Gear Remaining Life Prediction Method Integrated with Kernel Estimation and Random Filtering

Abstract

A kernel estimation and random filtering theory-based gear residual service life prediction method belongs to the technical field of mechanical reliability. The method comprises the following specificimplementation steps: 1, monitoring a gear degradation state in a main test gearbox in real time by using an acceleration sensor; 2, performing feature extraction on the gear degradation state; 3, performing non-parametric estimation on a probability density function of a continuous degradation state of the gear by using the characteristic that a kernel function does not make any assumption aboutdistribution of the data and starts from the data sample to obtain the probability density function of the degradation state of the gear based on the real-time state monitoring data; 4, updating a random filter recursive model parameter by using the real-time state monitoring data, and establishing a kernel estimation and random filtering-combined prediction model; and 5, predicting the remainingservice life of the gear through the kernel estimation and random filtering-combined prediction model. The kernel estimation and random filtering theory-based gear residual service life prediction method has the advantages that the degradation state and the real-time residual service life of the gear can be effectively predicted and a basis is provided for preventive maintenance of the gear.

Description

technical field [0001] The invention belongs to the field of mechanical reliability design, and in particular relates to a method for predicting the remaining life of a gear. Background technique [0002] Gears are the key components in the transmission system of mechanical equipment widely used in the machinery industry. When gear failures such as broken teeth, tooth surface fatigue, and gluing occur, it often causes catastrophic damage to the entire mechanical equipment. Taking the wind turbine as an example, the gear failure rate is the highest in the entire wind turbine, accounting for about 60%, and its maintenance cost is also high, accounting for about 40%. The problem that needs to be solved urgently, and in the whole maintenance plan formulation process, the remaining life prediction of the gear is the focus and difficulty. With the development of information sensing equipment, the real-time monitoring of the running state of the gear is carried out, and the large ...

AI Technical Summary

Problems solved by technology

The existing prediction methods have the following problems: First, the existing prediction methods need to assume the structure of the state degradation model, and it is necessary to assume that the samples used as the basis for judgment conform to a specific model structure. The relationship between the assumptions of these model structures and the actual physical model There is often a large gap; secondly, most of the parameter estimation problems involved in the prediction model cannot guarantee global convergence; finally, because the gear is in a changing environment, its state degradation model will change, and a single prediction model cannot Adapting to changes in the environment requires the combination of multiple forecasting models,

Images