Vertical shaft type impact crusher impeller and fracture detection method

Abstract

The invention belongs to the technical field of vertical shaft type impact crushers, and particularly discloses a vertical shaft type impact crusher impeller fracture detection method which comprises the following steps: acquiring digital current data during impeller fracture in real time through a Hall current sensor and a PLC (Programmable Logic Controller) control system, extracting a feature data set from the digital current data, after the optimal parameter gamma of the Gaussian kernel function is determined, the center and the radius of a hypersphere of the SVDD model are determined through an optimization algorithm, a threshold value T is set, and a real-time feature data set is obtained in real time through a Hall current sensor and a PLC control system; and calculating the distance d from the real-time feature data set to the center of the hypersphere through the model, if d is greater than T, judging that the impeller is fractured, otherwise, continuing detection. In order to solve the problem that the impeller of the vertical shaft crusher is easy to break and other parts are damaged, in the aspect of impeller breakage detection, an SVDD model detection system is constructed based on a current monitoring technology, and the state of the impeller is accurately monitored in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of vertical shaft impact crushers, and particularly relates to an impeller of a vertical shaft impact crusher and a fracture detection method. Background Art

[0002] Vertical shaft impact crushers are widely used in industries such as artificial sand making, pulverized coal preparation, and ore fine crushing. The main structure and working principle of a vertical shaft impact crusher: The structure mainly consists of a vertical bearing box, a turntable, and an impeller. The impeller and the turntable are fixed to each other to form an integral body. Impact anvils are installed on the inner wall of the crusher. Under the high-speed rotation of the crusher turntable and the impeller, the material to be crushed is thrown at high speed onto the anvil by centrifugal force, and then the anvil rebounds the material onto the blades of the impeller for counterattack. This process is repeated until the material is crushed into finished products to complete the crushing task.

[0003] When the crusher crushes iron ore, due to the relatively high hardness of the ore, the lining plate provided on the impeller chassis wears relatively fast, and even causes the impeller to break. The broken impeller fragments rotate at high speed in the crushing cavity and cause damage to other components of the crushing cavity. Summary of the Invention

[0004] The purpose of the present invention is to provide an impeller of a vertical shaft impact crusher and a fracture detection method to solve the problem that when the crusher crushes iron ore, due to the relatively high hardness of the ore, the impeller edge is prone to fracture, and the fractured edge lining plate will cause damage to the crusher.

[0005] To achieve the above purpose, the basic solution provided by the present invention is: A fracture detection method for an impeller of a vertical shaft impact crusher, including the following steps:

[0006] S1. First, the Hall current sensor is used to collect the current in real time during the rotation of the entire crusher, and the current is transmitted to the PLC control system, and the current is converted into digital current data through an A / D converter;

[0007] S2. The data processing unit of the PLC control system extracts features from the digital current data and combines the extracted features into a feature data set;

[0008] S3. The feature data is input into the trained SVDD model to obtain the distance d from the feature data set to the center a of the SVDD model hypersphere, and it is compared with the pre-set distance threshold T of the center a of the SVDD model hypersphere. If d > T, it is determined that the impeller is fractured. If d ≤ T, it is considered that the device is in a normal state and continuous monitoring is carried out;

[0009] S4. When the SVDD model determines that the impeller is broken, the SVDD model transmits an electrical signal to the PLC control system. The PLC control system immediately stops the front-end equipment of the vertical shaft crusher. After the vertical shaft crusher has crushed the ore already in the crushing chamber, the PLC control system stops the vertical shaft crusher and the back-end equipment.

[0010] The beneficial effects of the present invention are as follows: In terms of impeller breakage detection, the SVDD model detection system constructed based on current monitoring technology realizes real-time and accurate monitoring of the impeller state. This can not only timely detect potential impeller breakage risks, avoid damage to other components of the crusher caused by the high-speed rotation of broken impeller fragments, reduce equipment maintenance costs and downtime, but also maintain high accuracy and reliability under complex working conditions, effectively improving the safety and stability of equipment operation.

[0011] Solution 2, which is the optimization of the basic solution. In S1, calculate the mean value α and standard deviation δ of the digital current data, clean the outliers with α + 3δ as the boundary, and at the same time use the weighted moving average method to smooth the digital current data; in the cleaning link, according to the mean value and standard deviation, screen out the outliers to prevent interference with subsequent analysis; the smoothing process uses the weighted moving average method, assigns high weights to recent data, reduces short-term fluctuation noise, highlights the current trend, and enhances the accuracy and stability of fault detection.

[0012] Solution 3, which is the optimization of the basic solution. In S2, the features include the mean value, maximum value, minimum value, standard deviation, peak-to-peak value of the digital current, and the frequency spectrum of the current waveform. The frequency spectrum feature of the current waveform is calculated by fast Fourier transform of the current waveform spectrum; the impeller breakage changes the mechanical and electromagnetic characteristics of the equipment, causing fluctuations in the energy of specific frequencies in the current spectrum, triggering sudden changes in the amplitudes of certain frequencies or generating new characteristic frequencies. This frequency spectrum feature captures the subtle changes in impeller breakage, complements basic features such as the mean value, maximum and minimum values, constructs a comprehensive feature dataset, deeply mines the fault information contained in the data, and improves the accuracy and sensitivity of fault diagnosis.

[0013] Solution 4, which is the optimization of the basic solution. In S3, the training steps of the SVDD model are as follows:

[0014] Step 1: Combine the current data of the vertical shaft impact crusher in the normal operation and impeller breakage states into a feature dataset;

[0015] Step 2: Use the cross-validation method to determine the key parameter γ of the Gaussian kernel function;

[0016] Step 3: After obtaining the optimal parameter γ of the Gaussian kernel function, determine the values of the center a and radius R of the hypersphere of the SVDD model through the training set to obtain the optimal SVDD model; through the training of the model, the optimal model for detecting impeller breakage can be obtained.

[0017] Solution 5, which is the optimization of Solution 4. First, within the range [0.1, 10] of the parameter γ, experiments are conducted with a step size of 0.5 to obtain different trial values of the parameter γ. Then, each trial value of the parameter γ is input into the SVDD model, and the SVDD model is trained using the training set to obtain the hyper-sphere center a and radius R for different parameters γ. Next, the classification accuracy of the SVDD model for impeller fracture data on the validation set is calculated, and the optimal value of the parameter γ is obtained through the highest classification accuracy. The formula for calculating the classification accuracy is as follows:

[0018]

[0019] In the formula: ψ represents the accuracy; K represents the number of samples correctly judged as impeller fracture by the SVDD model; Z represents the total number of samples in the validation set.

[0020] Finally, the highest accuracy obtained by the SVDD model on the validation set is used to determine the optimal value of the parameter γ.

[0021] Solution 6, which is the optimization of the basic solution. After determining the optimal parameter γ of the Gaussian kernel function, the training set is input into the SVDD model. During the training process, it is necessary to minimize the objective function that includes the size of the hyper-sphere and the tolerance for abnormal data, while satisfying the constraint conditions to define the boundary between normal data and abnormal data. Finally, through multiple iterations, the values of the hyper-sphere center, radius, and relaxation variables are adjusted to optimize the SVDD model parameters and obtain the values of the hyper-sphere center a and radius R. The objective function τ to be minimized is as follows:

[0022]

[0023] Constraint conditions:

[0024] ||y i -a|| 2 ≤R 2 +ξ i , ξ≥0

[0025] In the formula: τ represents the objective function to be minimized; R 2 represents the size of the hyper-sphere; represents the tolerance for abnormal data; ||y i -a|| 2 represents the square of the distance from the data point y i to the hyper-sphere center a, and ξ i represents the relaxation variable. Through multiple iterations, the hyper-sphere center, radius, and relaxation variable are optimized to obtain a stable model, providing a solid core criterion for fault determination and efficiently and accurately identifying impeller fracture abnormalities.

[0026] Solution VII, which is the preference of the basic solution. In S3, the calculation formula of the distance threshold T from the center a of the SVDD model hypersphere is as follows:

[0027] T = μ + 2δ

[0028] In the formula: T represents the threshold; μ represents the average distance from the feature dataset to the center of the hypersphere; δ represents the standard deviation of the distance from the feature dataset to the center of the hypersphere; by setting the threshold, normal or abnormal current features can be effectively distinguished, improving the timeliness and accuracy of fault warning.

[0029] Solution VIII, which is the preference of the basic solution. In S3, the calculation formula of the distance d from the feature dataset to the center a of the hypersphere is as follows:

[0030]

[0031] In the formula: d represents the distance from the real-time feature dataset to the center a of the hypersphere; z k represents the k-th data point in the feature dataset; a k represents the k-th data point in the hypersphere center.

[0032] Solution IX, which is the preference of the basic solution, includes a chassis. A support seat is provided on the chassis, and a hammer is detachably connected to the support seat. Wear-resistant ceramics are pasted on the chassis, and side lining plates are bolted to the side of the chassis. Bolt holes for connecting the crusher turntable are opened on the chassis; by setting wear-resistant ceramics and optimizing the overall structure, the wear-resistant and damage-resistant performance of the crusher can be improved, preventing the side lining plates from wearing and breaking, and strongly supporting efficient and stable crushing operations.

[0033] Solution X, which is the preference of Solution IX. A through hole is opened at the top of the support seat, a connecting block is provided on the hammer, a threaded hole is opened on the connecting block, a threaded pin for threadedly connecting the threaded hole is embedded in the through hole, a groove for embedding the connecting block is opened on the support seat, and the threaded hole is directly opposite to the through hole; through the connection of the threaded pin and the connecting seat, it is convenient to replace the hammer. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is the top view of an impeller of a vertical shaft impact crusher according to the present invention;

[0035] Figure 2 is Figure 1 the sectional view taken along line A-A in

[0036] Figure 3 is the exploded view of the support seat and the hammer in an impeller of a vertical shaft impact crusher according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0037] The present invention will be further described in detail through specific embodiments as follows:

[0038] The reference numerals in the accompanying drawings of the specification include: 1. chassis; 2. support seat; 3. impact hammer; 4. wear-resistant ceramic; 5. side lining plate; 6. connecting block; 7. threaded pin.

[0039] Embodiment 1

[0040] A method for detecting the fracture of the impeller of a vertical shaft impact crusher includes the following steps:

[0041] S1. First, the Hall current sensor is used to collect the current in real time during the rotation of the entire crusher, and the current is transmitted to the PLC control system. The current is converted into digital current data through an A / D converter, and then the digital current data is cleaned and smoothed; Data cleaning: Calculate the mean and standard deviation of the digital current data, and identify and clean the outliers with the mean plus or minus three times the standard deviation as the boundary. Through data cleaning, the abnormal points caused by electromagnetic interference and instantaneous sensor failures can be effectively excluded, and the reliability of data quality can be improved; Data smoothing: The weighted moving average method is used to process the data, reducing short-term fluctuations and noise interference, making the data trend clear, and facilitating accurately grasping the current change law caused by the impeller fracture. Model number of the Hall current sensor: HBA100-LAH;

[0042] S2. The data processing unit of the PLC control system extracts the mean, maximum value, minimum value, standard deviation, peak-to-peak value and spectral characteristics of the current waveform from the preprocessed digital current data, and combines the extracted features into a feature data set. The spectral characteristics of the current waveform are calculated by performing a fast Fourier transform on the current waveform spectrum, and the main frequency components and amplitudes are extracted as spectral characteristics. The mean reflects the overall level of the current, the standard deviation quantifies the degree of fluctuation, the maximum and minimum values define the change interval, and the peak-to-peak value intuitively presents the fluctuation amplitude. These features present the basic outline of the current from multiple dimensions, reserving basic information for fault judgment;

[0043] S3. Input the feature data into the trained SVDD model, obtain the distance d from the feature data set to the center a of the SVDD model hyper-sphere, and compare it with the pre-set distance threshold T of the center a of the SVDD model hyper-sphere. If d > T, it is determined that the impeller is fractured; if d ≤ T, it is considered that the equipment is in a normal state and continuous monitoring is continued;

[0044] The calculation formula for the distance d from the feature data set to the center a of the hyper-sphere is as follows:

[0045]

[0046] In the formula: d represents the distance from the real-time feature data set to the center a of the hyper-sphere; z krepresents the k-th data point in the feature dataset; a k represents the k-th data point in the hypersphere center.

[0047] The calculation formula for the distance threshold T of the SVDD model's hypersphere center a is as follows:

[0048] T = μ + 2δ

[0049] In the formula: T represents the threshold; μ represents the average distance from the feature dataset to the hypersphere center; δ represents the standard deviation of the distance from the feature dataset to the hypersphere center.

[0050] The training steps of the SVDD model are as follows:

[0051] Step 1: Combine the current data of the vertical shaft impact crusher in the normal operation and impeller fracture states into a feature dataset;

[0052] Step 2: Use the cross-validation method to determine the key parameter γ of the Gaussian kernel function. The Gaussian kernel function has the ability of non-linear mapping, which can map complex low-dimensional current data to a high-dimensional space, highlight the potential structure and pattern of the data, efficiently process the complex non-linear relationship between impeller fracture and current, greatly improve the fault diagnosis accuracy and generalization ability of the model, and determine the optimal value of the key parameter γ of the Gaussian kernel function through the cross-validation method.

[0053] First, collect 300 groups of feature datasets, and randomly divide them into a training set and a validation set according to the ratio of 80% and 20%, that is, the training set contains 240 groups of data, and the validation set contains 60 groups of data. The validation set can effectively evaluate the generalization performance of the model and prevent overfitting. In the value range [0.1, 10] of the parameter γ, conduct experiments with a step size of 0.5 to obtain different trial values of the parameter γ. Then, input each trial value of the parameter γ into the SVDD model, use the training set to train the SVDD model, obtain the hypersphere center a and radius R of different parameters γ, and then calculate the classification accuracy of the SVDD model for impeller fracture data on the validation set. The optimal value of the parameter γ is obtained through the highest classification accuracy. Among them, the calculation formula for the classification accuracy is as follows:

[0054]

[0055] In the formula: ψ represents the accuracy; K represents the number of samples correctly judged as impeller fracture by the SVDD model; Z represents the total number of samples in the validation set;

[0056] Step 3: After determining the optimal parameter γ of the Gaussian kernel function, input the training set into the SVDD model. During the training process, it is necessary to minimize the objective function that includes the size of the hypersphere and the tolerance for abnormal data, while satisfying the constraint conditions to define the boundary between normal data and abnormal data. Finally, through multiple iterations, adjust the values of the center, radius, and relaxation variables of the hypersphere to optimize the SVDD model parameters and obtain the values of the hypersphere center a and radius R. Among them, minimize the objective function τ:

[0057]

[0058] Constraint conditions:

[0059] ||y i -a|| 2 ≤R 2 +ξ i , ξ≥0

[0060] In the formula: τ represents the minimized objective function; R 2 represents the size of the hypersphere; represents the tolerance for abnormal data; ||y i -a|| 2 represents the square of the distance from the data point y i to the hypersphere center a, and ξ i represents the relaxation variable;

[0061] S4. When the SVDD model determines that the impeller is broken, the SVDD model transmits an electrical signal to the PLC control system. The PLC control system immediately stops the front-end equipment of the vertical shaft crusher. After the vertical shaft crusher has crushed the ore already in the crushing cavity, the PLC control system stops the vertical shaft crusher and the back-end equipment.

[0062] Embodiment 2

[0063] As Figures 1 to 3 shown: A vertical shaft impact crusher impeller includes a chassis 1 and impact hammers 3. A support seat 2 is fixedly connected to the chassis 1. A through hole is opened at the top of the support seat 2. A connecting block 6 is fixedly connected to the impact hammer 3. A threaded hole is opened on the connecting block 6. A threaded pin 7 for threaded connection with the threaded hole is embedded in the through hole. A groove for embedding the connecting block 6 is opened on the support seat 2. The threaded hole is directly opposite to the through hole. Wear-resistant ceramics 4 are pasted on the chassis 1. The side of the chassis 1 is connected with side lining plates 5 by internal hexagonal bolts. Bolt holes for connecting the crusher turntable are opened on the chassis 1. The chassis 1 and the crusher turntable are connected by internal hexagonal bolts. A vertical bearing box is provided on the crusher. A motor is fixedly connected to the driving shaft of the vertical bearing box. The motor is electrically connected to a Hall current sensor. A V-belt pulley is sleeved between the driven shaft of the vertical bearing box and the turntable. The driving shaft is driven by the motor, the driven shaft is driven by the driving shaft, and the turntable is driven by the driven shaft through the V-belt pulley to rotate at high speed.

[0064] The implementation mode of this embodiment is as follows: When the vertical shaft impact crusher is running, the turntable rotates at a high speed to drive the impeller to rotate. The impact hammers on the chassis collide violently with the materials entering the crushing chamber at a high speed under the action of centrifugal force to achieve primary crushing. In addition, the materials not broken by the impact hammers rebound into the crusher after being struck by the impact hammers and collide with other materials that have not been fully broken to complete the crushing. During this period, the wear-resistant ceramics play a key protective role. Relying on the characteristics of high hardness and low friction coefficient, the wear-resistant ceramics greatly reduce the wear and fracture of the impeller caused by the materials.

[0065] The above are only the embodiments of the present invention, and common knowledge such as the specific structures and characteristics known in the solution is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the implementation effect of the present invention and the practicality of the patent. The protection scope required by this application shall be subject to the content of its claims, and the specific implementation manners described in the specification can be used to interpret the content of the claims.

Claims

1. A method for detecting fracture of an impeller of a vertical shaft impact crusher, characterized in that: The following steps are involved: S1. First, the current of the entire crusher during rotation is collected in real time through the Hall current sensor, and the current is transmitted to the PLC control system, and the current is converted into digital current data through the A / D converter; S2, a data processing unit of the PLC control system extracts features from the digital current data and combines the extracted features into a feature data set; S3, input the feature data into the trained SVDD model, obtain the distance d from the feature data set to the center a of the SVDD model hypersphere, and compare it with the preset distance threshold T of the center a of the SVDD model hypersphere. If d>T, it is determined that the impeller is broken. If d≤T, it is considered that the equipment is in normal state and monitoring continues; S4. When the SVDD model determines that the impeller is broken, the SVDD model transmits an electrical signal to the PLC control system, and the PLC control system immediately stops the front-end equipment of the vertical shaft crusher. When the vertical shaft crusher has crushed the ore in the crushing chamber, the PLC control system stops the vertical shaft crusher and the back-end equipment.

2. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 1, characterized in that: In S1, the mean α and standard deviation δ of the digital current data are calculated, and the abnormal values ​​are cleaned with α+3δ as the boundary. At the same time, the weighted moving average method is used to smooth the digital current data.

3. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 1, characterized in that: In S2, the features include the mean value, maximum value, minimum value, standard deviation, peak-to-peak value and spectrum of the current waveform of the digital current, and the spectrum feature of the current waveform is calculated by fast Fourier transforming the spectrum of the current waveform.

4. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 1, characterized in that: In S3, the training steps of the SVDD model are as follows: Step 1, combining the current data of the vertical shaft impact crusher in normal operation and impeller fracture state into a feature data set; Step 2: Use cross-validation method to determine the key parameter γ of Gaussian kernel function; Step 3: After obtaining the optimal parameter γ of the Gaussian kernel function, the center a and radius R of the SVDD model hypersphere are determined through the training set to obtain the optimal SVDD model.

5. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 4, characterized in that: First, within the range of [0.1, 10] of the parameter γ, the experiment was carried out with a step size of 0.5 to obtain different trial values ​​of the parameter γ. Then, each trial value of the parameter γ was input into the SVDD model, and the SVDD model was trained using the training set to obtain the center a and radius R of the hypersphere with different parameters γ. Then, the classification accuracy of the SVDD model for the impeller fracture data on the validation set was calculated, and the optimal value of the parameter γ was obtained through the highest classification accuracy. The calculation formula for the classification accuracy is as follows: Where: ψ represents the accuracy; K represents the number of samples correctly judged as impeller fracture by the SVDD model; Z represents the total number of samples in the validation set.

6. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 4, characterized in that: After determining the optimal parameter γ of the Gaussian kernel function, the training set is input into the SVDD model. During the training process, the objective function needs to be minimized, including the size of the hypersphere and the tolerance to abnormal data, while satisfying the constraints and defining the boundaries between normal data and abnormal data. Finally, the center, radius, and slack variable values ​​of the hypersphere are adjusted through multiple iterations to optimize the SVDD model parameters and obtain the values ​​of the hypersphere center a and radius R, where the objective function τ is minimized: Constraints: ||y i -a|| 2 ≤R 2 +ξ i ,ξ≥0 Where: τ represents the minimization objective function; R 2 represents the size of the hypersphere; Indicates the tolerance for abnormal data; ||y i -a|| 2 Represents data point y i The square of the distance to the center of the hypersphere, ξ i represents the slack variable.

7. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 1, characterized in that: In S3, the calculation formula of the distance threshold T of the center a of the SVDD model hypersphere is as follows: T=μ+2δ Where: T represents the threshold; μ represents the mean distance from the feature data set to the center of the hypersphere; δ represents the standard deviation of the distance from the feature data set to the center of the hypersphere.

8. A method for detecting fracture of an impeller of a vertical shaft impact crusher according to claim 1, characterized in that: In S3, the calculation formula of the distance d from the feature data set to the center a of the hypersphere is as follows: Where: d represents the distance from the real-time feature dataset to the center of the hypersphere a; z k Represents the kth data point in the feature data set; a k represents the kth data point in the center of the hypersphere.

9. A vertical shaft impact crusher impeller, characterized in that: The invention comprises a chassis (1), wherein a support seat (2) is provided on the chassis (1), an impact hammer (3) is detachably connected to the support seat (2), wear-resistant ceramics (4) are applied to the chassis (1), a side lining plate (5) is connected to the side bolts of the chassis (1), and bolt holes for connecting to the crusher turntable are opened on the chassis (1).

10. The impeller of a vertical shaft impact crusher according to claim 9, characterized in that: The top of the support seat (2) is provided with a through hole, the impact hammer (3) is provided with a connecting block (6), the connecting block (6) is provided with a threaded hole, a threaded pin (7) for threaded connection with the threaded hole is embedded in the through hole, and a groove for embedding the connecting block (6) is formed on the support seat (2), and the threaded hole is directly opposite to the through hole.

Images