Purpose
This paper aims to introduce a high-temperature grease design method assisted by back propagation neural network (BPNN) and verify its application value.
Design/methodology/approach
First, the grease data sets were built by sorting out the base data of greases in a large number of literatures and textbooks. Second, the BPNN model was built, trained and tested. Then, the optimized BPNN model was used to search the unknown data space and find the composition of greases with excellent high-temperature performance. Finally, a grease was prepared according to the selected composition predicted by the model and the high-temperature physicochemical performance, high-temperature stability and tribological properties under different friction conditions were investigated.
Findings
Through high temperature tribology experiments, thermal gravimetric analysis and differential scanning calorimetry experiments, it is proved that the high temperature grease prepared based on BPNN has good high-temperature performance.
Originality/value
To the best of the authors’ knowledge, a new method of designing and exploring high-temperature greases is successfully proposed, which is useful and important for the industrial applications.