Industrial internet turns Big Data boom into operator profits
GE Hitachi Nuclear Energy (GEH) and Exelon Generation’s Industrial internet platform project is using the latest analytics and sensor technology to optimize plant maintenance and increase asset life, Hao Dinh, Innovation Leader at GEH, said.
Across the energy industry, the integration of advanced analytical software into plant operations is allowing genenrators to profit from soaring Big Data activity by proactively optimizing performance, reducing outage durations, and hiking return-on-investment for assets.
Big Data analytics have evolved to enable the processing of large data streams-- including temperature and vibration indicators-- to produce real-time analyses, or “insights,” that can be used to manage and potentially reduce costs of operations while adding value to services.
GEH and Exelon Generation, the largest nuclear operator in the US, announced in September 2015 their joint development of Predix Cloud, a digital solution package that predicts asset performance, improves decision-making and reduces operating costs.
The technology is already being used to provide predictive assessments of key power plant components such as turbines, Dinh told attendees of Nuclear Energy Insider’s Nuclear Operations and Maintenance Efficiency Conference in November.
"We're collecting real-time data, associated to how well that turbine is running, then running analytics on it so that we can forecast in the future the high probability of when that turbine or equipment is going to fail...That is what we are shooting for, taking all this data, running analytics on it to come up with powerful insights," Dinh said. Operators are provided the analysis to make better decisions, he noted.
Applying advanced analytics to the predictive maintenance of assets could save industrial companies such as water and power suppliers up to 12% on scheduled repairs and lower overall maintenance costs by up to 30%, according to the Industrial Internet Insights Report 2015, published by Accenture and GE.
For the global report, researchers consulted firms in the Aviation, Wind, Power Generation, Power Distribution, Oil and Gas, Rail, Manufacturing, and Mining sectors. The study found that 73% of firms are already investing more than 20% of their overall technology budget on Big Data analytics and more than 20% of firms are investing more than 30%.
“Moreover, three-fourths of executives expect Big Data spending to increase this year,” the report said.
The study also clearly showed the gains yet to be made in manipulating big data for predictive operational purposes.
Some 35% of firms said their Big Data analytics capabilities were focused on the analysis of operations, while only 13% of firms said their Big Data analytics were focused on predictive measures.
As well as focusing on predictive maintenance gains, the Predix platform is designed to optimize nuclear maintenance operations and mitigate delays by digitising and analyzing the wide range of information feeds during outages and establishing performance improvements.
Working the Net
The Predix platform is using Watchtower remote data control and Lighthouse real time software applications in pilot projects at Exelon plants, to more effectively manage and maintain assets and raise operational efficiency.
Exelon told Nuclear Energy Insider it is too early to provide an update on the progress of the pilots, but GEH’s Dinh told the Operations and Maintenance conference that the power of the Industrial Internet is already providing influential insights to companies across many industries.
“Predix is enabling our customers across GE, [including] healthcare, aviation, power and water, and now nuclear to take advantage of these insights to predict potential issues or failures and resolve them before they become issues,” Dinh said.
There is an array of definitions for the Industrial Internet, which is a conceptual space where physical and digital data merge (Big Data) and which enables the storage and analysis of vast volumes of information arriving from a variety of sources at different speeds.
Big Data analytics and the Industrial Internet have opened the way for nuclear operators to move from being reactive to proactively predicting and preventing maintenance problems, equipment failures and costly declines in operational performance.
The power and utilities industry is already being targeted by software developers of Big Data management and analytical engines such as the Pi System from OSIsoft, LLC, which captures, processes, analyses and stores real-time data and events.
New digital analytical platforms will need to ensure the data collected from nuclear plant equipment is reliable to avoid the ‘Garbage In Garbage Out’ pitfall in which computers do not filter unintentional or nonsensical data.
The data sets must be cleaned and appropriately organised so that meaningful knowledge is gained from the analysis.
Value in action
GEH and Exelon incorporated the Internet of Things alongside Big Data, advanced analytics and intelligent machines in the Predix package to enable real-time and reliable data streaming from sensors fitted to equipment.
In the nuclear plant setting, Watchtower systems collect and analyze sensing information from equipment to determine whether the probability threshold of that equipment failing has been breached.
The added value is in interpreting and acting on that analysis within the context of the specific equipment, and enabling operators to decide if and when preventative maintenance tasks are needed and whether such measures could prolong the asset’s life and yield a higher return-on-investment.
Exelon is also piloting Lighthouse within operational optimization procedures, specifically the assessment of key performance indicators (KPI), used by the World Association of Nuclear Operators and the US’ the Institute of Nuclear Power Operations (INPO), to rate a plant’s operational performance against industry accepted standards.
Using the flexible, expansive cloud storage and analytical digital platform allows the vast volumes of real-time sensing data to be analysed alongside historical operating data and produce an INPO score for the particular KPI under scrutiny.
Analysis of a declining predictive score can reveal the previous occasion KPIs behaved sub-optimally and the steps taken to remedy or mitigate the decline, so the operator can take action to avoid a poor operational performance INPO rating.
Global Unit Capability Factors towards performance indicators
In the above chart from WANO, Unit Capability Factor is the percentage of maximum energy generation that a unit is capable of supplying to the electrical grid, limited only by factors within the control of plant management. A high unit capability factor indicates effective plant programs and practices to minimize unplanned energy losses and to optimize planned outages.
The Industrial Internet has the potential to identify and account for threshold errors by allowing greater correlation and differentiation between warnings arriving through different data streams, which should reduce the number of false negatives and positives issued by each piece of equipment.
While the power of the Industrial Internet lies in its ability to reveal crucial operational information from a multitude of data sets, humans need to decide how that information can benefit and add value to the organisation.
The next step is the development of solutions to assist with the potential glut of insights generated by advanced analytics and support- the interventions that will offer the greatest operational value for the operator.
By Karen Thomas