This paper presents an approach for combined mass and energy integration in process synthesis and illustrates it at the production of crude synthetic natural gas (SNG) from lignocellulosic biomass and its separation in a membrane cascade. Based on a general process superstructure, the design problem is decomposed into non-linear unit models whose energy and mass balances are used as constraints in mixed integer linear programming (MILP) that targets the maximum combined production of fuel, heat and power. The flowsheet structure and its operating conditions are thereby considered as complicating decision variables in an overall non-linear and non-continuous optimisation problem that is addressed with an evolutionary, multi-objective optimisation algorithm. In a process that uses its waste and intermediate product streams to balance the heat demand, such a formulation allows for identifying intensified, overall optimal flowsheets by considering all aspects of the process design.