Cereal particle crushing degree detection device and method based on acoustic emission sensor

Abstract

The invention discloses a device and a method for detecting the crushing degree of dry grain particles based on an acoustic emission sensor. The device is composed of an acoustic emission sensor probe, a waveguide rod, a signal conditioning module, a data acquisition module, a signal analysis module, a display module and a single-particle friction collision experiment table. Wherein the front end of the waveguide rod is in contact with particles, and the acoustic emission sensor is mounted on the section of the rear end of the waveguide rod and used for detecting acoustic signals transmitted from the waveguide rod; the signal conditioning module amplifies a signal output by the acoustic emission sensor through a pre-amplifier, and then the acoustic emission signal is acquired through a data acquisition module formed by an acoustic emission signal analyzer; the signal analysis module processes the collected signals by adopting a blind source separation algorithm and a deep learning method to obtain the damage degree of the particles and transmits the damage degree to the display module for display; influence factors and types of particle crushing are explored on a single-particle friction collision experiment table by adjusting particle sizes, particle materials, flat plate inclination angles, flat plate friction coefficients, collision speeds, collision positions and the like. The invention provides a novel grain nondestructive testing technology based on an acoustic emission sensor and a deep learning method.

Description

Technical Field

[0001] The invention belongs to the field of non-destructive testing of food, and in particular relates to a device and method for detecting the degree of crushing of grain particles based on an acoustic emission sensor. Technical Background

[0002] Grain processing is a key link in agricultural production. However, mechanical stress during processing may cause grain grains to break, affecting their processing and edible quality. Real-time monitoring of the degree of grain breakage during grain processing is crucial for optimizing the processing technology. At present, the detection methods for the degree of grain breakage mainly include visual detection, mechanical measurement and acoustic detection. Among them, acoustic emission technology has attracted much attention due to its simplicity, directness, high sensitivity and online monitoring capabilities. When an object is deformed or broken under the action of external or internal forces, the energy will be quickly released or redistributed, so the phenomenon of transient elastic waves generated locally in the object is called acoustic emission. The elastic waves generated by acoustic emission are emitted from the local material, propagated through the medium, and finally reach the surface of the material. By analyzing the acoustic emission signal and extracting the information contained in the acoustic emission signal, the location of the acoustic emission source and the changes in internal defects of the material can be determined. Acoustic emission technology monitors the material state by detecting the emission of stress waves inside the material, providing a new way to monitor the degree of grain breakage in real time during grain processing.

[0003] This paper proposes a device and method for detecting the degree of breakage of grain particles based on an acoustic emission sensor. The device uses an acoustic emission sensor to measure the stress wave of the object being measured, and simultaneously uses a deep learning blind source separation algorithm to identify the acoustic emission signals of particle crack propagation and wear, so as to realize the detection of the degree of breakage of grain particles.

[0004] The beneficial effects of the present invention are as follows: 1) The acoustic emission sensor detection device has a simple structure and low cost, and is suitable for various food processing environments. It is a new type of particle crushing degree measurement technology. 2) The acoustic emission method for detecting particle crushing is widely used. This method does not require a transparent observation window and can measure the degree of particle crushing in a variety of complex environments and non-transparent containers. For example, it can be applied to the detection of grain crushing degree at the outlet of the harvester during field harvesting, the detection of grain crushing degree in mechanical transportation such as pneumatic conveying, and the detection of grain crushing degree in a fluidized bed dryer. Summary of the invention

[0005] The object of the present invention is to provide a grain particle breakage detection device and method based on an acoustic emission sensor, characterized in that the particle breakage detection device is composed of an acoustic emission sensor probe, a waveguide rod, a signal conditioning module, a data acquisition module, a signal analysis module, a display module and a single particle friction collision test bench.

[0006] When designing the installation of the acoustic emission sensor probe and the waveguide rod, it is necessary to adjust the installation position of the acoustic emission sensor probe, the material of the waveguide rod, the shape of the waveguide rod and other parameters according to the size and resolution requirements of the object to be measured. The front end of the waveguide rod is the position for contacting particles, and the acoustic emission sensor is installed on the rear end section of the waveguide rod. The acoustic emission sensor is used to detect the acoustic signal transmitted from the waveguide rod. The acoustic emission sensor and the waveguide rod are bonded with a coupling agent to reduce the influence of external vibration.

[0007] The signal conditioning module amplifies the signal output by the acoustic emission sensor through a preamplifier, and then the data acquisition module composed of an acoustic emission signal analyzer acquires the acoustic emission signal.

[0008] The signal analysis module uses a blind source separation algorithm and a deep learning method to process the collected signals to obtain the degree of damage of the particles and transmits it to the display module for display.

[0009] The single-particle friction collision test bench is composed of a stepper motor, a push rod, a sample particle collection device, and a flat plate, and can be used to study the acoustic emission mechanism of single-particle friction collision. On the single-particle friction collision test bench, single particles with different degrees of wear (different crack sizes) are placed on a flat plate with a fixed tilt angle through a stepper motor. After the particles slide from the flat plate due to friction, they hit the front end section of the horizontal waveguide rod, generating a stress wave that propagates outward. An acoustic emission sensor is installed on the rear end section of the waveguide rod to collect the acoustic emission signals generated by the particles near the collision position. The influencing factors and types of particle breakage are explored by adjusting the particle size, particle material, flat plate tilt angle, flat plate friction coefficient, collision speed, collision position, etc.

[0010] The present invention provides a method for detecting the degree of grain crushing based on an acoustic emission sensor, which is characterized by comprising the following steps:

[0011] Step 1: In a single particle friction collision mechanism experiment or actual application, the acoustic emission sensor measures the stress wave signal of the measured particle in real time. The signal conditioning module processes the weak acoustic emission signal and transmits it to the signal analysis module through the data acquisition module for acoustic signal processing.

[0012] Step 2: Use the weak acoustic emission signal data collected in step 1 as training samples and test samples.

[0013] Step 3. Through various blind source separation signal processing methods such as wavelet analysis, empirical mode decomposition (EMD), variational mode decomposition (VMD), etc., the weak non-stationary acoustic emission signals are subjected to noise reduction and feature extraction. The acoustic emission signals generated by grain damage are divided into continuous acoustic signals due to particle collision movement and burst acoustic signals caused by particle cracks. The blind source separation method is a signal processing algorithm that does not require prior knowledge and can separate the original signal from a mixed signal of continuous signals and burst signals. The separation and extraction of burst acoustic signals are realized based on the signal decomposition results as the input of the intelligent recognition algorithm for the degree of particle breakage.

[0014] Step 4: Use deep learning methods such as long short-term memory networks to identify the features extracted from the separated signals and original signals in step 3 to obtain different degrees of particle fragmentation.

[0015] In the method for detecting the degree of grain crushing based on acoustic emission sensors, in step 3, the long short-term memory network is used for a given sequence x=(x 1 ,...,x T ), hidden vector sequence h = (h 1 ,...,h T ), the output vector sequence y = (y 1 ,...,y T ) from t = 1 to T, calculated by the following formula:

[0016] h t =H(W xh x t +W hh h t-1 +b h )

[0017] y t =W hy h t +b h

[0018] Among them, W, b, and H are the weight, bias vector, and hidden activation of the model respectively. The calculation formulas of these parameters are as follows:

[0019] i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i )

[0020] f t =σ(W xf x t +W hf ht-1 +W cf c t-1 +b f )

[0021] c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c )

[0022] o t =σ(W xo x t +W ho h t-1 +W co c t +b o )

[0023] h t =σtanh(c t )

[0024] Where W xo and W hi are the input-output gate matrix and the hidden input gate matrix respectively. Parameters i, f, o, c and σ represent the input gate, forget gate, output gate, cell activation vector and logistic sigmoid function respectively.

[0025] The device and method for detecting the degree of grain crushing based on an acoustic emission sensor provided by the present invention are simple to operate, suitable for a variety of complex food processing environments, and are a new type of online intelligent method for detecting the degree of grain crushing. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Schematic diagram of a grain particle breakage degree detection device based on an acoustic emission sensor in Example 1.

[0027] Figure 2 (a) and (b) are the front view and top view of the single particle friction collision test bench in Example 1, respectively.

[0028] Figure 3 This is a flow chart of the particle breakage degree detection method based on deep learning and blind source separation in Example 2.

[0029] Figure 4 This is the result diagram of analyzing the acoustic emission mixed signal using the EMD and VMD methods.

[0030] In the figure: 1 measured particle, 2 waveguide rod, 3 acoustic emission sensor probe, 4 signal conditioning module, 5 data acquisition module, 6 signal analysis module, 7 display module, 8 collection device, 9 stepper motor, 10 sample particle, 11 push rod, 12 plate DETAILED DESCRIPTION

[0031] The present invention provides a device and method for detecting the degree of grain crushing based on an acoustic emission sensor, which can realize defect detection of a detected object; the present invention is described as follows in conjunction with the accompanying drawings:

[0032] Example 1

[0033] like Figure 1 As shown, this embodiment provides a grain particle breakage degree detection device based on an acoustic emission sensor, and the grain particle breakage degree detection device based on an acoustic emission sensor is composed of a waveguide rod 2, an acoustic emission sensor 3, a signal conditioning module 4, a data acquisition module 5, a signal analysis module 6 and a display module 7.

[0034] A waveguide rod is inserted at the measuring position. In order to form an impact plane on the waveguide rod for the measured particle, the front end of the waveguide rod is a cylindrical structure with a cross-section, which is made of stainless steel 304. The acoustic emission sensor is installed on the cross-section of the rear end of the waveguide rod. The acoustic emission sensor is used to detect the acoustic signal transmitted from the waveguide rod. The acoustic emission sensor and the waveguide rod are bonded with a coupling agent to reduce the influence of external vibration. The measured particle 1 is driven by air (or other media) and hits the waveguide rod to generate elastic waves. The elastic wave is transmitted to the vicinity of the acoustic emission sensor and is detected by the acoustic emission sensor. Since the signal amplitude obtained by the acoustic emission sensor is small and contains environmental noise. In order to improve the signal-to-noise ratio of the collected signal, it is necessary to amplify, filter, digitize and other steps of the acoustic signal through the signal conditioning module, and the signal is transmitted to the computer through the data acquisition module for signal analysis and processing. In the signal analysis module, the particle crushing degree is obtained by extracting the characteristic parameters in the acoustic emission signal through the parameters obtained from the single particle friction collision experiment and the blind source separation method, and then transmitted to the display module for display.

[0035] Figure 2 It is a single particle friction and collision test bench. On the single particle friction and collision test bench, a single rod-shaped sample particle 10 with different wear degrees (different crack sizes) is placed on a flat plate 12 with a fixed tilt angle through a stepper motor 9. After the particle slides from the flat plate due to friction, it hits the front section of the horizontal waveguide rod, generating a stress wave that propagates outward. An acoustic emission sensor is installed on the rear section of the waveguide rod to collect the acoustic emission signals generated by the particles near the collision position. The influencing factors and types of particle breakage are explored by adjusting the particle size, particle material, flat plate tilt angle, flat plate friction coefficient, collision speed, collision position, etc.

[0036] Example 2

[0037] like Figure 3 As shown, this embodiment provides a method for detecting the degree of grain breakage based on an acoustic emission sensor, comprising the following steps:

[0038] Step 1: In a single particle friction collision mechanism experiment or actual application, the acoustic emission sensor measures the stress wave signal of the measured particle in real time. The signal conditioning module processes the weak acoustic emission signal and transmits it to the signal analysis module through the data acquisition module for acoustic signal processing.

[0039] Step 2: Use the weak acoustic emission signal data collected in step 1 as training samples and test samples.

[0040] Step 3. Use various blind source separation signal processing methods such as wavelet analysis, empirical mode decomposition (EMD), variational mode decomposition (VMD), etc. to perform noise reduction and feature extraction on weak non-stationary acoustic emission signals. The acoustic emission signals generated by grain damage are divided into continuous acoustic signals due to particle collision movement and burst acoustic signals caused by particle cracks. The blind source separation method is a signal processing algorithm that does not require prior knowledge and can separate the original signal from a mixed signal of continuous signals and burst signals. First, the intrinsic mode components (IMFs) of different scales are decomposed using two different methods, EMD and VMD. The time-frequency domain results of each IMFs after decomposition are as follows. Figure 4 As shown in the figure, the burst type acoustic signal is separated and extracted according to the signal decomposition result, which is used as the input of the particle fragmentation degree intelligent recognition algorithm.

[0041] Step 4: Use deep learning methods such as long short-term memory networks to identify the features extracted from the separated signals and original signals in step 3 to obtain different degrees of particle fragmentation.

[0042] In the method for detecting the degree of grain crushing based on acoustic emission sensors, in step 3, the long short-term memory network is used for a given sequence x=(x 1 ,...,x T ), hidden vector sequence h = (h 1 ,...,h T ), the output vector sequence y = (y 1 ,...,y T ) from t = 1 to T, calculated by the following formula:

[0043] h t =H(W xh x t +W hh h t-1 +b h )

[0044] y t =W hy h t +b h

[0045] Among them, W, b, and H are the weight, bias vector, and hidden activation of the model respectively. The calculation formulas of these parameters are as follows:

[0046] i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i )

[0047] f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f )

[0048] c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c )

[0049] o t =σ(W xo x t +W ho h t-1 +W co c t +b o )

[0050] h t =σtanh(c t )

[0051] Where W xo and W hi are the input-output gate matrix and the hidden input gate matrix respectively. Parameters i, f, o, c and σ represent the input gate, forget gate, output gate, cell activation vector and logistic sigmoid function respectively.

[0052] The above embodiments are only preferred specific implementations of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A device for detecting the degree of grain breakage based on an acoustic emission sensor, characterized in that: The particle breakage detection device is composed of an acoustic emission sensor probe, a waveguide rod, a signal conditioning module, a data acquisition module, a signal analysis module, a display module and a single particle friction collision test bench.

2. The grain particle crushing degree detection device based on acoustic emission sensor according to claim 1 is characterized in that: When designing the installation of the acoustic emission sensor probe and the waveguide rod, it is necessary to adjust the installation position of the acoustic emission sensor probe, the material of the waveguide rod, the shape of the waveguide rod and other parameters according to the size and resolution requirements of the object to be measured. The front end of the waveguide rod is the position for contacting particles, and the acoustic emission sensor is installed on the rear end section of the waveguide rod. The acoustic emission sensor is used to detect the acoustic signal transmitted from the waveguide rod. The acoustic emission sensor and the waveguide rod are bonded with a coupling agent to reduce the influence of external vibration.

3. The grain particle crushing degree detection device based on acoustic emission sensor according to claim 1 is characterized in that: The signal conditioning module amplifies the signal output by the acoustic emission sensor through a preamplifier, and then the data acquisition module composed of an acoustic emission signal analyzer acquires the acoustic emission signal.

4. The grain particle crushing degree detection device based on acoustic emission sensor according to claim 1 is characterized in that: The signal analysis module uses a blind source separation algorithm and a deep learning method to process the collected signals to obtain the degree of damage of the particles and transmits it to the display module for display.

5. The grain particle crushing degree detection device based on acoustic emission sensor according to claim 1 is characterized in that: The single-particle friction collision test bench is composed of a stepper motor, a push rod, a sample particle collection device, and a flat plate, and can be used to study the acoustic emission mechanism of single-particle friction collision. On the single-particle friction collision test bench, single particles with different degrees of wear (different crack sizes) are placed on a flat plate with a fixed tilt angle through a stepper motor. After the particles slide from the flat plate due to friction, they hit the front end section of the horizontal waveguide rod, generating a stress wave that propagates outward. An acoustic emission sensor is installed on the rear end section of the waveguide rod to collect the acoustic emission signals generated by the particles near the collision position. The influencing factors and types of particle breakage are explored by adjusting the particle size, particle material, flat plate tilt angle, flat plate friction coefficient, collision speed, collision position, etc.

6. A method for detecting the degree of grain breakage based on an acoustic emission sensor, characterized in that: The following steps are involved: Step 1: In a single particle friction collision mechanism experiment or actual application, the acoustic emission sensor measures the stress wave signal of the measured particle in real time. The signal conditioning module processes the weak acoustic emission signal and transmits it to the signal analysis module through the data acquisition module for acoustic signal processing. Step 2: Use the weak acoustic emission signal data collected in step 1 as training samples and test samples. Step 3: Through various blind source separation signal processing methods such as wavelet analysis, empirical mode decomposition (EMD), variational mode decomposition (VMD), etc., the weak non-stationary acoustic emission signals are subjected to noise reduction and feature extraction. The acoustic emission signals generated by grain damage are divided into continuous acoustic signals due to particle collision and burst acoustic signals caused by particle cracks. Blind source separation is a signal processing algorithm that does not require prior knowledge and can separate the original signal from a mixture of continuous and burst signals. The burst signal is separated and extracted based on the signal decomposition results and used as the input of the particle fragmentation degree intelligent recognition algorithm. Step 4: Use deep learning methods such as long short-term memory networks to identify the features extracted from the separated signals and original signals in step 3 to obtain different degrees of particle fragmentation. In the method for detecting the degree of grain breakage based on acoustic emission sensors, in step 3, the long short-term memory network is used for a given sequence x=(x1,...,x T ), hidden vector sequence h=(h1,...,h T ), the output vector sequence y=(y1,...,y T ) from t = 1 to T, calculated by the following formula: h t =H(W xh x t +W hh h t-1 +b h ) y t =W hy h t +b h in, W, b, and H are the weights, bias vectors, and hidden activations of the model, respectively. The calculation formulas for these parameters are as follows: i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i ) f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f ) c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) o t =σ(W xo x t +W ho h t-1 +W co c t +b o ) h t =σtanh(c t ) Where W xo and W hi are the input-output gate matrix and the hidden input gate matrix, respectively. Parameters i, f, o, c, and σ represent the input gate, forget gate, output gate, cell activation vector, and logistic sigmoid function, respectively.

Images