The global energy transition is forcing a massive reckoning within the power generation sector. As utilities move to repower aging coal plants, modernize infrastructure, or integrate renewable assets, they face a silent, pervasive crisis: the "data graveyard." For decades, critical information regarding equipment specifications, maintenance history, and operational nuances has been siloed across disparate Enterprise Asset Management (EAM) systems, unorganized personal drives, and the fading memories of long-tenured employees.
1898 & Co., a consultancy firm specializing in critical infrastructure, is at the forefront of a new movement to resolve this challenge using Artificial Intelligence. By transforming fragmented, analog, and digital "dark data" into a unified source of truth, utilities are not only improving efficiency but are fundamentally changing how they operate their facilities.
The Core Challenge: Institutional Memory vs. Modern Efficiency
The difficulty of maintaining legacy power plants is often underestimated. When a plant undergoes a repower, the documentation often fails to account for the gap between original construction and current reality.
"This is a common challenge that we’re seeing in industry," says Chris Wiles, principal AI solutions architect at 1898 & Co. "Whether it’s to repower an old coal plant or something brand new where all the data just didn’t get encapsulated from planning to design to construction, the disconnect is universal."
The problem is twofold: technical and cultural. Technically, the data is trapped in obsolete file formats, siloed software, and inconsistent naming conventions. Culturally, the most vital information—the "tribal knowledge" regarding which valve sticks when the temperature drops or which lubricant works best for a specific bearing—resides entirely in the minds of veteran staff.
"Quite often, it’s in a person’s head," Wiles notes. This dependence on individuals creates operational fragility; when a key staff member retires, they take the "operating manual" for the plant with them.
Chronology of a Digital Transformation: The Cleco Pilot
The ongoing pilot project at a Cleco power plant in Louisiana serves as a benchmark for this digital evolution. The facility, a complex mosaic of original infrastructure and modern upgrades, faced the classic struggle of fragmented records.
Phase 1: The Functional Audit
The project began not with software, but with a functional workshop. The goal was to identify "where the data isn’t." By mapping out the workflows of the plant staff, the 1898 & Co. team identified the specific pain points where technicians were forced to rely on "workarounds"—the informal networks of people who knew where to find answers when the official documentation failed them.
Phase 2: Establishing the Source of Truth
Cleco’s initial requirement was pragmatic: staff needed a way to locate piping and instrumentation diagrams (P&IDs) and other fundamental documents without wasting hours searching multiple drives. 1898 & Co. employed advanced vectorization techniques to digitize and index these records. This allowed the utility to create a centralized, searchable repository where data could be queried across all types, sources, and historical ages.
Phase 3: Narrative Reconstruction
With the records centralized, the team moved to fill the "narrative gaps." By using AI to analyze historical logs, maintenance records, and project files, 1898 & Co. is generating comprehensive system and subsystem narratives. These documents explicitly define the plant’s evolution, bridging the gap between its pre-repower state and its current, modernized configuration.
Supporting Data: Quantifiable Gains in Operational Value
While the Cleco pilot is still gathering long-term performance metrics, previous deployments provide a clear picture of the return on investment. Cooperative Energy, another utility client, provides a compelling case study on the financial impact of data reconciliation.
In their deployment, Cooperative Energy ingested data for 16,000 assets from three separate, disconnected plant drives. The primary objective was to reconcile three conflicting sets of Preventative Maintenance (PM) recommendations.
The $3 Million Result
By automating the ingestion and attribute population process, Cooperative Energy realized significant savings. "They saved about $3 million in unadjusted labor costs by being able to extract all the data from three plant drives and populate all the attributes and specifications for those 16,000 assets," Wiles explained.
The "Handwritten" Breakthrough
The value of AI in this context extends to capturing data that was never digitized. During the Cooperative Energy project, a veteran employee announced their retirement. They left behind dozens of pages of handwritten notes—a critical repository of institutional knowledge.
"All of their lubricants for all the assets in question were literally written down on sheets of paper, and they’d never been entered in any system," Wiles recalled. The AI tool was used to ingest these notes and cross-reference them with existing EAM data. In several instances, the AI identified conflicts between the handwritten notes and the digital EAM. Upon investigation, the team discovered that the handwritten guidance—the result of decades of trial and error—was actually more accurate than the generic manufacturer recommendations already in the system. This allowed for the correction of maintenance protocols before the employee left, effectively "offboarding" their knowledge into the digital system.
Implications for the Workforce and the Industry
A common fear regarding AI in industrial settings is the potential for workforce reduction. However, Wiles presents a different perspective: the goal is to shift labor from "data searching" to "higher-value work."
Reclaiming Time
The efficiency gains are substantial. Wiles estimates that by centralizing data, workers save between one and three hours per week. In a high-stakes environment like a power plant, this is not just about convenience; it is about safety and operational reliability. When a technician spends less time hunting for a diagram, they have more time to conduct actual inspections, improve maintenance quality, or address long-standing mechanical issues.
Enabling Future Operations
The broader implication is that AI is enabling operational changes that were physically impossible as recently as two years ago. The ability to synthesize decades of data into actionable intelligence allows utilities to:
- Improve Predictive Maintenance: By having clean, consistent data, AI models can better predict equipment failure.
- Accelerate Training: New employees can access the "institutional knowledge" of their predecessors through the centralized, AI-curated narrative of the plant.
- Ensure Compliance: Centralized, searchable records make regulatory audits significantly less burdensome.
"We’re building out technology pretty quickly that manifestly changes the way that you operate a plant for the better," Wiles asserts. The transition from reactive, siloed data management to proactive, integrated intelligence is becoming a competitive necessity.
Looking Ahead: POWERGEN 2027
The industry will have an opportunity to dive deeper into these findings at the upcoming POWERGEN 2027 conference in Salt Lake City. The session, titled "An AI case study: Improving asset data while setting the foundation for more efficient plant operations," will provide a granular look at the methodologies used by 1898 & Co. and the specific results observed at the Cleco and Cooperative Energy sites.
Event Details:
- Session: An AI case study: Improving asset data while setting the foundation for more efficient plant operations
- Speaker: Chris Wiles, 1898 & Co.
- Date/Time: Tuesday, Jan. 19, 2027, at 10:15 a.m.
- Location: Salt Lake City, Utah
As the power industry navigates the complexities of decarbonization and aging infrastructure, the ability to "mine" the past for data will likely define the success of the future. The integration of AI into legacy plant operations is no longer an experimental luxury; it is the infrastructure of tomorrow.
For more information on the conference, attendees are encouraged to see the full agenda and register for POWERGEN 2027 to participate in the conversation shaping the next generation of power generation.
