(255f) Decision Support Platform for the Automation of Processing Recipes and Process Scheduling Management Supported By Knowledge Management

Authors: 
Muñoz, E., Centro de Investigacion en Matematicas A.C.
Puigjaner, L., Universitat Politècnica de Catalunya
Capón-García, E., ETH Zurich
Nowadays industries seek smart solutions for deal with information and data management. The effective and timely use of enterprise wide optimization models requires robust and reliable data acquisition systems to extract model relevant parameters and to drive the proposed enterprise wide coordination strategies. This work aims to create a knowledge-based platform for systematic and standardized management of the general, site and master recipes within the process industry. Besides, the system deals with data extraction from process documentation, facilitating the management of data exchange at scheduling level with different process levels. The platform allows for the creation of a master recipe ready for the production planning and scheduling, as well as for the process management. As a result, the recipe management functionalities are supported by a traceable and reliable system. Thus, different information and data required for the scheduling task is delivered to the specific analytical program as required. Finally, the framework could easily react to production system changes or analytical modeling changes, with flexible data and information structures.

An ontological model based on ANSI/ISA-88 and ANSI/ISA-95 standards is the basis for the development of the proposed algorithms in the recipe management functions, since it allows information sharing, and semantic searches within available information are possible. Additionally, the recipes and procedural elements derived from the different recipe management functions are instantiated in the aforementioned ontological model. The algorithms have been programmed in Jython and embedded with the semantic model within a software platform.

On the one hand, recipe management activity consists of five main functions. Firstly, The definition of recipe procedural elements function, consist of defining process stages, operations and actions. Then, resulting recipe building blocks are necessary for the creation of the general and site recipes, and for the definition of the master recipe procedural elements (unit procedures, operations and phases). Even more, one of the most critical issues consists of mapping of the process stages, process operations and process actions defined in the site recipes into unit procedures, operations and phases contained in the master recipe.

On the other hand, four control activities are considered in this work, such as, recipe management activity creates, master recipes, production orders processing function from ANSI/ISA-95 standard providing additional information for the production planning and scheduling, process management providing the status of the process cell and the batch progress, as well as, information to the unit supervision for batch execution, and finally, production information management reporting all available information collected from the other control functions.

In addition, the detail of the master recipe has been specified in the semantic model in order to identify the alternative scheduling problem features related to mass balances, batch assignment, sequencing and timing constraints. The main advantage of this framework consists of using a common representation approach, namely the ontological model, which can be used to retrieve the required information for optimizing any scheduling problem.