Control and operation of heat exchanger networks using model predictive control

While, during the past decades, many strategiesfor heat exchanger network synthesis and design have beendeveloped, much less effort have been dedicated to developingonline optimal control strategies to tackle their complex anddistributed dynamics. Since past optimal control design effortspredominately revolves around centralized control ideas theytypically suffer from high computational demands. By con-trast, this paper proposes an agent based decentralized predic-tive control approach, where the computational demand isdistributed between several agents. This paper also exploresthe computational demand associated with the proposed ap-proach and compares it against a traditional, centralized, pre-dictive control approach