Methods for constructing and identifying formation identification models for electric drive heads used in drilling rigs.

Abstract

This invention discloses an electric drive head for drilling rigs, a method for constructing and identifying formation identification models, belonging to the field of coal mine tunnel drilling machinery. It includes a mounting base with a reducer mounted on it. A water supply device is located at the axial front end of the reducer, and a permanent magnet synchronous motor is located at the axial rear end. The permanent magnet synchronous motor includes a housing with a front cover and a rear cover at both ends. The permanent magnet synchronous motor body is located inside the housing, and its output shaft extends out of the front cover and is fixedly connected to the axial rear end of the reducer. The use of a permanent magnet synchronous motor direct-drive reducer replaces the hydraulic pump station, motor, and hydraulic components, reducing hydraulic oil usage and making it more environmentally friendly. The permanent magnet synchronous motor direct-drive reducer has high transmission efficiency and is more energy-efficient. It also facilitates precise control of speed and torque, as well as intelligent drilling, solving the technical problem of low efficiency in traditional hydraulic drilling rig rotary power heads in the prior art.

Description

Technical Field

[0001] This invention belongs to the field of coal mine tunnel drilling machinery, specifically a drilling rig electric drive head, a method for constructing a formation identification model, and an identification method. Background Technology

[0002] Traditional hydraulic drilling rigs use hydraulically driven rotary power heads, requiring a complex system including pump stations, valve blocks, and oil tanks. This system is characterized by low transmission efficiency, high energy consumption, large hydraulic oil usage, and high operating costs. Safe, efficient, and green mining is essential for the sustainable development of the coal industry, while digitalization, intelligentization, and greening are the development directions for coal mining. Electro-drive is the best solution for reducing hydraulic oil and component usage and improving transmission efficiency. Achieving electro-drive for the drilling rig power head is key to realizing the overall electrification of drilling rigs.

[0003] Therefore, there is an urgent need for a front-mounted water-driven electric power head for drilling rigs. The electric drilling rig power head uses a permanent magnet synchronous motor as the power source and a planetary reducer to enhance torque output. The entire system simplifies the energy conversion process, greatly improves transmission efficiency, and has a simple structure that is easy to implement intelligent control. Compared with hydraulic drilling rigs, it does not require frequent maintenance and reduces the use of oil. The motor can adaptively change power according to the load. Therefore, the electric drive power head has significant advantages in terms of high efficiency, convenience, reliability, multifunctionality, controllability, and economy. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method for constructing and identifying a formation identification model for an electric drive head for drilling rigs, so as to solve the technical problem of low efficiency of traditional hydraulic drilling rig rotary power heads in the prior art.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] An electric drive head for a drilling rig includes a mounting base, on which a reducer is mounted. A water feeder is mounted at the axial front end of the reducer, and a permanent magnet synchronous motor is mounted at the axial rear end of the reducer.

[0007] The permanent magnet synchronous motor includes a housing, with a front cover and a rear cover at both ends of the housing; the permanent magnet synchronous motor body is disposed inside the housing, and the output shaft of the permanent magnet synchronous motor body extends out of the front cover and is fixedly connected to the axial rear end of the reducer.

[0008] The permanent magnet synchronous motor is equipped with a water distributor; a first water inlet pipe and a first water outlet pipe are provided between the water distributor and the front end cover of the permanent magnet synchronous motor; a second water inlet pipe and a second water outlet pipe are provided between the water distributor and the rear end cover of the permanent magnet synchronous motor; and a third water inlet pipe and a third water outlet pipe are provided between the water distributor and the housing of the permanent magnet synchronous motor.

[0009] It also includes a controller electrically connected to the permanent magnet synchronous motor, the controller being used to identify the formation type and adjust the speed of the permanent magnet synchronous motor according to the formation type.

[0010] This invention also includes the following technical features:

[0011] The permanent magnet synchronous motor and the reducer, as well as the reducer and the water supply device, are connected by uniform splines.

[0012] Furthermore, a method for constructing a formation identification model, executed in the controller of the electric drive head of the drilling rig, includes the following steps:

[0013] Step 1: Real-time acquisition of the three-phase stationary coordinate system current signal of the pool synchronous motor. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and A;

[0014]

[0015] Step two, use the following formula to process the two-phase stationary coordinate system current signals obtained in step two. and Normalization and preprocessing are performed to obtain the normalized signal. and A;

[0016]

[0017]

[0018] in:

[0019] , These are the current signals in the two-phase stationary coordinate system, respectively. and The mean, A;

[0020] , These are the current signals in the two-phase stationary coordinate system, respectively. and The standard deviation, A;

[0021] Step 3, normalize the signal and Perform a short-time Fourier transform and normalize the signal. and The superposition of the spectra yields a two-dimensional feature matrix S;

[0022] Step 4: Construct a stratigraphic identification model;

[0023] The stratum identification model comprises an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer connected in sequence.

[0024] The input layer is used to input the two-dimensional feature matrix S;

[0025] The convolutional layer is used to extract the local spatial relationships of the two-dimensional feature matrix S through convolution operations, capture important time-frequency features, and output a feature map;

[0026] The pooling layer is used to downsample the feature map to obtain a smaller feature map;

[0027] The fully connected layer is used to expand the smaller feature map output by the pooling layer to obtain multiple one-dimensional vectors. The multiple one-dimensional vectors are integrated to output a one-dimensional vector containing the original prediction scores of each stratum label.

[0028] The output layer is used to pass... The function transforms the one-dimensional vector containing the original predicted scores of each stratigraphic label from the output of the fully connected layer into a probability distribution, thereby obtaining the final predicted probability of the current input two-dimensional feature matrix S belonging to each stratigraphic label. Output the final predicted probability The formation label corresponding to the maximum value;

[0029]

[0030] in:

[0031] This is the original prediction score of the stratigraphic identification model for the current input two-dimensional feature matrix S belonging to the k-th stratigraphic label;

[0032] To Perform exponential operations to convert the corresponding original predicted scores into positive numbers;

[0033] This is the original prediction score of the stratigraphic identification model for the current input two-dimensional feature matrix S belonging to the j-th stratigraphic label;

[0034] To Perform exponential operations to convert the corresponding original predicted scores into positive numbers;

[0035] This represents the total number of stratigraphic label categories;

[0036] For the first Predicted probability of stratigraphic labels;

[0037] Step 5: Using the two-dimensional feature matrix S obtained in Step 3 as input and the corresponding stratigraphic labels as output, the stratigraphic identification model constructed in Step 4 is trained using the cross-entropy loss function to obtain the trained stratigraphic identification model.

[0038] A formation identification method, executed in the controller of the electric drive head of the drilling rig, specifically includes the following steps based on the construction method of the formation identification model:

[0039] Step 1: Real-time acquisition of the three-phase stationary coordinate system current signal of the pool synchronous motor. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and A;

[0040]

[0041] Step 2: Apply the following formula to the two-phase stationary coordinate system current signal obtained in Step 2. and Normalization and preprocessing are performed to obtain the normalized signal. and A;

[0042]

[0043]

[0044] in:

[0045] , These are the current signals in the two-phase stationary coordinate system, respectively. and The mean, A;

[0046] , These are the current signals in the two-phase stationary coordinate system, respectively. and The standard deviation, A;

[0047] Step 3, normalize the signal and Perform a short-time Fourier transform and normalize the signal. and The superposition of the spectra yields a two-dimensional feature matrix S;

[0048] Step 4: Obtain the two-dimensional feature matrix from Step 3. The trained stratigraphic identification model, obtained by inputting into the stratigraphic identification model construction method, outputs the corresponding stratigraphic label;

[0049] Step 5: Based on the formation tags identified by the formation identification model in Step 4, determine the rotational speed of the permanent magnet synchronous motor. Adjustments will be made, with the adjustment amount being... .

[0050] It also includes an adaptive control system for calculating the speed adjustment of the permanent magnet synchronous motor. Its characteristic is that it includes an input layer, a hidden layer and an output layer arranged sequentially;

[0051] The input layer is used to input input variables, which include three-phase stationary coordinate system current signals and formation labels;

[0052] The hidden layer uses a nonlinear activation function. The output variable is obtained from the input variable;

[0053] The output layer is used to output output variables, including the speed adjustment amount of the permanent magnet synchronous motor. .

[0054] Compared with the prior art, the beneficial technical effects of this invention are:

[0055] (I) In this invention, a permanent magnet synchronous motor direct drive reducer is used to replace the hydraulic pump station, motor and hydraulic components, which reduces the use of hydraulic oil and is more environmentally friendly; the permanent magnet synchronous motor direct drive reducer has high transmission efficiency and is more energy-efficient; it is easier to achieve precise control of speed and torque and intelligent drilling, which solves the technical problem of low efficiency of the traditional hydraulic drilling rig's rotary power head in the prior art.

[0056] (II) The changes in the current signal of the permanent magnet synchronous motor in this invention accurately sense the stratum conditions at the bottom of the hole during the drilling process, and at the same time autonomously implement corresponding strategies: when encountering soft coal seams, increase the rotation speed to improve drilling efficiency; when encountering hard coal seams, reduce the rotation speed to prevent damage to the drill bit and drill rod; and when encountering hard rock impact, reduce the rotation speed and retreat to protect the drilling tools.

[0057] (III) The adaptive controller in this invention can autonomously adjust the control strategy based on real-time data, adapt to different geological conditions, has strong nonlinear mapping capability, can handle complex nonlinear relationships, and improve control accuracy. The controller has self-learning capability and can continuously optimize during operation. Attached Figure Description

[0058] Figure 1 This is a three-dimensional structural schematic diagram of the electrically driven force head of the water delivery device for drilling rigs in this invention;

[0059] Figure 2 This is a three-dimensional structural diagram of the electrically driven force head of the water delivery device for drilling rigs in this invention from another angle.

[0060] Figure 3 This is a two-dimensional structural schematic diagram of the electrically driven force head of the water delivery device for drilling rigs in this invention;

[0061] Figure 4 This is a schematic diagram of the control principle in this invention;

[0062] Figure 5 This is a flowchart of the stratigraphic identification method in this invention;

[0063] Figure 6(a) shows the loss function curves for training and validating the neural network controller model;

[0064] Figure 6(b) shows the difference in the controller's response between the neural network and the traditional PID controller.

[0065] The meanings of the labels in the diagram are as follows: 1-mounting base, 2-reducer, 3-water feeder, 4-permanent magnet synchronous motor, 5-water distributor, 6-first inlet pipe, 7-first outlet pipe, 8-second inlet pipe, 9-second outlet pipe, 10-third inlet pipe, 11-third outlet pipe;

[0066] 401 - Housing, 402 - Front cover, 403 - Rear cover.

[0067] The specific content of the present invention will be further explained in detail below with reference to the embodiments. Detailed Implementation

[0068] It should be noted that, unless otherwise specified, all components in this invention are those known in the art.

[0069] The following are specific embodiments of the present invention. It should be noted that the present invention is not limited to the following specific embodiments. All equivalent modifications made based on the technical solutions of this application fall within the protection scope of the present invention.

[0070] The present invention provides an electric drive head for drilling rigs, including a mounting base 1, a reducer 2 is provided on the mounting base 1, a water feeder 3 is provided at the axial front end of the reducer 2, and a permanent magnet synchronous motor 4 is provided at the axial rear end of the reducer 2.

[0071] The permanent magnet synchronous motor 4 includes a housing 401, with a front cover 402 and a rear cover 403 at both ends of the housing 401; the permanent magnet synchronous motor body is disposed inside the housing 401, and the output shaft of the permanent magnet synchronous motor body extends out of the front cover 402 and is fixedly connected to the axial rear end of the reducer 2.

[0072] A water distributor 5 is provided on the permanent magnet synchronous motor 4; a first water inlet pipe 6 and a first water outlet pipe 7 are provided between the water distributor and the front end cover 402 of the permanent magnet synchronous motor 4; a second water inlet pipe 8 and a second water outlet pipe 9 are provided between the water distributor and the rear end cover 403 of the permanent magnet synchronous motor 4; and a third water inlet pipe 10 and a third water outlet pipe 11 are provided between the water distributor and the housing 401 of the permanent magnet synchronous motor 4.

[0073] It also includes a controller, which is electrically connected to the permanent magnet synchronous motor 4. The controller is used to identify the formation type and adjust the speed of the permanent magnet synchronous motor according to the formation type.

[0074] In the above technical solution, a permanent magnet synchronous motor direct drive reducer is used to replace the hydraulic pump station, motor and hydraulic components, reducing the use of hydraulic oil and making it more environmentally friendly; the permanent magnet synchronous motor direct drive reducer has high transmission efficiency and is more energy-efficient; it is easier to achieve precise control of speed and torque and intelligent drilling, solving the technical problem of low efficiency of the traditional hydraulic drilling rig's rotary power head in the existing technology.

[0075] A permanent magnet synchronous motor is directly connected to a planetary reducer, with water placed at the front end of the reducer. The motor torque is amplified by the reducer to achieve high torque output. The water distributor 5, along with the first inlet pipe 6, first outlet pipe 7, second inlet pipe 8, second outlet pipe 9, third inlet pipe 10, and third outlet pipe 11, collectively cool the permanent magnet synchronous motor 4, ensuring effective heat dissipation. The controller detects the current of the permanent magnet synchronous motor to determine the stratum information within the borehole and outputs corresponding strategies to control the motor's speed, ensuring smooth drilling operations.

[0076] The permanent magnet synchronous motor 4 and the reducer 2, as well as the reducer 2 and the water supply device 3, are connected by a uniform spline.

[0077] In the above technical solutions, spline connections facilitate installation and disassembly.

[0078] A method for constructing a formation identification model, executed in the controller of an electric drive head for drilling rigs, includes the following steps:

[0079] Step 1: Real-time acquisition of the three-phase stationary coordinate system current signal of the pool synchronous motor. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and A;

[0080]

[0081] Step two, use the following formula to process the two-phase stationary coordinate system current signals obtained in step two. and Normalization and preprocessing are performed to obtain the normalized signal. and A;

[0082]

[0083]

[0084] in:

[0085] , These are the current signals in the two-phase stationary coordinate system, respectively. and The mean, A;

[0086] , These are the current signals in the two-phase stationary coordinate system, respectively. and The standard deviation, A;

[0087] Step 3, normalize the signal and Perform a short-time Fourier transform and normalize the signal. and The superposition of the spectra yields a two-dimensional feature matrix S;

[0088] τ is the time variable in the short-time Fourier transform integral, representing the instantaneous time point corresponding to the sliding of the window function on the time axis over the entire signal duration;

[0089] Step 4: Construct a stratigraphic identification model;

[0090] The stratigraphic identification model consists of an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer connected in sequence.

[0091] The input layer is used to input the two-dimensional feature matrix S;

[0092] Convolutional layers are used to extract the local spatial relationships of a two-dimensional feature matrix S through convolution operations, capture important time-frequency features, and output feature maps;

[0093] Pooling layers are used to downsample feature maps, resulting in smaller feature maps;

[0094] The fully connected layer is used to expand the smaller feature map output by the pooling layer to obtain multiple one-dimensional vectors. The multiple one-dimensional vectors are integrated to output a one-dimensional vector containing the original prediction scores of each stratum label.

[0095] The output layer is used to pass through The function transforms the one-dimensional vector containing the original predicted scores of each stratigraphic label from the output of the fully connected layer into a probability distribution, thereby obtaining the final predicted probability of the current input two-dimensional feature matrix S belonging to each stratigraphic label. Output the final predicted probability The formation label corresponding to the maximum value;

[0096]

[0097] in:

[0098] This is the original prediction score of the stratigraphic identification model for the current input two-dimensional feature matrix S belonging to the k-th stratigraphic label;

[0099] To Perform exponential operations to convert the corresponding original predicted scores into positive numbers;

[0100] This is the original prediction score of the stratigraphic identification model for the current input two-dimensional feature matrix S belonging to the j-th stratigraphic label;

[0101] To Perform exponential operations to convert the corresponding original predicted scores into positive numbers;

[0102] This represents the total number of stratigraphic label categories;

[0103] For the first Predicted probability of stratigraphic labels;

[0104] Step 5: Using the two-dimensional feature matrix S obtained in Step 3 as input and the corresponding stratigraphic labels as output, the stratigraphic identification model constructed in Step 4 is trained using the cross-entropy loss function to obtain the trained stratigraphic identification model.

[0105] In the above technical solution, the cross-entropy loss function is used to measure the difference between the model prediction results and the actual stratum labels, and gradient descent algorithms such as the Adam optimizer are used to iteratively update the internal parameters of the model, such as weights and biases, to gradually minimize the above differences until the model performance converges and stabilizes.

[0106] Generally, the size of the two-dimensional feature matrix S is chosen to be... To adapt to the input requirements of CNN.

[0107] Pooling layers can reduce data dimensionality, prevent overfitting, and retain significant features.

[0108] This invention provides a formation identification method, executed in the controller of the electric drive head of a drilling rig, and a method for constructing a formation identification model, specifically including the following steps:

[0109] Step 1: Real-time acquisition of the three-phase stationary coordinate system current signal of the pool synchronous motor. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and A;

[0110]

[0111] Step 2: Apply the following formula to the two-phase stationary coordinate system current signal obtained in Step 2. and Normalization and preprocessing are performed to obtain the normalized signal. and A;

[0112]

[0113]

[0114] in:

[0115] , These are the current signals in the two-phase stationary coordinate system, respectively. and The mean, A;

[0116] , These are the current signals in the two-phase stationary coordinate system, respectively. and The standard deviation, A;

[0117] Step 3, normalize the signal and Perform a short-time Fourier transform and normalize the signal. and The superposition of the spectra yields a two-dimensional feature matrix S;

[0118] Step 4: Obtain the three-dimensional feature matrix from Step 3. The trained stratigraphic identification model, which is input into the stratigraphic identification model construction method, outputs the corresponding stratigraphic labels.

[0119] Step 5: Based on the formation tags identified by the formation identification model in Step 4, determine the rotational speed of the permanent magnet synchronous motor. Adjustments will be made, with the adjustment amount being... .

[0120] It also includes an adaptive control system for calculating the speed adjustment of the permanent magnet synchronous motor. Its characteristic is that it includes an input layer, a hidden layer and an output layer arranged sequentially;

[0121] The input layer is used to input input variables, which include three-phase stationary coordinate system current signals and formation labels;

[0122] The hidden layer uses a non-linear activation function. The output variable is obtained from the input variable;

[0123] The output layer is used to output output variables, including the speed adjustment of the permanent magnet synchronous motor. .

[0124] In the above technical solution, the adaptive controller can autonomously adjust the control strategy based on real-time data, adapt to different geological conditions, has strong nonlinear mapping capabilities, can handle complex nonlinear relationships, and improve control accuracy. The controller has self-learning capabilities and can continuously optimize during operation.

[0125] Example:

[0126] This embodiment presents a formation identification method that acquires the three-phase stationary coordinate system current signal of a swimming pool synchronous motor in real time. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and For two-phase stationary coordinate system current signals and Normalization and preprocessing are performed to obtain the normalized signal. and ;

[0127] :[0.21,0.23,0.25,…,0.22] (Length 1024 points, sampling frequency 1024Hz, coverage 1s);

[0128] :[0.18,0.20,0.22,…,0.19] (length 1024 points, and...) (Synchronous acquisition);

[0129] Then, using a window of length 128, the normalized signal is truncated by sliding with a 50% overlap rate. and Then perform Fourier transforms on each window segment, and finally take the amplitude value of the Fourier transform result to obtain the result. and and Short-time Fourier transform amplitude value and ,Will and The two-dimensional feature matrix S is obtained by superimposing time and frequency.

[0130] For example: The 10th window, frequency point 15 (corresponding to) The amplitude of the short-time Fourier transform of is 0.68. If the amplitude of the short-time Fourier transform at the same position is 0.61, then S is... The resulting two-dimensional feature matrix S is as follows:

[0131]

[0132] The above two-dimensional feature matrix The trained stratigraphic identification model, which is input into the stratigraphic identification model construction method, outputs the corresponding stratigraphic label as soft coal seam.

[0133] The stratigraphic labels obtained from the stratigraphic identification model affect the rotational speed of the permanent magnet synchronous motor. Make adjustments;

[0134] The current is currently 30A, and the actual rotation speed is 1200 rpm, which is below the target lower limit for the flexible media layer. Based on the flexible media layer adjustment strategy, the adjustment amount... After adjustment, the motor current increased and was re-acquired at 33A, which is in line with the target current range for soft coal seams.

[0135] Verification example:

[0136] To verify the actual performance of the adaptive controller, simulation modeling and experiments were conducted, with the following operating conditions set:

[0137] Initial speed The controller response was tested at 1500 rpm under three typical geological conditions: soft coal seam, hard coal seam, and hard rock impact. The simulation results show that:

[0138] The control system achieves a steady-state response within 0.5 seconds, with an error convergence time less than half that of a traditional PID controller; the system overshoot is less than 5%, exhibiting no oscillation and good robustness; under different geological conditions, the attention mechanism enables the control strategy to demonstrate significant adaptability: improving performance in soft coal seams. The advance rate was automatically limited in hard coal seams; the control error converged smoothly, and the dynamic error approached zero, verifying the effectiveness of real-time error feedback in adjusting network parameters.

[0139] To train the neural network controller model, a total of 3000 sets of historical rotational speed, propulsion speed, current, and formation label data under different formation conditions were collected as the training set. The training settings are as follows:

[0140] Network structure: 3 input nodes, 2 hidden layers (16 neurons each), 2 output nodes. Activation function: Optimizer: Adam; Learning rate: 0.001; Training epochs: 200; Batch size: 64. Training results are as follows:

[0141] It exhibits fast convergence speed, with the MSE of the loss function decreasing to below 0.002 within 100 epochs. The final training accuracy is 98.2%, and the validation accuracy is 96.7%.

[0142] It performed best in hard rock impact samples. The average error is less than ±3%, and the average error between the control response time and the actual set value is within ±2%. Figure 6(a) shows that the MSE decreases rapidly and tends to converge over 200 epochs, verifying the excellent training effect of the model; Figure 6(b) shows the difference in error response between the traditional PID and neural network controllers, with the neural network control response being more stable and converging faster.

Claims

1. A method for constructing a formation identification model, executed in the controller of an electric drive head for a drilling rig, characterized in that, The electric drive head for the drilling rig includes a mounting base (1), a reducer (2) is provided on the mounting base (1), a water feeder (3) is provided at the axial front end of the reducer (2), and a permanent magnet synchronous motor (4) is provided at the axial rear end of the reducer (2). The permanent magnet synchronous motor (4) includes a housing (401), with a front end cover (402) and a rear end cover (403) at both ends of the housing (401); the permanent magnet synchronous motor body is provided inside the housing (401), and the output shaft of the permanent magnet synchronous motor body extends out of the front end cover (402) and is fixedly connected to the axial rear end of the reducer (2); A water distributor (5) is provided on the permanent magnet synchronous motor (4); a first water inlet pipe (6) and a first water outlet pipe (7) are provided between the water distributor and the front end cover (402) of the permanent magnet synchronous motor (4); a second water inlet pipe (8) and a second water outlet pipe (9) are provided between the water distributor and the rear end cover (403) of the permanent magnet synchronous motor (4); and a third water inlet pipe (10) and a third water outlet pipe (11) are provided between the water distributor and the housing (401) of the permanent magnet synchronous motor (4). It also includes a controller, which is electrically connected to the permanent magnet synchronous motor (4), and the controller is used to identify the formation type and adjust the speed of the permanent magnet synchronous motor according to the formation type; The method includes the following steps: Step 1: Real-time acquisition of the three-phase stationary coordinate system current signal of the permanent magnet synchronous motor. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and A; Step two, use the following formula to process the two-phase stationary coordinate system current signals obtained in step two. and Normalization and preprocessing are performed to obtain the normalized signal. and A; in: , These are the current signals in the two-phase stationary coordinate system, respectively. and The mean, A; , These are the current signals in the two-phase stationary coordinate system, respectively. and The standard deviation, A; Step 3, normalize the signal and Perform a short-time Fourier transform and normalize the signal. and The superposition of the spectra yields a two-dimensional feature matrix S; Step 4: Construct a stratigraphic identification model; The stratum identification model comprises an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer connected in sequence. The input layer is used to input the two-dimensional feature matrix S; The convolutional layer is used to extract the local spatial relationships of the two-dimensional feature matrix S through convolution operations, capture important time-frequency features, and output a feature map; The pooling layer is used to downsample the feature map to obtain a smaller feature map; The fully connected layer is used to expand the smaller feature map output by the pooling layer to obtain multiple one-dimensional vectors. The multiple one-dimensional vectors are integrated to output a one-dimensional vector containing the original prediction scores of each stratum label. The output layer is used to pass... The function transforms the one-dimensional vector containing the original predicted scores of each stratigraphic label from the output of the fully connected layer into a probability distribution, thereby obtaining the final predicted probability of the current input two-dimensional feature matrix S belonging to each stratigraphic label. Output the final predicted probability The formation label corresponding to the maximum value; in: This is the original prediction score of the stratigraphic identification model for the current input two-dimensional feature matrix S belonging to the k-th stratigraphic label; To Perform exponential operations to convert the corresponding original predicted scores into positive numbers; This is the original prediction score of the stratigraphic identification model for the current input two-dimensional feature matrix S belonging to the j-th stratigraphic label; To Perform exponential operations to convert the corresponding original predicted scores into positive numbers; This represents the total number of stratigraphic label categories; For the first Predicted probability of stratigraphic labels; Step 5: Using the two-dimensional feature matrix S obtained in Step 3 as input and the corresponding stratigraphic labels as output, the stratigraphic identification model constructed in Step 4 is trained using the cross-entropy loss function to obtain the trained stratigraphic identification model.

2. The method for constructing the stratigraphic identification model as described in claim 1, characterized in that, The permanent magnet synchronous motor (4) and the reducer (2) are connected by splines, as are the reducer (2) and the water supply device (3).

3. A stratigraphic identification method, based on the method for constructing the stratigraphic identification model according to claim 1, characterized in that, Specifically, the following steps are included: Step 1: Real-time acquisition of the three-phase stationary coordinate system current signal of the permanent magnet synchronous motor. , and and adopt The transformation converts a three-phase stationary coordinate system current signal into a two-phase stationary coordinate system current signal. and A; Step two, use the following formula to process the two-phase stationary coordinate system current signals obtained in step two. and Normalization and preprocessing are performed to obtain the normalized signal. and A; in: , These are the current signals in the two-phase stationary coordinate system, respectively. and The mean, A; , These are the current signals in the two-phase stationary coordinate system, respectively. and The standard deviation, A; Step 3, normalize the signal and Perform a short-time Fourier transform and normalize the signal. and The superposition of the spectra yields a two-dimensional feature matrix S; Step four, use the two-dimensional feature matrix obtained in step three. The trained stratigraphic identification model, which is input into the stratigraphic identification model construction method, outputs the corresponding stratigraphic label. Step 5: Based on the formation tags identified by the formation identification model in Step 4, determine the rotational speed of the permanent magnet synchronous motor. Adjustments will be made, with the adjustment amount being... .

4. The formation identification method as described in claim 3 further includes an adaptive control system for calculating the speed adjustment amount of the permanent magnet synchronous motor. Its characteristics are, This includes an input layer, a hidden layer, and an output layer, arranged sequentially. The input layer is used to input input variables, which include three-phase stationary coordinate system current signals and formation labels. The hidden layer uses a nonlinear activation function. The output variable is obtained from the input variable; The output layer is used to output output variables, including the speed adjustment amount of the permanent magnet synchronous motor. .

Images