Today’s technology makes it possible to access process data from machines and the machine periphery. But simply recording and storing data is not enough to generate value. In order to use machine data meaningfully, it must first be prepared in such a way that everyone can understand what the respective data point represents in the real world.
To make data understandable, it must be provided with context – that is, information that describes the data. We refer to this process as contextualization.
In this article, we provide insight into how exactly these tools work, why digitization projects become easier and faster to execute through such an approach – and what the process looks like in detail.
End-to-end solution for data contextualization and analysis
At ENLYZE, we have developed tools that make it possible to acquire new data points from any PLC or industrial PC and add context to them within minutes. This way, new machine data points can be added to the ENLYZE platform, relevant context can be attached to the data, and the data can be used for analysis and in dashboards.
We have paid particular attention to ensuring that our users can perform this process on their own and do not require external or internal IT specialists.
This approach offers tremendous advantages: The system can be continuously adapted to changing requirements. Because each user and process expert can make the changes independently, adaptations to the analysis dashboards are almost instantaneous. By moving process experts into the center and cutting dependencies on IT departments, the user’s motivation and adoption increase as everyone is empowered to work independently. Manufacturing anomalies and problems can thus be solved in a data-driven and efficient way.
Why are so few manufacturing companies digitized
Digitizing machines poses two major challenges:
- Reading the data from the machines
- Processing the data to make it usable
Reading out the data usually requires a combination of software and hardware. The hardware establishes the physical connection to the machine and the software acts as a translator for the respective communication protocol of the machine.
More on this topic and how we solve the problem with our edge device SPARK can be found in this article.
Once the physical connection and the appropriate protocol are in place, data can be read from the machine. The problem regarding reading out the data is thus solved.
Prepare data to make it usable for everyone
The data were originally only intended to be used internally by the machine itself and therefore often have cryptic designations. Often the designation of the data points consists only of a few letters and numerical values.
Thus, the data points cannot be easily interpreted by a human. For this, the data points must first be “translated”.
We call this “translation” contextualization. In this process, the data point is given an understandable name.
For example, the internal tag on the PLC
Ist Temperatur Heizzone 1 is translated to Actual Temperature Heating Zone 1. How exactly we arrive at this translation is clarified later in the text.
In addition, the data point is assigned a unit. In this example it is a temperature, therefore
°C°C is chosen. If necessary, the value can also be scaled. This will also be explained in detail later. After this information has been added to the data point, it is contextualized.
Only by adding the context, the data point, and what it represents is understood by everyone. This makes the data easily accessible to all users and anyone from production to controlling can use the data point for analysis or in dashboards.
The traditional approach of other vendors
Previously, contextualization was costly and time-consuming as knowledge from two different areas is needed. On the one hand, expert knowledge about the machine and the process is needed to correctly interpret the data, on the other hand, the IT knowledge is needed to establish the connection to the machines, to attach the context to the data points, and to store these data points permanently.
Today, the task is usually performed by engineering firms or automation technicians. As a rule, an implementation by external service providers takes 5-10 working days and costs accordingly.
However, with this traditional approach, there is a lot of friction between the process engineers and the IT experts involved.
The process engineers have to identify the data points that are needed and communicate them to the IT experts. Once the IT has found all the data points, they need to be validated together with the process engineers. After successful validation, the data points are then permanently stored in a system by the IT experts.
The entire process, from selection and validation to permanent storage, therefore requires at least two departments (or external partners), lots of manual work to identify the data points, and involves costs and effort for project management, coordination, and communication.
The ENLYZE approach
The contextualization solution offered by ENLYZE is based on automating manual tasks and enabling process engineers to perform the process withoutIT expertise, following the initial integration into the IT environment.
In this first step, the local IT experts and ENLYZE integrate our SPARK edge device into the network and connect it to the machine. From this point on, first data can be read out from the machine. Further information on SPARK and how it works can be found here.
All further steps, for the complete digitization of your machines, can now be performed location-independently from within the ENLYZE app and without IT support.
Contextualization in the ENLYZE App
After the connection with the machine has been successfully established, all relevant data points can be identified, contextualized, and permanently recorded with the ENLYZE App. This way, everyone can subsequently work with the data in a meaningful way and understand what is behind the respective data point.
The process is divided into 3 steps:
- Get an overview of all data points
- Explore data points
- Contextualize data points
1. Overview of all data points
In the ENLYZE app, all data points of the respective data sources (PLC, sensor, etc.) are listed automatically. In addition, all relevant information that the data source provides for the data points is also displayed (such as the PLC programmer’s comment, identifiers, etc.).
The listed data points can now be searched and filtered based on properties such as data type, data block, etc using simple strings. This allows the relevant data points to be found as quickly as possible.
2. Explore data points
Data points exported directly from the data source are often cryptically labeled. Because of this, it is difficult to find and identify the correct data points in the wealth of data points available.
For example, the meta-information of a data point of a PLC of type S7-300 looks like this:
Struct SP1 von Panel (Sollwert1)
(Interner) Datenpunkt Name:
Here the available meta-information of the data point is the only information, which the S7-300 makes available initially.
Due to this limited information, no data point can be identified clearly.
However, the following assumption can be made: It is a temperature, possibly that of heating zone 1.
To check this assumption we’ve developed our unique exploration feature.
The data point can be explored – continuously recorded and visualized – with one click: In addition to the existing meta information of the data point, the time history of the data point can then be used to uniquely identify it and quickly prove assumptions on the actual meaning of any data point.
While a data point is explored with the ENLYZE app, ideally 2-3 photos are taken of the HMI and the process parameter values displayed on it. These images (process parameter values from the HMI) can then be compared with the history of the explored values from the ENLYZE App, via a comparison of the timestamps. This procedure enables a clear assignment of all data points.
If the exploration reveals that the explored process parameter is a data point that should not be recorded, it can simply be discarded and will not be recorded any further.
3. Contextualize data points
During the contextualization process, a human-readable name, scaling factor, and unit are set.
Defining a human-readable name:
After a data point has been identified and uniquely assigned, the data point can be given a human-readable plain name.
In our example:
Ist Temperatur Heizzone 1
In this way, the cryptic name of the PLC becomes a human-readable plain name. Based on the plain name, everyone understands what the data point represents.
Scaling of data points:
However, because of the way PLCs are programmed, values often need to be scaled. In our example, the value 2200 is provided by the PLC (see image below). However, our experience tells us that temperatures above 250°C cannot occur in heating zone 1.
A comparison with the HMI images provides information. The temperature present at the given time was 220.00°C. Consequently, scaling with the factor 0.1 must be performed. Scaling can be set via the scaling factor in the ENLYZE App.
Adding a unit:
A unit must be defined for each data point. Only this defines the value of a data point unambiguously.
In our case, we are dealing with a temperature: that is why we add the unit °C.
The following shows how the contextualization of the data point is performed in the ENLYZE App:
All necessary information has now been added to the data point. A data point that was initially not uniquely identifiable has become the data point with the clear name Actual Temperature
Ist Temperatur HeizzoneHeating Zone 1, which specifies values in
°C°C, and needs to be scaled by a factor of 0,1. With this information, everyone in the company now understands what this data point represents and how to work with it.
By clicking the Contextualize button
Kontextualisieren this information is attached to the datapoint. From now on the data point is permanently streamed into the ENLYZE platform and you and your colleagues can use it for analyzing and building dashboards.
The process of contextualization thus assigns information to any data point that in turn uniquely identifies it, preservesit permanently, and makes it accessible to everyone.
The advantage is obvious. Process experts can independently identify the correct data points and embed a representation within the platform that can be understood by anyone. Only as a result of meaningful and quick analyses can be carried out by anyone.
Customize and set up the system without IT capacity.
Once our system is online and after just few initial steps of IT Setup, it is easy for the users themselves to make adjustments. Adjustments to the system, such as adding or removing a data point, can be done without (external) expertise. The steps previously required by IT and external consultants have been automated within our system.
This approach benefits our customers in several ways:
- IT resources are saved: IT capacity is only needed for initial integration
- Employees can adapt the system to their needs themselves
- Unnecessary coordination tasks between IT and domain experts for data selection are avoided ✅