Systems modelling and simulation are two important facets for thoroughly and effectively analysing manufacturing processes. The ever-growing complexity of the latter, the increasing amount of knowledge, and the use of Semantic Web techniques adhering meaning to data have led researchers to explore and combine together methodologies by exploiting their best features with the purpose of supporting manufacturing system's modelling and simulation applications. In the past two decades, the use of ontologies has proven to be highly effective for context modelling and knowledge management. Nevertheless, they are not meant for any kind of model simulations. The latter, instead, can be achieved by using a well-known workflow-oriented mathematical modelling language such as Petri Net (PN), which brings in modelling and analytical features suitable for creating a digital copy of an industrial system (also known as "digital twin"). The theoretical framework presented in this dissertation aims to exploit W3C standards, such as Semantic Web Rule Language (SWRL) and Web Ontology Language (OWL), to transform each piece of knowledge regarding a manufacturing system into Petri Net modelling primitives. In so doing, it supports the semantics-driven instantiation, analysis and simulation of what we call semantically-enriched PN-based manufacturing system digital twins. The approach proposed by this exploratory research is therefore based on the exploitation of the best features introduced by state-of-the-art developments in W3C standards for Linked Data, such as OWL and SWRL, together with a multipurpose graphical and mathematical modelling tool known as Petri Net. The former is used for gathering, classifying and properly storing industrial data and therefore enhances our PN-based digital copy of an industrial system with advanced reasoning features. This makes both the system modelling and analysis phases more effective and, above all, paves the way towards a completely new field, where semantically-enriched PN-based manufacturing system digital twins represent one of the drivers of the digital transformation already in place in all companies facing the industrial revolution. As a result, it has been possible to outline a list of indications that will help future efforts in the application of complex digital twin support oriented solutions, which in turn is based on semantically-enriched manufacturing information systems. Through the application cases, five key topics have been tackled, namely: (i) semantic enrichment of industrial data using the most recent ontological models in order to enhance its value and enable new uses; (ii) context-awareness, or context-adaptiveness, aiming to enable the system to capture and use information about the context of operations; (iii) reusability, which is a core concept through which we want to emphasize the importance of reusing existing assets in some form within the industrial modelling process, such as industrial process knowledge, process data, system modelling primitives, and the like; (iv) the ultimate goal of semantic Interoperability, which can be accomplished by adding data about the metadata, linking each data element to a controlled, shared vocabulary; finally, (v) the impact on modelling and simulation applications, which shows how we could automate the translation process of industrial knowledge into a digital manufacturing system and empower it with quantitative and qualitative analytical technics.