Connected sensors deployed in a smart way can detect trouble before it impacts processes or production, saving maintenance costs and avoiding downtime.
To avoid inefficient maintenance routines and the costs that go with that, manufacturers and network operators can use smart sensors and the principles of data science to optimise the maintenance process.
According to this report from McKinsey, IoT-based predictive maintenance helps to reduce maintenance costs of factory equipment by up to 40 percent. It can also “reduce equipment downtime by up to 50 per cent and reduce equipment capital investment by 3 to 5 percent by extending the useful life of machinery”.
Welcome to the world of true predictive maintenance.
The problem with time-based maintenance
For years, manufacturers have been practicing a time-based approach to equipment maintenance. The primary factor in planning any maintenance routine was simply the age of the machinery. The older the equipment, the more frequent the maintenance procedures would be.
A worldwide study from ARC group states that only 18% of equipment fails due to its age. The remaining 82% of failures occur randomly. Such findings show that an age-based approach to maintenance is not cost effective.
To avoid ineffective maintenance routines and the high costs that accompany them, manufacturers can use Industrial IoT and data science to their advantage.
A more efficient power grid
Breakdowns in power grids can cause interruptions in power distribution. This would cause enormous trouble to the everyday lives of almost all people, businesses and services in affected areas. This makes predictive maintenance especially important for such critical infrastructure.
Finland’s transmission system operator, Fingrid, has organised several innovation contests to find the best partners to digitalize the maintenance and monitoring of their operations.
Following their win in one of these contests in 2018, Haltian’s Thingsee wireless sensors are now used to measure the temperatures of connecting components in Finland’s electrical substations. That’s important, because a rise in temperature is a sign of increasing electrical resistance, which can be caused by dirt or corrosion. In addition, the sensors can be used to measure aspects such as humidity, air pressure, ambient light, presence, and distance.
Such sensors don’t totally remove the need for manual checks, but they do make monitoring much more effective and reduce the risk of problems.
Connected tools to halve maintenance
While there is an obvious use-case for power grids, sensor-based monitoring can also be used in a wide range of industrial applications.
The Ericsson Panda manufacturing plant in Nanjing uses Cellular IoT to connect a thousand devices, including high-precision screwdrivers. The modules transmit about 100 bytes of data every eight hours to indicate recent usage. The data is collected in a cloud solution for analysis.
By monitoring usage data, operations managers can pinpoint exactly when the tools need recalibration instead of working to an inefficient predetermined schedule. According to this case study, Ericsson anticipates the solution, that costs just USD $20 per unit, will cut maintenance work in half, save USD $10,000 every year, and achieve breakeven in just two years.
Predictive maintenance in other industries
Tools and Machineries.
Safety and telemetry systems are often used in industrial environments, generally requiring their own specific networks. Besides this, many less critical parameters can be gathered with new wireless technologies to form the basis for predictive maintenance.
There are two main categories of data:
- Usage: How often and how long is the tool used? Which functionalities are most often used? Who are the employees working on it?
- Status: What is the temperature (and variance) of the machine over time? Are there any irregular vibrations in the system?
Using the data to carry out a full root cause analysis will help to prevent future failures.
Retail and Distribution.
IoT is already providing valuable help for global supply chain operations, including asset tracking with Cellular IoT. But connected technology can further enhance the relatability of the distribution chains by predicting failures of critical assets.
A typical example is the cold chain. For many years, focus has been put on finding low cost solutions for measuring the temperature of goods transported. This helps to understand if the goods were kept at the right temperature and which goods may need to be destroyed.
While this is a good example of IoT in action, there is something that can be done to predict and avoid failures in the first place. Applying predictive maintenance to refrigeration systems to better understand when it is about to fail will save valuable time and money, and avoid wasting valuable resources such as food and medicines.
Roads, bridges and railways are critical infrastructures for society. Their maintenance is important to make sure that safety is upheld for their entire lifespan. In this case, IoT can make a big impact.
Mechanical parameters such as vibrations can be constantly monitored. Any anomaly in the recorded patterns can indicate a need for maintenance or it can advise the presence of an emergency situation. For example, at Düsseldorf Airport, 50 in-road NB-IoT sensors have been installed to monitor the status of the only bridge that gives access to airport’s fuel tank storage site.
Read more: Smart airport takes off in Germany
The challenge is to scale it up
Predictive maintenance is not a new concept. We are all familiar with predicting when a car is about to fail by way of irregular sounds or vibrations. What is new and what IoT brings to the table is the capability to collect data from thousands or millions of devices. We can learn from it, create and update models and advise actions, all automated and at an unprecedented scale.
It is clear that to succeed, companies will need to leverage all latest technologies in Big Data Analytics and Artificial Intelligence. However, even before that, their ‘things’ will need to get connected.