a rigorous process to manage safety and compliance, there was a big opportunity to make it more efficient and scalable across plants through automation and analytics. The opportunity was to fully digitize the data collected from tanks, automate integration of multiple data flows, and provide tools to plant personnel to analyze data, to highlight and assess risks and hazards for quick actions.”
In August 2021, Owens Corning set developing the loss prevention platform. A cross-functional team across Environment, Health, Safety (EHS), engineering, controllers, and plant leaders worked together with the IT analytics COE to lay out the vision and roadmap. To address the biggest pain points of the current process, the team identified three key requirements for the platform: It needed to provide timely access to data and proactive analytics that highlight risks; give operators the ability to share insights, actions, and learnings across plants; and be easy enough to be used by a diverse group of tank operators with varying degrees of technology experience. The team felt the best approach would be to leverage low-code tools.
“A key consideration was to enable the business team, with a citizen developer, to enhance, operate, and manage the solution long-term,” Melkote says. “Giving more control to business users to enhance the solution was a driver to choose a low-code technology platform.”
The resulting system is a single source of truth for loss prevention data, with proactive monitoring and analytics that alert plant leaders with real-time insights for hazard prevention. It uses a combination of advanced analytics and machine learning to perform relational data analysis beyond the raw data.
The development team, led by Muhammad Shoib, enterprise information architect at Owens Corning, took a minimum viable product (MVP) development approach, creating a proof-of-concept that was deployed at one plant within three weeks. That provided valuable feedback for fine-tuning the process and provided excitement among users and the business team. With their support, Shoib’s team scaled the solution and fully deployed it in 17 plants within three months.
“The MVP took two weeks to build and let users experience the solution to give feedback,” Shoib says. “We then moved to a pilot for one plant, which took eight weeks of iterative development. Deployment at the first plant took two weeks plus hyper care to fully operationalize the solution — to get end users, the product owner, and the citizen developer comfortable.”
Melkote says the pace of adoption and the minimal training required for end users was a pleasant surprise. “The iterative approach to developing an MVP, pilot, and the first plant implementation minimized any hiccups,” Melkote says.
Empowering business users
The approach undertaken for the project represents a significant shift for Owens Corning, which previously relied on a traditional approach to solution delivery, with the full scope solution delivered to the business for testing and validation. Melkote says getting the business tuned into an iterative method of co-developing the solution was a big change, but ultimately a worthwhile one.
“The MVP set the tone and the speed for what was possible that the users could touch and feel,” Melkote says. “They got onboard after the MVP to support the monthly release schedule of functionality.”
Since the platform has been deployed, Owens Corning has replaced personal tools and siloed information with a digital platform accessible across the enterprise, Melkote says. That, in turn, has enabled more efficient data flows and analytics that have increased the speed of decision-making such that the business can now solve more use cases with minimal IT involvement.
“Our business team now manages the solution,” Melkote says. “They felt empowered to enhance the solution at their pace. Dependence on IT to prioritize their enhancements is no longer an issue.”
Melkote now heartily recommends the MVP and low-code approach to her peers.
“Start small with an MVP, let your business partners experience the solution, see the value quickly,” Melkote says. “Put your business in the driver’s seat with the right roles; they manage the speed, functionality. Continue to nurture citizen developers, business product owners; be the best business partner. Let your business partners share stories and experiences and market the success stories.”
