The availability of large amounts of data from within and beyond the manufacturing process and the capability to harness it, organize it and convert it into actionable information in real or near real time forms the very foundation of Industry 4.0 for most manufacturers. Used effectively, the intelligence created through the use of this ‘big data’ can both positively influence strategy and impact the broader value chain for sustainable transformation.
Data analytics is the key to success in the modern manufacturing paradigm. The electronics industry in particular is looking at massive investments in big data analytics. The forecasted investment will hit $31.8 Billion USD by 2027 in the semiconductor and electronics industry segment alone.
Data has been transforming into ‘big data’ in every industry. The advent of automation tools, sensors which are able to communicate data over the internet through IIoT, quality/compliance data captured through mounted cameras/meters/sensors, data captured through process equipment and devices/tools used by operators, M2M communication enhancement, application integration across the layers within the IT infrastructure, data from upstream and downstream supply chain partners, data from finished products from beyond the fab– all contribute to creating the large amounts of ‘big data.’
In the electronics industry, owing to the automated or semi-automated nature of operations, data abundance hasn’t been an issue. So what does the term ‘big data’ imply for electronics manufacturing? What is the best possible way for electronics manufacturers to approach big data analytics?
In a modern SMT fab, every time a stencil is loaded or a squeegee makes a pass, data is generated. Every time a nozzle picks and places a component, data is generated. Every time a camera records a component or board inspection image, data is generated. The abundance of data in the electronics industry is a result of the long-existing and widespread process automation and proliferation of sensors, gauges, meters and cameras, which capture process metrics, equipment data and quality data.
Data is generated for every single transaction that happens on the shop floor; depending on the operation being executed there may be several data points for a given shop floor transaction. An image of a finished board for example would capture several data points which may lead to adjustment of nozzle settings or pick-up height, resulting in a reduced rejection rate post-visual inspection.
Big data exists in a ‘big’ way in electronics fabs; however, having large amounts of data and being able to leverage it to create tangible value are two separate things. In SMT and electronics the main challenge isn’t the availability of data, rather the ability to look at the data generated from the process as a whole, making sense of data pertaining to each shop floor transaction, then being able to use this data to generate information from a single point of truth instead of disparate unconnected point solutions and use the generated insight to make decisions which ultimately improve process KPIs, OEE, productivity, yield, compliance and quality.
MES translates big data into a big success for manufacturers
As we said, there is no shortage of data in SMT and electronics plants and their supply chains. The issue which plagues most fabs is the way in which a myriad of point solutions exist on the shop floor, existing in the process either due to due to legacy or custom solutions, or simply because there is not solution which overarches the entire process.
A modern MES is the answer. As an application it interconnects the IT/OT organization through integration at all levels, which ensures transactions captured through the numerous sensors, gauges, meters and devices are managed simultaneously at the edge and converted to a single-source, reliable, truth/instance, which then forms the basis of intelligence which improves decision making.
A modern MES data platform enables end-to-end visibility via data capture and analysis in an SMT line. Data becomes the centerpiece of the entire operation. Data from equipment is captured through either PLC/SCADA level integration or directly from the equipment. Data from sensors, which record process-level outputs (for example squeegee pressure, angle, speed, or nozzle performance) are used by the MES, combined with other process data for context and scope. Cameras provide another level of intelligence; the captured images of component placement is subsequently converted to actionable information by setting specific KPIs for performance, allowing process stakeholders to view operations output and performance in pre-configured forms through dashboards or on a mobile device.
Data turned into information with AI
Data pertaining to equipment availability, quality and performance (which forms the basis of OEE) and other important metrics pertaining to yield, waste, rejection and Go/No Go inspections may be displayed.
But customized dashboards and the ability to view all process related data is only the start. While real-time visibility allows for process personnel to make decisions faster, increasing production uptime, improving product quality, reducing re-work and capturing all compliance data to ensure adherence to GMP in regulated manufacturing operations, it is the insights gained through application of AI which provides the step-change in overall agility.
The real intelligence and insights generated by a MES have the potential to convert the entire operation from a reactive one to a proactive one using AI. Failure in equipment or outliers in quality and performance can be predicted and lines can be configured automatically based on shifts in demand patterns or inputs from either end of the supply chain.
McKinsey points out that almost 75% of surveyed companies have initiated an AI- related project. However, only 15% have realized meaningful, scalable impact. It is further highlighted that companies need to take a ‘smart data’ approach when it comes to leveraging AI from a Big Data perspective.
Given that, we have translated the ‘smart data’ approach for SMT and electronics manufacturing, through the lens of a MES, as follows:
- Define the process- For electronics manufacturing this would entail clearly outlining each operation and understanding the status quo, whether or not the process generates enough data and if yes, how is the data harnessed. What role do point solutions play in the overall process; and can be alleviated by a MES. Once the process is defined, with the level of integration needed, along with key areas of improvement, quality and oversight requirements, the MES can be deployed to create an enterprise view and eliminate dependency on point systems.
- Enrich the data– A modern MES data platform captures data from the ‘edge’ and accesses data from applications like historians. This allows for the creation of the necessary data context for relevancy and meaning. The MES enriches this data and allows the creation of charts, reports and trends, which enable faster decision making and reduces impacts of out of specification events.
- Reduce the dimensionality– The data collected through sensors and IIoT should be used by the MES along with engineering data to provide insights, which lead to tangible and verifiable process improvements. Data captured from the screen printing machines and squeegee movements for example, might indicate when, compared with past data, that a screen tear is imminent. Such predictive analytics facilitate activities such as replacement of the screen, allowing for downtime to be minimized. The MES might also re-route production orders automatically, allowing for overall throughput to remain unaffected.
- Apply machine learning– The MES data platform uses advanced analytics tools to create data models which target random and sporadic variations. When fully manifested, machine learning uses past and present data to predict future process outcomes. These projected outcomes allow for changes to be made, averting unexpected events which affect yield, quality and breakdowns. The intelligence created allows for both immediate and sustained improvements of the manufacturing process.
- Implement and validate the models– Once the entire process is controlled and driven by the MES, process models can be examined, to ensure that the process remains at the highest possible level of productivity in the given circumstances and data models of the process are validated through physical results. At its highest level, AI would allow for automated process orchestration and insights from the process to become predictive in nature, permitting possible failures to be detected and corrected in advance, drastically reducing costs pertaining to quality, breakdowns and compliance.
For SMT and Electronics manufacturers, the technology for improvements using data analytics is available today, through a modern MES. Instead of waiting for your customers, OEMs or the competition to drive needed changes, a modern MES allows you to be proactive in your big data/analytics journey.
A modern MES can take your modern equipment and process tools, the proliferation of IoT sensors and the mass of data generated, and bring it into a single, simple data platform for acquisition, transformation and analysis. It brings the ‘single source of truth’ needed for effective, competitive, profitable operations.
For SMT and Electronics, big data analytics means MES!
You must be logged in to post a comment.