Knowing the best times to perform maintenance on your equipment and machines is a delicate art. If you perform maintenance too infrequently, you’ll be at risk of unexpected breakdowns that create costly delays. Do it too frequently, on the other hand, and you’ll be wasting time and money on unnecessary repairs.
Fortunately, you can find better answers to this question by enhancing your workflows with two extremely valuable technologies: machine learning and the Internet of Things (IoT). You’ll save yourself from the tedious guesswork of finding the right maintenance schedule, and in the process you’ll also make smarter operations decisions.
What Is the Internet of Things (IoT)?
Although a lot of hype has been swirling around the IoT in recent years, the benefits that it can provide you are very real. Nevertheless, many people still aren’t quite sure what the IoT is or why they should use it.
Simply put, the IoT is a vast network of physical devices, vehicles and other items. Just as the Internet connects computers in different locations, the IoT extends this concept to thousands of different objects, from toasters to tractors. Each of these items is embedded with electronics, software, sensors, actuators and network connectivity.
Thanks to this connectivity, objects belonging to the IoT can communicate and exchange data. This opens up endless possibilities for “smart” devices that can share information in order to be more useful and effective. As a simple example, your alarm clock might send a message to your coffee maker as soon as you press the “alarm off” button so that the machine will prepare your coffee while you’re waking up.
The many applications of the IoT are proving to be a major catalyst for business digital transformations. Management consulting firm Bain & Company predicts that by 2020, B2B IoT segments will generate over $300 billion every year.
How to Leverage IoT for Preventive Maintenance
Complex systems such as aircraft engines and manufacturing facilities are examples of industrial IoT systems that have thousands of components. Each of these components must be operating correctly in order to obtain optimum levels of performance, productivity and returns on investment.
Attached to each component is a sensor emitting a signal that communicates the component’s remaining usable life and other key performance indicators (KPIs). These signals can be collected in real time and analyzed in combination with historical data and previous service requests.
The end product of this process is a set of predictions about which components are at highest risk of failure. You can then use these predictions to efficiently schedule repairs in conjunction with other IoT data.
For example, another IoT device may inform you that inventory for a critical repair part is low so that you can reorder the part before you run out. In addition, you can consult historical data such as the average technician response time or the average time to repair. This will help give you a sense of how long the maintenance might take, and plan ways to work around it.
By predicting the best times and ways to perform maintenance calls, you can proactively avoid equipment failures and hence costly downtime. According to consulting firm McKinsey & Company, using the IoT to make such predictions can reduce maintenance costs by up to 25 percent–not to mention slashing the number of unplanned outages in half.
How the Google Cloud Platform Is Used for Preventive Maintenance
Google Cloud Platform’s Cloud IoT Core is a service for collecting and managing data from IoT devices. Together with other Google Cloud Platform services, Cloud IoT Core is able to assist you in obtaining a major goal for manufacturing facilities: generating proactive alerts that warn of impending failures.
These alerts will prevent you from suffering lengthy, costly downtimes, and can be used in combination with a machine learning model in order to be more productive and efficient.
Applying this general concept to a specific example demonstrates how you can leverage machine learning and IoT devices to the benefit of your company. For instance, sites in the oil and gas sector typically consist of an oil field, with each field comprising multiple oil rigs.
Each of these oil rigs has countless numbers of diverse components, and each component has multiple sensors attached. These sensors transmit information about various parameters of interest in real time, such as the oil’s temperature, viscosity, and pressure.
Cloud IoT provides a framework to collect sensor data at scale, but collecting the data is only the first step. The next step is to analyze all of this information in aggregate in order to measure the health of the machines.
Sensors at these oil rigs send data in real time to Google’s Cloud IoT Core. This data is streamed into Cloud Pub/Sub, a service for passing messages between machines and devices. It’s then processed using Cloud Dataflow, which calls a machine learning algorithm hosted on Google’s Machine Learning Engine. If the algorithm predicts that the machine is about to fail, it sends a signal to users so that they can decide on the appropriate preventive maintenance action.