The work presents a novel framework for the synthesis and optimisation of complex design processes that combines superstructure-based optimisation, semantic models (in the form of ontologies) and analytical tools. The work addresses the representation and extraction of process synthesis knowledge during the optimisation process with the purpose to simplify and interpret design results. The simplification relies on a gradual evolution of the superstructure and corresponding adjustments of the optimisation search. The interpretation is accomplished with the use of analytical tools to translate data into descriptive terms understood by users. Means of analysis include dynamic ontologies populated by computer experiments and continuously upgraded in the course of optimisation. In such a way, knowledge is developed throughout the search. The systematic interpretation of the solutions yields to an understanding of the solution space and to a systematic reduction of the representation employed. The presented approach overcomes the inconclusiveness and difficulty of translation of the solutions usually found in classical stochastic optimisation approaches as well as reduces the experiments to be performed. The approach enables monitoring the search, which is carried out in terms of the extraction of design classes at each optimisation stage. The work explains the integration of the components of the framework and gives detail of its implementation. The framework is presented for the synthesis of isothermal reactor networks, essentially addressing the challenges of a multi-level optimisation problem approached with stochastic tools. However, the approach is not restricted to any particular type of application or optimisation method. The developments are illustrated with various examples from the literature and from industry. Results show how important features and patterns are retrieved at very early stages of process design and demonstrate how the approach reduces the complexity often involved in the final solutions delivering much more clear and simple design configurations as only important features with strong impact on the performance are represented. The designer is provided with optimal design patterns that translate into practical designs rather than complex structures