Key Take Aways

• By implementing reliable process monitoring in manufacturing processes, costs related to downtime and production errors can be significantly reduced.

• Live dashboards displayed on screens in the workshop are a proven means of reducing reaction times during disturbances and irregularities thanks to their visibility.

• Challenges in implementing live dashboards include:

  • the provision of real-time data from various data sources

  • the flexible design of dashboards according to the requirements of different user groups

• The visualization tool Grafana combined with the ENLYZE data platform provides a complete infrastructure for creating live dashboards

  • from the plant control to the TV screen

  • low IT effort for the factory IT

  • Dashboards can be individually created and iteratively improved by users themselves

Why Process Monitoring?

Process monitoring is an essential part of modern manufacturing processes because it ensures safe, reliable, efficient, and economical operation of production. Monitoring the production process allows companies to identify and rectify problems in a timely manner, thereby avoiding costly production errors and reducing waste.

Additionally, process monitoring can also provide valuable data that can be used to improve manufacturing processes and increase efficiency. By analyzing production data, manufacturers can identify areas for improvement and make adjustments to optimize their operations.

Options for Process Monitoring

In continuous manufacturing processes, the control (PLC) of the manufacturing system regulates the key process variables to a constant target value. However, disturbances and defects can interfere with automatic control, as can external factors such as ambient conditions, raw materials used, or even operator errors. Therefore, it is still necessary to continuously monitor the actual values in manufacturing processes.

There are various methods of process monitoring that can be differentiated, among other things, by the location of monitoring.

Reading values directly at the measuring point

In older manufacturing systems, alongside a PLC, analog measuring devices are often installed, such as pressure gauges for pressure measurement as shown in the illustration. With such measuring instruments, the measured value is not processed electronically in any way, but is displayed directly at the measurement location on an analog scale.

Similar situations often occur with more modern, retrofitted measuring systems. While measurement values are captured and processed digitally, they can only be displayed within the manufacturer's proprietary software.

What both cases have in common is that in addition to the analog instrument or the purchased measuring system, there are usually other displays along the production line that measure critical process variables. As a result, employees may only recognize disturbances when they look at the display that shows the noticeable measured value. The consequence is extended reaction times before corrective measures are taken.

Undetected disturbances have the potential to cause significant damage. This ranges from increased production scrap and rework to unplanned downtimes or even compromising working and environmental safety.

Reading values at the plant HMI

Modern production systems have one or more human-machine interfaces (HMI), usually in the form of a screen with a keyboard, through which the target values for the process are set. Additionally, the current actual values of these process parameters as well as disturbances are displayed on the HMI.

The process monitoring using the existing plant HMI offers several advantages compared to reading at the measurement point. By directly comparing the target values and actual values, deviations can be quickly identified. Furthermore, the process can be adjusted or even halted directly at the HMI when a disturbance requires it.

However, there are also disadvantages to using only the HMIs of the systems for process monitoring. HMIs are primarily not designed for continuous monitoring but for operating the system. Accordingly, the HMI user interface is structured. The set parameters are displayed, but not necessarily all process variables relevant for monitoring are visible at a glance. Adapting to different user groups and their requirements is also not provided for.

Furthermore, the low visibility of HMI screens is problematic. Due to their size and positioning, the screen and process variables can only be perceived by someone standing directly in front of it. Lacking options for process data storage means that anomalies can only be detected at the moment they occur and not afterwards.

Live dashboards on screens in the workshop

A newer approach to monitoring manufacturing processes involves visualizing process variables in real-time on dashboards. These live dashboards are constantly displayed on large screens placed at various points in the workshop.

Live dashboards differ significantly in their characteristics from the two previously described methods of process monitoring.

For example, screens can be installed in any location within the facility, independent of the location of the manufacturing system. It is thus possible to strategically place the dashboards to achieve high visibility for as many people as possible. This allows system operators to keep an eye on the process, even when they are needed elsewhere in the facility.

Additionally, charts and color coding can make process data easily interpretable. If, for instance, a display shows a speed of 25 RPM, only people directly involved in the process will judge whether this 25 RPM is good or bad. However, with a displayed speed of 25 RPM that is flashing red because a set limit has been undershot, even a layperson understands that an intervention in the process is needed.

Live dashboards thus allow for the simultaneous monitoring of a greater number of process variables by a larger number of people. This can drastically reduce the reaction time to take corrective measures during disturbances.

However, there are also challenges in process monitoring with live dashboards. The setup of live dashboards is often associated with high IT effort. Process data from different parts of a plant (and thus from different manufacturers) must be automatically captured and made available in real-time to a suitable visualization tool. Moreover, the dashboards need to be designed and brought to TV screens to achieve the desired visibility.

It is often the case that such dashboards in manufacturing companies are designed either by the IT department or external service providers due to the complexity of the associated setup processes. Future adjustments can only be implemented by them in these cases.

This is problematic because involving internal or external IT departments drives up the effort and thus the costs for the introduction and maintenance of live dashboards. Since the internal or external IT designs the dashboards but does not use them themselves, the design and subsequent adjustments must be worked out in often lengthy iteration cycles between production and IT.

Which method is the best?

Compared to live dashboards, all previously mentioned methods of process monitoring share the characteristic that they are usually already installed in the system and that monitoring of the production process is already possible through them.

The question that arises here is not whether live dashboards on TV screens or the other tools for process monitoring should be used. Rather, it is about whether live dashboards should be added to the existing systems.

Considering the clear advantages of increased visibility, thereby reducing reaction time, and flexible adjustment options, the question can clearly be answered with "Yes, I would like my live dashboards today!" However, this is conditional on overcoming the challenges mentioned previously.

For a successful implementation of live dashboards, two things are required:

• An infrastructure capable of providing data from various components from different manufacturers in real-time.

• A visualization tool that allows dashboards to be designed and displayed in the workshop without requiring extensive prior knowledge or significant IT effort.

Our Solution: Live Dashboards with the ENLYZE Platform and Grafana

At ENLYZE, we combine our data platform with the visualization software Grafana into a complete solution for creating live dashboards. Our edge devices read process data directly from machine controls and peripheral devices and send them via a secure VPN connection to the ENLYZE cloud. There, they are enriched with contextual information such as labels or physical units. The cloud provides the enriched data in real-time tools such as here, for example, Grafana.

Grafana is a web-based tool that does not require additional software to be installed on company computers, as the dashboards are built directly in the web browser. To bring a dashboard created in Grafana to the screens in the workshop, only commercially available televisions with internet access are needed.

Once the relevant machines are integrated into the ENLYZE platform, dashboards can be designed freely in the Grafana user interface. It is intuitive enough that no extensive prior knowledge is necessary to create one's own dashboards.

As illustrated in the image, the process variables to be monitored, the type of visualization, or limits can be defined via simple dropdown menus and text fields. Once the dashboard is completed, it is saved and can be accessed through any web browser by entering login credentials.

Practical Examples: Live Dashboards for Different User Groups

Due to Grafana's ease of use and the straightforward distribution via web browsers, live dashboards can be deployed in manufacturing without the factory IT needing to be involved after the initial setup. This means that dashboards can be designed and adjusted by the people who actually use them to monitor the process.

Furthermore, individual dashboards can be created for different groups of people, visualizing only the relevant process information for them. Below, we provide two practical examples of such individual dashboards.

Example 1: Dashboard for Plant Operators/Shift Managers

The task of monitoring a manufacturing process primarily falls to the employees who operate the system and thus directly control the process. A dashboard that supports them in this task answers typical questions at a glance, such as:

• Are critical parameters within the tolerance range?

• Are the different system components set correctly?

The dashboard below for a sample extrusion process illustrates the strengths of live dashboards. The data from multiple system components, here from three extruders with four dosers each, as well as a winder at the end of the process, is visible at a glance.

Even without knowledge of exact process specifications or extrusion processes in general, we can recognize that there may be a problem with Extruder C because there are process values marked in red. In practice, either the plant operator would notice the irregularity themselves, or they would be alerted by others looking at the dashboard.

The bar chart below the dashboard shows the total weight of the last produced rolls. The weight is calculated automatically from the extruder throughput and the current winding length of the winder. Such and other more complex calculations can be directly implemented in Grafana.

Example 2: Cross-Sectional Productivity for Management

In contrast to plant operators, individuals in management positions are usually more interested in the productivity of entire production areas than in the process variables of individual system components. Typical questions from factory management that live dashboards can answer include:

• Are all systems running? How effectively is my area currently producing?

• Is the time-critical order xyz already being manufactured?

• Were there any downtimes while I was not present during the night shift?

The dashboard below provides productivity information for each of the four systems in a production area. The designation of the currently manufactured product comes from the ERP or BDE system of the company, which, like the system control, is connected to the ENLYZE platform via an edge device.

The performance parameters throughput and line speed come from the system control and are provided with thresholds that are assessed by color markings to indicate whether a particular speed corresponds to a high or low performance level. The downtimes displayed in the right column are automatically calculated from the time series data of the line speed. Any time intervals in which the line speed falls below a certain threshold are classified as downtimes.

For employees with planning or management tasks, it is advantageous that Grafana dashboards can be accessed via web browsers. This allows them to track what is happening in manufacturing directly from their office without having to go to the production floor or wait for weekly reports.

Conclusion

Live dashboards are excellently suited for monitoring modern manufacturing processes due to their ability to provide viewers with information from various sources at a glance. They can complement already existing displays and make new process information available in real-time through their own calculations.