Administrative Adjustment Boosts AI Sensor Efficacy in Paper Manufacturing

Unlocking AI's Full Potential: A Strategic Shift in Industrial Maintenance
Initial Hurdles with Data Overload and the Need for a New Approach
Upon assuming his role as a reliability engineer at Domtar's Kingsport, Tennessee, paper mill, Matthew McLaughlin quickly encountered a significant challenge. Following a motor failure, he was tasked with analyzing a full day's worth of sensor data from 450 units to identify a solution. McLaughlin realized this was an impossible feat, estimating it would take 32 weeks to process just one day's data, underscoring the overwhelming data volume generated by advanced AI technologies in manufacturing environments.
Embracing Collaborative Support to Maximize Sensor Capabilities
To overcome the limitations of manual data analysis and fully utilize Waites Sensor Technologies' machine-learning capabilities, McLaughlin initiated a more proactive engagement with the tech provider. Domtar already had access to Waites' comprehensive support, including a dedicated analyst. By intensifying these interactions from monthly to weekly calls, and by sharing additional data from thermal imaging, infrared cameras, and ultrasonic meters, Domtar fostered an environment where the AI system could learn and adapt more effectively. This heightened collaboration led to noticeable improvements within a mere three months.
The Symbiotic Relationship Between Advanced AI and Human Acumen
Rob Ratterman, CEO and cofounder of Waites, explained that while AI performs the initial analysis, the diagnostic process requires the combined efforts of Waites analysts, on-site vibration experts, and Domtar's maintenance personnel. This blend of AI insights and human judgment is essential for accurate problem diagnosis and resolution. Ratterman emphasized that merely deploying a sensor system yields some benefits, but achieving transformative results, like those seen at Domtar, demands a dedicated individual to champion the new approach throughout the organization. McLaughlin's team now meticulously tracks every network and equipment action item, aiming for a resolution time of under 30 days. The AI-generated analytics also inform precise maintenance adjustments, such as optimizing lubricant viscosity, which in turn reduces energy consumption and operational costs.
Cultivating Trust and Integrating AI into Core Operations
McLaughlin's strategy extended to building trust among Domtar's long-standing employees by distributing daily email summaries of system alerts and findings. Initially, operations managers frequently contacted him regarding alerts, but as confidence in the AI system grew, these calls diminished, signaling a profound shift in trust. Domtar now relies on Waites data to inform significant business decisions and integrate it as a critical component of their planning processes. With 748 sensors in operation, Domtar's staff actively monitors equipment for subtle vibrations indicating potential issues, from motor bearing cracks to lubrication deficiencies, enabling proactive maintenance and preventing costly failures. This predictive capability has resulted in substantial savings, including a remarkable year-long absence of fan belt purchases, a testament to the power of AI-driven continuous monitoring and maintenance.
Measuring the Impact: Tangible Savings and Enhanced Reliability
While Domtar chose not to disclose the specific costs associated with its AI implementation, McLaughlin affirmed the system's value, citing an estimated saving of 1,546.65 hours in unplanned downtime. This significant reduction in downtime underscores the economic justification for the Waites system, demonstrating that the investment in AI-assisted predictive maintenance is indeed "worth every penny" when leveraged effectively through strategic administrative adjustments and human-technology collaboration.