Maintenance by the Book

Joseph Santamaria, CIO, PSEG [NYSE: PEG], Joseph Santamaria is Vice President – Information Technology and Chief Information Officer for PSEG Services Corporation, as of October 2012. He...  More >>

More posts by this contributor:

• Meeting the Changing Needs of Customers

Share your experiences through CIOReview

Not long ago, car owners followed the owner’s manual to ensure proper maintenance: Oil change every 3,000 miles, tune-up every 25,000 miles, tire pressure every other fill-up—everything by the book. Today, news cars are equipped with sensors and computers that tell you when and what to do to maintain the car for peak efficiency and performance.

For years, fossil generating stations also have been run by the book to ensure safe and reliable operation. However, remote sensor technology and big data are transforming the way we do business. Our business partner, PSEG Fossil, asked us, “What would be a better way to work?” The answer: If they could know before a part failed or if they could replace it only when it might fail, they could operate plants more often and reduce the cost to operate.

PSEG Fossil operates 15 fossil generating stations across five states, utilizing different fuel and other operating parameters. In order to take advantage of economies of scale, PSEG IT, along with our external partners GE (M&D Solution), OSI (Pi Software) and DLL Solutions (PI Integrator), was tasked with creation of a centralized monitoring station for the entire fleet. The idea was to ensure that knowledge gained from individual plants was not lost to geography, time or other barriers around the operation and maintenance of individual plants.

The Monitoring & Diagnostic (M&D) Center is an enterprise-wide analytical platform that supports near real-time, un-interrupted data-flow using the latest and most advanced software (GE SmartSignal and Efficiency Maps) that provides consolidated and centralized data and enlarged dashboard visualization. A high-availability environment is established for continuous backup of data and feed to central PI servers and subsequently to the M&D system. It required a redesign of a back-end communication network, re-visioning of information security policy, and architecture to deliver the real-time performance and analytical dashboards using cloud computing. The initiative encompasses asset models and efficiency maps from GE SmartSignal, CycleWatch, Dream Reports and Efficiency Maps.

The plants previously followed a planned maintenance schedule based on a “by the book” set of parameters for the operating machinery. Parts are ordered and replaced based on this schedule. The M&D Center uses more than 10,000 data points to monitor each plant and its components. Based on predictive modeling, it will recommend maintenance to ensure the plant is operating at maximum efficiency or that it will be ready to run when called to service. These data points, more than 90,000 across the currently implemented plants, build a big data pool that is being used to increase accuracy of the predictive model. A plant no longer needs to use an antiquated time-based formula to determine which parts to replace; this will reduce outages, reduce part costs and maximize plant output.

 ​Understanding where resources need to be positioned across the fleet is a tremendous improvement over past operating procedures 

The PSEG M&D Center was created to deliver on the promise of a new way to work: A single source of truth where executives, energy trader/schedulers, asset managers and plant managers can reliably find the information they need to make timely decisions regarding the operation of the fossil generating fleet, as well as the ability to look at performance data in real time by plant, by fuel, by fleet or by a parameter of your choosing, ensuring that decisions are timelier and more accurate.

No more local knowledge. Sharing information is how we do business at PSEG, but timing is often dictated by factors beyond our control. Automation allows the data to be available to all to ensure that the most effective course of action can be implemented based on the total knowledge base of the fleet. A part fails, a switch trips or a parameter is out of norm—all of these factors will potentially lead to a corrective action. The shared knowledge may recommend changing not just Part X that failed but also Part Y based on the sensor input and historical knowledge of the frequency of these parts failing together. The use of performance data from multiple plants and power generation units allows the models to be self-improving which will lead to more precise failure prediction and better efficiency predictability.

Performance

The M&D Center’s thermal model analysis application will constantly monitor unit performance to ensure that each plant is running at maximum available efficiency. This increase in unit performance will result in a direct increase in revenue per unit. In addition, if a plant is not running at maximum efficiency, then a potential shift in generation can be made to a different plant that is running more efficiently to meet the system demand until corrections can be made to bring the original plant back to expected production levels. This ability to shift generation as demand changes is increased significantly when you can predict how power generation units will perform.

Maximize Resources, Reduce Cost

Understanding where resources need to be positioned across the fleet is a tremendous improvement over past operating procedures. This allows the fleet to operate at peak capacity. The M&D Center provides this ability, creating a competitive advantage in a challenging power producing marketplace.

• Better power generation unit performance allows each unit to produce more power; more power means more revenue.
• Ability to predict failures before they occur, results in increased power generation unit up-time; up-time equals more opportunity to generate revenue.
• Ability to increase visibility into plant performance and maintenance requirements allows the operations managers to share resources across plants; reduced resources means reduced operational cost.