Abstract: A method for detecting a characteristic of a machine or component according to a time and characteristic value record includes the steps of converting a time and characteristic value record of data samples of the machine or component in a selected time window to a two-dimensional pixel bitmap in a specified pixel window, horizontal coordinates of pixel points characterizing a time sequence of data samples and vertical coordinates of pixel points characterizing quantified characteristic values of the samples, the quantified characteristic values being pixel values converted from characteristic values of individual samples using a maximum value in the window as a standard; marking the pixel bitmap according to an existing conclusion to form a sample set for training a machine learning intelligent algorithm model; and using a pixel bitmap sample in the sample set to train the algorithm model; and using the model to detect from newly collected data samples.

Abstract: A method for predicting a remaining useful life of a machine on the basis of a data record, according to the following steps. Step 1) using regression analysis to fit a mathematical model of a machine life curve reflecting a variable relationship between time and a characteristic value, and calculating the time needed for the life curve to reach a preset failure threshold, and step 2) repeating step 1 above with a portion of data randomly omitted, and obtaining statistically a probability distribution of the expected RUL according to a repetition result. The RUL corresponding to the maximum probability distribution is determined to be the most likely expected RUL of the machine. The above method avoids bias in the prediction of machine RUL using a single life curve model, and can significantly improve the reliability and accuracy of machine prediction.

Abstract: An intelligent identification method for a vibration characteristic of rotating machinery, the steps providing converting a speed or acceleration time domain signal of mechanical vibration to a frequency domain envelope spectrum by signal processing, extracting a frequency upper limit value fmax of the envelope spectrum; at least screening out a high energy harmonic with a frequency range within fmax/Nmax by amplitude comparison. Nmax is a frequency multiple upper limit multiple for performing a frequency multiple check on the high energy harmonic. Then, extracting at least one set of characteristic parameters, based on respective amplitudes and/or frequencies, of 1-fold to Nmax-fold frequency region peaks of each high energy harmonic. The 1-fold frequency region peak of the high energy harmonic is the high energy harmonic itself. Finally, inputting the at least one set of characteristic parameters of each high energy harmonic into a machine learning intelligent algorithm to perform training and calculation.

Abstract: A method for detecting a characteristic of a machine or component according to a time and characteristic value record includes the steps of converting a time and characteristic value record of data samples of the machine or component in a selected time window to a two-dimensional pixel bitmap in a specified pixel window, horizontal coordinates of pixel points characterizing a time sequence of data samples and vertical coordinates of pixel points characterizing quantified characteristic values of the samples, the quantified characteristic values being pixel values converted from characteristic values of individual samples using a maximum value in the window as a standard; marking the pixel bitmap according to an existing conclusion to form a sample set for training a machine learning intelligent algorithm model; and using a pixel bitmap sample in the sample set to train the algorithm model; and using the model to detect from newly collected data samples.

Abstract: An intelligent identification method for a vibration characteristic of rotating machinery, the steps providing converting a speed or acceleration time domain signal of mechanical vibration to a frequency domain envelope spectrum by signal processing, extracting a frequency upper limit value fmax of the envelope spectrum; at least screening out a high energy harmonic with a frequency range within fmax/Nmax by amplitude comparison. Nmax is a frequency multiple upper limit multiple for performing a frequency multiple check on the high energy harmonic. Then, extracting at least one set of characteristic parameters, based on respective amplitudes and/or frequencies, of 1-fold to Nmax-fold frequency region peaks of each high energy harmonic. The 1-fold frequency region peak of the high energy harmonic is the high energy harmonic itself. Finally, inputting the at least one set of characteristic parameters of each high energy harmonic into a machine learning intelligent algorithm to perform training and calculation.