CN103543210A - Pressurized pneumatic transmission flow type detection device and method based on acoustic emission technology - Google Patents

Abstract

The invention discloses a detection device and method for pressurized pneumatic conveying flow patterns based on acoustic emission technology. The easy-to-measure and reliable acoustic emission signal is used as an auxiliary variable to detect the flow pattern of the dominant variable gas-solid two-phase flow. The acoustic emission receiving probe is installed on the outer wall of the horizontal pipe of the pressurized pneumatic conveying system, which can fully and accurately obtain the collision and friction sound signals between the particles in the pipe and the pipe wall, and between particles and particles. The signal is processed by a two-stage signal amplifier. Amplified, and then sent to the computer through the data collector. The computer conducts Hilbert-Huang transformation analysis on the collected acoustic signals, and extracts the corresponding eigenvalues on this basis, and establishes the relationship between the flow pattern and the eigenvalues of the acoustic signal through the generalized regression neural network (GRNN). The trained neural network can be used to infer the flow pattern in the pipe according to the characteristics of the acoustic signal. The invention has the advantages of not invading the flow field, easy installation and disassembly of the device, sensitive detection, environmental protection and safety, real-time online and the like.

Description

Pressurized pneumatic conveying flow pattern detection device and method based on acoustic emission technology

technical field

The invention relates to the technical field of multiphase flow measurement, in particular to a pressurized pneumatic conveying flow pattern detection device and method based on acoustic emission technology.

Background technique

The flow pattern is an important engineering detection object in the gas-solid two-phase flow system, which affects the flow characteristics, heat and mass transfer characteristics and operation reliability of the system. At the same time, the accurate measurement of the two-phase flow parameters also depends on the understanding of the flow pattern. Under pressurized conditions, the gas-solid two-phase flow behavior is more complex and the stability decreases. It is particularly important to grasp the flow pattern and realize the accurate identification of the convection pattern to control the safe operation of the pressurized pneumatic conveying system.

The pressure (differential pressure) signal of the pressurized pneumatic conveying process carries a large amount of dynamic information, which is a comprehensive reflection of material characteristics, conveying shape, stability, geometric characteristics of conveying pipelines and energy exchange and other flow characteristics. Through pressure (differential pressure) signal processing (such as spectrum analysis, wavelet transform, Hilbert-Huang transform, etc.), the flow pattern and its changes in the pipe can be known. However, there are many shortcomings and limitations in the application of traditional pressure (differential pressure) sensors in pressurized pneumatic conveying. For example, the measuring point needs to invade the flow field, which will inevitably affect the flow in the pipe; the position is fixed and it is inconvenient to move; installation and maintenance requirements High, in order to ensure the accuracy and reliability of the measurement, after a period of use, the pressure measuring points and probes need to be cleaned regularly, and the pressurized pneumatic conveying system often has to be depressurized. The process is cumbersome and dangerous. Pressure also means waste of air source; especially for differential pressure sensors, any air leakage at any measuring point can easily cause overrange and damage the sensor.

In recent years, more advanced detection methods such as nuclear magnetic resonance, CT imaging, gamma ray, and optical fiber technology have been applied in multiphase flow, which has made flow pattern measurement develop to a certain extent. However, there are some shortcomings in the above-mentioned detection methods, such as nuclear magnetic resonance technology equipment is expensive and not easy to popularize; CT imaging data processing volume is large, and the process is complicated and cumbersome; microwave method is easily disturbed by the environment, and the accuracy is not high; There is no safety guarantee for personnel; fiber optic measurement technology has certain effectiveness and reliability, but it is an intrusive measurement that interferes with the flow field to a certain extent.

When a material or structure is deformed by force, it releases strain energy in the form of an elastic wave. Acoustic Emission (AE) technology uses a piezoelectric probe coupled to the surface of the material or component to receive the generated elastic wave. Convert elastic waves into electrical signals. Use subsequent circuits to process and display the detected electrical signals, and then obtain the internal conditions of materials or components. Therefore, acoustic emission signals are easier to obtain and can better reflect the complexity of gas-solid two-phase flow from a microscopic point of view.

The acoustic emission signal of the pressurized pneumatic conveying process reflects complex nonlinear and non-equilibrium dynamic characteristics, which can be deeply understood through the multi-scale analysis method of the acoustic signal. FFT spectrum analysis is the most basic analysis method in the field of signal processing. It can describe the frequency characteristics of the signal well, but it cannot provide time domain information and is not suitable for analyzing nonlinear and non-stationary signals. The wavelet transform is essentially a Fourier transform with an adjustable window. The signal in the wavelet window must be stable, without getting rid of the limitations of FFT spectrum analysis, and the basis function must be pre-selected by experience, which cannot be changed during the entire signal analysis process. . Power spectrum analysis, like wavelet analysis, also has inherent defects of Fourier transform. Chaos analysis is an effective means to study the nonlinear characteristics of gas-solid two-phase flow, but the extraction of characteristic parameters requires a long time series, and the calculation results vary greatly with the initial parameters. In recent years, Huang et al. proposed a new time-frequency multi-scale signal analysis method—Hilbert-Huang transform (HHT), which is a major breakthrough in the analysis of linear and stationary signals based on Fourier transform, especially suitable for nonlinear Stationary signal analysis.

There are two types of detection methods based on acoustic waves, active and passive. The former usually uses ultrasonic waves for active detection, while the latter analyzes the acoustic emission signals generated during the receiving process. The acquired signals often better reflect the essence of the process, and the measurement method is easier to implement.

Contents of the invention

The object of the present invention is to provide a pressurized pneumatic conveying flow pattern detection device and method based on acoustic emission technology. With this device and method, the easy-to-measure and reliable acoustic emission signal can be used as an auxiliary variable to perform the gas-solid two-phase detection of the dominant variable. The detection of the flow pattern avoids adverse effects on the flow in the pipe due to the intrusion of the measuring point into the flow field, and reduces the measurement error.

The pressurized pneumatic conveying flow pattern detection device based on acoustic emission technology involved in the present invention includes a pressurized pneumatic conveying pipeline test section, an acoustic emission receiving probe, a preamplifier, a main amplifier, a data collector and a computer; wherein, there are several The two acoustic emission receiving probes are respectively installed closely on the outer wall of the pressurized pneumatic conveying pipeline test section, and the acoustic emission receiving probes are sequentially connected with the preamplifier, the preamplifier, the main amplifier, the data collector and the computer.

As a further improvement of the above technical solution, the number of the acoustic emission receiving probes is four.

As a further improvement of the above technical solution, the installation positions of the acoustic emission receiving probes are evenly distributed along the same circumference of the outer pipe wall of the test section of the pressurized pneumatic conveying pipeline.

The method for pressurized pneumatic conveying flow pattern detection involved in the present invention comprises the following steps:

1) Data acquisition: Four acoustic emission receiving probes installed on the outer wall of the pressurized pneumatic conveying pipeline to receive the acoustic emission signals generated by the collision and friction between particles and the pipe wall, and between particles in the pressurized pneumatic conveying pipeline, The acoustic signal is amplified by amplifiers at various levels, and then sent to the computer through the data collector;

2) Perform Hilbert-Huang transformation analysis on the acoustic emission signal, and extract the corresponding eigenvalues, specifically:

The Hilbert-Huang transformation analysis process includes empirical mode decomposition EMD and Hilbert transformation.

First, decompose the collected acoustic emission signal to obtain a series of intrinsic mode functions Intrinsic mode function, IMF, and meet the following conditions: In the entire data segment, the number of extreme points and zero-crossing points must be equal or differ by at most one ; At any data point, the mean value of the envelope of the local maximum value and the envelope of the local minimum value is zero; by constantly eliminating the mean value of the upper and lower envelopes connected by the minimum value and maximum value of the signal, the original acoustic emission signal x(t) can be decomposed into:

$\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

In the formula, I _i (t) is the IMF component obtained by decomposition; r _n (t) is a constant or a monotone function;

Next, do a Hilbert transform for each IMF

${\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&pi;</span>}}{<span\; class="notranslate">\; \&Integral;</span>}_{<span\; class="notranslate">\; -</span><span\; class="notranslate">\; \&infin;</span>}^{<span\; class="notranslate">\; \&infin;</span>}\frac{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}}\mathrm{<span\; class="notranslate">\; d\&tau;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In the formula, L _i (t) is the Hilbert transformation of I _i (t).

Analytical signal Z _i (t)

${\mathrm{<span\; class="notranslate">\; Z</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{{\mathrm{<span\; class="notranslate">\; j\&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

in,

Amplitude $a_i\left(t\right)=\sqrt{I_i^2\left(t\right)+L_i^2\left(t\right)}---\left(4\right)$

Phase angle ${\mathrm{\&theta;}}_i\left(t\right)=\arctan \left(\frac{L_i\left(t\right)}{I_i\left(t\right)}\right)---\left(5\right)$

The instantaneous frequency of each IMF component is

${\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}}\frac{\mathrm{<span\; class="notranslate">\; d</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; dt</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

According to the energy distribution of IMF components in different frequency bands, the flow pattern is analyzed and judged, and the percentage of the energy of each frequency band in the total signal energy is used as the characteristic value to represent the information of the flow form. The energy characteristic value e is introduced and defined as:

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}}{\mathrm{<span\; class="notranslate">\; E.</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}}{\mathrm{<span\; class="notranslate">\; E.</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}}{\mathrm{<span\; class="notranslate">\; E.</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

In the formula, E _h , E _m , and E _l represent the energy of high, medium and low frequency bands respectively, and E represents the total energy. Correspondingly, e _h , _em and e _l represent the energy percentages of high, medium and low frequency bands respectively; energy calculation The formula is as follows:

${\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{<span\; class="notranslate">\; -</span><span\; class="notranslate">\; \&infin;</span>}^{<span\; class="notranslate">\; +</span><span\; class="notranslate">\; \&infin;</span>}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; dt</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

Utilizing the advantages of the Hilbert-Huang transform (HHT) which is suitable for nonlinear and non-stationary signal analysis, using HHT to analyze and process the acoustic emission signal in the pressurized pneumatic conveying process can get rid of the Fourier transform-based method which is only suitable for linear and stationary signals. Limitations of signal analysis.

3) Using the generalized regression neural network (GRNN), the energy characteristic value e obtained in step 2) is used as input, and the output corresponds to the flow pattern, that is, the suspension flow is (0,0,0,1), and the layered flow is (0 ,0,1,0), the dune flow is (0,1,0,0), the plunger flow is (1,0,0,0), the neural network is established and trained, the trained network can be completed Mapping from the energy characteristic value e space to the flow pattern space, so as to establish the correlation between the flow pattern in the pipe and the characteristics of the acoustic emission signal; when the flow pattern is unknown, the trained neural network can be used to reverse the flow pattern in the pipe and detect it online in real time Changes in the flow pattern in the pipe.

The generalized regression neural network (GRNN) used has the ability of local approximation, and the learning speed is faster; and there are few artificially adjusted parameters, and the learning of the network all depends on data samples, so that the network avoids human subjective assumptions to the greatest extent. The impact of the predicted results, the stability is better.

As a further improvement of the above technical solution, the sampling frequency of the data acquisition in the step 1) is set above 200kHz, and the sampling time lasts above 30s.

The device of the present invention is mainly aimed at the detection of pressurized pneumatic conveying flow pattern. The detection method of the invention is non-invasive, and can monitor the flow pattern in the pipe in real time on-line. It avoids the impact on the flow in the pipe and reduces the measurement error; at the same time, it avoids the inconvenient movement of the traditional pressure (differential pressure) sensor in the pressurized pneumatic conveying measurement, the need for regular loading and unloading to clean the probe, and the high pressure environment is easy to damage the sensor, etc. Disadvantages: Acoustic emission signals can be collected at different positions on the outer pipe wall of the conveying pipeline conveniently, and flow information in the pipe can be obtained more accurately and comprehensively. Compared with nuclear magnetic resonance, CT imaging, γ-ray and other methods, it has the advantages of better economy, environmental protection and safety.

Description of drawings

Fig. 1 is a schematic structural diagram of a pressurized pneumatic conveying flow pattern detection device based on acoustic emission technology.

Fig. 2 is a block diagram of a flow pattern detection method using a pressurized pneumatic conveying flow pattern detection device based on acoustic emission technology.

Fig. 3 is a flow chart of the software program of the present invention.

Detailed ways

Embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

As shown in Figure 1, a pressurized pneumatic conveying flow pattern detection device based on acoustic emission technology includes a pressurized pneumatic conveying pipeline test section, an acoustic emission receiving probe, a preamplifier, a main amplifier, a data collector and a computer. Among them, the four acoustic emission receiving probes are evenly installed on the same circumference on the outer pipe wall of the test section of the pressurized pneumatic conveying pipeline. By collecting the acoustic emission signals at different positions in the cross-section of the pipeline, it is possible to obtain a more accurate and comprehensive Pipe flow information. The acoustic emission receiving probe is connected with the preamplifier, the preamplifier is connected with the main amplifier, the main amplifier is connected with the data collector, and the data collector is connected with the computer.

As shown in Figure 2, the flow pattern detection method of the present invention adopts the idea of indirect measurement, and through the Hilbert-Huang transformation analysis and processing of the acoustic emission signal, the easy-to-measure process variable (auxiliary variable) and the difficult-to-measure process to be measured are established. The mathematical relationship between variables (leading variables) realizes the measurement of the process variable (leading variable) to be measured. Therefore, the acoustic emission signal is used as an easily measurable auxiliary variable in the present invention, and the leading variable is an important parameter in the pressurized gas-solid two-phase flow—flow pattern.

First of all, data collection is carried out. Four acoustic emission receiving probes installed on the outer wall of the pressurized pneumatic conveying pipeline receive the acoustic emission signals generated by the collision and friction between the particles in the pressurized pneumatic conveying pipe and the pipe wall, and between particles. The acoustic signal is amplified by amplifiers at various levels, and then sent to the computer through the data collector. The sampling frequency is set above 200kHz, and the sampling time lasts for more than 30s.

After the signal acquisition is completed, the percentage of the energy of each scale detail of the acoustic emission signal to the total signal energy is selected as the signal characteristic value, that is, the percentage of the energy of each scale detail to the total signal energy after the signal is processed by the Hilbert-Huang transform, as follows:

The Hilbert-Huang transformation analysis process includes empirical mode decomposition (EMD) and Hilbert transformation.

First, decompose the collected acoustic emission signal to obtain a series of intrinsic mode functions Intrinsic mode function, IMF, and meet the following conditions: In the entire data segment, the number of extreme points and zero-crossing points must be equal or differ by at most one ; At any data point, the mean value of the envelope of the local maximum value and the envelope of the local minimum value is zero; by constantly eliminating the mean value of the upper and lower envelopes connected by the minimum value and maximum value of the signal, the original acoustic emission signal x(t) can be decomposed into:

$\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

In the formula, I _i (t) is the IMF component obtained by decomposition; r _n (t) is a constant or a monotone function;

Next, do a Hilbert transform for each IMF

${\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&pi;</span>}}{<span\; class="notranslate">\; \&Integral;</span>}_{<span\; class="notranslate">\; -</span><span\; class="notranslate">\; \&infin;</span>}^{<span\; class="notranslate">\; \&infin;</span>}\frac{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}}\mathrm{<span\; class="notranslate">\; d\&tau;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In the formula, L _i (t) is the Hilbert transformation of I _i (t).

Analytical signal Z _i (t)

${\mathrm{<span\; class="notranslate">\; Z</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{{\mathrm{<span\; class="notranslate">\; j\&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

in,

Amplitude $a_i\left(t\right)=\sqrt{I_i^2\left(t\right)+L_i^2\left(t\right)}---\left(4\right)$

Phase angle ${\mathrm{\&theta;}}_i\left(t\right)=\arctan \left(\frac{L_i\left(t\right)}{I_i\left(t\right)}\right)---\left(5\right)$

The instantaneous frequency of each IMF component is

${\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}}\frac{\mathrm{<span\; class="notranslate">\; d</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; dt</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

According to the energy distribution of IMF components in different frequency bands, the flow pattern is analyzed and judged, and the percentage of the energy of each frequency band in the total signal energy is used as the characteristic value to represent the information of the flow form. The energy characteristic value e is introduced and defined as:

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}}{\mathrm{<span\; class="notranslate">\; E.</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}}{\mathrm{<span\; class="notranslate">\; E.</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}}{\mathrm{<span\; class="notranslate">\; E.</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

In the formula, E _h , E _m , and E _l represent the energy of high, medium and low frequency bands respectively, and E represents the total energy. Correspondingly, e _h , _em and e _l represent the energy percentages of high, medium and low frequency bands respectively; energy calculation The formula is as follows: ${\mathrm{E.}}_i={\&Integral;}_{-\&infin;}^{+\&infin;}{\left|I_i\left(t\right)\right|}^2\mathrm{dt}---\left(8\right)$

3) On the basis of a large number of experimental sample data, the generalized regression neural network (GRNN) is used, and the energy characteristic value e obtained in step 2) is used as input, and the output corresponds to the flow pattern, that is, the suspended flow is (0,0,0, 1), the layered flow is (0,0,1,0), the dune flow is (0,1,0,0), the plunger flow is (1,0,0,0), the neural network is established and its After training, the trained network can complete the mapping from the energy feature value e space to the flow pattern space, so as to establish the correlation between the flow pattern in the pipe and the characteristics of the acoustic emission signal; when the flow pattern is unknown, use the trained neural network to reflect Push the flow pattern in the tube, and detect the change of the flow pattern in the tube in real time.

As shown in FIG. 3 , a software program block diagram of a pressurized pneumatic conveying flow pattern detection device and method based on acoustic emission technology is described. The program is compiled based on automatic detection technology and computer data processing technology. When running, it first performs hardware initialization to check whether the hardware driver is normal and whether the hardware is connected normally. If it succeeds, perform the following operations. If the hardware initialization is successful, set the parameters of the software, including sampling frequency, number of sampling points (or sampling time), selected channel number, etc. After setting these parameters as required, data collection, processing and analysis can be carried out, and finally the identification of the flow pattern in the pressurized pneumatic conveying pipe can be completed.

There are many specific application approaches of the present invention, and the above description is only a preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, some improvements can also be made without departing from the principles of the present invention. Improvements should also be regarded as the protection scope of the present invention.

Claims (5)

1. A pressurized pneumatic transmission flow pattern detection device based on an acoustic emission technology is characterized in that: the device comprises a pressurized pneumatic conveying pipeline testing section (1), an acoustic emission receiving probe (2), a preamplifier (3), a main amplifier (4), a data acquisition unit (5) and a computer (6); wherein, a plurality of acoustic emission receiving probes (2) are respectively and tightly attached to the outer pipe wall of the pressurized pneumatic conveying pipeline testing section (1), and the acoustic emission receiving probes (2) are sequentially connected with a preamplifier (3), a main amplifier (4), a data collector (5) and a computer (6).

2. The pressurized pneumatic conveying flow pattern detection device based on acoustic emission technology of claim 1, characterized in that: the number of the acoustic emission receiving probes (2) is four.

3. The pressurized pneumatic conveying flow pattern detection device based on acoustic emission technology of claim 1, characterized in that: the mounting positions of the acoustic emission receiving probes (2) are uniformly distributed along the same circumference of the outer pipe wall of the pressurized pneumatic conveying pipeline testing section (1).

4. A method for detecting a pressurized air transport flow pattern using the pressurized air transport flow pattern detection apparatus according to any one of claims 1, 2, or 3, comprising the steps of:

1) data acquisition: four acoustic emission receiving probes (2) arranged on the outer pipe wall of the pressurized pneumatic conveying pipeline testing section (1) receive acoustic emission signals generated by collision and friction between particles and the pipe wall in the pressurized pneumatic conveying pipeline, and the acoustic signals are amplified by amplifiers at all stages and then are sent to a computer through a data acquisition unit;

2) performing Hilbert-Huang transform analysis on the acoustic emission signals, and extracting corresponding characteristic values, specifically:

the Hilbert-Huang transformation analysis process comprises Empirical Mode Decomposition (EMD) and Hilbert transformation;

firstly, decomposing the acquired acoustic emission signals to obtain a series of Intrinsic mode functions, IMFs, and satisfying the following conditions: the number of extreme points and zero-crossing points must be equal or at most one different over the entire data segment; at any data point, the envelope of the local maximum and the average of the envelopes of the local minima are zero; by continuously rejecting the mean of the upper and lower envelopes of the signal, which are connected by the minimum and maximum values, the original acoustic emission signal x (t) can be decomposed into:

<math> <mrow> <mi>x</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>I</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>r</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </math>

in the formula I_i(t) is the IMF component obtained by decomposition; r is_n(t) is a constant or monotonic function;

next, a Hilbert transform is performed on each IMF

<math> <mrow> <msub> <mi>L</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mi>&pi;</mi> </mfrac> <msubsup> <mo>&Integral;</mo> <mrow> <mo>-</mo> <mo>&infin;</mo> </mrow> <mo>&infin;</mo> </msubsup> <mfrac> <mrow> <msub> <mi>I</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&tau;</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mi>t</mi> <mo>-</mo> <mi>&tau;</mi> </mrow> </mfrac> <mi>d&tau;</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> </math>

In the formula, L_i(t) is I_i(t) Hilbert transform.

Analysis of the signal Z_i(t)

<math> <mrow> <msub> <mi>Z</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mi>I</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <mi>j</mi> <msub> <mi>L</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mi>a</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <msup> <mi>e</mi> <mrow> <msub> <mi>j&theta;</mi> <mi>j</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow> </math>

Wherein,

amplitude value $a_i\left(t\right)=\sqrt{I_i^2\left(t\right)+L_i^2\left(t\right)}---\left(4\right)$

Phase angle <math> <mrow> <msub> <mi>&theta;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>arctan</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <msub> <mi>L</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <msub> <mi>I</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow> </math>

The instantaneous frequency of each IMF component is

<math> <mrow> <msub> <mi>&omega;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mrow> <mn>2</mn> <mi>&pi;</mi> </mrow> </mfrac> <mfrac> <mrow> <mi>d</mi> <msub> <mi>&theta;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mi>dt</mi> </mfrac> <mo></mo> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow> </math>

Analyzing and judging the flow pattern according to the energy distribution of IMF components of different frequency bands, representing the information of the flow form by taking the percentage of the energy of each frequency band in the total energy of the signal as a characteristic value, introducing an energy characteristic value e and defining as follows:

$e_h=\frac{E_h}{E},e_m=\frac{E_m}{E},e_l=\frac{E_l}{E}---\left(7\right)$

in the formula E_h、E_m、E_lRepresenting the energy in the high, medium and low frequency bands, respectively, E representing the total energy, and, correspondingly, E_h、e_mAnd e_lRespectively representing the energy percentages of high, medium and low frequency bands; the energy calculation formula is as follows:

<math> <mrow> <msub> <mi>E</mi> <mi>i</mi> </msub> <mo>=</mo> <msubsup> <mo>&Integral;</mo> <mrow> <mo>-</mo> <mo>&infin;</mo> </mrow> <mrow> <mo>+</mo> <mo>&infin;</mo> </mrow> </msubsup> <msup> <mrow> <mo>|</mo> <msub> <mi>I</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mi></mi> <mo>|</mo> </mrow> <mn>2</mn> </msup> <mi>dt</mi> <mo></mo> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>8</mn> <mo>)</mo> </mrow> </mrow> </math>

3) adopting a Generalized Regression Neural Network (GRNN), taking the energy characteristic value e obtained in the step 2) as input, outputting the energy characteristic value e corresponding to the flow pattern, namely, the suspension flow is (0, 0,0, 1), the stratification flow is (0, 0,1, 0), the sand dune flow is (0, 1,0, 0), the plunger flow is (1, 0,0, 0), establishing a neural network and training the neural network, wherein the trained network can finish mapping from an energy characteristic value e space to a flow pattern space, so that the correlation between the flow pattern in the pipe and the characteristics of the acoustic emission signal is established; under the condition that the flow pattern is unknown, the trained neural network is used for reversely pushing the flow pattern in the pipe, and the change of the flow pattern in the pipe is detected on line in real time.

5. The method of detecting pressurized pneumatic transport flow patterns according to claim 4, characterized in that: the sampling frequency of data acquisition in the step 1) is set to be more than 200kHz, and the sampling time lasts for more than 30 s.

Images