A device and method for real-time detection of mineral powder grade

Abstract

The present disclosure proposes a device and method for real-time detection of mineral powder grade. The detection device includes an array coil, a signal detection circuit and a main control unit connected in sequence; one or more array coils are arranged above the mineral powder transmission device to be detected, and the signal detection circuit is used to detect the change in coil voltage when the mineral powder passes through, and the main control unit identifies the grade of the mineral powder according to the voltage change. The present disclosure realizes online monitoring of mineral powder or slurry based on the array coil, evaluates the grade level, and can play a role in preliminary judgment and classification of samples, which is convenient for subsequent refined mineral processing. It can greatly reduce the cost of mineral processing, and has the advantages of high production efficiency, good product quality, and no environmental pollution.

Description

Technical Field

[0001] The present disclosure relates to the technical field related to mineral powder grade detection, and more specifically, to a device and method for real-time detection of mineral powder grade. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] The grade of iron ore refers to the content of useful components or useful minerals in unit volume or unit weight of iron ore. Commonly expressed methods include mass percentage, g / ton, g / cubic meter, g / liter, etc. It is the main indicator to measure the economic value of the ore deposit. For example: the theoretical maximum grade of hematite (mainly Fe2O3) is 70%; the theoretical maximum grade of siderite (mainly FeCO3) is 48.3%; the theoretical maximum grade of limonite (mainly Fe2O3.H2O) is 62.9%; the theoretical maximum grade of magnetite (mainly Fe3O4) is 72.4%.

[0004] Magnetite is a highly magnetic iron ore. There are very few pure magnetite ores. It is generally a composite ore containing TiO2 and V2O5. In addition, it may also contain elements such as Ni and Co. Other impurities include quartz, silicates, and harmful compounds such as phosphorus and sulfur. The existing beneficiation method is to crush and grind the magnetite into very small particles of ore powder or ore pulp through multiple stages, and then gradually filter out single metals through magnetic separation, gravity separation, flotation and other procedures. The existing magnetic separation method is relatively rough, which simply separates impurities in the ore and cannot monitor the ore grade in real time, which is not conducive to subsequent refined beneficiation. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a real-time detection device and method for mineral powder grade, which realizes online monitoring of mineral powder or slurry based on array coils, evaluates the grade, and can play a role in preliminary judgment and classification of samples, which is convenient for subsequent refined mineral processing. It can greatly reduce the cost of mineral processing, and has the advantages of high production efficiency, good product quality, and no environmental pollution.

[0006] In order to achieve the above objectives, the present disclosure adopts the following technical solutions:

[0007] One or more embodiments provide a real-time detection device for mineral powder grade, comprising an array coil, a signal detection circuit and a main control unit connected in sequence; one or more array coils are arranged above a mineral powder transmission device to be detected, the signal detection circuit is used to detect changes in coil voltage when mineral powder passes by, and the main control unit identifies the grade of the mineral powder based on the voltage change.

[0008] One or more embodiments provide a method for real-time detection of mineral powder grade, comprising the following steps:

[0009] One or more array coils are arranged above the mineral powder transmission device to be detected;

[0010] Get the voltage value detected by the array coil when the mineral powder passes through the array coil;

[0011] Based on the detected voltage change data, the regression tree algorithm is used to identify the grade of mineral powder.

[0012] Compared with the prior art, the present invention has the following beneficial effects:

[0013] The present invention realizes online monitoring of mineral powder or slurry based on array coils, evaluates grade, and can play a role in preliminary judgment and classification of samples, which is convenient for subsequent refined mineral processing. It can greatly reduce the cost of mineral processing, and has the advantages of high production efficiency, good product quality, and no environmental pollution.

[0014] The advantages of the present disclosure and the advantages of additional aspects will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The exemplary embodiments of the present disclosure and the description thereof are used to explain the present disclosure but do not constitute a limitation of the present disclosure.

[0016] Figure 1 is a schematic structural diagram of the array coil of Embodiment 1 of the present disclosure;

[0017] Figure 2 is a schematic structural diagram of a detection device according to Embodiment 1 of the present disclosure;

[0018] Figure 3 is a schematic diagram of mineral powder passing through the array coil of the detection device of Example 1 of the present disclosure;

[0019] Figure 4 is a measurement principle diagram of a detection device according to Embodiment 1 of the present disclosure;

[0020] Figure 5 This is a schematic diagram of the active coil voltage acquisition principle of the detection device of Embodiment 1 of the present disclosure;

[0021] Figure 6 is a flow chart of a method for identifying the grade of mineral powder using a regression tree algorithm in Embodiment 1 of the present disclosure;

[0022] Figure 7 is a principle block diagram of a signal detection circuit in Embodiment 1 of the present disclosure;

[0023] Figure 8 is a pressure difference measurement circuit diagram of Embodiment 1 of the present disclosure;

[0024] Fig. 9 is a rectifier and filter circuit diagram of Embodiment 1 of the present disclosure;

[0025] Fig.10 is a schematic diagram of the regression tree model structure of Example 1 of the present disclosure;

[0026] Among them: 1. Base plate, 2. Shielding frame, 3. Active coil, 4. Conveyor belt. DETAILED DESCRIPTION

[0027] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0028] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0029] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof. It should be noted that, in the absence of conflict, the various embodiments in the present disclosure and the features in the embodiments can be combined with each other. The embodiments will be described in detail below in conjunction with the accompanying drawings.

[0030] Example 1

[0031] In the technical solutions disclosed in one or more embodiments, Figure 1-10 As shown, a real-time detection device for mineral powder grade includes an array coil, a signal detection circuit and a main control unit connected in sequence; one or more array coils are arranged above a mineral powder transmission device to be detected, the signal detection circuit is used to detect changes in coil voltage when mineral powder passes by, and the main control unit identifies the grade of the mineral powder based on the voltage change.

[0032] This embodiment detects the grade of the ore powder by setting an array coil and detecting the voltage change of the array coil caused by the ore powder, thereby realizing online automatic detection, which can greatly reduce the ore dressing cost and has the advantages of high production efficiency, good product quality and no environmental pollution.

[0033] In some embodiments, the structure of the array coil is as follows Figure 1 As shown, it includes a base plate 1, a plurality of active coils 3 arranged in an array on the base plate, a magnetic core is arranged at the center of each active coil, and a shielding frame 2 is arranged around each active coil to form a coil unit; each active coil is connected in series with a capacitor and then connected to the main control unit.

[0034] Specifically, the capacitor is located on an external circuit connected to each active coil 3, and the capacitor group corresponding to each active coil 3 is arranged under the active coil 3, and the capacitors are respectively connected in series with the active coils 3 through wires.

[0035] Optionally, the coil unit cell may be a quadrilateral, a square, a rectangle, etc.

[0036] Optionally, the structure of each coil unit cell is the same, that is, the size and shape of the shielding frame are the same, the number of turns of the active coil and the diameter of the coil are the same, and the size and structure of the magnetic core are the same.

[0037] In this embodiment, the sensing structure is divided into a number of identical cells according to the units of the coils. Each cell is filled with mineral powder for each detection, thereby ensuring that the quality and volume of the mineral powder in each cell or passing through the cell are the same.

[0038] Detection of ore grade: Materials of different composition, quality, structure and physical state show different electromagnetic characteristics under a certain alternating electromagnetic field. Based on this, the grade of ore powder can be detected in real time during the sorting process.

[0039] The measurement principle is as follows Figure 4 As shown. The resonant circuit of the detection coil and the capacitor in series has a resonant frequency of several megahertz. When an alternating current of a certain frequency is input to the detection coil, an alternating electromagnetic field will be generated around the coil. When metal enters the alternating electromagnetic field, the eddy current effect of the metal will cause magnetic field fluctuations, which will cause the resonant circuit to deviate from the resonance point, thereby changing the voltage at both ends of the detection coil. The stronger the magnetism of the ore, the greater the disturbance to the magnetic field and the greater the change in voltage.

[0040] When in use, by constructing a data set of voltage amplitude, phase changes and ore grade, a model is established and trained to achieve the purpose of accurate ore dressing.

[0041] During the detection process of this embodiment, the detected target passes over the detection coil, which is usually uniform mineral powder after drying, to ensure that the detected target has the same physical state and uniform composition.

[0042] Optionally, the mineral powder transmission device to be detected is specifically a conveyor belt 4, a specific setting structure, the array coil is fixedly arranged above the transmission device, and when the mineral powder passes through the array coil on the transmission device, it causes a change in the voltage of the array coil.

[0043] In this embodiment, multiple array coils are provided to improve the accuracy of detection. Figure 2 As shown,

[0044] like Figure 2As shown, a plurality of 4*4 cells are arranged on the conveyor belt 4, and each cell corresponds to a group of sensor coils. Under the action of the conveyor belt, the mineral powder in the cell moves to the left at a uniform speed. Figure 3 As shown in the figure, at a certain moment, when the mineral powder reaches the detection area, the voltage changes of the coils corresponding to the 16 cells can be measured, and then its grade can be determined. The time taken by the conveyor belt to move one cell side is a sampling cycle, and the voltage changes of the detection coils are recorded every sampling cycle. Figure 2 As shown in the figure, the mineral powder in each cell is tested 4 times. The changes in multiple sampling parameters are comprehensively considered to effectively reduce the detection error. Assuming that N groups of 4*4 cells are set on the conveyor belt, the mineral powder in each cell is tested 4N times, which greatly improves the accuracy of the detection.

[0045] In some embodiments, the main control unit includes a main controller, a signal detection part, a multi-way switch control part, and a waveform generating part; the main controller, the signal detection part, and the multi-way switch control part are connected in sequence, the multi-way switch control part is connected to the array coil through the signal detection circuit, and the main controller is connected to the array coil through the waveform generating part and the signal detection circuit.

[0046] Optionally, the main controller can use a single-chip microcomputer, using an STM32F407 series ARM single-chip microcomputer as the core control part.

[0047] Signal detection part: configured to collect the voltage of the subdivided grid one by one by controlling the multi-way switching switch; the CPU of this embodiment has 16 AD ports, which can realize the parallel collection of 16 channels of data, and each channel can realize serial collection through the switch. The coil voltage collection principle diagram is shown in FIG. Figure 5 shown.

[0048] The waveform generation part is configured to convert square waves of different frequencies emitted by the CPU into sine waves; the sine waves drive each active coil in the array coil after power amplification and amplitude stabilization circuits, providing a constant driving current for the active coils.

[0049] Furthermore, the main control unit also includes a data storage part and a human-computer interaction part.

[0050] The data storage part is used to store data, and is used to store the original data of zero point adjustment and other setting data.

[0051] The human-computer interaction part may include an audible and visual alarm, a display, etc., which are respectively connected to the main control; the audible and visual alarm realizes audible and visual alarm, and the display can display the type and shape of the metal object.

[0052] Optional, such as Figure 7As shown, the signal detection circuit includes a power supply, and a voltage difference measurement circuit, an overvoltage protection circuit, a voltage amplification circuit, a rectification and filtering circuit and an A / D conversion circuit which are connected in sequence, and the A / D conversion circuit is connected to a main controller.

[0053] The voltage difference measurement circuit can use OPA series operational amplifiers, such as Figure 8 shown.

[0054] The rectifier and filter circuit uses a high-precision, single-chip root-mean-square DC converter AD637, such as Fig. 9 shown.

[0055] In some embodiments, the main controller processes the detected voltage data to identify the grade of the ore powder. Specifically, the grade of the ore powder can be identified by using a regression tree algorithm based on the detected voltage change data. The identification method is as follows: Figure 6 As shown, the following steps are included:

[0056] Step 1: Obtain the voltage value after the mineral powder to be detected passes through the active coil, extract the voltage amplitude and phase, and perform preprocessing;

[0057] Step 2: Input the preprocessed voltage value into the trained regression tree model to obtain the grade of the mineral powder to be tested.

[0058] The regression tree model includes multiple regression trees, each regression tree corresponds to a division unit of the grade level, and a regression tree corresponds to a division of the input space (ie, feature space) and an output value on the division unit.

[0059] like Fig.10 As shown in the figure, each node in the figure is a regression tree, and the nodes are represented by rectangular boxes, such as node ①, node ②, node ③, node ④, etc. marked in the figure.

[0060] The regression tree uses a heuristic method to divide the input space. It will traverse all input variables to find the optimal segmentation variable j and the optimal segmentation point s, that is, select the jth feature x j and its value s divide the input space into two parts, and then repeat this division operation.

[0061] Optionally, the training process of the regression tree model includes the following steps:

[0062] Step 21, voltage amplitude, phase change and ore grade data set;

[0063] Step 22: Divide the data of the data set into a training set and a test set;

[0064] Classify the data, divide the models that do not exist in the data set into the training set, and divide the models that exist in the data set into the test set.

[0065] The model that is not in the data set becomes a model of the test set in the data set after training.

[0066] Step 23: data preprocessing, which includes normalization, sample label setting and material definition.

[0067] Among them, normalization: take a set number of test values, such as 20, and use the normal distribution standardization method for normalization, as follows:

[0068]

[0069] Among them, x represents a magnetic field data sample, Avg function represents the mean value of this sample detection, σ represents the standard deviation of the detection value, x i and Represents the value of the i-th detection value of sample x before and after normalization.

[0070] Sample label setting: In view of the changes in the voltage amplitude and phase at both ends of the detection coil, two labels, amplitude and phase, are set for each sample to indicate the changes in the coil voltage amplitude and phase.

[0071] Define model material: The material model definition adopts a regression tree-based model. One regression tree model corresponds to one division method of the sample space. The division method in this embodiment corresponds to the grade level.

[0072] Step 24: Taking the changes in voltage amplitude and phase as input and the corresponding ore grade as output, the constructed regression tree model is trained to obtain the parameters of the regression tree model.

[0073] After completing the parameter initialization for the regression tree model, the model is trained on the training set, and the objective function is as follows:

[0074]

[0075] Among them, L is the cross entropy loss function. The smaller the loss function L is, the closer the prediction is to the real sample; x is the sample label value, is the predicted value of the model. w is all the parameters to be learned in the model. α represents the regularization strength and is a manually adjustable parameter used to control the degree of model fitting.

[0076] When training the model, the output prediction value is calculated from the input parameters, and the objective function value is obtained from the above formula. The objective function value is updated with the parameter value according to the selected optimization algorithm to complete a round of model training.

[0077] Optionally, the optimization algorithm is the Adam optimizer, which has the advantages of adaptive learning rate and high robustness. The model is evaluated based on the model prediction accuracy, and the model is further optimized through artificial intelligence deep learning.

[0078] Example 2

[0079] Based on Example 1, this embodiment provides a method for real-time detection of mineral powder grade, comprising the following steps:

[0080] Step 1, setting one or more array coils above the mineral powder transmission device to be detected;

[0081] Step 2: Obtain the voltage value detected by the array coil when the mineral powder passes through the array coil;

[0082] Step 3: Based on the detected voltage change data, a regression tree algorithm is used to identify the grade of the mineral powder.

[0083] This embodiment detects the grade of the ore powder by setting an array coil and detecting the voltage change of the array coil caused by the ore powder, thereby realizing online automatic detection, which can greatly reduce the ore dressing cost and has the advantages of high production efficiency, good product quality and no environmental pollution.

[0084] Step 31, obtaining the voltage value after the mineral powder to be detected passes through the active coil, extracting the voltage amplitude and phase, and performing preprocessing;

[0085] Step 32: input the preprocessed voltage value into the trained regression tree model to obtain the grade of the mineral powder to be detected.

[0086] The regression tree model includes multiple regression trees, each regression tree corresponds to a division unit of the grade level, and a regression tree corresponds to a division of the input space (ie, feature space) and an output value on the division unit.

[0087] The regression tree uses a heuristic method to divide the input space. It will traverse all input variables to find the optimal segmentation variable j and the optimal segmentation point s, that is, select the jth feature x j and its value s divide the input space into two parts, and then repeat this division operation.

[0088] Optionally, the training process of the regression tree model includes the following steps:

[0089] Step 321, voltage amplitude, phase change and ore grade data set;

[0090] Step 322, dividing the data set into a training set and a test set;

[0091] Classify the data, divide the models that do not exist in the data set into the training set, and divide the models that exist in the data set into the test set.

[0092] Step 323: data preprocessing, the preprocessing includes: normalization processing, sample label setting and material definition.

[0093] The model that is not in the data set becomes a model of the test set in the data set after training.

[0094] Among them, normalization: take a set number of test values, such as 20, and use the normal distribution standardization method for normalization, as follows:

[0095]

[0096] Among them, x represents a magnetic field data sample, Avg function represents the mean value of this sample detection, σ represents the standard deviation of the detection value, x i and Represents the value of the i-th detection value of sample x before and after normalization.

[0097] Sample label setting: In view of the changes in the voltage amplitude and phase at both ends of the detection coil, two labels, amplitude and phase, are set for each sample to indicate the changes in the coil voltage amplitude and phase.

[0098] Define model material: The material model definition adopts a regression tree-based model. One regression tree model corresponds to one division method of the sample space. The division method in this embodiment corresponds to the grade level.

[0099] Step 324: Take the changes in voltage amplitude and phase as input and the corresponding ore grade as output, train the constructed regression tree model, obtain the parameters of the regression tree model, and test it with the data of the test set.

[0100] After completing the parameter initialization for the regression tree model, the model is trained on the training set, and the objective function is as follows:

[0101]

[0102] Among them, L is the cross entropy loss function, x is the sample label value, is the predicted value of the model. w is all the parameters to be learned in the model. α represents the regularization strength and is a manually adjustable parameter used to control the degree of model fitting.

[0103] When training the model, the output prediction value is calculated from the input parameters, and the objective function value is obtained from the above formula. The objective function value is updated with the parameter value according to the selected optimization algorithm to complete a round of model training.

[0104] Optionally, the optimization algorithm is the Adam optimizer, which has the advantages of adaptive learning rate and high robustness. The model is evaluated based on the model prediction accuracy, and the model is further optimized through artificial intelligence deep learning.

[0105] The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

[0106] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A device for real-time detection of mineral powder grade, characterized in that: It includes an array coil, a signal detection circuit and a main control unit connected in sequence; one or more array coils are arranged above the mineral powder transmission device to be detected, the signal detection circuit is used to detect the change of coil voltage when the mineral powder passes through, and the main control unit identifies the grade of the mineral powder according to the voltage change; The array coil comprises a bottom plate, a plurality of active coils arranged in an array on the bottom plate, a magnetic core is arranged at the center of each active coil, and a shielding frame is arranged around each active coil to form a coil unit; each active coil is connected in series with a capacitor and then connected to a main control unit; the coil unit is a quadrilateral; and each coil unit has the same structure; Specifically: the capacitor is located on an external circuit connected to each active coil, the capacitor group corresponding to each active coil is arranged under the active coil, and the capacitors are connected in series with the active coils through wires; The time taken by the transmission device to move one unit side length is taken as a sampling period, and the voltage change of the detection coil is recorded every sampling period, and the changes of multiple sampling parameters are comprehensively considered; The main control unit includes a main controller, a signal detection part, a multi-way switch control part and a waveform generation part; The main controller, the signal detection part and the multi-way switch control part are connected in sequence, the multi-way switch control part is connected to the array coil through the signal detection circuit, and the main controller is connected to the array coil through the waveform generation part and the signal detection circuit; Signal detection part: configured to collect coil cell voltages one by one by controlling a multi-way switch; The main controller uses a regression tree algorithm to identify the grade of the ore powder based on the detected voltage change data. The regression tree model includes multiple regression trees, each of which corresponds to a division unit of the grade level. A regression tree corresponds to a division of the input space and the output value on the division unit.

2. A device for real-time detection of mineral powder grade according to claim 1, characterized in that: The waveform generating part is configured to convert square waves of different frequencies sent by the main controller into sine waves.

3. A device for real-time detection of mineral powder grade according to claim 1, characterized in that: The main control unit also includes a data storage part and a human-computer interaction part; the human-computer interaction part includes an audible and visual alarm and a display.

4. A device for real-time detection of mineral powder grade according to claim 1, characterized in that: The signal detection circuit includes a power supply, and a voltage difference measurement circuit, an overvoltage protection circuit, a voltage amplifier circuit, a rectifier and filter circuit, a peak detection circuit and an A / D conversion circuit which are connected in sequence, and the A / D conversion circuit is connected to the main controller.

5. A method for real-time detection of mineral powder grade, characterized in that: The steps include: One or more array coils are arranged above the mineral powder transmission device to be detected; The array coil comprises a bottom plate, a plurality of active coils arranged in an array on the bottom plate, a magnetic core is arranged at the center of each active coil, and a shielding frame is arranged around each active coil to form a coil unit; each active coil is connected in series with a capacitor and then connected to a main control unit; the coil unit is a quadrilateral; and each coil unit has the same structure; Specifically: the capacitor is located on an external circuit connected to each active coil, the capacitor group corresponding to each active coil is arranged under the active coil, and the capacitors are connected in series with the active coils through wires; The time taken by the transmission device to move one unit side length is taken as a sampling period, and the voltage change of the detection coil is recorded every sampling period, and the changes of multiple sampling parameters are comprehensively considered; Get the voltage value detected by the array coil when the mineral powder passes through the array coil; The main control unit includes a main controller, a signal detection part, a multi-way switch control part and a waveform generation part; The main controller, the signal detection part and the multi-way switch control part are connected in sequence, the multi-way switch control part is connected to the array coil through the signal detection circuit, and the main controller is connected to the array coil through the waveform generation part and the signal detection circuit; Signal detection part: configured to collect coil cell voltages one by one by controlling a multi-way switch; Based on the detected voltage change data, the regression tree algorithm is used to identify the grade of the mineral powder; The regression tree model includes multiple regression trees, each regression tree corresponds to a division unit of the grade level, and a regression tree corresponds to a division of the input space and an output value on the division unit.

6. A method for real-time detection of mineral powder grade according to claim 5, characterized in that: According to the detected voltage change data, the regression tree algorithm is used to identify the grade of the mineral powder. Specifically: Obtain the voltage value after the mineral powder to be detected passes through the active coil, extract the voltage amplitude and phase, and perform preprocessing; The preprocessed voltage value is input into the trained regression tree model to obtain the grade of the mineral powder to be tested.

7. A method for real-time detection of mineral powder grade according to claim 5, characterized in that: The training process of the regression tree model includes the following steps: Voltage amplitude, phase change and ore grade data set; Divide the data of the dataset into training set and test set; Data preprocessing, including normalization, sample label setting and material definition; Taking the changes in voltage amplitude and phase as input and the corresponding ore grade as output, the constructed regression tree model is trained to obtain the parameters of the regression tree model, which is then tested using the data from the test set.

Images