Abstract
The processing quality of the block hole system affects the working performance of the marine diesel engine block directly. Choosing an appropriate combination of process parameters is a prerequisite to improving the accuracy of the block hole system. Uncertain fluctuations of process parameters during the machining process would affect the process reliability of the block hole system, resulting in an ultra-poor accuracy. For this reason, the RBF method is used to establish the relationship between the verticality of the cylinder hole and process parameters, including cutting speed, depth of cut, and feed rate. The minimum cylinder hole verticality is taken as the goal and the process reliability constraints of the cylinder hole are set based on Monte Carlo, a reliability optimization model of processing parameters for cylinder hole is established in this paper. Meanwhile, an improved particle swarm algorithm was designed to solve the model, and eventually, the global optimal combination of process parameters for the cylinder hole processing of the diesel engine block in the reliability stable region was obtained.
Highlights
The block, which is a large thin-walled structure component, is one of the most important parts of the marine diesel engine
The minimum verticality of the diesel engine block cylinder hole is taken as the optimization objective, and the Hooke-Jeeves algorithm is combined with the particle swarm optimization algorithm to solve the established reliability optimization model
When the cutting depth and feed rate are small, the chip takes away a lot of heat, which reduces the residual stress of the cylinder hole, With the increase of cutting speed and feed rate, the heat carried by the chip is limited, and the cutting thermal effect is enhanced, which leads to the continuous increase of residual stress on the surface of the cylinder hole the variable interval with relatively stable reliability fluctuation is selected as the reliability stability region constraint of the diesel engine block cylinder hole machining, as shown in Table 6, where XR(L) and XR(U) respectively represent the reliability stability region interval the lower limit and upper limit
Summary
Q is the number of output nodes, wki is the adjustment weight between the k-th output layer and the i-th hidden layer nerve. The RBF model is trained through the cylinder hole machining test result data, and the test data with sample number 16–25 in Table 2 is used as the test sample to verify the prediction accuracy of the trained RBF model. M is the number of Validation sample points, yn is the actual value of the sample points,yn is the predicted value calculated by the approximate model, and y is the average value of the test sample point set. M is the number of test sample points, yn is the actual value of the test sample points, and yn is the predicted value of the approximate model. The established RBF model is relatively accurate and meets the accuracy requirements of the approximate model
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
