Method for calculating residual life of ion tube

Through the ion tube aging characteristic curve calculation method based on LSTM, the problem of the inability to accurately calculate the remaining life of the ion tube in the prior art is solved, the accuracy and reliability of the calculation are improved, the operation and maintenance costs are reduced, and the absolute and relative costs are provided. Lifetime information reduces the operation and maintenance pressure of the equipment.

CN119047329APending Publication Date: 2024-11-29HUNAN NORMAL UNIVERSITY
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411253281.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-09
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The existing technology cannot accurately calculate the remaining life of ion tubes, resulting in increased equipment operating costs and misjudgment of life, and the inability to provide maintenance and replacement requirements information in a timely manner, affecting the effect of air pollution control.

Method used

The residual life calculation is performed based on the aging characteristic curve of the ion tube, and the model is trained and the absolute and relative remaining life information is provided in combination with actual application requirements.

Benefits of technology

It improves the accuracy and reliability of the remaining life of the ion tube, reduces operation and maintenance costs, reduces life anxiety, and restores discharge performance through effective maintenance and extends the service life of the equipment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119047329A_ABST
    Figure CN119047329A_ABST
Patent Text Reader

Abstract

The invention discloses a method for calculating the residual life of an ion tube, which comprises the following steps of: acquiring running state data of the ion tube to obtain an aging characteristic curve of the ion tube; training a long-short-term memory neural network by adopting an ion tube aging characteristic curve; and calculating the residual life of the ion tube by adopting the trained long-short-term memory neural network. According to the invention, the long-short-term memory neural network is adopted to calculate the residual life of the ion tube, and the designed method has the advantages of intelligence, accuracy and capability of meeting actual application requirements.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method for calculating the remaining life of an ion tube. Background Art

[0002] The ion tube is a dielectric barrier discharge device used in the field of air pollution control. In the actual application of the ion tube, there is a common problem of anxiety about the life of the ion tube. Users cannot obtain the remaining life of the ion tube in a timely and accurate manner, and can only obtain information about the failure of the ion tube indirectly, such as the weakening of the air pollution control effect and the increase in user complaints. In order to maintain the operation of the equipment, owners usually adopt strategies such as increasing the frequency of equipment maintenance and frequently replacing ion tubes, which will lead to a significant increase in the cost of equipment operation and maintenance, and will also increase the misjudgment rate of the remaining life of the ion tube, and will further increase the owner's anxiety about the life of the ion tube. Therefore, accurately calculating the remaining life of the ion tube can increase the reliability of the ion tube operation, reduce the operation and maintenance costs, and reduce anxiety about the life of the ion tube.

[0003] In the actual application of ion tubes, timely and effective maintenance can partially restore the discharge performance of ion tubes. Maintenance personnel can significantly improve the performance of ion tubes in treating atmospheric pollutants and improve user experience by cleaning the surface of glass tube media and removing dust attached to stainless steel nets. Therefore, timely and accurate information on maintenance needs for ion tubes is an effective measure to reduce anxiety about the life of ion tubes. However, there is also the problem of nitrogen oxide deposition in continuously running ion tubes. At the end of the life of ion tubes, white nitrogen oxides are evenly distributed on the inner wall of the glass tube. This substance cannot be removed through maintenance, but it will affect the discharge of the ion tube and will eventually cause the ion tube to fail. Therefore, timely and accurate information on the replacement needs of ion tubes is also an important measure to reduce anxiety about the life of ion tubes.

[0004] The existing solution (patent 201910704923.8) obtains the remaining life information of the ion tube by recording the working voltage and working time of the ion tube. This solution does not consider the state characteristic changes, aging process and maintenance of the ion tube itself. The positive impact on the life of the ion tube is only converted into the voltage-time integral considering the working voltage. Its essence is to calculate the remaining life of the ion tube based on the cumulative working time. Based on this, in order to improve the accuracy and reliability of the calculation of the remaining life of the ion tube, this solution uses a long short-term memory neural network (Long Short-Term Memory, LSTM) to calculate the remaining life of the ion tube. The designed method has the advantages of being intelligent, accurate and meeting the needs of practical applications. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an intelligent, accurate and practical application-satisfactory method for calculating the remaining life of an ion tube.

[0006] The technical solution of the present invention to solve the above problem is: a method for calculating the remaining life of an ion tube, comprising the following steps: S1. Obtaining the ion tube operating status data and obtaining the ion tube aging characteristic curve; S2. Using ion tube aging characteristic curve to train long short-term memory neural network; S3. Use the trained long short-term memory neural network to calculate the remaining life of the ion tube.

[0007] The effect of the present invention is that, in view of the problem of calculating the remaining life of an ion tube, compared with the traditional method, the present invention provides an intelligent, accurate and practical application-satisfactory method for calculating the remaining life of an ion tube. The advantage of this calculation method is that, based on the aging characteristic curve of the ion tube, a long short-term memory neural network is used to calculate the remaining life of the ion tube, and the relative remaining life and absolute remaining life of the ion tube are given according to the actual application requirements, thereby improving the reliability and practicality of the calculation of the remaining life of the ion tube and reducing the operation and maintenance cost of the ion tube. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 Flow chart of the calculation of the remaining life of the ion tube in the present invention.

[0009] Figure 2 Figure 1 is a diagram of a data acquisition device for acquiring ion tube operating status data in the present invention.

[0010] Figure 3 Schematic diagram of the calculation of the remaining life of the ion tube in the present invention. DETAILED DESCRIPTION

[0011] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0012] The present invention provides a method for calculating the remaining life of an ion tube, the flow chart of which is as follows: Figure 1 As shown, it includes the following steps: S1. Obtain the ion tube operating status data and obtain the ion tube aging characteristic curve.

[0013] The ion tube is a dielectric barrier discharge device used to generate negative ions. The ion tube consists of an inner electrode, an outer electrode, and a sealed glass tube medium. Power frequency high voltage is applied between the inner and outer electrodes to form a stable discharge on the surface of the glass tube. As an example, the parameters of the ion tube analyzed in this patent are: tube diameter 38mm, tube length 250mm, closed quartz glass tube thickness 1mm, outer electrode is stainless steel mesh, and inner electrode is perforated aluminum plate. Figure 2 The ion tube operation state data acquisition device shown in the figure obtains the state data of the ion tube during continuous operation. The data acquisition device includes an industrial frequency transformer, an STM32 single-chip microcomputer and its power supply, a data storage card, a voltage and current sensor, an operation state acquisition circuit, a signal monitoring port, a connection port of the tested ion tube, etc. Among them, the industrial frequency transformer is used to generate a high voltage for the normal operation of the external tested ion tube. The STM32 single-chip microcomputer obtains the voltage and current signals of the low-voltage side of the industrial frequency transformer through a voltage sensor (PT) and a current sensor (CT), and calculates the corresponding total voltage and measurement voltage signals through a built-in BP neural network model, and finally stores the total voltage and measurement voltage signals in the data storage card. The operation state acquisition circuit uses a capacitor C1 and C2 series voltage division method to obtain the total voltage during the continuous operation of the ion tube, and uses a capacitor Cm and the ion tube in series to obtain the measurement voltage during the continuous operation of the ion tube. Therefore, the total voltage and measurement voltage of the ion tube obtained represent the working voltage and discharge charge of the ion tube respectively, and are output to an external oscilloscope through a signal monitoring port. Further based on the total voltage and measurement voltage of the ion tube, the Lissajous figure of the ion tube can be displayed on the oscilloscope. Through the signal monitoring port, the voltage and current signals on the low-voltage side of the power frequency transformer corresponding to the total voltage and the measured voltage can also be obtained on an external oscilloscope. These signals can be used to analyze the fitting characteristics of the BP neural network model built into the STM32 microcontroller, and the BP neural network model can be trained and updated according to the quality of the fitting characteristics. As an embodiment, the data acquisition parameters of the ion tube operation status data acquisition device constructed by this patent are: sampling frequency is 25kHZ, each sampling time is 60ms, and sampling is performed every 30 minutes. If the theoretical life of the tested ion tube is 20,000 hours, the length of the finally generated operation status data is 6×10 7 The calculation process is as follows: 20000h×2 times / h×3 cycles / time×500 points / cycle=6×10 7 Points The time point of ion tube failure is determined by obtaining and analyzing the Lissajous figure of the ion tube through an external oscilloscope or directly analyzing the measured voltage signal, and the total life time of the ion tube is obtained. The measured voltage signal curve of the ion tube during continuous operation obtained by the data acquisition device is calibrated to obtain the ion tube aging characteristic curve.

[0014] S2. Use the aging characteristic curve of ion tube to train long short-term memory neural network.

[0015] The structure of the constructed long short-term memory neural network is: input layer, LSTM layer, Relu activation layer, fully connected layer, regression layer, and the network hyperparameters are: the number of neurons in the input layer is 1000, the number of neurons in the LSTM layer is 800, the Relu activation layer uses the Relu activation function, and the number of output neurons in the fully connected layer is 50. The obtained ion tube aging characteristic curve is split and normalized according to the constructed long short-term memory neural network structure, and the long short-term memory neural network is trained and tested to obtain a trained long short-term memory neural network.

[0016] S3. Use the trained long short-term memory neural network to calculate the remaining life of the ion tube.

[0017] The remaining life calculation process of the ion tube is as follows: during the operation of the ion tube, at a sampling frequency of 25 kHz, the collected ion tube measurement voltage signals of two consecutive power frequency cycles are input into the long short-term memory neural network trained by step S2, and the measurement voltage signal prediction curve of the ion tube is calculated. The failure time of the ion tube is further obtained from the calculated measurement voltage signal prediction curve or the corresponding Lissajous diagram, which is the remaining life of the ion tube in this period, in hours. Figure 3 The figure shows the schematic diagram of the calculation of the remaining life of the ion tube, including two processes: offline training and online calculation of the long short-term memory neural network (LSTM). In the offline training process of LSTM, the ion tube operation state data acquisition device is used to obtain the ion tube operation state data, and the ion tube state recognition model based on the convolutional neural network (CNN) is used to obtain the time point of ion tube failure, so as to calibrate the ion tube operation state data and obtain the ion tube aging characteristic curve, and finally use the aging characteristic curve to train the long short-term memory neural network model. In the online calculation process of LSTM, for the ion tube in actual operation, the operation state data of the ion tube at a certain time point is first obtained, and the ion tube operation state prediction curve is obtained by calculation of the long short-term memory neural network, and the remaining life of the ion tube at that time is obtained by combining the ion tube state recognition model.

[0018] For the ion tubes in actual operation, considering the positive impact of effective maintenance on the remaining life of the ion tubes, the remaining life of the ion tubes defined in this patent includes absolute remaining life and relative remaining life. Absolute remaining life refers to the remaining life of the ion tube calculated by the LSTM neural network after the completion of the most recent maintenance during the operation of the ion tube minus the cumulative working time of the ion tube after maintenance, in hours; relative remaining life refers to the remaining life calculated by the LSTM neural network during the operation of the ion tube, in hours. The relative remaining life is only valid for a single maintenance cycle, while the absolute remaining life is the length of time the ion tube is away from complete failure. The relative remaining life is used to provide maintenance demand information for the ion tube, and the absolute remaining life is used to provide replacement demand information for the ion tube. As an embodiment, if the theoretical life of a new ion tube is 20,000 hours, the absolute remaining life of the ion tube when it is first put into use is equal to 20,000 hours minus the operating time, and the corresponding relative remaining life is calculated by the LSTM neural network. Assuming that after a maintenance of an ion tube is completed, the LSTM neural network calculates that the remaining life of the ion tube is n hours, then the absolute remaining life of the ion tube is n hours, and the relative remaining life is also n hours; if the ion tube works for m hours after this maintenance, the absolute remaining life is nm hours, and the relative remaining life is calculated by the LSTM neural network. Therefore, the relative remaining life is relative to a maintenance cycle, while the absolute remaining life is the length of time until the ion tube completely fails.

Claims

1. A method for calculating the remaining life of an ion tube, characterized in that: The following steps are involved: S1. Obtaining the ion tube operating status data and obtaining the ion tube aging characteristic curve; The ion tube refers to a dielectric barrier discharge device for generating negative ions, which is composed of an inner electrode, an outer electrode and a closed glass tube medium. Power frequency high voltage electricity is applied between the inner and outer electrodes to form a stable discharge on the surface of the glass tube; the acquisition of the ion tube operation status data refers to the use of an ion tube operation status data acquisition device to acquire the status data of the ion tube during continuous operation, and the data acquisition device outputs a fixed power frequency high voltage electricity to achieve normal discharge of the ion tube and acquire the total voltage and measurement voltage signal during the continuous operation of the ion tube; the total voltage and measurement voltage of the ion tube refer to the power frequency high voltage applied to the ion tube and the voltage on the measurement capacitor connected in series with the ion tube, which represent the working voltage and discharge charge of the ion tube respectively. Based on the total voltage and measurement voltage of the ion tube, the Lissajous figure of the ion tube can be obtained; the time point of failure of the ion tube is determined by analyzing the Lissajous figure of the ion tube or directly analyzing the measurement voltage signal, the total life time of the ion tube is obtained, and the measurement voltage signal curve of the ion tube during continuous operation acquired by the data acquisition device is calibrated, thereby obtaining the aging characteristic curve of the ion tube; S2. Using ion tube aging characteristic curve to train long short-term memory neural network; The structure of the long short-term memory neural network (LSTM) is: input layer, LSTM layer, Relu activation layer, fully connected layer, regression layer, and the network hyperparameters are: the number of neurons in the input layer is 1000, the number of neurons in the LSTM layer is 800, the Relu activation layer adopts the Relu activation function, and the number of output neurons in the fully connected layer is 50; the obtained ion tube aging characteristic curve is subjected to data splitting and normalization processing according to the constructed long short-term memory neural network structure, and the long short-term memory neural network is trained and tested to obtain a trained long short-term memory neural network; S3. Use the trained long short-term memory neural network to calculate the remaining life of the ion tube.

2. The method for calculating the remaining life of an ion tube according to claim 1, characterized in that: In the step S3, the remaining life of the ion tube is calculated using the trained long short-term memory neural network, and the process is as follows: during the operation of the ion tube, the collected ion tube measurement voltage signals of two consecutive power frequency cycles are input into the long short-term memory neural network trained by step S2, and the measurement voltage signal prediction curve of the ion tube is calculated, and the failure time of the ion tube is further obtained from the calculated measurement voltage signal prediction curve or the corresponding Lissajous diagram, that is, the remaining life of the ion tube in this period, in hours; the remaining life of the ion tube includes an absolute remaining life and a relative remaining life, and the absolute remaining life refers to the remaining life of the ion tube calculated after the most recent maintenance is completed during the operation of the ion tube minus the accumulated working time of the ion tube after maintenance, in hours; the relative remaining life refers to the remaining life calculated by the constructed long short-term memory neural network during the operation of the ion tube, in hours; The relative remaining life is only valid for a single maintenance cycle. The absolute remaining life is the length of time until the ion tube completely fails. The relative remaining life is used to provide maintenance demand information for the ion tube, and the absolute remaining life is used to provide replacement demand information for the ion tube.

Citation Information

Cited By

  • Multi-modal neural network calculation method for predicting residual life of ion tube

    CN121388490A

  • A multi-modal neural network computing method for ion tube residual life prediction

    CN121388490B