Nowadays, manufacturers must be increasingly flexible to quickly produce a high mix of on-demand, customer-specific, low volume product types. This requires flexible assembly lines with operators that are well-supported in their constantly changing assembly task, while producing high-quality, first-time-right, zero-defect products. Information coming from various supporting systems, such as ERP, MES and operator support systems, needs to be combined by the operator that configures the assembly line with materials, instructions and machine initialization settings. In this paper, we present a knowledge model that captures the main concepts and their relations in flexible manufacturing to deal with these challenges. This model is constructed by integrating existing manufacturing ontologies and can be used as the basis for information collection, exchange and analysis in information systems used in flexible manufacturing. The model supports (1) easy definition of recipe-based manufacturing instructions for engineers and operators, and (2) flexible, modular and adaptive support for human/cobot instructions. We also describe a demonstration set-up with an existing operator support system (OPS) in which the recipe concept is used in the engineering process to easily reuse existing modular components for assembling different product types.