The variation in stream conditions over time inevitably adds significant complexity to the task of integrating non-continuous processes. The Time Averaging Method (TAM), where stream conditions are simply averaged across the entire time cycle, leads to unrealistic energy targets for direct heat recovery and consequently to Heat Exchanger Network (HEN) designs that are in fact suboptimal. This realisation led to the development of the Time Slice Method (TSM) that instead considers each time interval separately, and can be used to reach accurate targets and to design the appropriate HEN to maximise heat recovery. However, in practise the HENs often require excessive exchanger surface area, which renders them unfeasible when capital costs are taken in to account. An extension of the TSM that reduces the required overall exchanger surface area and systematically distributes it across the stream matches is proposed. The methodology is summarised with the help of a simple case study and further improvement opportunities are discussed