Beyond Predictive Maintenance

By now, all our readers know what the concept of “predictive maintenance” encompasses and what advantages it offers over traditional or reactive maintenance. The Spanish company EM&E (Escribano Mechanical & Engineering) intends to go one step further with its weapon stations, adding integrated learning and monitoring, which will allow first and second echelon maintenance personnel to perform tasks hitherto reserved for higher Levels. This project was initiated at the proposal of the MALE (Army Logistic Support Command) and its success will be key to the future Technical Life Cycle Support Office (OTACV) of the VCR 8×8 Programme.

During the Cold War, the presence of tens of thousands of Warsaw Pact battle tanks was one of the main threats facing NATO Member States. The same was true of Eastern air forces and their navies, given the sheer numbers of platforms they maintained in service. To counter these threats and knowing that it was impossible to match the Soviets and their allies in numbers, technology was chosen to overcome overwhelming numerical inferiority. As a result, armoured vehicles designs (and all other platforms, weapons and weapons systems) became increasingly sophisticated, incorporating better firing directions, more powerful engines, complex transmissions and a host of electronic equipment.

Keeping a smaller number of platforms in service – not to deny the advantages that the technologies born of the Second Offset Strategy provided – always meant less resistance to attrition. In other words, in the event of an engagement, every casualty, whether a loss in combat or a breakdown that could not be fixed on the ground, would be felt much more in the Western ranks than in the Soviet ones.

Predictive maintenance arose from the need to maintain a high level of operability in means which, due to their cost and relatively small number, were not – and are not – expendable. Unlike the Soviets, for whom aircraft engines or tank propellants were something to be mass-produced and, if necessary, replaced at the drop of a hat in the context of a large-scale war in which they were guaranteed numerical superiority, things were quite different in the West. Having tanks and armoured vehicles whose cost increased exponentially as each new generation incorporated more and more complex systems and components, every casualty was a major and in some cases irreplaceable loss, regardless of whether it was due to mechanical failure or range.

This was clearly demonstrated on several occasions, for example in Iraq on 19 March 2003. Then, operating from the Saudi base in Arar, a column of 718M Pinzgauer vehicles moved deep into the Middle Eastern country with the mission of creating chaos behind enemy lines. In command was Lieutenant Colonel Pete Blaber, whom we incidentally discussed in the article on Operation Anaconda. Blaber and his men, after covering 600 kilometres, were to take control of the Haditha Dam – some 300 kilometres west of Baghdad – to prevent the Iraqis from destroying it and flooding the routes approaching the Iraqi capital from the south, which was the axis of the main US advance.

Having taken the objective with the help of the 75th Ranger, they were to proceed to the second part of the mission, which was to simulate an Allied offensive from the western flank to create confusion among the Iraqi commanders. To make the attack more realistic, H1 Air Base was captured intact on 24 March and a dozen M1 Abrams were transported in several USSOCOM-modified C-17A Globemaster. Advancing with hardly any logistical support, in less than two weeks of fighting more than half of the M1 Abrams were rendered inoperable, most of them due to damage, except for one that overturned in a ditch and had to be abandoned (prior self-destructing to avoid falling into enemy hands). In other words, in 15 days of operations, and without suffering a single casualty from an enemy grenade or missile strike, the unit had lost half of its strength and its combat capability was similarly reduced.

Of course, this is just one example out of thousands. In any case, it is very illustrative and should make us think about what would happen in operations of longer duration and to an army that did not have the resources of the American army. If the latter has always stood out for anything, it has always been for its logistical capacity and the high ratio between the number of uniformed troops in the rear and those at the front.

A second factor is that the complexity of today’s weapon systems multiplies the number of possible errors, especially those involving non-mechanical components, such as software or electronics. In this sense, a vehicle such as the future VCR 8×8 Dragon could be rendered unusable for combat not only because the engine or gearbox blows, or because several wheels get punctured, or because its main weapon gets stuck, or anything else we can imagine. It could also be rendered inoperable if the battlefield management software, or the one that controls the weapon station, “crashes” for whatever reason, or if any of the many processors it incorporates melts down.

Logically, the greater the number of subsystems, the greater the chance that something like this will happen. Hence, greater the chance that neither the crew nor the maintenance teams closest to the front line will be able to do anything about it. At least not without on-site access to expert knowledge and without all the documentation and instructions necessary to carry out their task, however complex it may be, being available anywhere and under any conditions.

As Rafael López Mercado said in his article on predictive maintenance “in the United States, thanks to a mentality open to change, which makes them capable of assimilating the lessons learned and implementing the solutions discovered, the implementation of predictive maintenance was promoted from all branches of their military forces”. This is, no more and no less, the task ahead for our Army. Some steps have already begun to be taken and will be important for “Fuerza 35”. Others will be more difficult to implement, but we cannot shrink back before the  challenge they pose.

In the case that concerns us today and related to EM&E weapon stations, such as the RWS Guardian 30, they intend to go a step further in predictive maintenance, something fully assimilated from the early stages of design and for which they have integrated numerous sensors in their RWS. Their idea is to host the necessary technical documentation in the cloud, combine predictive maintenance and interactive learning with augmented reality, and make use of Big Data. In short, they want to find a way to keep their stations running no matter what happens. Let’s see how.

Interactive learning and predictive maintenance in the civilian world

One of the characteristics of the times we are living in is that unlike in the past, civilian industry is ahead of military industry in many aspects. Many companies have been implementing augmented reality and interactive learning systems in which, using VR glasses and QR codes or natural identification systems that detect the user’s hands and the shapes of what is in front of them, they make it possible to instruct their operators.

At best, these systems allow the trainee to explore the parts of the system to be instructed on by himself, to move them in virtually all directions, to access technical documentation on site, and so on. Logically, while the training process is in progress, all this is done under human supervision, as it is always necessary to have the advice or guidelines of someone who knows the system thoroughly. However, given the possibility of visualising in real time on another screen what the VR glasses show to the trainee’s eyes, there is no obligation for the expert to be physically present at the place where the course is given.

This system has its advantages, especially in a globalised and interconnected world where, for example, very few countries produce advanced machine tools. Imagine, then, the case of a Spanish, Japanese or German company that has managed to export a state-of-the-art milling system to the United States. As the reader will understand, this system, installed in a factory in Missouri, is more profitable the longer it is in operation. In the event of a breakdown, if the production company is forced to relocate staff there (which is not at all unreasonable, but rather quite frequent), it will take days at best before it can return to operation.

The options in these cases have always been scarce. One, very basic and in demand, is to set up a subsidiary in the country to which the machines are exported, guaranteeing 24/7 maintenance. However, this is a costly solution that not only reduces the profit margin of the exporting company. This is often a problem as it depends more on subcontractors than on technicians from the parent company, as the latter can rarely dispense with its staff to send them to other latitudes, let alone deal with several incidents at the same time, in teams spread over several countries. The solution is to train the operators of these machines in their maintenance through interactive learning, reducing to a minimum the number of occasions when it is actually necessary to transfer the company’s own personnel.

Of course, none of this implies that local personnel will be able to undertake every possible repair. Not even all the maintenance tasks that complex systems such as those described above require. However, it is possible to solve a large number of incidents, especially if an engineer from the manufacturing company is available to the operator on site, who, via telematics, can supervise the actions to be taken, make recommendations, guide the process, solve doubts, resolve setbacks, etc. It could be argued that this way of working is nothing more than a ruse by the manufacturers to save on costs, but the truth is that it is a system that works, is increasingly in demand and fulfils the ultimate objective: to keep these machines running for as long as possible.

Moving from the civilian to the military world, we can see that there are hardly any differences between the machines we have used as examples (milling machines, cutters, 3D printers, large lathes, stamping machines, presses…) and modern weapon systems. If, returning to the case of the VCR 8×8 Dragon, we look at the whole, we will see that it has multiple subsystems, each of which can fail, causing the vehicle to lose its combat capability or to be degraded. Thus, if the optronic systems fail, if the engine gives out or if the turret ring gets stuck or the electric motor that moves it breaks down, we will find ourselves with a weapon system that is as expensive as it is useless, as it would not be able to fight with any guarantees.

Given that the conditions in which it must serve, both on manoeuvres and once in theatre, are particularly demanding, we continually risk failures of all kinds. This makes them overly dependent on the maintenance capabilities available, especially those of the I and II Level, which are those that allow a system to be returned to combat in the shortest possible time. However, this is not always easy, as the means or knowledge required for certain tasks are rarely available at these levels. This is where both interactive learning and predictive maintenance come into play.

Predictive maintenance can be summed up as the ability to perform the corrective or replacement tasks necessary for the proper functioning of a system before failure occurs. In a sense, every time we take the car to the garage, replace the oil or change the filters, we are doing this very thing. Nevertheless, in scenarios where what is at stake is not a call for a tow truck or a failed holiday due to a mechanical breakdown, something more is needed. This is where predictive maintenance can play its part, as it provides real-time knowledge of the condition of multiple components, making it possible to bring forward those repair or replacement operations so that mechanical failure never occurs, or at least minimises the chances of it happening.

Besides, if we think about complex platforms, such as a supertanker or a container ship, we see that the number of people on board is much smaller than one might expect, given the displacement of these vessels. If a monster like the “Emma Maersk”, with a length of almost 400 metres, can be handled by little more than a dozen people, it is not thanks to magic, but to extreme automation and predictive maintenance. This makes it possible for most of the ship’s systems to be sending data about their status 24 hours a day, so that one person can monitor and take the necessary preventive measures before they hang up. It is also because over the years, thanks to the experience gained and a better understanding of the vast amounts of data collected by multiple sensors installed in a large fleet like the one this company manages, they know with certainty what the failure alerts are for each component and are much more thorough with maintenance tasks. The same applies to fleets of trucks or aircraft. Transport companies, with dozens or even hundreds of vehicles in their inventory, need to minimise the number of incidents, which is why they have developed a long experience in the use of predictive maintenance systems.

Consequently, if we think about it, the problems they have to face are not too different from those faced by an army such as the Spanish army, with an armoured fleet numbering hundreds of vehicles. Without going any further, in the case of the future VCR 8×8 Dragon, we are talking about 348 units in the first phase alone, to be built by the TESS Defence consortium, made up of Santa Bárbara Sistemas, Indra, SAPA Placencia and Escribano Mechanical & Engineering.

Predictive maintenance and the Guardian 30

At the end of the previous section, we referred to the Army’s VCR 8×8 Dragon programme. As we know, this vehicle will mount – pending the resolution of the appeal filed by Pap Tecnos – Escribano’s Guardian 30 RWS. With this in mind, the Dirección General de Armamento y Material (DGAM) – responsible for a programme that if completed will be valued at almost 4,000 million euros-, has begun work on the life cycle of future vehicles, including the turrets. To this end, it is considering the creation of the Oficina Técnica de Apoyo al Ciclo de Vida (Technical Life Cycle Support Office, OTACV), equipped with the personnel and resources necessary to ensure operational availability and that the costs of maintenance throughout the life cycle do not exceed the forecasts. As Infodefensa explains:

“The office would work in four main areas: maintenance engineering, upgrades and modifications, advanced maintenance and spare parts management and materials logistics. Its tasks would be very diverse: operational readiness; equipment adjustment, calibration, and alignment; support to logistics bodies; equipment maintenance; predictive maintenance management; fault investigation and fault identification; technical documentation control and management; configuration control; obsolescence management; digital twin operation; equipment and system modifications and actions management; retrofits management; spare parts and components supply; and training activities management.”

Achieving all of this is beyond the capabilities of MALE or DGAM and will be impossible without the involvement of the companies that make up the consortium in charge of producing the new vehicles, as well as the weapons and other systems installed in them. In the specific case of EM&E, the task is to implement the predictive maintenance and interactive learning mentioned above. In the opinion of its engineers, to combine them to achieve a whole that is much more useful to the Army’s objectives. In order to do this, it will be essential to deepen a series of lines of work, such as:

• Improve data analytics capability: to create a knowledge base on which to base predictive maintenance. As we have said, when the RWS come into service, we will be talking about hundreds of units, each with multiple sensors and all transmitting in real time. Powerful analytical tools will be needed to extract the really useful information from such a jumble of data.

• Implement augmented reality systems: so that the RWS can display accurate data on the status or functionality of the system to the user in real time, putting synthetic information in front of what the eye sees thanks to the use of augmented reality glasses such as Google Lens. This sounds a bit cryptic, but it means that the company will have to create a robust, intuitive, and functional interface that allows the data produced by the system to be transformed into useful information for the maintenance operator, the engineer in charge and even the crew if they need to intervene on the fly.

• Virtualising systems (digital twin): the products will have an embedded simulator that will allow an immersive reality experience, reproducing the physical environment and creating the necessary situational awareness so that the person in charge of supervising the system has virtual access to all the elements that make it up and can not only evaluate them, but also manage them from a distance.

From the combination of these three lines of work, they expect the creation of an integrated maintenance and learning system that has, at least on paper, many advantages over the current ways of working.

For instance, the system should be able to indicate to the levels when and how maintenance operations should be carried out through a checklist that will allow a single work methodology to be established. Even better, in the event that for whatever reason the checklist changes (e.g. after a new analysis of the data sent by the sensors installed on the turrets), the checklist would be updated in real time in the cloud, making it available to all the units and levels involved.

This last aspect related to the cloud is crucial because what will ultimately make it possible for the model EM&E is working on to be implemented is the rapid development of information and communication technologies in recent years. Also, this shall be combined with the emphasis that the armed forces are placing on adapting to this phenomenon. It will be thanks to improvements in communication that operators will be able to receive advice from in-house experts should they need it. Also, when the time comes, this same expert will be able to take control of the system from a distance, carrying out any operations deemed necessary and accessing all relevant information for maintenance or repair. As the company explains, in certain cases “it is better that one of our engineers can see the system through his own eyes, even from a distance, than through the eyes or impressions of others”.

However, this dependence on communications and cloud access capacity is also an Achilles’ heel. Not only is – and will be – bandwidth limited, but systems can go down, communications infrastructures, including satellites, can be attacked, and so on. The company knows this and plans to have backup systems that allow operators to access information even in environments where communications have been degraded.

Beyond the bottleneck of communications, there are many other complications that EM&E is trying to overcome by investing in R&D and innovation. The company has been including sensors in its products for years, something that in cases such as the Guardian 30 has been taken to the extreme, including multiple systems capable of measuring every relevant piece of data. More importantly, they have been collecting data on the main parameters that influence operation (CPU, GPU, humidity, clearances, deviations, vibrations, operating temperatures…) since the very first design phase. Thanks to the analysis of the data collected, they have been able to create a series of constantly updated databases, which will allow a better understanding of the needs of each weapon station on a day-to-day basis and optimal maintenance planning. All in all, this work will take years to bear its best fruits.

Furthermore, some of the stations that the company produces, and which have been exported in large quantities, already have sufficiently precise data on performance and reliability or to organise maintenance with full guarantees. For instance, in the case of those delivered to the Spanish Navy (34 turrets so far) or to the Peruvian Navy whose process is still underway. As for the Guardian 30, although during these years of design and testing they have been able to collect a significant volume of information, it will be when it enters mass production and begins to be installed on the barges of the VCR 8×8 Dragon when the RWS will be able to have all the relevant data.

Given the number of sensors per RWS, the number of turrets that will eventually be produced and the multiple scenarios in which they will have to operate as time goes by, the accuracy of the analysis will become more and more precise. Therefore, it will positively affect the life cycle of the weapon stations themselves. However, this will require both the company and the Technical Life Cycle Support Office to work hard to develop the Big Data technologies needed to deal with gigabytes and gigabytes of information and as they say to separate the wheat from the chaff.

Having all the above data and being able to analyse it is a huge advantage, but it is of no use if the maintenance staff on the other side of the screen are not properly trained or do not have the resources to do their job. For all these reasons, beyond predictive maintenance, the company considers interactive learning and the use of online resources to be essential. Thus, operating manuals will be included in the systems themselves, as well as in the traditional paper format. The company says it does not believe in them, even though it recognises their importance and the need for the end user to know them in depth. In this vein, they argue that better results will be obtained through simulation and augmented reality, which allow the user to learn how to use the system without the need for manuals.

This approach, in which manuals and all technical information are fully digitised and stored in the cloud – backed up in the physical memory of the weapon stations themselves in case communications fail – means that the files can be constantly updated. Besides, it implies that the user can access them in any situation not only for consultation, but also as a form of interactive support. In fact, the challenge is not so much to codify maintenance guidelines on paper, but to create the necessary tutorials and aids so that any operator, with little or no specific training on the system in question, can undertake multiple tasks.

Incidentally, although we are paying almost all our attention to hardware, the same applies to software, which is becoming increasingly important for the operation of modern weapons systems. For example, the system makes it possible to know how often a programme has been interrupted and to act accordingly for the future.

Moreover, we had the opportunity to see the system in operation with the company’s Director of Technology, José Carlos Hidalgo, and to verify, as shown in the images, that the system guided the user through each step of the process. In addition, at the same time, the user can consult not only the real-time data collected by the sensors, but also the characteristics of each component. However, the work ahead, until each maintenance operation has its own tutorial in PDF, with explanatory videos, etc., is enormous. Indeed, it will take years and a lot of money to bring this about, although the company is determined to do so at all costs.

In Hidalgo’s opinion, there is also an important psychological aspect, which maximises the advantages of this new way of carrying out maintenance: it makes those in charge of carrying it out even more interested in their work. It may seem a trivial matter, but the fact is that being able to “play” with the virtual components, having all the information at hand and turning otherwise routine operations into something more playful and interesting has its advantages and helps to involve staff in pursuit of a common goal: to provide the best possible maintenance.

Last but not least, we must once again stress the concept of industrial and technological sovereignty, which would never be such if foreign companies had to be used to provide support for the life cycle. In this sense, the fact that both the barge, the RWS and the multiple subsystems (combat, optronics, communications, etc.) are Spanish allows total freedom and independence also in terms of maintenance. This is something that would not be possible in the same way if foreign companies were selected, even if the usual industrial considerations were negotiated and all the strings were tied up on paper. In other words, the fact that the technologist is Spanish allows the user to carry out any maintenance operation without limitations, something that would otherwise be impossible and something of which we have plenty of examples in the Armed Forces.

Conclusions

The Spanish Armed Forces have long since opted for quality over quantity when it comes to their weapons systems. In the case of the Spanish Army and programmes such as the VCR 8×8 Dragon, this commitment will be pushed to the limit, as these systems are much more complex than those they are intended to replace, which will necessarily limit the number of them in service. In this context, the need to maintain each of them in operational condition regardless of the circumstances is an imperative. The only possible way to do this is to combine predictive maintenance with profound changes in the training and support provided to those in charge of carrying out these important tasks. This means relying on augmented reality, simulation, the cloud, and the analysis of large volumes of data.

Escribano openly acknowledges that it is not a pioneer in any of these technologies, but rather draws on the experience gained by many other companies over the last two decades. However, the company believes they can combine them and develop solutions that offer significant added value. Regarding their weapon stations and especially the VCR 8×8 programme, they will be able to significantly improve life cycle support.

It will not be easy, as achieving the goal they have set for themselves involves the sensorisation each product, including natural or artificial markings that allow datasheets, informative videos or manuals to be displayed. Also, developing these contents by customising the information offered to suit the exact needs of the end user, simulating each maintenance operation for each part, designing secure systems that allow the information to be stored both in the cloud and in the weapon stations themselves, coordinating with the manufacturers of the barges and each of the other subsystems that make up such a complex vehicle, etc. In short, a major challenge that they are facing with a view to being able to do two things: 1) Minimise the chances of their weapon stations failing and; 2) If necessary, ensure that the crews or second levels have at their disposal all the resources and support necessary to solve almost any problem in situ. This is the only way to guarantee maximum availability for their products and thus ensure that combat capabilities always remain high.

Beyond what this project means for the Madrid-based company, it is clear that its success will also depend on the commitment of other actors. Without going any further, the Army itself will have to adapt to a new way of working. A process that always encounters resistance within any  organisation, but which is essential if the aim is to get the best out of new vehicles that are far from being simple mechanical assemblies.

For the time being, the company has already showcased its project at the Toledo Forum, which was hosted by MALE. In the coming months and years, they will continue to work, as they see this as a differentiating element compared to the competition. They are also confident that the possibility of offering better support for the life cycle of their products will eventually tip the balance in future tenders. If they deliver as promised, we have no doubt that this will happen.