Introduction
IoT has been a catalyst for Digital Transformation for multiple industries which use equipment intensively like manufacturing, telecom, and mining. Machine data is being used significantly to increase efficiency in operations and create new revenue streams for the OEMs. The new recurring revenue streams could be realized by converting existing hardware products sales into Hardware as a Service model or by creating new data analysis products which analyze the data generated by their hardware products already in the field. The easiest path to monetizing any organization’s device is to start with the installed base.
Why Retrofit
Usually, the equipment cost is only around 15-20% of TCO, the rest goes into installation, system integration, networking, and most importantly, training the human resources. Unfortunately, this sometimes makes it prohibitively expensive to replace the hardware.
Lifespans of the large, heavy aging machinery are sometimes 30-40 years, making it expensive to replace before the planned remaining useful life. Furthermore, unplanned outages are unavoidable when the equipment is replaced. Therefore, loss of productivity through this unplanned downtime is sometimes a deal-breaker.
How do we start
As with almost everything in life, we need to start with the end objective in mind – we need a clear understanding of the business outcome for rolling out IIoT. It could be as simple as measuring the OEE at a machine level to increase productivity or a more complicated use case like identifying production line setup issues by comparing against a golden dataset. We then work backward to determine the visualizations and reports required to give us the desired insights and the sensor data points needed to create the dashboards. The data aggregation and storage strategies are then identified to enable the business use cases.
The nuts and bolts
Retrofitting this equipment could be as simple as adding an adaptor to the existing diagnostics serial ports to send the data over the cellular or local network to the IIoT system. In some cases, the firmware of the control system might need to be upgraded to send out the data collected from the sensors to the IIoT system over the network. We have also encountered instances where the machine’s log files had to be collected and parsed offline to get the required information. We could also use gateways that implement legacy protocols like OPC, Modbus, BACnet and send the data to the IoT cloud. In some cases, additional sensors like vibration sensors had to be fitted to the equipment. Finally, in some situations, we have other data entry tablets to capture inputs from the operator when it is not possible to access the machine data economically.
Our Case Studies
We have worked with an Electronic Contract Manufacturing Company where we had setup systems to automate & digitize OEE calculation at an equipment level at their assembly lines. We enabled the customer to link causality of less than desirable OEE with available machine data. We also established a workflow system to track and close low OEE instances.
We have worked with a leading Control System OEM involved in semiconductor manufacturing tools. We developed systems to store and analyze high-frequency sensor data ( 3-4 million sensor values per second) generated by the control system. This data was to be analyzed to identify issues in the tool after a service visit or after its use for long periods to identify drifts in setup parameters.