Rum was a cyclic counting process of short-term load-time history, and
Rum was a cyclic counting technique of short-term load-time history, along with the load amplitude was extrapolated soon after fitting the amplitude distribution in accordance with the amplitude-frequency histogram. On the other hand, this method only considers the influence of load amplitude on fatigue damage and ignores the imply load, as well as the final load spectrum is just not ideal. Subsequently, scholars such as Nagode performed rain flow counting processing on random loads, utilised mixed two-parameter Weibull distribution to match the load amplitude and applied it to forklift parts to attain parameter extrapolation [11,12]. Around the basis of one-dimensional amplitude extrapolation, Nagode extended the concept of two-dimensional rain flow matrix, BMS-8 custom synthesis working with mixed Weibull distribution and mixed regular distribution to match the amplitude and imply on the load respectively. In accordance with the joint probability density function of the load, a parameter rain flow extrapolation technique based on mixed distribution is proposed [13]. As a result of wide application range of statistical distribution and good extrapolation effect, the extrapolation of rain flow load spectrum primarily based on mathematical statistics has steadily attracted people’s interest and has been broadly utilized. Although the mixed distribution includes a excellent fitting effect around the load, the parameter estimation procedure is comparatively cumbersome, along with the calculation time becomes the key issue of this strategy. For the weak periodicity of the load signal generated during the service on the extruder, a uncomplicated linear model is difficult to get superior prediction accuracy. With all the improvement of artificial intelligence, deep studying procedures have flourished and are steadily applied to resolve corresponding difficulties inside the engineering field. Below the guidance of deep mastering theory, a lot of variant models of neural networks are proposed, which can understand complicated nonlinear information well. At present, as one of many significant branches of deep mastering, recurrent neural network (RNN) has accomplished many successes within the fields of personal computer vision [14], speech recognition [15] and organic language processing [16]. Due to the memory capacity of RNN model to time series information, it is widely used in information prediction. Even so, because of the troubles of gradient disappearance and gradient explosion, LSTM [17] is proposed as a variant of RNN model. Compared with all the conventional neural network, LSTM network can capture the traits inside a longer time series. Hence, this paper applies LSTM network to 5MN metal extruder, and greater prediction accuracy is obtained via load information prediction to optimize the compilation of load spectrum. two. Methodology two.1. Rain Flow Counting JNJ-42253432 Autophagy approach Load cycle counting is definitely the central element of statistical processing of random load-time history. The essence of counting is to study the load characteristic value and frequency of random load-time history in the viewpoint of fatigue harm. You will discover 3 most normally made use of cycle counting approaches in engineering [18]: horizontal cross strategy, peak cycle approach and rain flow counting approach. Frendahl and Socie [19] confirmed by way of a large quantity of information analysis and investigation that the fatigue life predicted by the rain flow cycle counting technique would be the most constant with all the actual fatigue life on the mechanical structure. The counting approach is divided into single parameter counting process and double parameter counting system. Simply because the single parameter counting process only considers t.
