Compress

Reliable feed pump technology is an essential prerequisite for the expansion of deep geothermal energy, i.e. the use of geothermal energy from the earth's crust at depths of more than 400 m, particularly in view of the planned geothermal and mine heat storage projects for the conversion of existing district heating systems in the Ruhr metropolis. New technical approaches to increase the efficiency and service life of these pumps as well as prediction systems for imminent pump failures during operation are of great interest. Due to the prevailing environmental conditions in which these types of pumps operate, efficiency and service life are sometimes greatly reduced, for example due to increased wear and deposits. In addition, there are frequent sensor failures, which also leads to a direct reduction in efficiency, as the pumps are only operated at a greatly reduced output for safety reasons. Consequently, a scientific investigation of computer-aided optimization of maintenance intervals and an improvement of sensor technology in the field of pumping technology in deep geothermal energy are indispensable.

The aim is to significantly minimize the immense costs caused by frequent pump changes and the associated long system downtimes. The prerequisite for this is the detection of sources of error, such as material wear, deposits and thermal loads, by monitoring ongoing pump operation in combination with computer-aided prediction models for planning optimized maintenance intervals (predictive maintenance). For this purpose, it is necessary to characterize the relevant operating conditions as well as the wear parts of the deep pumps used and to monitor the operation of the individual pump components using sensors or by evaluating operating data. Given the prevailing temperature levels, hydrochemical conditions and borehole and pump geometries, this poses increased challenges for the sensors, signal transmission and processing. The innovative core of this project lies in the technical implementation of intelligent pump monitoring with a connection to a condition monitoring system (CMS), which is to provide statistical predictions about the condition of a borehole pump using machine learning. This requires both intelligent embedded systems and their communication technology link for the central storage of recorded operating data and the realization of predictive maintenance.

Federal Ministry of Education and Research (BMBF)

13FH0I41IA

• Timon Sachweh
• Alexander Stein