DELMIA Virtual Production for Production Performance
The DELMIA Portfolio is grouped into distinct domains within manufacturing to offer solutions to enable your entire digital factory.
DELMIA - Process Rules Discovery (RDY)
Process Rules Discovery helps companies meet their production objectives in terms of quality, yield, environmental footprint, etc. Process Rules Discovery typically saves valuable time and money in rework.
The objective of Process Rules Discovery is to extract Best Practices and Risk Situations from historical data.
The results of the analysis expressed as rules will enable the user to gain knowledge about the process/product and to take decisions on how to improve yield by applying Best Practices and preventing Risk Situations.
Process Rules Discovery is a totally unique manufacturing intelligence product that allows to improve yield by examining past performance in order to predict future performance.
Based on historical data, Process Rules Discovery uses patented pattern-recognition technology to extract from a set of process and product characteristics which combinations have been successful and which ones generated issues (in terms of quality, yield, environmental footprint…) These best operational practices and risk situations will be displayed in laymen’s terms facilitating communication across the company to the shop floor.
The real challenge is to provide a simple solution to complex issues. Engineered by scientists but not for scientists, Process Rules Discovery is an algorithm-based application that offers a simple solution by providing data analysis that is straightforward, flexible and easy-to-deploy. Process Rules Discovery’s patented algorithms extract, optimize and allow editing of rules that are easy to understand and easy to communicate. It conveys this information through Best Operational Practices (BOP) enabling collaborative intelligence across your company. Rather than relying on subjective opinions or complex models, Process Rules Discovery allows the user to make decisions based on facts and historical lessons quickly. This ensures wide-spread adoption of BOPs throughout manufacturing operations, new product introductions and continuous improvement teams. Process Rules Discovery is designed to help you solve two important problems:
Accelerating Industrialization The industrialization problem could be summed up by the following sentence: “how do I go from a design or a prototype to a finalized product, ready to be shipped to my clients?” Very often, industrialization – the process used to create a viable product from a new design – can be complex, long and costly. Through its advanced technology, Process Rules Discovery is able to isolate the most important factors in an industrialization process, and pinpoint a possible solution to the many issues that can delay it. Process Rules Discovery therefore offers your company a competitive advantage, because you can commercialize new products more quickly and more efficiently than your competitors.
Optimizing the Production Process Once a product has gone through the industrialization process and has been in production for a longer period of time, new and unexpected problems can appear. For instance, due to new circumstances (a new key supplier, new machines, etc.), the level of defective products can rise to unacceptable levels. In a complex industrial process, it is often very hard to find the root cause of these production problems. Using Process Rules Discovery, you can determine the source of these issues and drastically reduce the number of products that are considered defective, thereby increasing the profitability of a production unit. Even if a plant has a well-tuned production process, Process Rules Discovery is able to suggest possibilities to refine and optimize this process even more.
Simple user interface
Rule analysis and optimization
A model of variables
Project management and reporting
Publication of Best Practices
Product Key Customer Benefits
The historical data that is analyzed is considered facts. Therefore the resulting rules are always supported by data. This can permit for example to have experts agree on the explanation of a problem. Value: all the results are proven /supported by the data.
Empirical approach The analysis is based on actual data from real life, either production data or experiments. Value: this succeeds when no theoretical model exists (or is too complex to maintain).
Analysis on qualitative and quantitative data The analyzed data set can contain text values (qualitative data) as well as numerical or date/times values (quantitative data). The resulting rules can handle a mix of variables of different types. Value: other alternate analysis approaches are often stuck in such situations (for instance, only quantitative data is supported).
Cope with non-monovariate problems The resulting rules provide explanations by combining several parameters, taking into the multidimensional aspect of the problem. Value: this essential feature enables to solve complex problems were the root cause may not be straightforwardly linked to a single parameter.
Handles non-standard statistical distributions or small amount of data The analyzed data does not need to follow a normal distribution. Analysis can be conducted with a very limited data set (that would be considered non-statistically significant). Value: other alternate analysis approaches are often stuck in such situations.
Handles missing data Parameters or records (rows) do not have to be ignored because some values are missing. On the contrary Process Rules Discovery will use all the existing values. Value: other alternate analysis approaches are often stuck in such situations.
Interactive Experiment Planner Process Rules Discovery has the ability to generate experiments to explore previously untested process configurations (still within existing operating space). Value: this approach is different from usual DOE as the user can focus on an area of the domain space and tell the system how many experiments he can afford to conduct.
User Friendly Ease and speed of use, user friendly, intuitive Patented visualizations Value: does not require extensive statistical training or to have a statistician working with the software.
Interactive Explore the unknown through “what-if” analysis. Build rules from scratch and test them against the data. Incorporate business constraints and human expertise. Value: the results can be adjusted and optimized by the business experts according to their own expertise without any statistic assistance.
Knowledge sharing Process Rules discovery permits to formalize the knowledge in the form or rules, whether these rules are the results of an analysis or the direct input from the experts or a mix of both. Value: knowledge capitalization permits to facilitate knowledge transfer or technical transfer. "Black box" techniques generate models that cannot be interpreted.