Defining an intelligent machine is no easy task. In the early days of computing, it was thought to be a computer whose answers to questions were indistinguishable from those of a human being. In this sense, the word "machine" was taken to mean "computer." If you search for "intelligent machine" on Google, you’ll still find lots of futuristic work in this vein.
The McGraw-Hill Dictionary of Scientific & Technical terms defines it like this: "A machine that uses sensors to monitor the environment and adjust its actions to accomplish specific tasks in the face of uncertainty and variability." Cited examples include industrial robots equipped with sensors and self-guiding vehicles that rely on vision rather than painted lines.
Within engineering, we think of an intelligent machine as a mechanical system that can take care of itself: a machine capable of accurate self-diagnosis that can quickly communicate its condition to an operator so that the problem can be resolved as soon as possible. It could be anything from a high-end car to a complex machine on a factory floor. This is the subtext to ‘uncertainty and variability’: as well as reacting to changes in its environment, an intelligent machine must look after itself, so that it can continue to work at maximum efficiency.
This is not to suggest that an intelligent machine is maintenance-free – that really is a futuristic dream – but it uses its in-built intelligence to detect potential problems and streamline maintenance intervals and procedures. All mechanical parts are prone to failure, of course. The trick is to detect this proactively, as part of a planned condition monitoring regime, and take action in advance, rather than waiting for the machine to fail and then spending time and money repairing it.
Intelligent machines will rely on several critical factors. The most important, by a long way, is information. Without data, there can be no intelligence or diagnosis. This data needs to be gathered, transmitted for analysis, and processed, which, in turn, requires sensors, data transmission and computing power. At SKF, we already have extensive experience in all these areas, and are ready to take it to the next level.
SKF Insight Introduction
The immediate answer might at first simply appear to be enhanced condition monitoring, adding an array of sensors to a machine in order to read its vital signs, then transmit them over WiFi to a central point. But a far more effective solution now exists in the shape of SKF Insight: it collects and transmits process data independently from inside the very heart of a machine using a fundamental engineering component: the bearing.
SKF Insight turns a simple bearing into a diagnostic powerhouse, by embedding into it a tiny, self-powered wireless sensor that transmits real-time information about process conditions. It takes condition monitoring far beyond what was previously possible. The technology, launched at Hanover in 2013, required three years of intensive research, including making the sensors smaller, overcoming power generation challenges and developing unique packaging for the sensors and electronics.
Conventional condition monitoring detects the early signs of failure by measuring vibrations caused by changes on the bearing’s surface. But this means that damage has already begun to occur. Rather than identifying this deterioration, SKF Insight detects the conditions that cause bearing failure before they can have an effect, and makes this information instantly available to operators.
Miniature electronic circuits, powered by the motion of the bearing itself, transmit this process data via a wireless link. There is no need to supply external power. This makes the technology supremely unobtrusive, because there are no wires ‘in’ to provide power, or wires ‘out’ to deliver the signal. This means it will work in places that would previously have been impossible. Just imagine trying to take signals out of a rotating gearbox, for example: it would be a complete mess, with entangled wires everywhere. With SKF Insight, signals can be taken from anywhere, and we are already developing solutions in challenging applications in wind turbines and steel manufacturing.
We developed SKF Insight because we know that bearings rarely fail in service under normal operating conditions due factors such as subsurface fatigue. Instead, the cause of failure is usually misuse or neglect: insufficient lubrication, for example, or running the bearing under conditions outside those originally specified. Insight’s embedded sensor measures the critical parameters that cause early bearing failure, such as lubricant contamination, or temperature, allowing operators to take corrective action while the machinery is still operating. The direct result is that expensive, disruptive failures are avoided, which reduces the total cost of asset ownership and extends machine operating life. It also makes it simpler for engineers to gain a far more detailed appreciation of the varied causes that can affect the calculation of bearing life.
By applying sensors directly within the bearing, SKF Insight identifies the risk of failure before even microscopic damage occurs.
SKF algorithms and diagnostics can identify duty excursions, lubricant contamination and lubrication problems, allowing operating conditions to be modified, and so avoid damage before it occurs.
By integrating SKF Insight with asset diagnostic and bearing health services, we can send information on actual operating conditions to cloud servers for remote diagnostics, enabling a better understanding of the risk of future damage and failure.
SKF Insight gives maintenance engineers a powerful new tool to keep machinery in prime condition, giving them capabilities way beyond traditional condition monitoring. It means that maintenance can be carried out at exactly the right time (we can even call it “adaptive maintenance”), rather than being guided by a strict schedule that is unrelated to the actual condition of the machinery or its components.
The intelligent wireless technology inside the bearing allows bearings to be configured in smart networks, which communicate via wireless gateways. The gateway can be local to the machine or to the plant.
System information is provided to the customer for analysis using SKF @ptitude Analyst, or sent via the SKF cloud to a remote diagnostic centre. From here, dashboards and reports can be supplied to the plant operator, machine manufacturer, SKF or any other authorized person with internet access. The inclusion of SKF in the list of ‘recipients’ is an important one, as its assistance in gathering and interpreting the data will be vital thanks to the deep bearing and machine knowledge existing in SKF.
Because the bearings are self-contained they can be used right in the heart of a machine, where it was previously impossible to embed sensors. This is a huge step forward in real time condition based maintenance, and provides a vastly improved understanding of the operating environment. Having such a deep knowledge of operating conditions – in real time – could even make it possible to upgrade a machine, extending its life or power rating beyond its initial specification.
The sensors communicate through each other, and the wireless gateway, to create a ‘mesh network’, providing both machine-wide and plant-wide information.
SKF Insight makes condition monitoring more widely applicable, especially where it might previously have been considered impossible. Because of this, it is being tested in industries including wind power, rail and steel manufacture.
SKF Insight also offers huge potential benefits to industries like wind energy, where the cost of maintenance is astronomical. In some offshore wind applications, changing the main bearing on a wind turbine is so expensive that it undermines the business case for building the turbine in the first place. Used here, intelligent bearings could monitor loads and lubrication conditions in service, giving plenty of time to prevent the development of damaging process conditions.
We are already working with customers to develop SKF Insight for wind turbine monitoring. It measures dynamic bearing information in the true operating state, then wirelessly communicates it to remote monitoring centers or local maintenance crews. The solution being considered will monitor bearing speed, vibration, temperature and lubrication. Most importantly, it can be retro-fitted, so could enhance the operational potential of both new turbines and the many thousands that are already in operation worldwide.
A similar solution, further in the future, is being developed for wheel end bearings in the rail segment. These critical components are normally changed at set intervals, regardless of their condition. SKF Insight creates a cost effective way of collecting condition monitoring data so that service bearing life, and change-out intervals, are determined based on actual, rather than predicted, operating conditions.
The ability to monitor and transmit information on operating conditions will bring about a revolution in bearings, in terms of maintenance planning, total cost of ownership and maximizing machine efficiency. Bearings have long been considered the heart of rotating machinery. Now, by imbuing them with intelligence, SKF Insight makes them the brain as well. It goes beyond traditional condition monitoring, into what might be called ‘future reliability’ – identifying potential problems before they occur, and taking immediate corrective action.
SKF Insight is already being put to work in high-end applications such as wind turbines. But consider the machine that we spend most of our time with: the car. Think about all the problems that could be avoided with this kind of advance intelligence, and you can see why the technology embodied within SKF Insight is truly a revolution – joining both bearings and condition monitoring.