An electric rotary drilling rig intelligent drilling method based on electric-mechanical multi-physical field perception and weighted voting decision

Abstract

The application discloses an electric rotary drilling rig intelligent drilling method based on electric-power multi-physical field perception and weighted voting decision, and belongs to the field of intelligent control of engineering machinery and artificial intelligence technology. The application comprises the following steps: a data acquisition module listens to an electric rotary drilling rig bus in a read-only mode and is additionally provided with vibration and acoustic sensors, and acquires three-phase current of a pressurized motor, three-phase current of a power head motor, a push rod position, a drilling depth and optional force sensor signal in real time, wherein the power head is driven by a permanent magnet synchronous motor through a speed reduction mechanism; a stratum feature calculation module calculates electric drive specific work, current harmonic distortion rate, vibration and acoustic features and other characteristic quantities with physical definitions based on the above signals; a stratum fusion decision module outputs a current depth stratum type and confidence through a weighted voting decision strategy; an adaptive drilling controller generates target current and speed instructions of the motor according to the stratum type through a feed-forward-feedback compound control strategy; the application also has a boulder precursor emergency protection capability based on current sudden change detection and an online optimization function of specific energy consumption. The application constructs a direct force perception model by using motor current signals of the electric rotary drilling rig, and combines electric-power physical field fusion decision, and has good engineering practicability and intelligent upgrading value.

Description

Technical Field

[0001] This invention relates to an intelligent drilling method for electric rotary drilling rigs based on electro-mechanical multi-physics field perception and weighted voting decision-making, belonging to the field of intelligent control of engineering machinery and artificial intelligence technology. Background Technology

[0002] Rotary drilling rigs are core equipment widely used in pile foundation construction. In recent years, with the advancement of electrification in construction machinery, electric rotary drilling rigs, which use permanent magnet synchronous motors driven by reduction gears and pressure applied by electric push rods, have gradually been adopted in the industry. Compared with traditional hydraulic drive systems, electric rotary drilling rigs have certain technical advantages in response speed, control accuracy, and signal sampling quality.

[0003] In the field of stratigraphic identification, relevant technical solutions have been researched and explored. For example, Chinese patent application CN121296109A discloses a method for determining stratigraphic information based on interpolation of geological exploration information from adjacent boreholes, which can output stratigraphic prediction results before construction. Chinese patent application CN118653822A discloses a stratigraphic identification method based on the pressure characteristics of rotary drilling rigs. This method constructs a training dataset by collecting historical drilling data and establishes a mapping relationship between pressure characteristics and stratigraphic categories. Chinese patent application CN120968420A discloses an electric drive head for drilling rigs used in coal mine tunnel drilling and a method for constructing a stratigraphic identification model.

[0004] The aforementioned technical solutions have room for improvement in practical engineering applications. Methods based on interpolation of geological exploration information struggle to achieve real-time detection of local geological anomalies within the borehole during drilling. Data-driven methods are heavily reliant on historical annotation data, requiring data re-collection and model training for application at new sites. Some solutions still limit their signal sources to traditional hydraulic parameters. Furthermore, existing solutions do not utilize the inherent motor current signal of the electric rotary drilling rig as an independent physical quantity for formation identification, nor do they fully leverage the rapid response characteristics of the electric drive system for emergency protection against abnormal conditions such as isolated rocks and stuck drill bits. This invention aims to provide an improved intelligent drilling method that integrates electro-mechanical multi-physics fields to further enhance the real-time performance and accuracy of formation identification. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] This invention aims to provide an intelligent drilling method for electric rotary drilling rigs based on electro-physical multi-physics sensing and weighted voting decision-making, in order to solve or improve at least one of the following technical problems:

[0007] First, how to utilize the inherent motor current signal of the electric rotary drilling rig to construct a feed force estimation method that directly reflects the interaction between the drill bit and the formation, replacing or supplementing the traditional hydraulic pressure signal, so as to reduce the dependence on historical data training;

[0008] Second, how to construct a dual-source mutual verification system by combining the motor current estimation force with the direct measurement value of the force sensor, so as to realize the self-diagnosis of sensor faults and system fault tolerance;

[0009] Third, how to improve the accuracy of formation identification through a weighted voting decision-making strategy based on the comprehensive perception of multi-physics field signals, and how to utilize the rapid response characteristics of the electric drive system to achieve emergency protection against isolated rocks or stuck drill bits.

[0010] (II) Technical Solution

[0011] To achieve the above objectives, the present invention adopts the following technical solution:

[0012] A smart drilling method for electric rotary drilling rigs based on electro-mechanical multi-physics sensing and weighted voting decision-making includes the following steps:

[0013] Step S1: During the drilling process of the electric rotary drilling rig, multi-physics field signals are synchronously acquired through the data acquisition module. The multi-physics field signals include the three-phase current of the pressurized permanent magnet synchronous motor, the three-phase current of the power head permanent magnet synchronous motor, the position sensor signal built into the electric push rod, the drilling depth sensor signal, and at least one of the following sensor signals: the three-axis acceleration sensor signal installed in the power head housing, the microphone acoustic signal installed near the power head, and the force sensor signal installed at the connection between the electric push rod and the drill rod; wherein, the power head is driven by the permanent magnet synchronous motor through the reduction mechanism.

[0014] Step S2: Based on the signals acquired in Step S1, at least two physically defined feature quantities are calculated in real time using the formation feature calculation module. These feature quantities include at least two selected from the following set: electromagnetic thrust estimated based on the pressurized motor current, electromagnetic torque estimated based on the power head motor current, electric drive specific work, current harmonic distortion rate, triaxial acceleration time-frequency characteristics, and acoustic Mel frequency cepstral coefficients. The calculation of these feature quantities uses physically deterministic algorithms and does not require training based on historical drilling data.

[0015] Step S3: Input the feature quantities calculated in Step S2 into the formation fusion decision module, and output the formation type and confidence level corresponding to the current drilling depth through a weighted voting decision strategy. The weights of the weighted voting decision strategy are configured as adjustable parameters, which can be adaptively adjusted according to different sensor signal quality.

[0016] Step S4: Input the formation type output in step S3 into the adaptive drilling controller. The controller generates the target current command for the pressurized permanent magnet synchronous motor and the target speed and limiting current command for the power head permanent magnet synchronous motor through a feedforward-feedback composite control strategy, and sends them to the actuator through the motor driver.

[0017] The key technical aspects of the above steps will be explained in detail below.

[0018] 1. Estimation of electric drive thrust

[0019] In this step, the electromagnetic torque of the pressurized permanent magnet synchronous motor can be calculated using the motor's electrical parameters:

[0020] T_m = k_t × I_q × k_temp

[0021] Where: k_t is the motor torque constant, in N·m / A; I_q is the q-axis torque current, in A, calculated from the three-phase current of the pressurized motor by Parker transformation; k_temp is the temperature compensation coefficient, dimensionless, and its expression is k_temp = 1- α_T × (T - T_0), where α_T ranges from 0.0010 to 0.0012 degrees Celsius.

[0022] The relationship between electromagnetic thrust, motor electromagnetic torque, planetary roller screw lead, and transmission efficiency can be expressed as: F_current = (2π × T_m × η_total) / Ph - F_fric

[0023] It should be noted that the efficiency coefficient η_total is not a fixed constant; it varies with load. Typical values ​​for engineering applications can be found in the following ranges: 0.88 to 0.92 for rated load and above, 0.80 to 0.85 for medium load, and 0.70 to 0.75 for light load or no-load conditions. In practical applications, a comprehensive efficiency coefficient mapping table matching the specific model can be obtained through pre-calibration tests.

[0024] With a force sensor configured, the system simultaneously acquires the direct force sensor measurement value F_sensor and the current estimation value F_current, forming a dual-source redundant force sensing system. The cross-validation index between the two is defined as:

[0025] C_force = 1 - |F_current - F_sensor| / (F_current + F_sensor + ε)

[0026] Where ε is a very small positive real number to avoid the denominator being zero. When C_force is greater than the preset threshold (0.9 can be taken as an example value), it is determined that the dual-source sensing results tend to be consistent, and the actual feed force is taken as the weighted average of F_current and F_sensor; when C_force is continuously lower than the preset threshold (0.7 can be taken as an example value) and the duration exceeds the set window (10 seconds can be taken as an example value), it is determined that there is an abnormality in the current transformer or force sensor, and the system automatically switches to the single signal source working mode and issues a fault prompt.

[0027] 2. Torque estimation of the power head

[0028] The electromagnetic torque of the power head motor can be calculated using the motor's electrical parameters:

[0029] T_elec_motor = k_t_head × I_q_head × k_temp_head

[0030] Where Telec_motor is the electromagnetic torque at the motor end, I_q_head is the q-axis torque current of the power head motor, and k_t_head is the torque constant of the power head motor.

[0031] The relationship between the electromagnetic torque at the motor end and the torque at the output end of the power head can be expressed as:

[0032] T_output = T_elec_motor × i × η_gear

[0033] Where i is the transmission ratio of the reduction mechanism, and η_gear is the overall efficiency of the reduction mechanism and its gear transmission system, which is dimensionless and typically ranges from 0.90 to 0.96.

[0034] In practical implementation, the elastic deformation ε of the housing can also be obtained by a strain sensor installed on the housing of the reducer. Another torque estimate can be obtained by the pre-calibrated torque-strain function relationship T_strain = f(ε). The two signals are then fused (for example, by weighted averaging or Kalman filtering) to output the fused torque, so as to reduce the cumulative error that may be generated by a single calculation path.

[0035] 3. Calculation of specific work of electric drive

[0036] The actual feed force F_act is determined using the aforementioned electric drive thrust estimation model or force sensor measurements. The instantaneous drilling speed V is calculated by differentiating the borehole depth D with respect to time t, i.e., V = dD / dt. The borehole cross-sectional area A is determined by the drill bit diameter D_bit, i.e., A = π × (D_bit / 2)^2.

[0037] The specific work of an electric drive is calculated using the following formula:

[0038] e_drive = F_act / A + [2π × (n / 60) × T_output] / (A × V)

[0039] Where n is the measured rotational speed of the power head in revolutions per minute (r / min), (n / 60) is the revolutions per second, and 2π×(n / 60) is the angular velocity in radians per second. This formula uses the International System of Units (SI) to ensure dimensional consistency. The normalized specific work e_corr is determined by the ratio of the electric drive specific work to the drill bit correction factor k_tool, i.e., e_corr = e_drive / k_tool. The drill bit correction factor k_tool is a dimensionless calibration value, which can be obtained through pre-calibration tests and set according to the different drill bit types.

[0040] 4. Current harmonic distortion rate

[0041] The harmonic distortion rate (THD) is defined as the ratio of the square root of the sum of the squares of the effective values ​​of all harmonic currents to the effective value of the fundamental current. The formula is as follows:

[0042] THD = sqrt(I_2^2 + I_3^2 + I_4^2 + I_5^2 + I_6^2 + I_7^2) / I_1

[0043] Wherein, I_1 is the effective value of the fundamental current, and I_2 to I_7 are the effective values ​​of the 2nd to 7th harmonic currents, respectively. This index can quantitatively reflect the stability of the load and can be used as an auxiliary criterion for judging the degree of formation fluctuation.

[0044] 5. Multiphysics Signal Feature Extraction

[0045] After the vibration signal is bandpass filtered (passband frequency range 0.5 Hz to 200 Hz), the kurtosis feature K = (1 / N) × Σ_{i=1}^{N} [ (x_i - μ_x) / σ_x ]^4 and the peak factor CF = max|x_i| / RMS(x_i) can be calculated, where μ_x is the mean, σ_x is the standard deviation, and RMS(x_i) is the root mean square value.

[0046] After pre-emphasis, framing, and windowing, the acoustic signal is processed by short-time Fourier transform to calculate the time spectrum, and the Mel frequency cepstral coefficients are extracted as acoustic feature vectors.

[0047] 6. Weighted Voting Decision-Making Integration Strategy

[0048] This invention employs a weighted voting decision-making strategy to perform decision-level fusion of the outputs of multiple feature quantities. The fusion expression is as follows:

[0049] C_final = argmax_j ( Σ_{m=1}^{M} ω_m × I(C_m = j) )

[0050] Where: C_final is the final output stratigraphic type; m is the feature index; M is the total number of features involved in the fusion; ω_m is the preset weight corresponding to the m-th feature, satisfying Σ ω_m = 1; C_m is the identification result corresponding to the m-th feature; I(·) is the indicator function, whose value of 1 indicates that the condition is true, and whose value of 0 indicates that the condition is false. The fusion decision confidence is calculated according to the following formula:

[0051] Confidence = max_score / Σ ω_m

[0052] Here, `max_score` represents the highest weighted score. The weighted voting decision-making strategy is adjustable: if no force sensors are configured on-site, the system estimates force based solely on current, reducing the number of features involved in the fusion; if no vibration and acoustic sensors are configured on-site, the system can identify features based solely on electric drive specific power and current harmonic distortion rate; in practical applications, the number of features involved in the fusion and the weight parameters corresponding to each feature can be dynamically adjusted based on sensor configuration and signal quality level.

[0053] 7. Adaptive Drilling Control Strategy

[0054] The feedforward quantity generation module obtains the target current command of the pressurization motor, the target speed of the power head motor, and the limiting current command of the power head motor from a table in the preset mapping rule base based on the formation type output by the formation fusion decision module.

[0055] The feedback control module employs a digital PID controller to perform closed-loop tracking of the actual current of the pressurizing motor (reflecting the actual thrust) and the actual speed of the power head motor. The controller output, after proportional conversion, serves as the current or voltage command for the motor driver.

[0056] Multivariable coupling is achieved through a mechanism of sharing feedforward target values: when the stratum type changes, the target current command of the pressurized motor, the target speed of the power head motor, and the limiting current command of the power head motor are switched synchronously, thereby achieving coordinated adjustment of thrust, speed, and limiting.

[0057] Emergency detection of boulder or drill jam precursors is performed according to the following mechanism. The detection module monitors the dynamic changes of the following indicators in real time: the rate of increase of electric drive specific power, the amplitude of the jump in current harmonic distortion rate, and the abrupt change rate of the q-axis component of the pressurized motor current. When any of the above indicators exceeds a preset threshold (for example, the electric drive specific power increases by more than 3 times within 1 second, the current harmonic distortion rate jumps from below 8% to above 20% within 2 seconds, or the abrupt change rate of the q-axis component of the pressurized motor current exceeds 50% per second), the system determines that a boulder or hard interlayer contact may occur, and automatically executes the emergency control protocol, which specifically includes: reducing the power head motor limiting current to a set percentage of the rated value (reference value is 70%), reducing the pressurized motor current to the corresponding value of the low-speed gear, issuing an audible and visual warning signal in the cab, and executing a reverse retraction command in models equipped with an electric actuator.

[0058] 8. Online optimization of specific energy consumption

[0059] Specific energy consumption is defined as the ratio of total input power to instantaneous drilling speed, calculated using the formula SEC = P_input / V. The total input power P_input is estimated using the output mechanical power and efficiency of each drive motor: For any motor, based on its output torque T_m (in N·m) and measured speed n_m (in r / min), the motor's output mechanical power (in kW) is obtained using the formula P_out,m = T_m × n_m / 9550; combined with the motor's operating efficiency η_m under the current conditions (obtained through offline calibration or online driver identification), its corresponding input power is calculated using the formula P_input,m = P_out,m / η_m; finally, the input power of the pressurization motor, power head motor, and other drilling-related drive units is summed to obtain the total input power P_input.

[0060] When the formation type is stable and continuous drilling exceeds a preset time (30 seconds for reference), the optimization program is initiated. Centered on the current power supply motor current command and the power head speed command, several disturbance combinations are generated at a set step size (reference step size: power supply motor current command change ±5%, power head speed change ±2 r / min). Each disturbance combination is maintained in a predetermined steady state for a predetermined time (reference 10 seconds), and the average SEC value during this period is recorded. After traversing all combinations, the current and speed combination that minimizes the SEC is selected, and the control parameter mapping table for this formation type at this site is dynamically updated. This optimization process is executed synchronously with normal drilling operations without interrupting construction.

[0061] Beneficial effects

[0062] Theoretical analysis shows that the present invention is expected to achieve at least one of the following beneficial effects:

[0063] First, it reduces the reliance on historical data training for stratigraphic identification. The electric drive specific work in this invention is defined and calculated based on the mechanical principle of rock fracturing specific energy, which has clear physical meaning and interpretability. Its identification boundary threshold can be obtained through a small number of calibration experiments, without the need for large-scale historical datasets for model training.

[0064] Second, it achieves dual-source redundancy of electric and force sources and system fault tolerance. This invention constructs a dual-source sensing system of electromagnetic thrust estimation and direct measurement by force sensors. It realizes self-diagnosis of faults in current transformers or force sensors through cross-checking indexes, and can automatically switch to the other signal source to continue working when a single signal source is abnormal.

[0065] Third, it helps improve the emergency response speed in sudden operating conditions such as isolated rocks and stuck drill bits. This invention utilizes the fast response characteristics of the current loop in the electric drive system to design an emergency protocol based on current mutation rate detection. Compared with traditional hydraulic systems, it has a faster response speed and is expected to be able to achieve protective actions such as load reduction and torque limiting in a more timely manner.

[0066] Fourth, the control strategy is well-suited to the response characteristics of the electric drive system. The feedforward-feedback composite controller and the online optimization strategy based on specific energy consumption in this invention can adjust the motor control parameters in real time according to the formation type. Its control command update cycle can match the response time of the motor driver, and it is expected to achieve a balance between drilling efficiency and energy consumption.

[0067] Fifth, the incremental hardware costs of the project deployment are controllable. The three-phase current signals of the pressurized motor and the power head motor relied upon by this method are obtained from the existing current detection circuit in the electric rotary drilling rig motor driver. They can be obtained by listening to the existing bus in a read-only manner, without the need to install additional dedicated pressure sensors to obtain the core force signals. If it is necessary to add vibration, acoustic or force sensing dimensions, a small number of mature industrial sensors can be added without structural modifications to the original control system. Attached Figure Description

[0068] Figure 1 This is an overall flowchart of the method of the present invention.

[0069] Figure 2 This is a schematic diagram of the framework for multi-physics signal acquisition and feature extraction in the method of this invention.

[0070] Figure 3 This is a block diagram illustrating the principle of the electric drive thrust estimation model in the method of this invention.

[0071] Figure 4 This is a decision framework diagram of the weighted voting decision-making strategy in the method of the present invention.

[0072] Figure 5This is a structural block diagram of the adaptive drilling controller in the method of the present invention.

[0073] Figure 6 This is a state transition diagram for the emergency response to isolated rocks or stuck drill bits in the method of the present invention. Detailed Implementation

[0074] To make the technical means of implementing the present invention easier to understand, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the relevant invention and are not intended to limit the invention.

[0075] Example 1: Simulation Verification Environment Description

[0076] This embodiment relies on the MATLAB / Simulink simulation platform to construct an electromechanical coupling simulation model of the drilling process of an electric rotary drilling rig, aiming to verify the feasibility of the technical solution of the present invention through simulation.

[0077] (I) Simulation Platform and Tools: The MATLAB / Simulink software platform was used to build the system model and perform simulation calculations. The Simscape Electrical module library was used to establish the field-circuit coupling model of the pressurized permanent magnet synchronous motor and the power head permanent magnet synchronous motor. A drill bit-rock interaction model based on Hertzian contact theory was used to describe the mechanical relationship between the drill bit and the rock. The Embedded Coder module was used to model a dual closed-loop controller with current and speed loops. MATLAB scripts were used for signal synchronization acquisition, feature extraction, fusion decision-making, and control command issuance.

[0078] (II) Simulation Parameter Settings: The simulation settings in this embodiment adopt the configuration parameters of a typical medium-sized electric rotary drilling rig. Specifically, the rated power of the pressurized motor is not less than 50 kW, the rated torque is not less than 300 Nm, and the rated speed range is between 1000 rpm and 2000 rpm; the rated thrust of the electric push rod is not less than 300 kN; the rated power of the power head motor is not less than 150 kW, the rated torque is between 150 kNm and 500 kNm, and the measured output speed of the power head after being driven by a multi-stage reduction mechanism is between 5 rpm and 25 rpm (the corresponding total transmission ratio of the reduction mechanism is usually in the range of 20 to 60). It should be noted that the above parameters are only illustrative and not intended to limit the invention in any way. In actual applications, they can be adaptively adjusted according to the specific parameters of different electric rotary drilling rig models. The signal sampling frequencies are set as follows: current signal 2000 Hz, triaxial accelerometer signal 2000 Hz, and acoustic sensor signal 20000 Hz. The drilling depth ranges from 0 to 60 meters.

[0079] (III) Simulated Stratigraphic Sequence Setup: The simulation model sets up a five-layer typical stratigraphic sequence, from top to bottom in terms of depth: 0-8 meters is soft soil (uniaxial compressive strength UCS approximately 5 MPa), 8-18 meters is sand (UCS approximately 15 MPa), 18-28 meters is gravel (UCS approximately 30 MPa), 28-42 meters is strongly weathered rock (UCS approximately 60 MPa), and 42-55 meters is moderately weathered rock (UCS approximately 90 MPa). An isolated boulder (UCS approximately 120 MPa, diameter ranging from 0.2 to 0.5 meters) is placed at a depth of 32 meters as a verification sample for abrupt changes in the working condition. The thickness, depth, and boulder size mentioned above are exemplary settings and can be adjusted accordingly based on the drilling plan of the specific project in actual applications.

[0080] (iv) Expected Simulation Results: Theoretical analysis shows that the expected range of electric drive specific work will vary depending on the hardness of the formation within different geological strata. For example: in soft soil, the expected value of electric drive specific work is between 0.5 MPa and 1.0 MPa; in sand, it is between 1.0 MPa and 1.8 MPa; in gravel, it is between 1.8 MPa and 2.8 MPa; in strongly weathered rock, it is between 2.8 MPa and 4.2 MPa; and in moderately weathered rock, it is between 4.2 MPa and 6.5 MPa. The above parameter ranges are theoretical calculations based on typical working conditions. In actual engineering, adjustments can be made according to the specific machine model, geological characteristics, and sensor configuration.

[0081] Under the boulder condition, theoretical analysis indicates that the pressurized motor current is expected to surge significantly upon contact with the boulder, the current harmonic distortion rate is expected to jump, and the electric drive specific power is expected to rise from approximately 3.5 MPa to approximately 10.8 MPa within a short period (approximately 1.2 seconds). After the system triggers the emergency protocol, the limiting current of the power head motor is expected to decrease to a set percentage (e.g., 70%) of the rated value within milliseconds (e.g., less than 2 milliseconds), and the relevant parameters are expected to return to the normal range after the boulder passes.

[0082] Example 2: Implementation of Data Acquisition and Signal Synchronization

[0083] This embodiment provides an engineering implementation plan for the synchronous acquisition of multi-physics field signals in step S1. An industrial-grade edge computing terminal is used as the core acquisition device. The signal access and acquisition methods are as follows:

[0084] The bus interface is configured in read-only listening mode and is connected to the bus network of the electric rotary drilling rig through the diagnostic interface. The sampling frequency is set to 2000 Hz to collect information such as the equivalent value of the pressurized motor current, the speed of the power head motor, the equivalent value of the power head motor current, and the drilling depth.

[0085] The triaxial accelerometer is connected via the IEPE analog input channel, with a sampling frequency set to 2000 Hz and a measurement range of ±50g. The waterproof microphone is connected via the audio input channel, with a sampling frequency set to 20000 Hz and a protection rating of at least IP66. The force sensor is connected via the bridge input channel, with a sampling frequency set to 500 Hz, and a measurement range selected according to the system's output force range (e.g., 500 kN). The depth sensor is connected via the encoder interface, with a resolution set to 0.01 meters and a sampling frequency set to 10 Hz.

[0086] All sensor signals are synchronized via GPS or Network Time Protocol, with the synchronization error preferably controlled within 1 millisecond.

[0087] Example 3: Implementation of Formation Interface Detection

[0088] This embodiment provides a specific implementation method for the formation interface detection in step S3. The method employs a strategy combining sliding window linear regression and second-order difference threshold detection, and includes the following sub-steps:

[0089] Sub-step 3.1: Maintain the depth-specific power sliding window. The window length L is adjusted by engineering within the range of 10 to 20 sampling points based on the characteristics of abrupt changes in the field strata.

[0090] Sub-step 3.2 involves performing linear regression on the data sequence of depth D and ratio e within the window, and using the least squares method to solve for the regression parameter e = a + s_i × D, thereby obtaining the slope s_i.

[0091] Sub-step 3.3: Slide the window with a step size of 1 and repeat sub-step 3.2 to obtain the slope sequence S = [s_1, s_2,…, s_m], where m is the number of times the window slides.

[0092] Sub-step 3.4 calculates the second difference Δ²S = diff(S, n=2) of the slope sequence, and the result reflects the acceleration characteristics of the rate of change of specific work.

[0093] Sub-step 3.5: Calculate the dynamic threshold Th = mean(Δ²S) + k × std(Δ²S), where k is an empirical coefficient, which can be taken as a reference value in the range of 2.0 to 3.0.

[0094] Sub-step 3.6: Detect the maximum value of the higher-order difference |Δ²S| sequence. If the maximum value exceeds Th, then determine that the depth point corresponding to the maximum value is the location of the formation interface.

[0095] Sub-step 3.7: Calculate the significance index Sig = |Δ²S_max| / std(Δ²S), which serves as a quantitative evaluation index for the confidence level of the interface.

[0096] Example 4: Parameter Configuration Example of Weighted Voting Decision-Making Strategy

[0097] This embodiment provides a specific weight configuration example for the weighted voting decision strategy in step S3. In this example, the three features involved in the fusion and their weight settings can refer to the following scheme: the weight of the identification result based on the electric drive specific power threshold method is 0.4, the weight of the identification result based on the current harmonic distortion rate and thrust fluctuation rule engine is 0.3, and the weight of the identification result based on the vibration and acoustic classifier is 0.3.

[0098] The aforementioned weighting configuration can be dynamically adjusted based on site geological conditions, sensor signal quality, and drilling tool operating conditions during implementation. For example, in the absence of a force sensor, the weights of thrust-related characteristics can be appropriately reduced while the weights of current harmonic characteristics and vibration characteristics can be increased accordingly; in environments with low noise, the weights of acoustic characteristics can be appropriately increased.

[0099] Example 5: Implementation of Power Head Torque Redundancy Acquisition

[0100] This embodiment provides an engineering implementation plan for multi-source acquisition of the power head torque in steps S1 to S2. In addition to estimating the electromagnetic torque using the aforementioned power head motor current, redundant measurements can be performed simultaneously using strain sensors on the power head housing to reduce the cumulative error that may be introduced by calculating a single physical quantity. Strain sensors (e.g., strain gauges or strain plates) are installed on the surface of the power head reducer housing or gearbox housing to collect the housing's elastic deformation ε in real time. By accurately calibrating this deformation and the power head output torque, a torque-strain function relationship, T_strain = f(ε), can be established. Calibration can be completed through loading tests before the equipment leaves the factory, or it can be performed on-site by collecting data from known formations during drilling. The estimated motor current value and the strain estimate are fused using methods such as weighted averaging or Kalman filtering to output the fused torque estimate.

[0101] Industrial applicability

[0102] The intelligent drilling method for electric rotary drilling rigs based on electro-mechanical multi-physics field perception and weighted voting decision-making provided by this invention is feasible for practical engineering deployment.

[0103] The three-phase current signals of the pressurized motor and the power head motor relied upon in this method are derived from the existing current detection circuit in the electric rotary drilling rig's motor driver. These signals can be obtained by listening to the existing bus in a read-only manner, eliminating the need for additional dedicated pressure sensors to acquire the core force signals, thus reducing deployment costs. To further enhance vibration and acoustic sensing, a triaxial accelerometer (installed in the power head housing) and a waterproof microphone can be added. For dual-source electro-force cross-verification, a force sensor (installed at the flange connecting the electric push rod and drill rod) can be added. All of these sensors are mature industrial products, and the incremental hardware costs are within an acceptable range for engineering design.

[0104] The signal processing algorithms involved in this method are all supported by mature computing toolchains and can be deployed on industrial-grade edge computing terminals, enabling real-time inference and decision-making within millisecond timescales. Each module has a moderate computational load and can achieve low-latency operation on embedded platforms.

[0105] This method listens to the existing bus in read-only mode to obtain some core signals, supports data interface with existing vehicle controllers, does not require structural modification of the original control system, and is feasible for pre-installation on new electric rotary drilling rigs and for upgrading existing models.

Claims

1. A smart drilling method for electric rotary drilling rigs based on electro-mechanical multi-physics field perception and weighted voting decision-making, characterized in that, Includes the following steps: Step S1: During the drilling process of the electric rotary drilling rig, multi-physics field signals are synchronously acquired through the data acquisition module. The multi-physics field signals include the three-phase current of the pressurized permanent magnet synchronous motor, the three-phase current of the power head permanent magnet synchronous motor, the position sensor signal built into the electric push rod, the drilling depth sensor signal, and at least one of the following sensor signals: the triaxial acceleration sensor signal installed in the power head housing, the microphone acoustic signal installed near the power head, and the force sensor signal installed at the connection between the electric push rod and the drill rod; wherein, the power head is driven by the permanent magnet synchronous motor through the reduction mechanism. Step S2: Based on the signal acquired in step S1, at least two physically defined characteristic quantities are calculated in real time by the formation characteristic calculation module. The characteristic quantities include at least two selected from the following set: electromagnetic thrust estimated based on pressurized motor current, electromagnetic torque estimated based on power head motor current, electric drive specific power, current harmonic distortion rate, triaxial acceleration time-frequency characteristics, and acoustic Mel frequency cepstral coefficients. Step S3: Input the feature quantities calculated in step S2 into the formation fusion decision module, and output the formation type and confidence level corresponding to the current drilling depth through a weighted voting decision strategy; Step S4: Input the formation type output in step S3 into the adaptive drilling controller. The controller generates the target current command of the pressurized permanent magnet synchronous motor and the target speed and limiting current command of the power head permanent magnet synchronous motor through the feedforward-feedback composite control strategy, and sends them to the actuator through the motor driver.

2. The method according to claim 1, characterized in that, The specific power of the electric drive is calculated based on the actual feed force, borehole cross-sectional area, measured rotational speed of the power head, estimated output torque of the power head, and instantaneous drilling speed. The actual feed force is obtained by converting the electromagnetic thrust through a lead screw drive or by direct measurement by a force sensor.

3. The method according to claim 2, characterized in that, When simultaneously acquiring the current signal from the pressurized motor and the force sensor signal, the cross-verification index between the estimated electromagnetic thrust and the measured force sensor value is calculated. Based on the comparison result of this index with a preset threshold, it is determined whether the dual-source sensing is consistent or, if a deviation occurs, it automatically switches to a single signal source and issues a fault prompt.

4. The method according to claim 1, characterized in that, The estimated output torque of the power head is obtained in a redundant manner: at least the electromagnetic torque is calculated using the current of the power head motor and converted by the transmission ratio of the reduction mechanism to obtain the first estimated torque value. At the same time, the deformation of the housing is measured by the strain sensor installed on the housing of the reducer and converted by the calibration curve to obtain the second estimated torque value. The two estimated torque values ​​are then fused and output.

5. The method according to claim 1, characterized in that, It also includes emergency detection and control of boulder or drill jamming precursors: real-time monitoring of the rise rate of electric drive specific power, the jump amplitude of current harmonic distortion rate and the change rate of pressurizing motor current. When any indicator exceeds the preset threshold, at least one of the following protective actions is automatically executed: power head motor current limiting, pressurizing motor deceleration and cab alarm.

6. The method according to claim 1, characterized in that, It also includes online optimization of specific energy consumption: in drilling intervals with stable formation types, the ratio of total input power per unit footage to instantaneous drilling speed is calculated, a disturbance is applied with the current control parameters as the center, the parameter combination that minimizes this ratio is recorded, and the recommended control parameter table for the corresponding formation is dynamically updated.

Images