Method for monitoring misalignment fault of rotor system of friction nanometer generator

Abstract

A misalignment fault monitoring method for a rotor system of a friction nano-generator comprises the following steps: designing a triboelectric structure according to the size of a bearing structure in the rotor system of the friction nano-generator in an offline stage, and collecting a friction current signal under the operation condition of the rotor system by presetting different misalignment fault degrees; the sample set is used for training a deep learning network structure model; and in the online stage, friction current signals collected in real time are analyzed and judged through the trained deep learning network structure model. The rotor system misalignment fault diagnosis device monitors and diagnoses the rotor system misalignment fault on the premise of not additionally installing a sensor, can achieve quantitative evaluation of the misalignment degree, and has the advantages of being low in cost, easy to manufacture, easy to integrate, small in transmission path interference and the like. By means of a deep learning model, a diagnosis conclusion can be given in a short time.

Description

Technical Field

[0001] The present invention relates to a technology in the field of triboelectric nanogenerators, specifically a method for monitoring the misalignment fault of the rotor system of a triboelectric nanogenerator. Background Art

[0002] Triboelectric nanogenerators show broad prospects in the fields of energy harvesting, biomedicine, etc., and also provide a way for in-situ monitoring of mechanical components, that is, obtaining triboelectric signals through the friction interface pairs in mechanical components and establishing the mapping relationship between triboelectricity and physical quantities without the need to rely on other sensors. Summary of the Invention

[0003] In view of the deficiencies of the prior art that additional sensors need to be installed, resulting in installation space limitations and signal interference, the present invention proposes a method for monitoring the misalignment fault of the rotor system of a triboelectric nanogenerator, which can monitor and diagnose the misalignment fault of the rotor system without installing additional sensors, and can quantitatively evaluate the degree of misalignment, with the advantages of low cost, simple manufacturing, easy integration, and little interference from the transmission path. With the help of a deep learning model, a diagnostic conclusion can be given in a short time.

[0004] The present invention is realized through the following technical solutions:

[0005] The present invention relates to a method for monitoring the misalignment fault of the rotor system of a triboelectric nanogenerator. According to the bearing structure size in the rotor system of the triboelectric nanogenerator, a triboelectric structure is designed. By presetting different degrees of misalignment faults, the triboelectric current signals under the operating conditions of the rotor system are collected in the offline stage as a sample set for training the deep learning network structure model; in the online stage, the triboelectric current signals collected in real time are analyzed and judged through the trained deep learning network structure model.

[0006] The design of the triboelectric structure means: adding copper foils outside the outer ring of the bearing in the rotor system to form interdigital electrodes, and installing them in the bearing housing of the rolling bearing in the rotor system to form a triboelectric bearing.

[0007] The addition of copper foils means: arranging arc-shaped copper foils alternately between adjacent outer rings of the bearings. Technical Effects

[0008] The present invention generates triboelectric signals through the internal bearings of the rotor system, and realizes the mapping between triboelectric signals and rotor misalignment through deep learning, without installing additional sensors, and can automatically sense the misalignment fault of the rotor system. Description of the Drawings

[0009] Figure 1 It is a flow chart of the present invention;

[0010] Figure 2 It is a schematic diagram of a three-span rotor system;

[0011] In the figure: servo motor 1, torque and speed sensor 2, coupling 3, bearing pedestal 4, triboelectric bearing 5, interdigital electrode 6;

[0012] Figure 3 It is a schematic diagram of the interdigital electrode structure;

[0013] Figure 4 It is a neural network architecture diagram;

[0014] Figure 5 It is a schematic diagram of the triboelectric signal. Specific implementation manners

[0015] As Figure 1 shown, this embodiment relates to a method for monitoring the misalignment fault of a rotor system, which specifically includes:

[0016] Step 1: Build a rotor system, and design a triboelectric structure according to the bearing structure size in the rotor system: In this embodiment, a three-span rotor system as Figure 2 shown is built, and the interdigital electrode as Figure 3 shown is pasted on the outer surface of the bearing outer ring.

[0017] Step 2: Simulate different misalignment fault degrees by placing cube gaskets with different thicknesses on the coupling, and obtain triboelectric current signals under different fault degrees and different working conditions.

[0018] Step 3: Build a ResNet18 deep learning structure network model as Figure 4 shown, and input the triboelectric current signal after data normalization into it for network training. When the optimal solution or the maximum number of iterations is reached, a misalignment fault diagnosis model for the rotor system used in the online stage is obtained.

[0019] The ResNet18 deep learning structure network model includes: seventeen consecutive convolutional layers, average pooling layers and fully connected layers.

[0020] Step 4: Use the misalignment fault diagnosis model of the rotor system for the misalignment fault of the rotor system running in online monitoring.

[0021] Through specific actual experiments, an interdigital electrode is placed in the bearing seat to form a triboelectric bearing, and three states of normal, misalignment fault degree of 0.1 mm, and misalignment fault degree of 0.5 mm are set. The experimental data obtained at a rotational speed of 250 rpm are respectively as Figure 5As shown, it can be seen that as the misalignment fault degree deepens, the amplitude of the friction current signal is continuously decreasing. The signal for each state is collected for 50 seconds, sliced without repetition at intervals of every 1024 data points, and 80% of the data is selected as the training set, with the remaining as the test set. After normalizing the signal, it is input into the ResNet18 network for training to obtain a training model. Finally, the test set data is used to verify the diagnostic accuracy of the model, and it is found that it can reach 96.2%.

[0022] Compared with the prior art, this method introduces triboelectric nanogenerator technology into the rotor system to realize in-situ monitoring of rotor misalignment faults, eliminating the economic cost of installing additional sensors, providing a low-cost monitoring solution, and verifying the effectiveness of the present invention through experiments.

[0023] The above specific embodiments can be locally adjusted in different ways by those skilled in the art without departing from the principles and purposes of the present invention. The protection scope of the present invention is subject to the claims and is not limited by the above specific embodiments, and all implementation solutions within its scope are subject to the present invention.

Claims

1. A method for monitoring the misalignment fault of the rotor system of a triboelectric nanogenerator, characterized in that, In the offline stage, design a triboelectric structure according to the bearing structure size in the rotor system of the triboelectric nanogenerator. By presetting different misalignment fault degrees, collect the triboelectric current signals under the operating conditions of the rotor system as a sample set for training the deep learning network structure model; in the online stage, analyze and judge the triboelectric current signals collected in real time through the trained deep learning network structure model. The design of the triboelectric structure mentioned above means: adding copper foils outside the outer ring of the bearing in the rotor system to form interdigital electrodes and installing them in the bearing housing of the rolling bearing in the rotor system.

2. The method for monitoring the misalignment fault of the rotor system of a triboelectric nanogenerator according to claim 1, characterized in that, The addition of copper foils mentioned above means: arranging arc-shaped copper foils alternately between adjacent outer rings of the bearings.

3. The method for monitoring the misalignment fault of the rotor system of a triboelectric nanogenerator according to any one of claims 1 or 2, characterized in that specifically It includes: Step 1: Build a rotor system and design a triboelectric structure according to the bearing structure size in the rotor system: Build a three-span rotor system and paste the interdigital electrodes on the outer surface of the outer ring of the bearing. Step 2: Simulate different misalignment fault degrees by placing cube gaskets with different thicknesses on the coupling to obtain triboelectric current signals under different fault degrees and different operating conditions. Step 3: Build a ResNet18 deep learning structure network model, input the triboelectric current signals after data normalization into it, and conduct network training; when the optimal solution or the maximum number of iterations is reached, obtain the misalignment fault diagnosis model of the rotor system for online monitoring. Step 4: Use the misalignment fault diagnosis model of the rotor system for the misalignment fault of the rotor system running in online monitoring. The ResNet18 deep learning structure network model mentioned above includes: seventeen consecutive convolutional layers, average pooling layers and fully connected layers.

Images