THE ROAD TO INDUSTRY 4.0 SERIES (PART 2)

This article was first published in Instrumentation Monthly and UK Manufacturing.

The process of manufacturing involves vast quantities of data and information. The materials and components consumed in the production of goods, assembly instructions, customer specifications, machinery information such as vibration and temperature are all being constantly produced in the flow of manufacturing activities. Data is also being garnered at much faster speeds as manufacturers increasingly deploy sensors to gather further evidence on the efficiency of the processes and the status of the machines in the assembly line.

However, recent research showed that only a third of this data is put to work – the rest goes unleveraged due to obstacles such as siloed data or inefficient data management practices. That means a lost opportunity.

Digital applications such as dashboard and analytics, artificial intelligence (AI), augmented and virtual reality and computer vision offer businesses the potential to improve efficiency, reduce costs and mitigate risks, resulting in potential performance improvements of up to 20%.

In this second article of our ‘Industry 4.0’ series, John Lawson discusses how manufacturing organisations can harness their shop floor data to realise its full value and use digital applications to achieve maximum return on investment.

In the first article, our experts explored techniques to deliver rapid solutions to help companies upgrade legacy applications and obtain the benefits of Industry 4.0.

While the people, processes, systems and machinery on the shop floor produce large amounts of data, there is often an issue with how organisations gather and manage them. Typically, this is done via spreadsheets and paper reports – manual methods that are prone to error and time-consuming.

Bringing new data acquisition capabilities is critical in this context. The adoption of Internet of Things (IoT) sensor solutions integrated into existing PLC (Programmable Logic Controller) and Supervisory Control and Data Acquisition (SCADA) kits, for example, can enable organisations to acquire more data, in real time, enhancing the visibility and processing of shop floor and supply chain events.

The use of harmonisation and visualisation tools is another important step in this journey. Harmonisation tools help to unify disparate data fields, formats, dimensions, and columns into a combined dataset, decreasing both the time spent creating accurate insights and the cost of data analysis.

Visualisation tools allow for insight generation and better integration of data points into downstream business processes. Digital dashboards can pull and display key metrics such as output, scrap rate, cycle time and overall equipment effectiveness in a visual way and at one single place, offering opportunities for leaders across the organisation to collaborate and work more efficiently.

More advanced dashboard versions have functionalities that support continuous improvement processes, including root cause analysis and debottlenecking. Enhanced data visibility and analytics also benefit product development, helping engineers and product teams consider the impact of improved functionalities on the production chain prior to launching new products.

Better data utilisation can also help manufacturers optimise maintenance activities. For example, in the aerospace and aviation industry, wearable technology solutions can replace manual tasks such as verifying an engineer's certification or logging their work time. This increases the utilisation of the skilled engineer and helps to improve efficiency, ultimately reducing costs.

Due to their wireless connectivity, IoT sensors placed in machines and other manufacturing infrastructure can pull data that was previously difficult to gather, such as temperature, vibration, humidity, light, radiation and CO2 level.

Many manufacturers that handle sensitive raw materials now place small devices in containers at warehouses and production lines to detect and immediately alert managers when readings exceed safety and quality thresholds. This affords operators a quicker reaction time, reducing damage and waste and ultimately improving the reliability and utilisation of their assets.

Real-time access to this data is also helping prevent toxic leaks and other incidents, critical in industries such as chemicals or oil and gas. Beyond the shop floor, IoT devices advance inventory and distribution management by monitoring finished products in transit.

While a growing number of companies are embracing IoT in factory settings, they face some obstacles to implementation. One common pain point is managing the integration of Industrial IoT (IIoT) applications with existing technology; frequently businesses do not establish the basic principles when designing their reference architecture structure.

To enable cross-enterprise data integration, ERP, MES and PLC data must be combined in digital applications such as real-time production debottlenecking and performance dashboards, computer vision-enabled quality management and selected machine learning (ML) control applications. Selecting the right IIoT platform and a corresponding data design, which allows data coming from the various systems to be integrated, is key when maximising IoT's usefulness.

Machine learning and other AI-based technologies can help unlock further insights from the growing amounts of data being collected across shop floors and supply chains.

While AI refers to a broad set of computerised capabilities that emulate some of humans’ cognitive abilities, ML is a subset of AI that enables machines to learn from data to deliver forward-looking intelligence, without being directly programmed to do so.

Presently, one of the most powerful applications of ML in manufacturing is in the predictive process control space and setting prescription for optimal operation. This is a critical area because machine failures and unplanned equipment downtime can cost manufacturers millions of dollars every year.

Predictive maintenance software powered by ML algorithms harness the data collected by machine sensors to monitor performance 24/7 and predict technical faults, avoiding unexpected stoppages or breakages. These solutions have been proven to achieve a reduction in machine downtime and overall maintenance costs of 10% to 20%.

Enhanced Vision System solutions now incorporate AI and ML to provide high levels of automation and increased accuracy in quality control and inspection. The technology enables digital "reading" in manufacturing environments and can support the production process by performing visual tasks traditionally done by humans – for example when selecting parts, performing quality assurance or detecting defects. Vision Systems combined with AI can complete tasks within a shorter time frame and with greater accuracy than human operators, often removing errors and the potential for cognitive bias. The net effect being improved efficiency, higher run rates and reduced costs associated with downstream scrap and rework. In an increasingly challenging market, it is essential that manufacturing organisations realise the full value of shop floor data. Digital and Technology solutions such as those highlighted above, are key to providing manufacturing companies with the opportunity to kick their performance into a higher gear whilst achieving maximum return on investment.