(580a) An Ontology Supported Integration Framework for Models and Data in Biorefining

Process modelling and simulation are vital tools to plan, evaluate, assess and develop different alternatives in the biorefinery design for production of value added chemicals, materials, fuels, and energy from biomass. However, modelling for biorefining processes poses a number of technical challenges due to the complexity of the characteristics of biomass feedstock, and associated processing technologies evolving multiple pathways, including thermochemical-, catalytic-, biochemical- and hybrid-conversion routes. Due to the complexity and variety of these processes, numerous software tools have been developed to obtain the best-of-breed software for each application area, to fulfil the requirements of particular engineering tasks. During the process of developing a new model, the best suited modelling tools for different parts of the process are employed. However, to understand process design as a continuous work process from an integrated perspective requires the use of tools from a diverse range of sources and disciplines simultaneously. With the current increasing number of Computer Aided Chemical Engineering (CAPE) tools, the emergence of interoperability between heterogeneous modelling tools and process simulators was demonstrated by Braunschweig et al. (2000), which has ultimately led to the CAPE-OPEN initiative.

The CAPE-OPEN has conceptualised and developed a standard to define a set of interface specification as a method pertaining to the interoperability of the mainstream flowsheet applications (Pons 2010). The interoperability is facilitated by a standard middleware service, which enables the use of a single set of common interface as a communication tool (Braunschweig et al. 2004; Morales-Rodríguez et al. 2008). Having the common interface increases the potential reusability of existing models without having to invest in the software development, which further reduce the developing time, efforts, and the need for expertise (van Baten & Pons 2014). However, to fully exploit the potential reusability of existing models, the task to become aware of and have access to all of those models from a variety of sources is a challenge, yet required in order to identify the most suitable model to achieve the best solution for a particular engineering problem. As a consequence, Yang et al. (2008) indicated the potential misuse of the models that can further lead to inefficient or even wrong solution to the problem.

This paper is intended to address the challenges that previous research, CAPE-OPEN framework has faced and hence to minimise user intervention in selecting a model for the purpose of model integration. To enable automated discovery process in order to reduce the reliance of user intervention, the semantic representation is used to describe the models and data in terms of their input and output parameters, preconditions, and functionality (Koo & Cecelja 2015). Additionally, the Semantic Web Services (SWS) are employed to create a semantic profile of models and data representing biorefining processes. To this end, the domain ontology assures unified description of models by guiding the model registration and hence instantiation of ontology through ontology parsing, whereas the SWS assures the process of model discovery, its invocation and consequently integration. To demonstrate proposed approach, the framework is built for the models and data residing in web-based semantic repository in the domain of biorefining. In the domain ontology, the models are characterised in terms of functionality, modelling methods, model granularities, required software environment, etc. and their input and output parameters are characterised in terms of type (i.e. material, utility, economic and emissions), compositions and associated parameters, etc. The classification and characterisation of parameters used in matching process are not limited to numerical values (i.e. value for the flow rate), but also includes descriptive (i.e. type of material, energy) or even composite (i.e. range of flow rate with minimum and maximum values) types. The options of the discovered models to the requesting model, the input-output (I/O) matching technique is employed to identify the models that matches to a certain degree of request, which permits for a high degree of flexibility by allowing a partial matching rather than a full (exact) matching between I/O parameters. The discovered models from matching process are then ranked according to a set of preferences and the best option that satisfies the matching criteria is proposed to the user. To verify the performance of model discovery and selection, models representing biochemical- and thermochemical conversion processes are used to demonstrate as a case study with an aim to evaluate the best suited model whilst avoiding wrong design solution.

References:

Braunschweig, B.L. et al., 2000. Process modeling : The promise of open software architectures. Chemical engineering progress, 96(9), pp.65â??76.

Braunschweig, B. et al., 2004. CAPE web services: The COGents way. Computer Aided Chemical Engineering, 18(C), pp.1021â??1026.

L. Koo, F. Cecelja, â??Model Integration Using Ontology Input-Output Matching,â? Computer Aided Chemical Engineering, 2015, volume 37, pp. 2567-2572.

Morales-Rodríguez, R. et al., 2008. Use of CAPE-OPEN standards in the interoperability between modelling tools (MoT) and process simulators (Simulis® Thermodynamics and ProSimPlus). Chemical Engineering Research and Design, 86(7), pp.823â??833.

Pons, M., 2010. How to make use of CAPE-OPEN? In 2010 AIChE Annual Meeting, 10AIChE.

van Baten, J. & Pons, M., 2014. CAPE-OPEN: Interoperability in Industrial Flowsheet Simulation Software. Chemie Ingenieur Technik, 86(7), pp.1052â??1064. Available at: http://doi.wiley.com/10.1002/cite.201400009.

Yang, a. et al., 2008. A multi-agent system to facilitate component-based process modeling and design. Computers and Chemical Engineering, 32(10), pp.2290â??2305.