AI-based optimization for industrial cooling supply systems: increase energy efficiency, reduce operating costs, and improve operational flexibility—while keeping implementation effort low.
Cooling supply systems in industrial buildings are often designed and operated using conventional control methods (e.g., on/off control or PID controllers). Many important influences—such as weather, internal loads, or energy markets—are not taken into account, even though these factors strongly affect energy use and operating costs.
Why AI?
Industrial cooling systems are dynamic, multidimensional, and highly site-specific. AI methods can continuously identify the best operating strategy within this complex solution space—unlocking measurable efficiency gains.
Three companies, three challenges
The partners represent different industrial applications and demonstrate how versatile AI can be in practice.
Use case 1: Complex process cooling at Merck
Merck KGaA operates a complex system comprising wet cooling towers, an ice storage unit, and absorption chillers. Parts of the system are driven using waste heat from a combined heat and power (CHP) unit. The key challenge lies in the strong dependency on ambient temperature and the dynamic control of the ice storage system to cover peak loads.
Use case 2: High cooling water demand at Bosch Rexroth
Bosch Rexroth AG operates several wet cooling towers connected in parallel, with separate tanks for supply and return flow. Cooling water demand fluctuates significantly. To avoid losses, the systems must be precisely coordinated.
AI based control offers the opportunity to automatically operate at energy optimal setpoints.
Use case 3: Hybrid cooling in a data centre at Equinix
Equinix (Germany) GmbH runs a modular hybrid cooling system with prioritised free cooling. When outdoor temperatures rise, the system automatically switches to mechanical cooling. AI based control can determine this transition more precisely, improving efficiency in data centres.
