Marek Drewniak

Online energy efficiency assessment in serial production - statistical and data mining approaches

Due to increasing energy costs and the demand to lower CO2 emissions in the production, companies try to apply energy efficient technologies in green (new installations) and brown (modernized installation) production systems. Hence, the procedures of energy efficiency assessment have to be integrated with production systems on the machine control level. Presented work compares the energy efficiency assessment based on statistical analysis with an approach based on data mining and cluster analysis. Experimental results have shown limitations and possible application fields for both methods. Even though the presented methodology was conducted on compressed air consumption monitoring system, it can also be easily applicable for other types of consumed energy.

Determination of the machine energy consumption profiles in the mass-customised manufacturing

This paper presents an original methodology and related algorithms that are dedicated to monitoring energy efficiency in discrete production stations. The learning phase of the algorithm is executed during the regular production process and is based on observations of the behaviour of the production station and energy consumption measurements. The collected information is then processed using data-mining procedures in order to find the clusters that reflect the energy consumption profiles that are specific for different variants of production. The profiles are used to monitor energy efficiency and detect anomalies. The main benefit of the proposed approach is its flexibility. No additional calibrating operations or technological knowledge are required. Although the presented proof by research results are focussed on pneumatic air installations, the proposed methodology can also be used for other media. The output of the proposed solution is a very precise estimation of energy consumption in reference to a given variant of production. It allows for the accurate detection of compressed air consumption anomalies. Such anomalies can be caused by technical (machine faults) and technological (human errors) problems. The proposed methodology can be applied for the optimisation of energy consumption and for the detection of machine maintenance problems that are visible through abnormal compressed air consumption.

Application of OPC UA Protocol for the Internet of Vehicles.

Nowadays, Advanced Driver Assistance Systems (ADAS) support drivers of vehicles in emergency situations that are connected with vehicular traffic. They help to save people’s lives and minimise the losses in accidents. ADAS use information that is supported by a variety of sensors, which are responsible for tracking the vehicle’s surroundings. Unfortunately, the range of the sensors is limited to several dozen metres and even less in the case of obstacles. This shortens the time for a reaction and, therefore, there may not be enough time to avoid an accident. In order to overcome this drawback, vehicles have to share the information that is available in ADAS. The authors investigated different vehicle-to-vehicle communication possibilities. Based on an analysis of the state of the art, the authors present an original concept that is focused on applying the OPC UA (IEC 62541) communication protocol for services that correspond to the Internet of Vehicles concept.

Application of Decision Trees for Quality Management Support

The quality management process is one of the most important manufacturing activities. Although it can be implemented in a dedicated IT system, because of the easy access to production data in Manufacturing Execution Systems, it seems that an MES is a particularly convenient place for its implementation. This paper describes the concept of using decision tree methods to analyse the relationship between the production path and quality problems. The proposed method is based on an information model that is compliant with the ISA95 standard. Because of this, it can be applied not only in the case presented in the research part, but is also applicable for the problem of quality analysis in other types of discrete production. The authors present the information model that was used, the proposed method of analysis and the results for the simulation data. The simulation scenario was created as a simplification of the actual production process of electronic devices performed by AIUT company.

Improving KPI Based Performance Analysis in Discrete, Multi-variant Production

Discrete production systems face the challenge of moving from a mass to a mass-customised production model. Classic methods for analysing Key Performance Indicators (KPI) that are based on a statistical approach are difficult to apply in the case of short series, multi-variant production. A new approach for KPI analysis that is based on machine learning and data mining methods has to be applied. The authors propose a new approach that is based on K-means clustering that can be useful for performance analysis in the case of short series, multi-variant production. The presented research is focused on discrete production systems with KPI data traceability on the work cell level. The main advantage of the presented solution is its ability to automatically estimate a number of technological variants that affect a given performance indicator.

Estimation of the Number of Energy Consumption Profiles in the Case of Discreet Multi-variant Production.

An analysis of energy efficiency at the machine level has become an important element of contemporary control and measurement systems. The results of such an analysis can not only be used as information about energy consumption but can also be used for predictive maintenance. The authors present a novel approach that is dedicated to the classification of machine-level energy efficiency that can be applied in the case of multivariate production. The concept was proven by research on the use case of an assembling station that consisted of a number of pneumatic devices. The proposed approach does not require detailed analysis about the production technology that is being used and also does not require additional knowledge about the order of the production variants that are being executed. The algorithm is based on observing the behaviour of the machine and then clustering the machine cycles that are observed.

Knowledge integration via the fusion of the data models used in automotive production systems

ABSTRACT In this paper a novel information model that can be used in Manufacturing Execution Systems is presented. The model is based on the fusion of ISA95, AML and OPC UA. ISA95 is used to define, unify and describe the details of a product and production technology. It also enables communication with ERP systems. The AML standard allows information about the production facilities to be presented. The OPC UA address space represents different parts of an information model while the OPC communication protocol enables it to be linked to actual production systems. The proposed concept is illustrated using an actual example of a production line for electronic devices.

An OPC UA server as a gateway that shares CAN network data and engineering knowledge

In order to connect existing electronic devices that are equipped with fieldbus communication capabilities into the emerging Cyber Physical Systems (CPS), new models that combine live data from the underlying devices together with the metadata information that describes them have to be created. The new solutions should be based on widely accepted communication standards and should be easy to implement even in the case of limited resources. The presented work focuses on an OPC UA server that is used as a gateway to share the information that is available in a Control Area Network (CAN). It joins the live data from physical devices with the meta information that supports their proper understanding. The engineering knowledge is taken directly from the CANoe framework that is used to develop CAN-based systems. The information is available to any OPC UA client according to the OPC UA data model. The focus of this paper is on a flexible model to present the data and meta information from the underlying physical systems. The proposed solution is based on the object-oriented based communication model that is supported by the OPC UA standard. The above-mentioned idea is illustrated by a simple use case of a CAN-based application from the automotive industry.

Data Mining Techniques for Energy Efficiency Analysis of Discrete Production Lines

Machine-level energy efficiency assessment supports the rapid detection of many technological problems related to a production cycle. The fast growth of data mining techniques has opened new possibilities that permit large amounts of gathered energy consumption data to be processed and analyzed automatically. However, the data that are available from control systems are not usually ready for such an analysis and require complex preparation – cleaning, integration, selection and transformation. This paper proposes a methodology for energy consumption data analysis that is based on a knowledge discovery application. The input information includes observations of the production system behavior and related energy consumption data. The proposed approach is illustrated on the use case of an energy consumption analysis that ws prepared for an automatic production line used in electronic manufacturing.

Model of a Production Stand Used for Digital Factory Purposes

In the first part of this paper a typical realisation of a production stand is presented along with a description of stages of its implementation, layers (fields) of realisation and their interconnections and dependencies. In the second part a digital factory concept is presented. The concept is used to support the transfer of knowledge during construction and maintenance of production stands by preparation of a model which includes components, roles and interfaces of a stand. The approach allows to easily and comfortably get access to resources related to specific layers, depending on demands and needs of the user. Remote access which is based on data in a cloud forms a system of services which are available from web browser level.