SGA Awards recognizes individuals or work groups who have developed innovative processes or programs.
We created a tool to stand between our data analytics and asset management systems to create work orders based upon data analytics alarms and our asset management system.
The Event Frame Management (EFM) tool was created to stand between our data analytics and asset management systems to create work orders based upon data analytics alarms. The OSI PIsoft data analytics system tracks multiple sensors and data points from our compressor fleet and places them into a data historian. When a data point is outside of parameters it logs an event frame with an alarm that is sent to our reliability engineers/specialists. Inside the Event Frame Management tool, the recorded event frames are captured electronically and reviewed by the reliability team based upon criticality. If an alarm requires onsite technicians to repair or look further into the alarm, with the push of a button the EFM tool will create an inspection work order in our asset management system with criticality and timeframe for scheduling the inspection.
Work order information captured upon investigation helps refine data analytics. Reliability Engineers and Specialists no longer have to create work requests manually for these efforts and additional data from work order completion is helpful for review and future enhancements.
As a result of this EFM tool, we have significantly reduced the amount of time required to build work orders for onsite review of critical alarms. In addition, we are able to capture additional data in the work orders from onsite inspections pertinent to the alarm that we generated from the OSI PIsoft program. Our reliability team has welcomed the reduction in effort to generate work orders that were previously done by hand. Our field technicians also have a standardized format with a work order to review the issue called out by the PIsoft system. The field technicians also have an embedded problem code in the work order to help guide them in their onsite inspection.
With completed EA-identified work order data, Southern Star can now review PI Event Frames against field inspection data. EAM provides historical work order data that can be used to reveal repeat issues on a particular asset or solutions for similar assets based upon similar PI Event Frames. EAM work order data can also be used to refine critical setpoints that create PI Event Frames to extend predictive maintenance while still ensuring asset reliability and efficiency.
From a business perspective, we now have a complete loop of information regarding our reliability alarms. When an inspection work order is generated from the EFM tool the information and problem code from the event frame alarm is embedded in the work order. The technician uses this information to guide their onsite inspection. Preset codes for Failure, Cause, and Action to repair are part of the work order closing process for the technician. As event frames are reviewed against completed work orders that information is being used to find recurring patterns of failure as well as refine critical setpoints for alarm creation. In addition, data trends are being used to improve reliability on assets of similar kind and run parameters. We now see an increase in reliability across our entire footprint as a result of this innovation.
Our team internally developed a smart records search tool to improve the searchability of PDF records. The challenge has been that we have millions of records varying in quality, age, and styles; as a result, significant effort goes into finding useful records as some were not previously searchable. Through collaboration with our technical teams, there was an opportunity to use Machine Learning and Artificial Intelligence to improve the searchability of records. We started with a proof of concept that yielded positive results. Two main objectives within our design were using Machine Learning (ML) in order to categorize visually similar records, and Artificial Intelligence to extract printed and handwritten text from them.
By training the ML model and leveraging AI, we improved search results, optimized the review process and narrowed down to target records more efficiently. One prime example with significant information is handwritten notes where traditional OCR is unable to extract the text. Our team had previously attempted to find the same records without the smart records search tool but were unable to find the same rate of success. With the prototype now successful, we continued to develop and improve the search tool by adding and training more records to increase model accuracy. The response has been very positive and eye-opening from traditional methods used; this has shifted our team’s momentum to be more innovative in trying to tackle problems that impact our day-to-day work.
The primary benefit of the smart records search tool to our business is that we have been able to optimize the current records review process; significantly less time was spent on parsing through large sets of documents, which allowed our team to allocate more time to technical analysis and project optimization. With the tool now further developed and used operationally, this has resulted in over $10M dollars in cost avoidance to date for our company. In addition, the tool is also being leveraged to help other teams with their own respective document search efforts; this has promoted greater interdepartmental collaboration and transfer of knowledge.
Stress corrosion cracking (SCC) is a form of environmentally assisted cracking and in general is the result of stress, environmental and chemical conditions including pH, in a material that is susceptible to cracking. Circumferential SCC (C-SCC) occurs when longitudinal stress, typically from ground movement or localized bending, is the major stress component. Limited assessment alternatives available for gas pipelines make management of the CSCC threat extremely challenging, potentially allowing anomalies to grow to a severity impacting the safe and reliable operation of a pipeline.
Selective digging approaches have been successful in identifying circumferential SCC. However, this approach is insufficient to predict the severity of C-SCC or confirm that all C-SCC indications have been examined. Since the major stress that results in C-SCC is longitudinal or along the axis of the pipeline, critical defects are typically greater than 80% of wall thickness and hydrostatic testing, which increases hoop stress, is not an effective method for detection.
In 2018 Xcel Energy and Novitech Inc. undertook a multi-year project to develop an advanced MFL ILI system to reliably detect and quantify C-SCC in susceptible pipeline systems. The resulting ILI system can successfully discriminate C-SCC occurrences from other pipe anomalies such as metal loss, as well as rank C-SCC severity into three categories: subcritical, significant, and severe.
The resulting ILI system has been implemented on over 200 miles of natural gas transmission pipelines 6, 8, and 10-inches in diameter and has successfully confirmed 78 C-SCC anomalies out of the 81 identified by the tool (3 false positives). The ILI system has a POD and POI >90% for C-SCC >30% in depth and 0.8 inches of circumferential width.
This innovation provides natural gas pipeline operators a reliable diagnostic system for the identification, management and repair of C-SCC. Prior to the development of this technology no assessment options existed to allow natural gas pipeline operators to identify and repair C-SCC anomalies prior to them reaching a severity that impacted the safe and reliable operation of the pipeline.
For Xcel Energy, in one specific scenario, this technology allowed the company to return a pipeline to full operation that was previously isolated at a reduced pressure with the only alternative being to pipeline replacement.