Digital systems can assess turbocharger health and enable shore-based staff to determine maintenance schedules, increasing a unit’s usable life
For the engineer, the turbocharger is not only a critical component but also something of ’blackbox’. Its maintenance is often left to specialists from shore, either the OEM or a service shop. While ship staff can observe the operating parameters and gauge its performance, they often do not have a deep enough insight into the health of turbochargers to anticipate an imminent failure.
We asked a cross section of stakeholders for their thoughts on turbocharger operation and maintenance processes and whether they felt predictive maintenance could be a valuable aid – both as a means of obtaining a more complete picture of the turbocharger’s health and in scheduling maintenance based on an assessment of that health, rather than running hours. Some operating engineers expressed scepticism that computer software could properly determine the health of a turbocharger; others were more welcoming, seeing it as an inevitable technological development that will benefit the industry.
PT Shipmanagement managing director Patrick Toll believes the root cause of most turbocharger failures lies in engine maintenance: valves breaking off or carbon build-up due to poor combustion.
For him, the difference between the exhaust gas temperature at the turbocharger inlet and the outlet is a good indication of how the equipment is performing. The lower the differential, the poorer the turbocharger’s performance. He cites this is an example of condition-based maintenance conducted by the crew.
Mr Toll firmly believes digitalisation should not take responsibility away from the crew: “Some cruise ship companies are investing in vessel control centers at shore, but we believe that even if office has all the data, the crew should decide what to do.”
He does however see positives in data analytics, noting that while traditionally onshore office would receive an engine performance record only once a month. “Taking the data from the alarm and monitoring system and running it through an algorithm can give a sense of how the parameters are trending,” he explains. “We in the office can tell the chief engineer that he should check the combustion system in the main engine or the turbocharger. But how to act on the advice is the choice of the crew,” he says.
Mr Toll says digitalisation should provide the tools for the crew to make the right decisions, but not take responsibility away from them. “We hear about condition-based maintenance, then predictive maintenance – the topics are flying so quickly across the table,” explains Mr Toll, injecting a note of caution.
A question of trust
DNV GL discipline group leader Thomas Knödlseder reasons that any scepticism toward predictive maintenance stems in-part from the fact that it requires great expertise. “Only a few can do it,” he says “And even then [it only works] at the end of a long process of data acquisition, health assessment and context. You need to know a lot.”
Mr Knödlseder points to the fact that even once the data has been collected, perhaps over a two-year period, it still needs to be analysed. Only then can the analysis answer questions such as: “How soon should a particular maintenance be done? What if I change the operational pattern now? How will that affect the scheduling?”
Mr Knödlseder says the decision to go with a predictive maintenance system would be based on how critical timely maintenance was perceived to be by an operator: “[Do you want] maintenance at the right time, not too early or too late, [or] at the lowest possible cost?” he asks. “Data, in and of itself, does not amount to much,” he explains. “It’s a question of how efficient and capable the algorithms are.”
Identifying the root causes
Regarding the turbocharger itself, Mr Knödlseder says the accumulation of foreign material is a key cause of turbocharger failure. Such accumulation creates an imbalance on the rotor which can be avoided by washing. If dirt accumulates it creates uneven running, which could lead to a bearing failure. “Then you have a catastrophic failure; you have blades flying all around the engine room,” he says, adding that today’s turbocharger bearings come with protecting shields to mitigate this problem. Still, he warns: “From detection of this problem to the catastrophic failure, you have roughly two minutes. [All] operating engineers can possibly do is shut down the turbocharger at that time.”
Any early warning system would need to monitor the parameters of the turbocharger and try to identify markers and correlations allowing a prediction to be made prior to that two-minute window. “Too often, we find the reason for the catastrophic failure only after the failure has happened,” he adds.
Mr Knödlseder says the while turbocharger’s can and do fail, it is not possible to say if these units have a higher failure rate than other equipment on board.
Acknowledging that most failures are usually operational in nature, Mr Knödlseder says OEM often provides the best guidance. “If everything is done by the book then there is a high chance that failures will not happen,” he says.
Mr Knödlseder, however, acknowledges that if a design issue does arise, then class would act on it and take it up with the manufacturer. “A couple of years ago, there were some occurrences that were turbocharger-specific, but the makers reacted very quickly to take correction action,” he says.
He adds that using condition monitoring to support maintenance decisions is not a new concept, but notes that: “We want to ensure any assessment is trustworthy. We ask the basic questions: Who has done this analysis? Did they have the right data? Who collected the data, how often, and where? Having said that, we do approve condition monitoring programmes for vessels without insisting on invasive inspections and [this] has been successful.”
But would he trust a computer to make that assessment through an algorithm? Mr Knödlseder continues: “If I am going on a long drive and my GPS tells me that there is congestion ahead that would delay me by 45 minutes, I may decide to overrule the GPS even if it was correct. Why didn’t I listen to my own GPS? That’s because my experience has convinced me that I know better. I did not know enough about where its data came from. And if I had a bad experience with my GPS in the past, then I tend not to trust it,” he says.
“Many companies are offering such diagnostics based on concepts that have not been tried out before. They [need to] convince the old hands in the industry,” he adds.
ABB’s head of global service product management, Thorsten Bosse, considers predictive maintenance to be a highly ambitious target. He says that for more than 100 years the design of a turbocharger and its respective maintenance have been targeted to prevent failures occurring in the first place. Hence, experience around failure prediction is limited. Further, he says, the physics of failure can be very demanding and difficult to understand. Still, Mr Bosse is optimistic that over time, this will change, citing the example of a crack in material: “The time between the onset of crack propagation and potential rupture of a highly-loaded material can be extremely short. It is very, very difficult to predict this. The difference now is that we will have data available in the future – a lot of data and from different sources. We can use artificial intelligence, pattern recognition and statistical algorithms on the data to move towards predictive maintenance. Data-based intelligence will help us to move maintenance from preventive towards predictive,” he says.
For Mr Bosse, failure prediction is a lighthouse target in the journey from preventive to predictive maintenance, and in between there may be numerous approaches that can be taken. ABB’s Digital SIKO, for instance, uses a concept called ‘exposure-based maintenance’ for the turbocharger. It offers flexible, exposure-based exchange recommendations for the rotating components of a turbocharger to optimise operational costs and assure safety and reliability. The data collected through the monitoring system is assessed by Digital SIKO. It considers the history of the machine, taking into account the operational loads the turbocharger has seen. With that data, the actual exposure of the turbocharger to those loads can be understood, which can be different from the load profile assumed when designing the turbocharger.
Digital SIKO is a further development of ABB’s safety concept, SIKO. “The core is the turbocharger’s rotor, the most demanding component,” says Mr Bosse. “Simply speaking, the safety concept ensures the rotating components do not burst, which may lead to catastrophic failures. Through Digital SIKO, we can say that under certain conditions and certain load profiles, we can run a particular turbocharger for a certain number of hours before performing maintenance. The product makes a judgment, for instance, on whether the turbocharger can continue to operate until the next dry dock without requiring replacement of the turbine shaft or the compressor.”
The digital assessment is combined with a physical assessment of the rotating components, which goes beyond a standard inspection, before any decision is made, he explains.
However, Mr Bosse acknowledges that there are concerns among users over sharing data. “Customers are careful. Questions like “who owns the data?” and “what is its value?” are important. More importantly, do I give my data to any potential service provider or just to the OEM, which I trust to make sense of the data, given their core expertise and design authority. And what do I get back in return?”
“This is understandable. It is a transition phase. We all need to get used to this. Digitalisation, including in maintenance, makes sense and benefits everyone,” says Mr Bosse.
High performance in demanding service
On the issue of turbocharger maintenance, TSI’s international sales executive Rhys Cleary gave his opinion to the following two questions:
1: Are turbochargers more prone to catastrophic failures? Is enough monitoring and information available on their operation?
Turbochargers work by utilising the exhaust gas produced from the combustion process. The negative side to this is that the main rotating components of the turbocharger find themselves bathing in a toxic mix of unburnt carbon, NOx and SOx. Although fuel quality and combustion efficiency can be monitored and adjusted, the damaging compounds will always be present.
A turbocharger’s worst enemy is an unbalanced rotor. The issue that needs to be identified is the cause of the imbalance. During a major overhaul the rotor is balanced to OEM specification; but uneven carbon build-up can cause an unbalance. A major cause of failure in turbochargers is foreign object damage. These kinds of failures tend to be catastrophic, unpreventable and come without warning. A simple piece of carbon dislodging from the exhaust manifold entering the turbocharger could damage the blades, unbalancing the rotor. This imbalance can quickly grow into a catastrophic failure.
It is worth noting the comparison between a common house fuse and the bearings in a turbocharger. People are quick to blame the bearing for the failure of the turbocharger; normally though, like a fuse going in a plug, this is the result rather than the cause. An imbalance in the rotor will normally be the cause.
2: Is predictive maintenance – by which we mean scheduling maintenance as per the running condition of the turbocharger – a priority for turbocharger manufacturers?
OEMs set out maintenance procedures: normally an intermediated inspection followed by a major overhaul. The intermediate inspection is a visual inspection mainly of the bearings, where most OEMs have introduced a method of calculating wear from a sacrificial face that can be measured to understand the used/remaining life. Some turbochargers can be inspected visually with the use of an endoscope, but these do not give a good enough overview to establish the current condition of the complete equipment. Regular visual inspection by experienced engineers will allow for a ‘history’ of information to build. In special situations, it may be possible to extend a service interval until the next earliest convenient port. This is not recommended however unless it is the only option.
This is mainly due to the lack of a warning sign. Such signs rarely show, and even they do, they tend to happen moments before a failure. An effective response from alert and highly experienced ship staff is required. The speeds at which the turbocharger turns ensures that by the time the warning signs emerge, failure is imminent in the case of foreign object damage.
When carbon builds-up, surging, high exhaust temperatures, high pressures and poor performance are the result. This is when an inspection by an experienced engineer can prevent further damage. Often, simple cleaning of components such as nozzle rings and re-balancing of the rotor is enough.