TURBOMACHINERY OEMS, CONTROLS VENDORS AND SOFTWARE PROVIDERS ARE HARNESSING DIGITIZATION TO REDUCE COSTS, IMPROVE EFFICIENCY AND SPEED TIME TO MARKET
BY DREW ROBB
For decades, manufacturers earned their revenue primarily through selling equipment. Digitization is changing that by turning transactional sales into an ongoing relationship, said Enrique Herrera, Industry Champion, OSIsoft
“We’re seeing companies embed digital subscription services into pumps, tractors and other heavy equipment,” he said. “These systems can alert a customer to an impending failure and avoid a costly breakdown.”
Plants that implement digitization can lower operating expenditures and gain a better return on capital by reducing unplanned outages. Maintenance employees become more effective and vendors can increase revenue by selling value-added services. The consumer sector is rife with success stories, such as Amazon, Alibaba, Uber, eBay and Airbnb. These companies have figured out how to use information and analytics to dominate their fields.
This strategy is now finding its way into the turbomachinery industries. Companies, such as GE, Mitsubishi Hitachi Power Systems, Siemens Energy, Schneider Electric, Emerson, OSIsoft, MAN Diesel and Turbo, and ABB are creating digital value to respond to customer behavior. “A digital focus is critical for near- and long-term growth.” said Scott Bolick, Head of Product Management and Portfolio Strategy, GE’s Power Digital Solutions. “The solutions are real, it’s happening today and early adopters are seeing tangible outcomes.
Take the case of a poorly performing asset. Currently, it might be taken down for several days of troubleshooting to find out what is wrong. With digitization and analytics, it may be possible to know six months ahead of time what the issue is and take steps to either eliminate or minimize downtime. “There is a huge amount of data generated by sensors inside turbines and auxiliary equipment. Yet only 2% of the data generated is ever used,” said Bolick. “The goal of digitization is to gain more insight from this information. We have reached the tipping point for digital adoption.”
• The average power plant produces 2 TB per day of data
• Tens of thousands of wind turbines output a data point every 10 seconds
• 600 million smart meters are already in the field.
• Many other sources are generating a massive amount of information.
Yet hardly any of it is ever analyzed, said Bolick. The hope is that as much as 75% of all unplanned downtime could be eliminated by gaining greater insight from this information.
“The Industrial Internet of Things (IIoT) enables predictive maintenance and proactive replacement of assets,” said Thomas Trepanier, Senior Vice President of Product Management at ABB Enterprise Software. “It makes it possible to optimize power generation for various market conditions through real time analysis, thereby reducing operating costs and increasing the fleet’s overall output.” There are already 13 billion connections worldwide that comprise the consumer internet, also known as the Internet of Things (IoT), said Trepanier.
IIoT makes up part of that total via sensors embedded in turbomachinery and its supporting auxiliary equipment. But that number is expected to skyrocket. Many costly failures and unscheduled shutdowns could have been prevented by analyzing available data, said Moritz von Plate, CEO of condition monitoring vendor Cassantec, a provider of condition-based prognostic solutions for industrial asset management.
But changes must occur before that can happen. Industrial data inside proprietary software programs, PLCs or other systems tend to be tied up in information silos. This lack of information sharing has to be overcome if the vision of digitization is to be fully realized. Software specialists must be trained on how the gathered insights relate to the real world of the turbine, compressor, pump, or the entire plant. “You have to combine analytics with engineering know-how to achieve the best results,” said von Plate.
Among the many inefficiencies that exist among current maintenance practices, von Plate cited the wind turbine industry where the assets are dispersed. Maintenance requires sending people and expensive cranes. Good diagnostics could greatly reduce costs by knowing in advance which wind turbines need the most attention so better planning of the deployment of people and cranes can be carried out.
Inefficient power plant fans are another example. The fan manufacturer’s ’s advice was to shut down the problem fans. But that would have led to revenue losses. “The data indicated that the fans could run for another year without a breakdown, although there would be some deterioration in their performance,” said von Plate. “With knowledgeable plant staff armed with full data and good analytics, you have the tools to make correct decisions.”
Another aspect of digitization related to the demographics of the industrial workforce. Going back two or three decades, power plants and utilities retained large, competent in-house maintenance teams. Today these teams are down to bare bones. To make matters worse, said von Plate, at half the utilities, more than 50% of current professional are due to retire within five years. “We have to document the tacit knowledge of these asset operators and automate these actions so that future generations can benefit,” said von Plate.
But the shift to digitization will not necessarily go smoothly. Manufacturers will need to integrate software, supervisory control and data acquisition (SCADA) and other systems into their products. This requires applications that allow users and engineers to more easily access data and analyze the vast amounts of data that accumulate. That is why so many are touting the benefits of various platforms.
These platforms solve the integration dilemma in the same way that the iPhone’s platform makes it easy for millions of apps to run on smartphones. Application developers use the Apple platform and write all their software to run on it. This greatly simplifies the integration picture. Similarly, in the industrial space, various platforms are emerging:
Predix is built for developing, deploying, operating and monetizing Industrial Internet applications. Through acquisitions, investment in startups, internal development and partnerships, GE is putting together a platform where the different industrial software components can converge and be analyzed.
The general concept is that Operational Technologies (OT) and Information Technologies (IT) need to be more tightly coupled to drive operational efficiencies, cost reduction and enhance the user experience. GE partnerships include IT heavyweights, such as Oracle, Intel, Cisco, Capgemini, Deloitte Digital, Infosys and Wipro Limited.
“Intel and GE have been working on IoT standards and initiatives for several years to accelerate deployment, and those efforts are now paying off,” said Doug Davis, senior vice president and general manager of IoT at Intel. Elements such as Asset Performance Management (APM), Operations Optimization (OO), Business Optimization (BO), Cybersecurity, Edge Computing and Advanced Controls are included in Predix.
The immediate result of these initiatives are the Digital Power Plant and Digital Wind Farm. GE has already signed 15 Digital Power Plant customers including Exelon, RasGas, Bord Gais, Crestwood, PSEG, Sapphire, Engro and Saif. Software Modeling is used to incorporate high-fidelity, physics-based models to create digital twins of plant components. For example, performance, combustion and life models can be used to operate gas turbines closer to design boundaries to meet a desired outcome in terms of efficiency or output.
“Power plant operation involves many trade-offs, such as output, availability, reliability, maintenance costs, flexibility, wear and tear and efficiency, but you can’t have all of them,” said Bolick. “So you have to use modeling and analytics to determine the best way to achieve immediate and long-term goals.”
OSIsoft’s PI System is used by about 65% of the Fortune 500’s industrial companies. It captures data from sensors, machinery and other pieces of equipment to provide insight into ongoing operations. The customer base includes 75% of U.S. nuclear reactors, Kellogg’s, the City of Halifax, Nova Scotia (saving 38 million liters of water a day because of leak detection), over 1,000 utilities, eBay, Major League Baseball, Shell, Chevron, Saudi Aramco, Caterpillar and Johnson & Johnson.
“Over 1.5 billion sensors in more than 127 countries are feeding data into PI System servers,” said Herrera. “Rockwell Automation integrates PI System into its control products, as do Flowserve, Dell, Cisco and Mitsui.” OSIsoft’s software makes it easy to port data to applications, such as the ESRI ArcGIS mapping system or the Microsoft Azure cloud, as well as visualization tools that allow people to analyze data in the PI System.
TransCanada CPG, for example, has created dashboards that gives its executives real-time insight into gas flows, pricing, potential problems and other information. Because they had this information, Columbia could reroute gas deliveries during the polar vortex in 2014 and avoid disruptions, added Herrera. Since 2011, the company believes it saved over $9.8 million in event prevention.
The Ovation platform has been steadily marching forward in functionality for many years. It recently added high-fidelity embedded simulation (Figure 4). “As users seek improved performance in areas such as unit and fleet availability, flexible operations and workforce effectiveness, virtual simulation tools provide them with unprecedented insight into their operations,” said Bob Yeager, president, Power & Water, Emerson Automation Solutions.
In the virtual power plant of the future, the high-fidelity Ovation simulator runs in parallel with the Ovation control system. Operators can test control logic changes in advance based on real-time plant data to ensure these changes won’t disrupt operations or cause a forced outage. Other Emerson technologies, such as pattern recognition, Pervasive Sensing and analytical tools will enhance the virtual experience. The Ovation system is used in thermal power plant applications, and as a turbine control system on steam turbines, gas turbines, boiler feed pumps and compressors by OEM’s or as a retrofit system.
“A developing trend we’ve seen is the operator/owner’s desire to identify when a machine is deviating from normal operation and will soon require maintenance — well before the first alarm is received,” said Yeager. “The goal is for all maintenance outages to be planned, completely avoiding unplanned or forced outages and the significant disruption they cause.”
For this to be possible, once a machine event is detected, it’s critical to predict how long the machine will be able to run until it reaches alarm level. Doing so provides a window in which the outage can be planned for with reasonable certainty. Accordingly, Ovation has integrated machinery health, which includes an algorithm called Peak Vue, aimed at predicting machinery failures. But digitization goes beyond the machine level by helping determine the optimal process conditions for running an entire plant or the optimal startup sequence in order to maximize performance and costs, while minimizing wear and tear on the equipment.
ABB’s Enterprise Asset Management
ABB stresses that digitization enables tools like enterprise asset management (EAM) software to bring more efficiency to the work cycle through capabilities, such as electronic work packages. This makes it easier to take assets out of and back into service more rapidly. Enhanced efficiency tied to an EAM system with features such as electronic work packages can lead to expected cost savings for a two-unit power plant of about $3.5 million per year, according to DataGlance. ABB is assembling its ABB Ability platform to connect facilities to the power of the IIoT and turn data insights into action. Trepanier cited European utility EDF reducing outage windows from 75 days to 30 days due to software efficiency and digitization.
MHPS’s Partnership with OSIsoft
Instead of erecting its own platform, MHPS has formed a partnership with OSIsoft to leverage the combination of PI System and MHPS’s power plant design and operational expertise. “Most of our customers already use the PI System so why tell them there is something else they have to buy and train their 20 people to use,” said Paul Browning, President and CEO, Mitsubishi Hitachi Power Systems Americas. “OSIsoft’s ecosystem of software encompasses over 250 partners providing services and applications for the PI System to facilitate process efficiency, asset health and energy management through real-time decision support and advanced analytics,” he said. “MHPS already use the PI System and many of those applications for our own monitoring and diagnostics services.”
This partnership highlights an easier pathway to digitization. According to an MHPS survey, almost 75% of power plants in the U.S. are already remotely monitored, with approximately two thirds of those having at least one major system remotely monitored by the original OEM of that equipment, usually a gas turbine. Outside the U.S., slightly over half of plants are remotely monitored. Further, about 70% of GTs, and over 50% of steam turbines are remotely monitored. For those already comfortable with remote monitoring, therefore, it is not too much of a leap to embrace the digital platform concept of not just remotely monitoring a few major systems but the entire process and the major systems within. The PI System will serve as the core for analytics services MHPS will use at power plants world-wide, with cloud-based and onpremises options. In addition, the PI System will continue to form the digital data infrastructure for existing total plant remote monitoring and diagnostics services, both for MHPS and other manufacturers’ equipment in gas turbine combined cycle and conventional steam power plants. MHPS adds several of its own elements as part of its digital power plant initiative. MHPS Information and Communication Technology (ICT) combines digital technology with customer collaboration and total plant design, operation and maintenance experience. It offers a range of data-driven solutions to meet increasing needs for operational flexibility, more challenging duty cycles and reduced maintenance as the power industry evolves. High bandwidth data communications, data security and analytics are other important elements, so engineering analysis and fleet-wide comparisons can take place in real time without a large team of experts located at each power plant. In parallel, sensor technology is evolving, steadily increasing the quality and quantity of available operational data.
“Monitoring of a large fleet with advanced pattern recognition software becomes less manpower intensive since the software is constantly performing analysis and alerting personnel when anomalies emerge, or ideally inform personnel of anomalies before they happen,” said Browning.
Schneider Electric’s Demand Side Platform
Schneider Electric’s cloud-based software services sit on top of new and existing microgrid control systems. This Demand Side Operation platform provides forecasting, ancillary services, optimization and weather services to make consumption, production or storage decisions easier, said Drew Gravitt, National Sales Manager for Economic Optimization at Schneider Electric. The platform incorporates connected products, advanced algorithms, analytics and controls in a service model. Schneider is partnering with the likes of Duke, Intel and Microsoft as part of this initiative. Schneider Electric also leverages the Microsoft Azure cloud platform to deliver its digital services, apps and analytics on computers and mobile devices. One such digital service, Conext 2 Advisor, is a secure monitoring and control solution for managing and optimizing the performance of solar power plants.
MindSphere is a cloud-based approach to the development of industrial applications, and analyzing large volumes of production data. It provides a foundation for applications and data-based services from Siemens and third-party providers. “MindSphere offers customers a development environment in which they can integrate their own applications and services,” said Aymeric Sarrazin, Vice President of Strategy, Siemens. For example, Siemens offers a zero degradation service. Although any plant suffers performance degradation over time, Siemens works with its customers to reduce degradation. It achieves this by conducting modernization and upgrade projects to neutralize any performance loss. Siemens also offers digital services, running on MindSphere, which offer plant-specific analysis to optimize performance and reduce downtime. The MindSphere ecosystem integrates, manages, analyzes and visualizes data to enable experts to predict and prevent failures, and uncover opportunities for performance improvements, as well as energy and cost savings. Different roles While there are many platform options, most companies will be happy to develop their own software and couple it with other tools to achieve the goals of digitization.
MAN Diesel & Turbo, for example, has been using such tools in its HOFIM electrically driven subsea compressor, transmits real-time performance data that can be checked from a remote control station. “We deliver all machines and trains ready for remote diagnosis,” said Harald Stricker, Senior Vice President Engineering, Turbomachinery, MAN Diesel & Turbo. “You now have a thousand signals from a machinery train, compared to about 50 signals some decades ago. The next steps are clearly about more intelligent functions and interconnections.”
A few decades ago, automobile drivers would tinker with their vehicles every weekend, check fluid levels, tire pressure and adjust the carburetor and spark plugs to ensure optimum operation. “Nowadays, a warning light tells you that an inspection is needed,” said Terry Maxey, Managing Director, Global Power Generation Lead, Accenture. “A computer provides the diagnosis so you don’t even need to open up the engine. That’ where the industry is going.”