Every piece of a produced part has real shapes and dimensions that differ from their desired specifications. This can be due to the production process itself, where some causes are predictable and others are not. In order to assure the highest quality of production, these variances must be analyzed and checked to assure that the differences remain within fixed limits.
“Measuring” the characteristics of the piece is the most obvious way to assure that the variance is under control. That makes data collection a crucial step in evaluating the output of the production process. The analysis of collected data helps the organization recognize how effective their process is and where to make necessary improvements.
The reduced prices in data storage systems and the accessibility of network infrastructures have made a global impact on data systems. Traditional difficulties posed by different types of signals from various sensors and the many protocols used by the elaboration units are now reduced. This is due to more “intelligent” sensors and the adoption of communication standards across the globe. Currently, the trend is to move the elaboration functions as close as possible to the sensors, making them objects capable in themselves to communicate with one another, on a network. Elaborated data are then sent to central storage systems where they can be stored for retrieving and further processing.
The amount of data collected is constantly growing and its usage is no longer restricted to the simple conformity evaluation of a part (e.g. good or bad). Due to the advance in technology, we can now provide data to feed the execution of several industrial support systems:
- Historians for data traceability
- Statistical Process in Control Analysis
- Production Planning systems
- Manufacturing Execution Systems that help manufacturing decision makers
- Enterprise Resource Planning systems
- Data Warehouse systems
- Data Mining and Machine Learning systems with the ability to optimize the production process and support reactive or predictive maintenance policies.
As the amount of data increases, so must the processing infrastructure. The need to store massive amounts of data has promoted more innovative techniques for structural storage such as the NoSql databases and distributed file systems. Correspondingly, new technologies have become available to increase the number of distributed processing methods and algorithms.
Due to this change, presentation techniques must also be adapted to effectively communicate the relationship between data and generating events to the user. Presentation techniques can now be considered a science of itself. One example is the use of descriptive statistical methods and presentation techniques as derived from the discipline of info-graphics.