Tackling the challenges of data 19 April 2022

Digitalisation projects aim to connect sensors to machines and then use that data to monitor, analyse, and control the process. The currency in this market is data, and only if it is organised and managed properly can any gains be made. While these days it is becoming increasingly easy to obtain large quantities of data, using this data to actually add value to a process is more challenging. Jody Muelaner offers a beginner’s guide

Data is used in production systems to improve productivity and quality in many ways. A range of different information systems can be applied to store, process and serve this data. These systems inform human decision making, as well as controlling automated machinery. Many of these systems were in common use way before the idea of Industry 4.0 was conceived. Below are some typical classes of information systems.

Material requirements planning (MRP) provides the most basic and essential data required by virtually all manufacturing operations, such as production planning, scheduling and inventory control.

Manufacturing resource planning (MRP II) is a more sophisticated version of MRP. While MRP is primarily concerned with materials, MRP II also includes finance and human resources. When people refer to MRP, they are often now actually talking about MRP II.

‘Historian’ stores time series data from sensors and instruments, which can be graphed to understand trends. Statistical process control (SPC) is an important and widely used application for a historian.

Human machine interface (HMI) is a control panel for a machine which enables a human operator to view data and issue commands.

Supervisory control and data acquisition (SCADA) enables real-time control and monitoring interactions between machines, HMIs and the historian. Using SCADA, an HMI can control multiple machines and view data related to multiple devices.

Internet of Things (IoT) technologies use standard internet protocols to connect devices. This can perform similar functions to SCADA, but standardised communication protocols greatly simplify the connection of different types of device from different vendors. Networking technologies such as Wi-Fi, ethernet and 5G are used. Big data analytics are also used to gain insights and make predictions from data. It is anticipated that the reduced cost and complexity of connecting devices and analysing large datasets will produce a step change in manufacturing intelligence, referred to as the fourth industrial revolution or Industry 4.0.

Manufacturing execution systems (MES) include functions such as operation scheduling and data collection. In some ways they can be seen as coming between and overlapping with MRP and SCADA.

Enterprise resource planning (ERP) describes the integration of a range of information systems related to manufacturing. These might include MRP, MES, PLM and CRM. An ERP system may be a monolithic software suite which handles all of these functions, or a core ERP system which interfaces with specialised applications from multiple vendors. Typically, only the top management interacts with the ERP and most people in the organization interact with one of the component systems which feeds into it.

Product data management (PDM) is a version control system for product specifications such as drawings. Although this is typically viewed as separate to the manufacturing systems, the MRP must receive bills of materials from the PDM.

Product lifecycle management (PLM) is an extension of PDM that also includes all the information related to a product, including planning the manufacturing process. This therefore greatly increases the interactions with the systems controlling the manufacturing processes.

These functions are clearly overlapping, and often they will not exist as discrete systems. It can be useful to visualise how the different factory information systems relate to each other in terms of an automation pyramid (pictured).


When starting out with a digitisation project, the range and complexity of systems can be overwhelming. The key is to keep it simple, starting with what you already know and thinking about how you will add value to the process. If you have an established process, you probably already know many of the important parameters and may currently track them on a dashboard.

A good place to start is to consider how a digital system could automate the analysis and actions that a human would carry out based on this data. The next step might then be to look at how algorithms could predict important outcomes, and what additional data might be required for this to be accurate. Systems are often best created by starting simple and increasing complexity gradually.

Simon Carr, general manager at Industry Forum, which helps SME manufacturers improve their performance, has three main pieces of advice for manufacturers looking to adopt Industry 4.0 and digitalisation: Don’t run before you can walk; understand the difference between data and information; and invest in training. He says: “Before implementing sophisticated, automated technologies, it is important that manufacturers ensure that the production operating system is working as smoothly as it can be. Large and leading companies may be able to make rapid developments, but not every business can make significant change as quickly. For many there is a requirement to make the most of the current process in order to create the opportunity for investment. Among other practices, we recommend OEE (overall equipment effectiveness) training, which can help businesses to understand where its data comes from and how to utilise it, justifying the investment in new technology.”

Kaizen also can use a number of approaches and tools, such as value stream mapping, which documents, analyses and improves information or materials flow required to produce a product or service, and total quality management (TQM). Carr adds: “The ideal factory is one that finds a balance between digital transformation and non-digital improvement equipment, with the former absorbed into the organisational culture.”

Data is raw and unstructured, such as sensor outputs. Information is data that has been structured and contextualised. Data only become useful once it has been transformed into information; much of the work in digitisation is in planning how to do this. Carr continues: “There is a big difference between output data and input data, the latter being more beneficial to prevent quality/performance problems. The key point is how the data is used, which should either be to respond to a ‘warning’ before it becomes a problem, or for excellent root cause analyses in almost forensic examination. The analysis should always lead to something perhaps less exciting, like updated standards and control plans.”

Jody Muelaner

This material is protected by MA Business copyright
See Terms and Conditions.
One-off usage is permitted but bulk copying is not.
For multiple copies contact the sales team.