Five steps to realizing the Internet of Things

September 9, 2019 – Industry 4.0 intends to closely connect people, machines, objects and systems across the company in order to reach higher flexibility, robustness and optimal resource allocation (i.e. a “waste-free value stream”).

The possibilities of Industry 4.0 are endless. The challenge is how to implement these technologies so as to get the best results on your bottom line.

By breaking the process into steps, you can digitalize your manufacturing in a systematic way as you harness the power of the Internet of Things (IoT), a collection of physical devices such as sensors and actuators that use the Internet to send and receive data.

STEP 1: Define the problem you are trying to solve and determine the influencing factors in your system. Is that a quality problem or is it machine downtime? Do we know the utilization capacity of the machine? Are we using our capital investments optimally? Can we profitably run production if demand goes up or down? Why are two identical lines producing at different rates?

STEP 2: Choose your data sources. You can collect data from the PLCs, motion parameters, command log, error codes or cycle counts. Some controllers may use proprietary communication protocols or the builder could have locked out access to the controller, and you would need to make inferences of what’s happening in that “black box” through external observations. Sensors can be convenient and non-disruptive means of capturing machine operating parameters, establishing performance benchmarks and also monitoring environmental conditions.

STEP 3: Connect to all your inputs. In a typical factory, there is usually equipment from multiple vendors, various protocols, various data formats, new machines, old machines, highly automated lines and manual assembly stations. You will need a gateway that connects to a wide variety of data sources. You will then have to clean up that data to minimize bandwidth, secure it and transport to your analytics stack, whether located on premises or on the cloud.

STEP 4: Analyze the data. In its simplest form, with a few parameters, graphical visualization may be sufficient to see trends or anomalous behaviour. As we get into big data – with a high volume, variety and velocity of data – you will need to consider machine learning tools to build, train and deploy models of your system. Depending on the problem definition in step one, these results can be used in various ways, such as predictive maintenance, data analytics, visualization and notification, OEE optimization or business intelligence.

STEP 5: Evaluate the improvement potential and execution. You would now use that data gathered to have your Kaizen team carry out a blitz to improve the current process or use machine-to-machine communication to make self-adjustments in real time.

Employing IoT sensors
Sensors can make an existing “black box” machine be more transparent in terms of knowing about the internal processes taking place inside the machine. Sensors are the sensory organs of the smart factory, or the factory of the future. To cover as many IoT applications as cost-effectively as possible, sensors must communicate wirelessly, make efficient use of energy and simplify what has, to date, been complex data collection.

Data is the raw material of the factory of the future. It is the basis for continuous quality improvements, higher productivity and greater availability of machines and plant. Data collection, however, is cause for much consternation at many manufacturing organizations. First, they have to find out which data is relevant for which frequency for the respective application, and second, to date there has been a lack of cost-effective and rapidly implementable sensor solutions to make production economically transparent.

New sensor requirements
On the road to the Industrial Internet of Things – the manufacturing-specific arm of the IoT – companies typically begin with a pilot project, which, if successful, is rolled out to further plants and sites. Further use cases then follow. In terms of standardization and running costs, the most diverse, economical and future-proof sensors are therefore required. MEMS (micro-electromechanical system) sensors, which unite multiple measurement functions on the smallest possible space, are an important building block of such solutions. MEMS sensor technology is already a fixed component in many vehicles, as well as in fitness trackers, smartphones or virtual reality glasses. Together with wireless and energy-efficient data transmission, they are therefore also solid candidates for the Industrial Internet of Things.

Measuring temperature and acceleration, MEMS sensors can, for example, be used in the production environment to detect overheating and increased vibrations, which indicate the threat of motor or bearing damage. IoT users can identify switch positions and record motor service life performance via magnetic field measurements. However, users have so far missed industrial sensors that combine wireless data transmission with long battery life at acceptable cost. Ideally, sensors also contribute to the simplification of data collection, as this aspect is responsible for up to 50 per cent of the time consumed in brownfield IoT projects. New sensor solutions achieve this through integrated onboard functions.

One such MEMS sensor is the SCD (Sense Connect Detect), developed by Bosch Rexroth. Upon activation, it immediately delivers measured values for temperature, acceleration, magnetic field/current and lighting.

Sensor connectivity and technology is changing so quickly that a step-wise approach is required. And by breaking the many technological opportunities into these tangible steps, you will be able to benefit from a digital transformation of your production. A good way to start is by investing in a future-proof platform that can be expanded as the technological development progresses.

_____

Asvin Parsad is the business development manager for Industry 4.0 at Bosch Rexroth Canada.

This article originally appeared in the September 2019 issue of Manufacturing AUTOMATION.