The Sibyl Platform is a complete Industry 4.0 solution for the maintenance departments in the industry.
It offers a set of services such as:
- Fault Detection
- Failure Prediction
- Fault identification
- Estimation of Remaining Useful Life of the mechanical equipment
which utilizes advanced machine learning algorithms and statistics.
Analysis of each module
Fault Detection allows the immediate intervention of maintenance engineers in the equipment, while the fault event is in progress. Utilizing the sensors installed in the equipment and making a combined use of multiple detection methodologies, the detection of the fault is achieved in the early stages of its development, with increased success rates. Since it does not rely on any training process, it can be applied from the first moment of operation of the Platform. It relies on state of the art techniques for detecting outliers in the senssorial input with the ability to support Big Data and fast data streams.
Failure Prediction, uses supervised machine learning and neural networks algorithms, that are trained in identifying recurring patterns before failure events. Patterns can be found either in the senssorial signals or in the maintenance logs (i.e. failure events sequences of minor importance that precede major failure events). In cases where maintenance records are limited, an innovative method of converting the senssorial input into artificial events is applied, allowing the application of the aforementioned algorithms.
Combined or alternatively to Fault Detection, Fault Identification can be also used in order to classify the detected faults into a set of predefined failure modes. Fault Identification uses several machine learning and neural networks techniques, based either on motif discovery or neural network training on historical data for classification.
Estimation of the Remaining Useful Life of the equipment
The Estimation of the Remaining Useful Life of the equipment can be configured to either estimate the remaining production cycles of the machine, or the remaining operating time, depending on the use case (i.e. whether it is a production machine that produces concrete items (e.g. a cold-forming press) or products with continues flow (e.g. an oven in the ceramics industry), respectively). Historical data of senssorial measurements until the “end of life” of the equipment, along with operation parameters of the equipment, are fed into a neural networks algorithm in order to train appropriate prediction models. The “end of life” of the equipment is a subjective concept, which adjusts according to the use case requirements. For example, it can be defined as the complete destruction of the equipment, or the duration until the production of products that do not meet the standards, or the duration until the occurrence of a major failure.
In addition to the four basic services presented, which can be applied to any field of use, the utilized architecture of the Platform enables the incorporation of additional services, that can be tailored to the needs of each customer.
How are the technicians been supported?
The Sibyl can assist the everyday inspection of the shop-floor machinery by the maintenance engineers, as it provides easy to read visualizations of the results, alerts on mobile devices with the option of providing feedback to improve the accuracy of the utilized algorithms and communication with any existing CMMS, ERP, MES, DSS, SCADA software, with the aim of immediate intervention of engineers in the equipment to prevent a failure or to repair a malfunction in progress.
Connection to AIMMS Software
The Sibyl platform can be connected to AIMMS software expanding in this way, the capabilities for automated procedures on technicians. For the connection sensors must be setuped on machinery, so that data from the sensors would flow in Sibyl.
The flow is presented below.
Except all above, Sibyl is only one part of the automated procedures that might occur. Moreover Sibyl might be connected with AIMMS platform and again with an ERP to provide with the maximum automation and productivity.