Planning a pump system is a difficult task. Many different load profiles (scenarios) are given and have to be supported by the system. You can choose from a wide range of pumps and fittings to include them in the system. Many configurations of the chosen components have to be considered for this task. Therefore, an exponential number of possible systems can be imagined, more than a system designer can all think of. From this set of possible systems, it is hard to find the optimal system with minimal power consumption, acquisition costs or maintenance costs.

Partners of this project developed a mathematical optimization program to compute this optimal pump system [1]. But such an optimization program is hard to read and to understand.

Therefore, a domain specific language was designed to generate this optimization program. The language includes among others pump specifications, a selection of components to be used, their connections and different load profiles. The generated program is then solved by external software tools and evaluated to show the optimal system in a simulation model. The generated optimization program is analyzed to achieve user friendly feedback about feasibility of a system with the specific requirements.

This paper presents some features of the language to proof that it is possible to specify a pumping system optimization scenario in a readable manner. Pumps are described by their characteristics, connection types and their costs. Characteristics are automatically linearized for a detailed model of the available operation points of the pump.

If it is not possible to configure a system based on the given constraints, an error report will be generated indicating the problem in the original specification. For example, there would be an error message if the required pressure cannot be achieved by only using the given pumps.

Another approach of this work is to evaluate the results of the optimization program solver. The result of the solver is a set of variables with their values. From this information a simulation model next to a layout of the cost optimal pump system are generated.

In summary this work makes the power of mathematical optimization methods usable for a wide range of users. It is easy to plan an energy optimal system with the presented tool which supports all requirements. Compiler technologies can analyze the planned system and give the user a readable feedback for his work. Also it is possible to generate faster solvable optimization programs than normal crafted programs.