TRANSFORMING MAINTENANCE, REPAIR AND OVERHAUL (MRO)

DIGITIZATION, ADDITIVE MANUFACTURING AND ANALYTICS ARE

REVOLUTIONIZING THE INDUSTRY

3D printers at a Siemens facility in Sweden

By Drew Robb

The annual International Gas Turbine Institute (IGTI) Turbomachinery Expo often has a dominant theme. For 2018, the focus turned to digitization, additive manufacturing (AM or 3D printing) and analytics. Each day’s keynotes and an entire track looked at every aspect of the digital revolution. The overriding message: there is value in adopting digital technologies, insight to be gained by investing in analytics, and better performance to be found by adopting additive manufacturing. However, attendees were cautioned to avoid overgeneralized digitization efforts. They were advised to find high-potential value targets to showcase the benefits.

Experts from companies, such as Delta Airlines, Lufthansa, GE, Siemens, Mitsubishi Hitachi Power Systems, Bechtel, Strategic Power Systems, Ansaldo Energia, PSM, Equinor (formerly Statoil), Air France, MTU Aero Engines and Pratt & Whitney were front and center in the digital push.

Digital keynote

The opening keynote set the tone for the week. Jaroslav Szwedowicz, ASME’s Gas Turbine Segment Leader, introduced the concept of Maintenance, Repair and Overhaul (MRO) in the Light of Digitization. He noted trends impacting the gas turbine (GT), such as the fluctuating power output of renewable generation. As a result, GTs need to be more flexible, being able to operate in peaking and partial load modes, as well as base load.

“IGTI is involving more stakeholders in MRO and digitalization,” said Szwedowicz. “We have also added an AM track for the first time.” Before introducing the guest speakers, he offered a history lesson, and part of the reason Lillestrom in Norway hosted this year’s show. Some 115 years ago, inventor Jens Elling built the first ever GT in Lillestrom. This 1903 model could provide 8 kW of power operating at 12,000 rpm. It comprised a six-stage radial compressor, intercooling, variable diffuser vanes, steam injection, and a single stage centripetal turbine with a turbine inlet temperature of 400°C. “All these years later, the GT industry continues to innovate,” said Szwedowicz.

Dr. Zuozhi Zhao, CTO of Siemens Power and Gas, began with a more recent historical reference. Over the past five years, he has observed a radical shift in the industry. Renewables and the introduction of batteries are bringing about challenges and opportunities in MRO. Digitization is being looked upon as an answer to changing industry dynamics. However, it is a broad term that encompasses big data, cloud computing, edge computing, analytics, automation and more. As such, there are many different interpretations of digitization.

Zhao said it is important to define clearly what you are talking about to know how to harness digitization properly. He was positive about the future. Yes, automation, robotics and artificial intelligence (AI) will take over many functions performed by people today, he said. But world electricity demand is going to double by 2045. That will generate plenty of new work opportunities. “Digitalization is required to address the complexity of modern energy systems,” said Zhao.

Billions in investment

Siemens has spent over 10 billion euros over the past decade in acquiring and developing software for digitization and automation. In the past year, the company has spent 1.2 billion in digital R&D. Siemens, a company known for its hardware prowess, now boasts 25,000 software engineers. One result of increasing emphasis on software is Siemens MindSphere. It is a cloud operating system that connects plant systems with digital, Internet of Things (IoT) and analytics applications.

“High-quality sensors are so cheap that you can deploy as many as you want,” said Zhao. “When you combine them with modern compute power, a cloud infrastructure and digital twins, you can enable a lot of things.” AM, he added, makes the digital twin a reality. He believes the digital twin is the answer to how to run plants and GTs with so many starts, stops and cycling while maintaining reliability.

A digital twin is an organized collection of physics-based methods and advanced analytics that is used to model the present state of every asset in a digital power plant. It can help perfect designs, improve plant operations and determine the best way to build power plants. This is achieved by model-based simulation of various scenarios to isolate the ideal approach. They can also be used by taking feedback from the operational side and looping back to model the reason for problems. This can lead to much better designs.

For example, Siemens has been using this technology to create 3D printed turbine blades that can run at 1,000 mph and temperatures at 1,250°C. These blades have a greatly improved internal cooling geometry designed by AM. The company is also using AM to provide spare parts on demand. Siemens has been active in 3D printing since 1989. It has integrated 3D printing with computer-aided design (CAD) software, controls technology, automation, and turbomachinery systems to be able to digitally model all processes.

It is deploying the technology to manufacture tailor-made products and create structures that cannot be achieved with conventional methods. Components currently manufactured in solid form can be produced in an additive manufacturing process with the required cavities already integrated and as strong as their current counterparts. They are also lighter and require less material. The benefits of AM are said to be the ability to bring products to market, reducing R&D and prototyping loops by 75%, and repairs being performed in a tenth of the usual time. Further, metal powder from the process can be recycled and used again, which leads to a reduction in scrapped material.

3D printed burner nozzles have already been made available for SGT-700 and SGT-800 GTs. They allow co-firing at higher combustion temperatures thanks to an improved burner tip design. Burner fronts were traditionally manufactured with 13 components and 18 welds. The AM versions consist of one component and two welds. They are also ten times faster to produce.

Zhao expressed enthusiasm about the way analytics capabilities are developing. Instead of merely describing past events or providing data-based proactive decision support, they have evolved to the point where analytics can explain why events happened, help to mitigate risk and predict what is likely to happen next with components, equipment, plants and entire fleets. “Informing plant managers, fleet managers and top management about how to operate in the future is the real value of analytics,” said Zhao. “Digitalization can assist in rapidly changing operating profiles based on price fluctuations, keep fleet availability high, lower risk and help managers make better decisions.”

3D printed burner fronts from Siemens

GE software

GE’s digital transformation began a few years ago to enhance physical assets with digital technologies. The company’s software business now augments how it services hardware around the world. “It can be difficult to make change happen due to the traditional way of operating,” said Russell Irving, Digital Twin General Manager and Chief Engineer, GE Global Research. “The digital side must never forget who we work for: we are the enabler, but the MRO network is the owner we serve.”

As such, it is up to those dealing with digitization initiatives to find out what customers want. This is typically maximum uptime at lowest cost, or greater profitability. The goal is to facilitate this through software and analytics. But a cultural shift is required. Irving said this is harder to achieve than technology development. But the rewards can be significant.

He cited the example of the gas turbine development cycle. It used to take 10 to 13 years to develop a new GT. Digitization has shortened that to three years for certain models, using GE Fastworks. Irving also talked about using digital twins for each asset. This, he said, increases productivity and attains better business outcomes. But it is better to begin on a small scale rather than trying to model everything. “Focus on outcomes, see what needs to change and make a twin of one module or one component that is the key to achieving that objective,” said Irving.

Gains include being able to determine if an asset can be run hotter, when inspections should be done and when to replace components. According to Irving, this can save tens of millions in unnecessary service overhauls per customer. “As we know more about the asset, we can have higher confidence on when and how to act with speed,” said Irving. “We can really change the face of MRO if all assets are connected and we attain system-level consciousness.”

Digital twins of turbomachinery can aid in design, manufacturing, and maintenance. He ended by comparing how you buy a pair of sneakers online compared to the industrial purchasing process. In the consumer world, you can track your order all the way from packaging to final delivery. The current service capabilities of the industry cannot do that. It rarely is possible to tell the customer where their parts are, and even when they will arrive. Digitization is the way to change that.

Digital panel

A panel on digitization showcased digital challenges and opportunities. Frode Abotnes, Vice President of the Technical Multifield Center at Equinor (formerly Statoil, the Norwegian state-run oil and gas agency), said digitization has been an enabler of his company’s high-value and low-carbon strategy.

One problem he faces, though, is having a lot of data but no real way to harness and analyze all of it. The information sits in information siloes that do not communicate. Another issue his company is addressing is how to find graduates with experience in both rotating equipment and software programming. He envisions the MRO field shifting in a similar way as the lift-sharing phenomenon that turned the taxi system on its head. “We can learn from the consumer world how to give better service and develop smoother digital work processes,” said Abotnes.

Pascal Decoussemaeker, Senior Product Manager at GE Power in Switzerland, and the European Turbine Network’s Asset Management Chair talked about root cause analysis. Sometimes there is a need to look back many years. That is why it is vital to have enough data available, yet you do not know exactly what data you might need. This means you must stockpile a lot of information from multiple sources. New computing platforms and the cloud make it possible to retain sufficient information to be able to perform insightful root cause analyses.

He believes that any cultural conflicts that may exist will resolve gradually as more digital natives enter the workforce. They will expect companies to operate like the consumer sector, and drive change. Gone are the days, he added, when you simply followed the GT OEM manual. Now you must optimize and adjust using digital asset management to reduce cost. The downside is an increase in risk. This is giving rise to outcome-based MRO where a specific amount of availability, reliability and performance is paid for as part of the contract.

“Machine learning lets us perform analytics rapidly and harness pattern recognition techniques to observe trends,” said Decoussemaeker. “Instead of needing to refer data to experts for interpretation, we can use algorithms to reach conclusions faster to assist operators.” How about the cost of digitization? He accepted that it required investment. But he considers that such initiatives should be effective enough to bring the overall cost down so there will be no need to pass on any costs to customers.

Shawn Gregg, General Manager Propulsion Engineering at Delta Air Lines, emphasized that domain expertise is the foundation. Digital technology lays on top of that and adds value. But it cannot be done without deep understanding of the turbomachinery and MRO domain. That is why Delta takes newly recruited graduates and puts them on the shop floor for a few weeks to tear apart and rebuild engines. They need to understand the physical reality of what they are dealing with if they are going to create workable models and productive applications.

Analysis & optimization

Sal DellaVilla, Founder and CEO of Strategic Power Systems (SPS), presented on data capture and analytics in a competitive power generation market. “We are in a new world of big data and analytics,” he said. “Owner operators demand efficiency, flexibility, durability, environmental friendliness and reliability from OEMs.”

He observed that each OEM has a different standard on how to test and validate machines. And the standard is changing in the face of digitization and new modes of plant operation. In the past, perhaps 100,000 hours were needed for a turbine design to be fully validated in the field. Some now promote 8,000 as being enough. He stressed that digitization is nothing new. Data and analytical tools have always been in use. But now the pace is accelerating. We can look at assets in near real time in detail to see what is really going on.

SPS offers the Operational Reliability Analysis Program (ORAP). It captures data from globally operating power plants. This provides a picture of the performance of the entire fleet of turbines. Individual plants can see how their numbers compare to the fleet. He gave the example of a plant in Southern California running aeroderivative GTs, which experienced “an event.”

The OEM’s remote monitoring and diagnostics facility then called to notify the plant that an event had occurred. They also wanted to know the cause and what action was taken. In this case, digital data failed to add any value. What is really required is predictive analysis. The plant operations and maintenance team addressed the event and returned the unit to service. “Operators these days are being fed enormous amounts of data,” said DellaVilla. “We need analytics to summarize, trend and provide tentative conclusions to help operators act rapidly.”

Unless action can be taken at the plant, he said, monitoring and diagnostics are of limited value. It is all about how quickly a fault may develop and what can be done about it in enough time to prevent problems. With the need for GTs to be available to start at short notice, ORAP numbers favor aeroderivatives which have average start times of 6.2 minutes compared to heavy duty GTs, which range from 17 to 30 minutes, depending on which class. Further, ORAP numbers indicate that recent changes to operating duty are exerting an impact on part replacement and maintenance patterns.

Michael Winter, Pratt & Whitney’s Chief Engineer for Technology, called for higher precision in manufacturing. He views AM as a tool to realize that. “If we can improve quality within 3D printed microstructures, we can achieve greater precision without incurring exorbitant costs,” he stated. “The form and fit of 3D printed parts is easy; it is not so easy to obtain the right function. We have to better understand the microstructure to handle the standard of AM parts.” His company is already putting AM parts into production aircraft engines. He listed benefits such as low-cost manufacturing of complex components, rapid prototyping, easier tooling and faster repairs.

Christian Verhoeven, Technology Executive New Product Introduction & Additive Manufacturing, GE Power Services, extended the list of AM advantages to include higher performance and efficiency, advanced cooling, lower part counts, weight reduction, faster time to market, lower lead times and the elimination of defects in assembly. The CT7 jet engine, for example, has modules that have had the number of their parts reduced from 300 to 1. In another engine, 855 parts were consolidated to 12, with a 5% weight reduction.

“Power turbines are benefitting from revolutionary cooling designs,” said Verhoeven. “We are now putting more 3D printed components into the hot gas path of the turbine.” The GE H-class, for instance, has been upgraded with AM parts. This brought about a 5 MW boost and an efficiency gain of 0.2% in simple cycle. In combined cycle, this rose to 8 MW and 0.4%, respectively.

Mitsubishi Hitachi Power Systems (MHPS) is another company investing heavily in 3D printing and digitization. Masahito Kataoka, MHPS General Manager for Large Gas Turbine Engines, said the latest parts it has developed include ring segments with channels, vanes with closed cooling, blades with complex channel cores and a new blade configuration, as well as combustor, compressor and seal parts. “We are applying AM to our gas turbines to create performance and manufacturing improvements,” said Kataoka.

This evolution, though, is being inhibited by size restrictions in 3D printers. The industry needs bigger machines, which are now under development. Another challenge is materials. Traditional casting materials and alloys are not useful for AM. A search is on for the best metal powders to produce consistent 3D printed parts. “These technologies are critical in boosting performance and achieving 65% combined cycle efficiency,” said Kataoka.

Vladimir Navrotsky, CTO of Power Generation Services at Siemens, said AM parts have improved the life expectancy of the combustion system in the SGT-9000 HL series, while reducing NOx levels and achieving more rapid design validation and reduced NOx levels. “Our development time for prototyping has been reduced by 75%,” he said. “We are now working on stator parts for gas turbines, which will lengthen the time between overhauls.”

GT lifecycle

Another panel delved into the gas turbine life cycle through a data analytics lens. DellaVilla spoke again, this time zeroing in on data ownership. He made it clear that the data generated from power plants is owned by the plant. OEMs and companies, such as SPS, share data. But it is essential to protect the rights of the data on behalf of the owner. “We need to develop a data-sharing community of GT owners to look at blades’ health, temperature, conditions, repair cycles and predict life,” said DellaVilla. “If you have a big problem, it is useful to know who else is having that issue.”

But a communication gap often exists between operations & maintenance personnel and plant owners. Those on the ground floor may wish to extend the scope of an overhaul. Owners want to get the facility running again and make money. They hope to avoid spending more money and adding time. This disconnect can lead to inefficiency. Digitization and analytics can assist in resolving such conflicts. But only if they are carried out with a full understanding of the mission of the plant, the physics of machine operation and the necessities of maintenance.

“You can’t do AI or predictive analytics without that domain knowledge,” said DellaVilla. “On the other side of the coin, if you don’t develop a sound business case, digital technologies might end up a bad investment.”

Perspective on how this applies to airlines came from Rudolf Pariso, Chief Digital Officer for Air France Industries Maintenance. He saw many facets to digitization. In the consumer space, he said, companies need a very good B-to-C portal. He believes the B-to-B side must follow that example to enable seamless interaction across the supply chain and make it easy to do business.

He expressed reservations about the way airlines currently interact with OEMs on engine matters. This entails sending in forms with observations from borescopes and other inspection data. If a crack is spotted, its specifics are communicated to the OEM to see whether that engine can continue in service or not. It can be a slow process to obtain a decision. Meanwhile, an engine sits on the ground idle. “We need to be faster between stakeholders,” said Pariso. “Another issue can be part availability: the unavailability of a small part, such as a cheap bracket may result in a $20 million engine being grounded.”

He considers that digitization can improve supply chain operations, reduce inventory where desired and improve profitability. But the industry is not used to investing heavily in this direction. It is more oriented to building facilities and adding physical assets. “We need to develop these new skills and combine data science with the engineering disciplines,” said Pariso. “Work scoping in ten years will be very different due to digitization and will bring a lower cost of ownership.”

But challenges remain. In the past, the requirements for in-flight data collection were modest: only a few snapshots of flight parameters during the journey. Today, constant data collection is the order of the day. That equates to five orders of magnitude more of data collection, said Pariso. But the rewards are substantial. Air France has decreased the number of operational interruptions by 50% and has lowered its component inventory by 10%. It is now experimenting with AI, using 10 years of historical supply chain data to streamline part delivery and smooth engine repairs.

Doctor Atul Kohli, Senior Fellow for Heat Transfer, Analytical Methods at Pratt & Whitney surprised the audience by citing a figure of $9 billion. That, he said, was the amount his parent company was investing over a five-year period on digitization. But complexity had to be understood, such as how engine conditions vary due to ambient conditions, air quality, temperature, pressure, particulate and other factors. It is no small matter to capture all that data accurately, let alone harness it effectively in analytics. “Engineers should not be waiting for numbers or struggling to compile data,” said Kohli. “They should be spending their time deriving insight from numbers assembled through digitization.”

MTU Aero Engines has already integrated design and maintenance data to bring about greater cooperation between both disciplines. “We have established a database for operational and part data that is available to our designers,” said Friedhelm Kappel, MTU Aero Engines. “When we inspect parts, we feed that data back into the design phase.” Easier said than done. Teething troubles include finding a way to connect a mountain of data with specific cases and findings. That, he added, is where AI comes in to analyze data automatically.

Benoit

Digital transformation

The digital theme was continued on the final day of the show. Jeffrey Benoit, Vice President of Product Management & Marketing for PSM Ansaldo Energia Power Services explained that Ansaldo Energia digitally monitors over 300 GT’s, ST’s and generators from various OEMs around the world in support of long term service agreements. The company has developed a digital solution that autonomously tunes the combustion system of industrial frame gas turbines.

Manual tuning is time consuming, can create a lean blow out situation, emissions can go too high and that can result in fines. PSM’s autonomous tuning system, on the other hand, is said to increase reliability and reduce cost. It captures information from successful and unsuccessful tunes to optimize tuning. Under this system, the number of tunes per week is higher initially, but falls sharply over time as the system learns the right way to tune that machine.

“Digitization is the way to manage complexity, transform data and maximize asset performance,” said Benoit. “In the case of AutoTune, it allows you to move to the margins, safely, and gain more operational flexibility while maintaining emissions and stay within the limits of combustion dynamics.”

Google is an example of smart digitization, said Bernhard Kruger-Sprengel, Lufthansa Technik. Google gathers massive amounts of data from vehicles and cell phones but has learned to only use relevant data for its traffic analysis. It simplifies the process by focusing only on where a vehicle is and whether it is moving. It ignores everything else: type of car, fuel type, number of passengers, the color of car, temperature readings and thousands of other parameters to provide valuable traffic service. In aviation, however, things are a little different. You cannot rely only on one or two parameters. Data analysis alone does not answer all MRO questions. There are too many related data. It is a multi-dimensional problem.

“The big challenge is the relationships that exist between data, which means you cannot easily generate conclusions or predictions with AI,” said Kruger-Sprengel. “You have to reconcile these models with the physical world.”