Twin-Control pp 125-136 | Cite as

Data Monitoring and Management for Machine Tools

  • Tobias FuertjesEmail author
  • Christophe Mozzati
  • Flavien Peysson
  • Aitor Alzaga
  • Mikel Armendia
Open Access


Data monitoring is a key feature in current manufacturing industry. It provides the chance to know how the machine tool and the machining process are performing and opens further opportunities to improve process performance at various levels (maintenance, process control and even feedback to design). Twin-Control proposes a monitoring and data management architecture for machine tools. It consists of a modular local monitoring equipment that is flexible to cover most of the machine tool configurations. Monitored data at local level is uploaded to a cloud platform to perform fleet-level analysis. The proposed architecture has been working for almost two years, providing an excellent source of real data used as a basis for the rest of developments of Twin-Control.


Monitoring Data management Fleet Cloud Analytics 

7.1 Introduction

Twin-Control concept combines the development of holistic simulation models with the knowledge of the performance of the real machines and processes. To deal with this second part, a data monitoring infrastructure must be implemented so that required information is acquired, managed and analyzed properly.

The approach used in Twin-Control consists in the installation of a local monitoring hardware that acquires internal variables of the machine, collects data of additional sensors and uploads all data to a cloud platform [1]. ARTIS Genior modular is used for the local monitoring, and PREDICT’s KASEM® is used as cloud platform for data analysis. A fleet-level data analysis will be performed by integrating all the information coming from the different machines.

This chapter is structured as follows. After a brief introduction, an overview of the equipment to be monitored and integrated in Twin-Control is presented. The third-section presents the different architectures proposed to cover all the use cases. Finally, the conclusions are presented.

7.2 Monitored Equipment

Figure 7.1 presents all the machine use cases monitored in Twin-Control project. As it can be observed, a total of 12 machine tools have been monitored. Each of the industrial validation scenarios includes three machines. For the aerospace validation scenario, located at MASA aerospace structural manufacturer installations (Agoncillo, Spain), three GEPRO machines were selected as use cases, named as Use cases 1–3. The machines present a similar architecture, but they differ in the number of axes and spindles. For the automotive validation scenario, located at RENAULT automotive component manufacturer (Cleon, France), three COMAU Urane machines were selected, named as Use cases 4–6.
Fig. 7.1

Map of Twin-Control monitored and connected machine tools

Three more use cases have been monitored to test and validate Twin-Control developments during the project:
  • Use case 7: CHIRON milling machine, located at ARTIS (Egestorf, Germany), to validate monitoring tools and test new features (NC Simulation).

  • Use case 8: GORATU 5 axis milling machine, located at TEKNIKER (Eibar, Spain), to validate machine tool characterization procedure.

  • Use case 9: EMAG machine, located at the ETA-FACTORY (Darmstadt, Germany), to test and validate energy efficiency models.

Finally, three additional machines, located at relevant research locations, have been taken to implement Twin-Control features and present them to the scientific and industrial community:
  • Use case 10: Starrag EcoSpeed milling machine, located at the AMRC (Sheffield, UK), to disseminate process models and monitoring developments.

  • Use case 11: GMTK vertical lathe, located at the recently opened Basque Advance Manufacturing Centre (Zamudio, Spain), to show Machine Tool characterization procedure, among other features.

  • Use case 12: a second EMAG machine, located at the ETA-FACTORY (Darmstadt, Germany), to promote energy efficiency features.

The different use cases show a heterogeneous mixture of CNC models and communication-field buses, which represent the typical state for machine tool users. With this variety, it is difficult to determine a unique local monitoring architecture. In the next section, the different implemented architectures are detailed.

7.3 Implemented Monitoring Architecture

In each use case, monitoring hardware has been installed. Figure 7.2 presents a diagram showing the generic monitoring architecture used in Twin-Control project. However, depending on use case specifications and requirements, this generic architecture is adapted.
Fig. 7.2

Generic monitoring architecture applied in Twin-Control project

The modular configuration provided by ARTIS is perfectly suited to cover the heterogeneous applications. Figure 7.3 shows the integrated hardware in two of the industrial use cases.
Fig. 7.3

Monitoring hardware installations in Twin-Control: a COMAU Urane machine from automotive validation scenario; b GEPRO 502 machine from aerospace validation scenario

ARTIS Genior modular (GEM) is the data acquisition module that can get real-time CNC/PLC data (around 100 Hz sampling, depending on the application). Also, an HMI visualization with any PC which is connected to the plant network is possible. For the real-time connection with the CNC, two different protocols are available. For the aerospace use cases, with FAGOR 8065 CNCs, an ARTIS FAGOR-CANopen protocol has been developed. This protocol allows a real-time monitoring of 32 bytes per sample. For the Twin-Control project, 28 bytes have been used to process data and 4 bytes have been reserved to identification and process information. To increase the number of possible real-time monitored signals, the CANopen sample rate was reduced to 4 ms. By combining two CANopen channels, real-time monitoring of 56 bytes (24 variables) with a sample rate of 8 ms is provided. For the machines located at the automotive validation scenario, mounting SIEMENS 840D CNC, PROFIBUS protocol is used to connect ARTIS GEM with machines PLC. This protocol enables system to exchange up to 16 CNC sensor signals in parallel. For some other use cases, as GORATU machine located at TEKNIKER, two of these devices are installed to increase the real-time monitoring variables up to 32.

Depending on the use case and its requirements, different variables have been configured to be monitored. Also, for each variable, it has been defined if real-time information is needed. As a sample, the real-time variables monitored in the two industrial validation scenarios are presented in Table 7.1.
Table 7.1

Example of real-time (10 ms sampling) variables monitored in machines from aerospace and automotive validation scenarios

Aerospace validation scenario Gepro 502

Automotive validation scenario COMAU Urane

Target position X

Target position X

Target position Xb

Target position Y

Target position Y

Target position Z

Target position Z

Target position B

Target position A

Real position X

Target position B1

Real position Y

Real position X

Real position Z

Real position Xb

Real position B

Real position Y

Real power X

Real position Z

Real power Y

Real position A

Real power Z

Real position B1

Real power B

Real position B2

Spindle torque S

Real position B3

Torque X

Real power X

Velocity Y

Real power Y

Machine true power

Real power Z


Real power A


Real power B1


Spindle torque S1


Spindle torque S2


Spindle torque S3


Spindle speed S2


Machine true power

Using the ARTIS GEM device, visualization of monitored data in machine tool HMI (Fig. 7.4) is possible for Windows XP-based controllers, using Ethernet connection and the GEM-Visu software.
Fig. 7.4

GEM-Visu for visualization of monitored signals

ARTIS Online Process Recorder (OPR) is connected to the GEM for data storage purposes (240 GB capacity). It is also capable of monitoring non-real-time data using OPC, as a second data source. In addition, GEM and OPR could exchange real-time information using CANopen, for example, in case of critical alarm information remotely detected inside the data management system (KASEM®).

ARTIS OPR shows a special feature in the EMAG machines located at the ETA-Factory (Darmstadt, Germany). Since the energy models are based on ON-OFF switching signals of the Bosch Rexroth control and the machine power aggregates, an OPC-UA interface has been developed. The energy efficiency models, developed in MATLAB/Simulink, are running at the OPR, and the signals are read from the control via OPC-UA. By using ARTIS GEM as additional software, it is also possible to monitor the energy model results in the GEM and visualize the results in the GEM-Visu (Fig. 7.5).
Fig. 7.5

Monitoring installation at the EMAG machine (ETA-Factory): hardware installation and screenshot of ARTIS HMI displaying energy consumption results

As an example, Table 7.2 presents the non-real-time variables monitored by the OPR devices installed in the industrial validation scenarios.
Table 7.2

Example of non-real-time (1 s sampling) variables monitored in machines from aerospace and automotive validation scenarios

Aerospace validation scenario Gepro 502

Automotive validation scenario COMAU Urane



Spindle speed S1

Process mode (G 00/G 01)

Spindle speed S2

Hydraulic system status

Spindle speed S3

High pressure cooling status

Hydraulic system status

Low pressure cooling status

Spindle temperature S1

Workpiece counter

Spindle temperature S2

FMD signals

Spindle temperature S3

Part info for RENAULT

Backlash X

RMS (15–1000 Hz)

Backlash X1

RMS (15–5000 Hz)

Backlash Y

RMS (5% around RPM tool 4)

Backlash Z

RMS (5% around RPM tool 12)

Backlash B1

RMS (5% around RPM tool 15)

Backlash B2


Backlash B3

ARTIS True Power (TP) module sends power measurements from installed hall sensors to the GEM. CANopen protocol is used to make sure that the signals could be transferred under real-time conditions. The hall sensors are equipped directly at the main power supply of each machine (Fig. 7.6). Through the measuring of all three phases of the main power supply based on the TP module, it is possible to get the machine true power under real-time conditions.
Fig. 7.6

Hall sensors installed in the main electrical input for energy monitoring purposes in GEPRO 502 machine from aerospace validation scenario

ARTIS Vibration Measurement (VM) module has been also installed in different use cases. Processes accelerometer measurements real-time (internal sample rate of 25 kHz) and sends indicators (e.g. RMS) to GEM through CANopen protocol. For example, automotive use cases were provided with a vibration monitoring with sample frequencies of 25 kHz. To allow a vibration monitoring with sample rates of 25 kHz, a new firmware for the VM module was developed and installed. This firmware stored the vibration data with a sample frequency of 25 kHz at the OPR. Inside the OPR, an automatic analyzing function was developed, which calculate the RMS values for five frequency bands and send the result continuously via GEM to the GEM module (Table 7.2).

ARTIS DDU-4K-Wisy (Fig. 7.7) is a new device from MARPOSS Monitoring Solutions that has been installed in the Starrag EcoSpeed machine, located at the AMRC (Sheffield, UK). It consists in a tool holder with embedded force and temperature measurement capabilities that is used for in-process force monitoring. The DDU 4K-Wisy is based on a HSK 32 tool holder which is equipped with eight strain gage rosettes each containing two mutually perpendicular grids. Via this measurement configuration four different measuring values are captured: torque, axial force and the axial tool holder deflection in two directions (perpendicular). To acquire temperature-compensated signals, each pair of strain gages is interconnected to four separate Wheatstone bridges. To log the generated data, a wireless transmission on the ISM radio band (around 2,4 GHz) is used along with a corresponding radio receiver which either can be connected to ARTIS monitoring devices via CAN bus or to a standard PC via USB.
Fig. 7.7

ARTIS DDU-4 K-Wisy wireless force monitoring device from MARPOSS monitoring solutions

7.4 Cloud Data Management

Fleet-wide proactive maintenance functionalities consist of the construction of model coupling data analytics and expert knowledge. Raw data and machine-level computed information are therefore a major input of the cloud platform which is powered by PREDICT’s KASEM® solution. The platform is accessible via the Internet and hosted on a secure server.

Data transfer from machine to the platform requires some flexibility to match with IT and production constraints. Indeed, in factories, machine tools are connected a production Ethernet which is used to exchange synchronization information between machines, gantry, robots and various management and control systems. Depending on the size and the age of the factory, networks can be close to saturation and a continuous transfer of high sampling rate data could impact the network availability and introduce losses of information packets. In addition, from cyber-security point of view, IT department is often reluctant to “connect” production network to the Internet because bad-intentioned people could get into the network and take control of equipment for instance.

In view of theses constraints, various data transfer architectures have been applied according to use cases. In case of laboratory and pilot machine tools, OPR has a direct connection to the cloud platform. For the industrial use cases, intermediate servers have been used, and these servers are connected to the production network and enable local data storage and edge computing. For the aerospace scenario, data from the OPR is automatically uploaded to the local KASEM® platform inside MASA facilities through MASA network; then this server, also connected to the Internet, manages the data transfer to the cloud as depicted in Fig. 7.8. In this case, local server also acted as a partition between machine network and Internet.
Fig. 7.8

Data exchange configuration for the aerospace validation use case

For the automotive scenario, IT constraints were more important, and to enable the data transfer a specific architecture has been developed in collaboration with the IT architects. This architecture is also based on a local platform hosted in a RENAULT data centre in the Renault Technocentre (Guyancourt, France) and connected to the production network as shown in Fig. 7.9. Server S1 hosts KASEM® solution, and knowledge base is hosted by server S2. Server S3 is a file storage server used for binary files. Finally, an intermediate Renault/Nissan system is used for secure transfers between the local and fleet platforms. In this case, architecture management—installation, configuration and update—can be made only from user connected to Renault network.
Fig. 7.9

Data exchange architecture for automotive use case

About every 10 min, each machine—OPR module—sends and receives real-time data from/to the local or fleet platform which represents around 5 MB transferred in batch mode. About every 60 min, each machine—OPR module—sends and receives non-real-time data from/to the local or fleet platform which represents around 6 MB in batch mode. The total amount of data stored per machine could be 600 GB per year. Inside KASEM® platform, there is a specific service able to index, read and archive binary with a life cycle management of acquisition channels. Regarding transfer protocols, WebDAV and FTP are the default options for upload stream. Once the Internet connection is set up and running, it is possible to download and see a report generated by KASEM® with the Genior modular HMI Plug-in.

Cloud platform is accessible at, with restricted access to authorized users. A screenshot of the home page of KASEM® platform is shown in Fig. 7.10. This interface provides the chance to visualize collected time series, computed information and stored reports, download data, and even advanced tasks like indicator generation via algorithm implementation.
Fig. 7.10

Home page of KASEM server showing the connected use cases in the data management system

7.5 Conclusions

This chapter presents the monitoring infrastructure installed in a total of 12 machines used for implementation of Twin-Control at different levels: industrial evaluation, scientific validation and dissemination.

At local level, the monitoring architecture is based on a modular configuration. ARTIS devices are used and configured according to each use case specifications and requirements. The modular system can synchronously acquire internal variables from the CNC/PLC (real-time and non-real-time data), as well as signals from additional sensors like accelerometers, hall sensors and force transducers. This flexible approach provides the opportunity to include a wide range of equipment in the monitored fleet and is perfectly adapted to the usual configuration of end-users, with very different machines in their installations.

The data monitored at local level is uploaded from ARTIS OPR modules to a cloud platform. The PREDICT’s KASEM® platform has been storing information from most of the use cases during almost two years in Twin-Control project. Hence, the proposed architecture has been validated and has provided an excellent source of real data used as a basis for the rest of developments of Twin-Control.

The capabilities of the installed infrastructure are not limited to data monitoring and storage. In following chapters, other capabilities like advance indicator calculation, fleet analytics and even the possibility to integrate model-based control actions in the machine will be covered.


  1. 1.
    Prado, A., Alzaga, A., Konde, E., Medina-Oliva, G., Monnin, M., Johansson, C.-A., Galar, D., Euhus, D., Burrows, M., Yurre, C.: Health and performances machine tool monitoring architecture. In: E-maintenance Conference, Luleå, Sweden, 17–8th June 2014Google Scholar

Copyright information

© The Author(s) 2019

Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

Authors and Affiliations

  • Tobias Fuertjes
    • 1
    Email author
  • Christophe Mozzati
    • 2
  • Flavien Peysson
    • 2
  • Aitor Alzaga
    • 3
  • Mikel Armendia
    • 3
  1. 1.MARPOSS Monitoring Solutions GmbHEgestorfGermany
  2. 2.PREDICTVandoeuvre-lès-NancyFrance
  3. 3.IK4-Tekniker, C/Iñaki Goenaga, 5Eiba, GipuzkoaSpain

Personalised recommendations