The world has undergone many changes in our life time, and has undergone even more beyond it. From globalisation to a push to dematerialisation, people are finding different ways of doing business every day. The difference is that the time between these changes in business is decreasing so rapidly that many companies are unable to keep up with the pace, and they pay the ultimate price.
What is the fourth industrial revolution?
According to a 2016 report, the average lifespan of a company on the S&P 500 list in 1965 was 33 years. This had shortened to 20 years by 1990, and is expected to reach 14 years by 2026. Many large retailers have closed down due to disruption by e-commerce. The manufacturing sector has been completely disrupted by the fast pace of change in a digital world. This sector is not alone. Energy, logistics and housing are but a few others who have seen a tremendous shift in how business is conducted. If we want to take a very basic example, we can look at how Uber changed the way taxi companies operating around the world. This change in the way we work, live, and communicate is what we call the fourth industrial revolution.
The fourth industrial revolution is the synergy of the physical, biological, and information worlds. This takes the physical capabilities of what we have developed as a human race and merges it with the insights and connectivity that we are able to draw from the information realm, in such a close-knit way, that we are able to treat them as one entity. Think about how augmented reality brings completely new possibilities to field workers, how artificial intelligence is able to change the processes and skills in so many industries, how 3D printing can change the manufacturing process.
If we want to understand the fourth industrial revolution, we need to start with knowing about previous industrial revolutions:
- First industrial revolution: The transition of society, between 1760 to 1840, to new manufacturing processes, mainly due to steam power, machine tools, and the mechanised factory system.
- Second industrial revolution: Also known as the Technological revolution, this phase was a massive jump in industrialisation from late 1800’s to the early 1900’s. The major contributors to this phase were mass production techniques, electricity, and the internal combustion engine. This exponential growth was also due to the speed at which heavy machines are able to make more heavy machines.
- Third industrial revolution: This revolution refers to the way that everyday life, as well as individual tasks, were made simpler by the introduction of computer technology.
- Fourth industrial revolution: Due to the ever increasing computing power at almost negligible prices, the cyber and physical worlds will merge to change the way in which we live, work, and relate to one another.
This fourth industrial revolution is the effect of Internet of Things, Internet of Systems, and cyber-physical systems on the way we do business. Companies are faced with the decision to continue the way in which they currently do business, and risk being disrupted by competitors, or to evolve and innovate with the times, to ensure that they stay ahead of the game.
How does this apply to Industry 4.0 and Predictive Maintenance?
With regard to industry, the jump between the third and fourth industrial revolutions would be harnessing the power of data to optimise the manufacturing process and is known as Industry 4.0. If we are able to combine the collective knowledge an organisation has about their manufacturing processes, the insights could be mind-blowing. This could range from making the end product quicker, better or more cost-effectively, or it could mean having better control over your industrial assets. This begins by creating an interconnected web of sensors within a manufacturing environment and ends with the proof that data is arguably the most important asset that an organisation can possess.
Being able to introduce wireless sensors to an industrial environment, and to be able to interconnect all of these sensors, whilst recording the data, can provide many advantages. One of the most profound is the ability to analyse the data and apply machine learning, to be able to find out exactly what happened in the time leading up to failures. This is invaluable information. This can allow the diagnosis of a machine before it breaks down.
Think of it as the equivalent of measuring the core temperature of someone who feels ill. If their temperature is measured and is above a certain predetermined threshold, roughly 37 degrees Celsius, it is indicative of a fever, or something else. This means that they are sick and that certain actions should therefore follow. This is exactly what happens with machines in predictive maintenance. Data is measured by sensors set up in specific locations around the machine. These sensors might measure properties such as vibration, deflection, strain, temperature, or sound. These sensors are also connected, mostly wirelessly, to a central hub which records all of this information. When the machine fails, the specific type of failure is noted, and the properties of all the measured data are analysed using machine learning to determine which of these properties might have shown signs pointing to the failure before the actual failure occurred. Once enough of these cases of failure have been analysed, and the machine learning algorithms have started to get an idea of what to look for in specific types of failures, the algorithm might start being able to predict failures. This in itself is a game-changer.
Imagine not having to take your car in for its scheduled service. Imagine only having to take your car in for a service when it was actually about to break down. No need to organise a borrowed car, or to take the morning off work. This would be a huge convenience, especially financially. Now imagine the need to take heavy machinery out of operation to replace parts that cost a lot of money, and may not actually need to be replaced. This is due to the lifetime of a part being based on the meantime to failure of parts just like it. Using the predictive maintenance you are able to see the true health of that part, and not what it should be. Here we see that the cost is two-sided. On one side, you have the chance of wastefully replacing a part that still has life left in it, and on the other hand, your entire production process is on hold whilst you do this. This can all be avoided if a predictive maintenance system is in place. If you can see by the data that your machine is in need of repair, you know that the downtime is completely necessary. It is the equivalent of checking your engine's temperature to see if it has a fever.
Another benefit to information being collected on a product is the ability to gain insight into how products are used. This insight can lead to more knowledge about the market and eventually a better customer journey. Let us use a very simple example. Imagine you sell electric bicycles and are able to collect data on the average trip length. This would allow you to finetune your battery pack to best suit the expectations of your users. If your users are running the battery flat every day, it might mean that your batteries do not have the capacity that the market wants, or it might mean that the batteries are not performing to specification. On the contrary, it might mean that you have over-specified your bicycle batteries and that you might be able to fit a smaller battery at a lower cost. If you can see that the average trip length is growing with time and that there is a general trend of longer cycling distances in the market, it can help you preempt this trend and fit larger capacity batteries before your users realise that they needed them.
Useful cases to illustrate the power of predictive maintenance can be found in the rail industry. Here it was shown that the health of both the trains, as well as the tracks are able to be monitored remotely, reducing the risk of oversight by maintenance workers. This also reduces unnecessary maintenance closures and unnecessary expenditures. In a slightly more applied case, power outages can be prevented in the utility industry by using drones equipped with sensors to map utility networks and use machine learning to identify trees that are about to fall onto utility lines.
How does this fit into digital transformation?
The term digital transformation can be defined as the jump that companies make to bridge the gap between the third and fourth industrial revolution. Industry 4.0 falls perfectly in line with this as it is merging the physical and information worlds to give one complete ecosystem, but the shift to Industry 4.0 is not something that should exist in a silo if you are looking to undergo true digital transformation. It is something that should complement a broader culture of digital transformation in your business. It is something that should exist in the DNA of your organisation. Transformation on this scale is something that needs to come from an organisational level. Many attempts have failed due to lack of discipline, lack of momentum, or lack of buy-in from the broader organisation. Many successful pilot programs have failed due to the inability to roll out into the organisation.
One of the first questions people might ask themselves is: ‘Is this for me, and how do I start?’. It may be the case that Industry 4.0 or predictive maintenance might not even be for you at this moment. If you don’t have high maintenance costs, or your production process efficiency is already through the roof, then this investment might not be as exciting right now, but if you are able to leverage the power of data, this can most certainly be an interesting venture. Please bear in mind that this is a big commitment. Many companies try to transform, but most fail. It is imperative to have a company-wide push.
The availability of budget for these projects is generally regarded as one of the biggest anxieties which companies face, but this is soon overshadowed by the amount of visibility that organisations obtain over their equipment and the savings that they are able to realise from reduced downtime and maintenance spending.
Overall the effects of predictive maintenance are far more than just the reduction of maintenance costs and breakdowns. Predictive maintenance is a commitment to a better and more efficient way of working. It is a commitment to a higher product quality and the ambition to push your organisation to the next level by taking the actual health of your equipment (organisation) into your own hands.
How can I start predictive maintenance at my organisation?
1. Lay the foundation
One of the first things you should be looking at is ensuring you have a strong, united leadership vision. This project is not going to go very far if it does not have a deep-seated drive for Industry 4.0 and predictive maintenance. If you do not already have a strong preventative maintenance culture and ability in your company, this is what you should build up to start from. You need to have the philosophy, the culture, and the trained technical skills.
2. Make sure you start with the right challenge
The machinery you choose for your first pilot is crucial to the success of your predictive maintenance project. Put some thought into how you can get the best results. Machinery that is simple but crucial to your production could be a good start, although avoid picking your most important piece of equipment. Trying to bite off too much at the beginning by using machinery that is too complex could be a deal-breaker. The success of this pilot is the momentum you are going to need to scale up.
3. Start collecting the data
Install the right sensors in the right places. This can be a challenge, but skilled maintenance teams should know what needs to be measured. This data needs to be stored in the cloud, easily accessible.
4. Create machine learning algorithms
Start getting the right people to incorporate machine learning algorithms to get insight out of your data. This will allow you to start being able to predict when failures will occur. The insights are only as good as your data so make sure that step 3 is solid!
5. Integrate these components
The next step is to bring everything together. You couple your sensors with your algorithms and ensure that automated reports are generated. It is important to ensure that these reports coincide with physical results.
6. Iteration is king
Ensure you continuously improve! You will always have learnings from your project that can be used to improve. Learn from your mistakes and ensure you continue your culture of continuous development. When you decide to scale things up and roll out to another piece of equipment, make sure you transfer your learnings and improve what you have already done.
Once you have a working pilot, the momentum is on your side. You can use this success to further prove the worth of the initiative. If you then take this and roll it out in methodical and measurable steps, making sure not to take on too much at once, you will begin to see the bottom line results which will gain the attention of other leaders in your company and further allow your pilot to break into the broader organisation. If your company possesses the culture of preventative maintenance already, this is the logical next step into taking your organisation into Industry 4.0.
The bigger picture
Finally, the data that is collected from products can add value that transcends that of product value alone. Yes, this data can directly improve the service that end customers receive and the brand value that comes with it, but to take this even further, this data can be used to innovate entire business models. We will unpack this and zoom into the concept of data-driven business model innovation in an article to be published very soon.
Like to have a chat on strategy in the meantime? Find me at email@example.com
Schrijf je in voor onze nieuwsbrief en laat ons jouw gids zijn in een complexe digitale wereld.