Plant Operations – Climbing the Levels of the Smart Factory

With the rise of the Industrial Internet of Things (IIoT), sensors and devices have become increasingly accessible, enabling seamless connectivity to the internet, production equipment, and each other. This connectivity is further enhanced by the global rollout of 5G infrastructure, which supports extensive data collection from numerous devices, driving efficiency and innovation in industrial processes.

Level 1 – Available data

The first level of a Smart Factory is collecting data related to production. This shows which aspects of production could be improved, where issues are occurring, and allows the impact of any changes to be assessed in order to optimize each stage.

Information collected may include data from sensors within machines, tracking information as parts move through a facility, quality control results, machine failure rates, and so on. At the first level of Smart Factory, this information is likely to be fragmented, stored in multiple locations and unstructured, but the data is available.

If an error occurs, production engineers can access the data to investigate the problem. More generally, analysts and engineers can utilize data to discover basic insights, but this process is time-consuming, and results may not be readily accessible to decision makers. Many factories are at this current stage of data collection and connectivity.

Level 2 – Connected and accessible data

Building on level 1, a level 2 factory will connect data –usually to a centralized location–  to provide a ‘single source of truth’. Since the data is connected, it can be collected continuously –time-series data– often in real-time and is readily accessible to data analysts and engineers.

The collection of data from connected sources is one of the key technologies of a Smart Factory and is an example of IIoT (although the internet is not strictly required). Data from multiple inputs, whether from sensors embedded in machinery or RFID tags on production parts, can be transmitted automatically, and often wirelessly, to the centralized data location.

This brings us to the second key technology underpinning a Smart Factory –big data storage. Depending on the size of the facility (or facilities) and the number of data sources within it, the amount of data collected can be very large, and for many companies this data is stored in local datacenters, in the cloud, or –increasingly– using a hybrid implementation of both. The other major benefit of big data –High Performance Computing (HPC)– comes in the next level (3).

By making the data connected and accessible, data analysts and engineers can work towards the goal of more efficient and flexible production. They can look at various stages in production where time or resources can be reduced, aim to eliminate bottlenecks, identify machinery that requires maintenance, and ensure that when an issue in production does occur, investigation can happen more quickly. From the information gathered, dashboards can be developed to track key production metrics and be made available to decision makers.

Level 3 – Active Data

Although supported with computers and data processing software, levels 2 and 3 both require the input and expertise of data analysts to gain value from the data gathered. Level 3 introduces the next key technology of a Smart Factory –Artificial Intelligence (AI) and Machine Learning (ML). 

In level 3, data is actively processed and analyzed using AI with minimal human input. AI models can be created and trained using ML to generate insights or predictions with high accuracy. Real-time alerts can then be sent to engineers and other relevant personnel to allow action to be taken proactively. In level 3, factory personnel do not need to actively query or monitor the data arriving from the various sensors and tracking equipment. Due to the computationally intensive nature of ML and AI, High Performance Computing (HPC) is often used, either on-premises, in the cloud, or in a hybrid implementation.

Predictive maintenance is a key example of this. Through analysis of machine sensor readings, product results, and other inputs, the AI model can predict the likely failure of a machine or component. The system then notifies the relevant technician to fit the right spare part before the problem results in unplanned downtime.

By analyzing historical production data, a predictive system can identify areas of production that contribute most (or least) to successful production runs, resulting in faster production runs, with reduced waste, lower product faults and/or reduced downtime. The system can then recommend various optimization suggestions for review by production engineers.

Level 4 – AI-Driven Automation

The final level of a fully realized Smart Factory – level 4 – builds upon the recommendations and insights generated by the AI systems described in level 3. Using autonomous, connected systems within the factory, changes can be implemented with little or no human involvement. Issues can be detected and resolved, and efficiencies can be identified and implemented automatically. For a system to be entrusted to make decisions autonomously, there must be sufficient production data with which to train and make-reliable the ML models using big data.

This degree of automation is currently rare and often, even where the ability to implement changes automatically exists, human oversight is beneficial. As technologies and systems mature and as the amount of data available to train models becomes sufficiently large, more factories are expected to reach this highly advanced level of ‘Smart’. In the short term, stages of manufacturing that are particularly dangerous or that require continuous human operation will benefit most.

---

Taking advantage of the latest market intelligence  

The companies best placed to deal with the challenges and maximize the opportunities in these markets are those who are well prepared and take advantage of tactical industry intelligence, updated in real-time by industry experts, that provides the latest information from across the globe.   

As well as identifying the different categories of products and grasping the different business strategies, they must also engage with the main areas of change, growth, and risks in the industry.    

In addition, people responsible for employees’ professional development should use training services that lead to industry certifications and proof of expertise in the industry. 

High-performing companies recognize their executives, sales, marketing, and service staff can be more effective when armed with up-to-date information about the challenges their clients face. Cambashi Industry Insights provides tactical industry intelligence from across the globe, updated in real-time by industry experts, including:  

For more information visit our Industry Training.

---

Read more about other industries in our articles