High-precision identification method for dynamic parameters of long cantilever of drill jumbo based on layered identification and nonlinear coupling modeling

Abstract

The invention discloses a high-precision identification method for dynamic parameters of a long cantilever of a drill jumbo based on layered identification and nonlinear coupling modeling, and belongs to the technical field of engineering machinery and control. The method is based on a multi-body dynamic modeling theory, and is combined with a dynamic parameter identification method, a least square optimization algorithm and a model online identification algorithm to realize high-precision identification of inertial parameters, friction coefficients and nonlinear coupling terms of all joints of the mechanical arm. By establishing a kinetic model containing nonlinear disturbance and proposing a layered identification strategy to cope with different stages of no-load operation, drilling impact and the like, nonlinear identification of impact-deformation coupling is realized, and the robustness and the real-time performance of kinetic parameter identification are improved. The method is suitable for optimization and fault diagnosis of the rock drilling operation control system under complex working conditions.

Description

Technical Field

[0001] The present invention relates to the field of engineering machinery control technology, and in particular to a high-precision identification method for the dynamic parameters of a long cantilever of a rock drilling rig based on hierarchical identification and nonlinear coupling modeling, which can be used in the fields of automated control, condition monitoring, and intelligent maintenance. Background Art

[0002] The three-arm drilling rig is a heavy-duty machine widely used in mining, tunneling and other engineering fields. Its mechanical arm system is usually composed of multiple joints and connecting rods. This equipment needs to operate continuously for a long time under various harsh working conditions such as high vibration, high dust concentration, drastic temperature and humidity changes, and limited space. Figure 1 As shown, the mechanical structure is prone to wear, looseness and even deformation, which in turn affects the dynamic characteristics of the system.

[0003] To improve drilling rig efficiency, ensure construction safety, and optimize operational processes, research on dynamic parameter identification for three-arm drilling rigs has important theoretical significance and engineering application value. By building a precise dynamic model and efficiently identifying key parameters, not only can the design accuracy and robustness of the control system be improved, but it can also provide a reliable data foundation for subsequent fault diagnosis, energy consumption optimization, and intelligent operation and maintenance.

[0004] In actual operation, the robotic arm of a three-arm drilling rig typically utilizes a long cantilever structure to achieve large-scale drilling operations and complete complex drilling trajectory planning tasks. This structure faces dynamic disturbances such as sudden load changes, external impacts, and support platform instability. This structure presents significant technical challenges in dynamic modeling and control. First, long cantilever robotic arms typically possess a high degree of freedom to meet the requirements for flexible operation along complex spatial trajectories. However, the large mass associated with the long arm span, the joint elasticity caused by the lightweight structure, and the hysteresis of the heavy-duty actuator introduce strong nonlinearities and strong coupling into the system dynamics. These factors make it difficult for traditional theoretical dynamic models to accurately reflect the actual system behavior. This is especially true when performing high-precision control, path tracking, or adaptive adjustment, as model errors can significantly impact control performance and operational efficiency. Second, due to the long cantilever length and frequent external disturbances in the operating environment (such as rock reaction forces and platform vibration), residual vibration effects are non-negligible. This residual vibration arises not only from the material properties of the connecting rod itself but also from joint clearances and joint elasticity. This can significantly increase the position error of the end effector and affect drilling accuracy.

[0005] Currently, research on the dynamics modeling of robotic arm systems primarily relies on the Newton-Euler or Lagrangian methods to establish theoretical equations. However, in complex equipment like three-arm drilling rigs, key parameters such as inertia (such as mass, center of mass position, moment of inertia), friction coefficient, and driver response characteristics often deviate significantly from their designed values ​​due to manufacturing tolerances, assembly errors, material aging, and changes in lubrication conditions. Furthermore, the coupling effects associated with the coordinated motion of multiple arms further complicate modeling.

[0006] Existing dynamic parameter identification methods primarily rely on least squares or recursive least squares methods based on Lagrangian modeling. Given a known structure, these methods construct regression models by collecting experimental data such as joint angles, velocities, accelerations, and drive torques, and estimate physical parameters such as joint mass, inertia, and center of mass position. These methods are effective for single-arm or simple-structured robotic arm systems but are difficult to apply to three-arm drilling rigs. Due to the strong coupling between the three-arm system structures, improper excitation signal design can render some parameters unobservable. In complex operating scenarios like rock drilling, nonlinear friction, impact, and high-frequency disturbances are significant, and ignoring these characteristics can lead to severe model mismatch. Furthermore, traditional methods are mostly offline and lack online update mechanisms, making them unable to adapt to model changes under varying operating conditions, limiting their practicality and robustness in dynamic environments.

[0007] In summary, the following common problems exist in existing technologies: 1. The design of excitation signals lacks systematicity, resulting in poor parameter observability, low regression matrix rank, and prone to parameter coupling and degradation. 2. The commonly used simplified linear friction model makes it difficult to accurately characterize the actual friction behavior of the rock drill arm under frequent starts and stops and impact loads. 3. Identification strategies are mostly static offline modeling, lacking online update mechanisms, and are unable to adapt to the frequently changing dynamic environment of rock drilling operations. 4. The application of dynamic models is limited, serving only controller design and failing to extend to scenarios such as state estimation and equipment health monitoring. Summary of the Invention

[0008] The purpose of the present invention is to provide a high-precision identification method for the long cantilever dynamic parameters of a drilling rig based on hierarchical identification and nonlinear coupling modeling, to solve the problems of low parameter identification accuracy and poor robustness in the prior art, and to improve the dynamic response capability and stability of the robotic arm control system.

[0009] A high-precision identification method for the dynamic parameters of a long boom of a drilling rig based on hierarchical identification and nonlinear coupling modeling includes the following steps: S1: pre-identification processing; S2: Establish the dynamic equations of the robotic arm; S3: Adopting a hierarchical identification strategy to deal with the nonlinearity of impact-deformation coupling; S4: Reconstructed linear regression model; S5: data collection and working condition classification; S6: Lumped disturbance modeling and joint identification.

[0010] Furthermore, step S1 is specifically as follows: S1.1: Build a 3D model of the long cantilever; S1.2: Simplified long cantilever model; S1.3: Determine DH parameters and coordinate system.

[0011] Furthermore, step S2 is specifically as follows: S2.1: Get the coordinate transformation matrix; S2.2: Establish pose matrix; S2.3: Compute Jacobian matrix expressions; S2.4: Use the Lagrangian method to establish the theoretical dynamic equations.

[0012] Furthermore, step S3 is specifically as follows: S3.1: Perform offline identification of static structural parameters; S3.2: Implement online updating of time-varying disturbance parameters.

[0013] Furthermore, step S5 is specifically as follows: S5.1: Divide the working conditions into stages; S5.2: Implement data collection strategy.

[0014] Furthermore, step S6 is specifically as follows: S6.1: Construct a nonlinear friction model; S6.2: Establish an impact equivalent stiffness model; S6.3: Establish structural elastic deformation terms.

[0015] Compared with the prior art, the present invention has the following advantages: 1. Use system excitation trajectory design to improve parameter observability The present invention introduces structured system excitation trajectories (such as PRBS, mixed frequency excitation, etc.) during the data acquisition stage, effectively stimulating the system's various-order dynamic responses, improving the rank and identification stability of the regression matrix, thereby significantly enhancing the observability and identifiability of dynamic parameters (such as mass, inertia, center of mass position, etc.), and avoiding the parameter coupling and degradation problems that occur in traditional identification.

[0016] 2. Introduce nonlinear friction modeling to improve model approximation accuracy To address the significant frictional effects of a drilling rig's robotic arm under conditions such as low speed, heavy load, and frequent starts and stops, this paper constructs a nonlinear friction model that incorporates Coulomb friction, viscous friction, and the Stribeck effect. Compared to traditional linear models, this approach more realistically reflects the dynamic relationship between joint output and load, improving model accuracy during startup, reversing, and impact phases.

[0017] 3. Support offline and online recognition modes to adapt to complex operation scenarios This paper proposes a hierarchical identification strategy, where static structural parameters are estimated offline and dynamic disturbance parameters are updated in real time using online algorithms (such as extended Kalman filtering or recursive least squares), thus balancing identification accuracy and operational efficiency. This mechanism can flexibly adapt to various operating conditions, including no-load testing, on-site rock drilling, and sudden disturbances, and has broad engineering applicability.

[0018] 4. System integration for controller design, state estimation, and health monitoring The precise dynamics model established not only serves as the basis for controller design (such as inverse dynamics and model predictive control), but also seamlessly integrates with state observers and remaining life prediction modules. By updating internal and external disturbance parameters such as friction, impact, and elasticity in real time, it facilitates equipment health monitoring, anomaly detection, and fault warning, possessing significant engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is the construction drawing of the drilling rig; Figure 2 It is a schematic diagram of the robotic arm structure; Figure 3 It is a simplified diagram of the joint coordinate system; Figure 4 It is a connecting rod model diagram; Figure 5 This is the linear velocity simulation result diagram; Figure 6 This is a flow chart of the method for identifying dynamic parameters of the long cantilever of the rock drilling rig in this application. DETAILED DESCRIPTION

[0020] The following is combined with Figure 2-6 This application is described in further detail.

[0021] A high-precision identification method for the dynamic parameters of a long boom of a drilling rig based on hierarchical identification and nonlinear coupling modeling includes the following steps: S1: Pre-identification processing S1.1: Build a 3D model of the long cantilever Use Solidworks 3D modeling software to model the long boom of the drilling rig. Figure 2The figure shows a three-dimensional model of a long boom. The long boom of a drilling rig consists of a base 1, a pitch joint 2, a planar swing joint 3, a boom telescopic support rod 4, a boom 5, a boom inner arm 6, a tilt joint 7, a rotation compensation joint 8, a pitch compensation joint 9, a telescopic rod 10, and a telescopic arm 11. The long boom is mounted on the body of the three-arm drilling rig via the base 1. The rotation compensation joint 8 and the pitch compensation joint 9 ensure that the drill boom is perpendicular to the tunnel section to be excavated for drilling operations. The telescopic rod 10 allows the drill boom to be extended to different positions to accommodate drilling operations at different heights. The planar swing joint 3 is defined as the first joint, the boom 5 as the second joint, the telescopic arm 11 as the third joint, the tilt joint 7 as the fourth joint, the rotation compensation joint 8 as the fifth joint, and the pitch compensation joint 9 as the sixth joint.

[0022] S1.2: Simplified long cantilever model To obtain the motion equation of the drill arm end effector, we must first establish the joint coordinate system of the drill arm, simplify each component into a simple link and hinge, then establish the coordinate system of each joint at the appropriate position, fix the first coordinate system at the first joint, and so on to establish the coordinate systems of the remaining five joints. Among them, the component between the plane swing joint 3 and the pitch joint 2 is converted into the first link I, the upper arm inner arm 6 is converted into the second link II, the upper arm inner arm 6 and the flip joint 7 are the third link III, the flip joint 7 and the rotation compensation joint 8 are the fourth link IV, and the rotation compensation joint 8 and the pitch compensation joint 9 are the fifth link V. Figure 3 For the simplified robotic arm joint coordinate system diagram, define q as a generalized vector matrix, where The summary of generalized variables, kinematic pairs, and generalized variable descriptions for each joint is shown in Table 1.

[0023] Table 1 Generalized variables and descriptions of long cantilever

[0024] S1.3: Determine DH parameters and coordinate system According to the structure of the robotic arm, a DH parameter table is established to determine the link length, joint offset, joint angle and link torsion angle.

[0025] Create a DH table and determine the DH parameters of each joint, as shown in Table 2: Table 2 DH parameter table

[0026] S2: Establish the dynamic equations of the robotic arm S2.1: Get the coordinate transformation matrix After establishing the DH parameter table, substitute the parameters of each joint of the drill arm into the matrix change formula between each joint to obtain the coordinate change matrix between adjacent joints of the drill arm, as shown in formulas (2-1)-(2-6).

[0027] (2-1) (2-2) (2-3) (2-4) (2-5) (2-6) Where, , .

[0028] S2.2: Create a pose matrix The pose matrix of the drill arm end effector coordinate system {6} relative to the base coordinate system {0}, that is, the drill arm forward kinematics matrix, can be established as shown in formula (2-7): (2-7) Where, is the transformation matrix of the end effector relative to the base coordinate system, Represents the state vector of the drill arm end relative to the base coordinate system, Indicates the position of the drill boom end relative to the base coordinate system.

[0029] S2.3: Compute Jacobian matrix expression In order to calculate the center of mass velocity and angular velocity of each link, a code was written in MATLAB to use the vector product method to calculate the ideal Jacobian matrix. Then, a link model identical to the actual long cantilever structure was established in Adams to verify whether the calculated Jacobian matrix was correct. Figure 4 shown.

[0030] Add kinematic pairs and drive functions at the joints in sequence, add marker points at the end positions, apply drive functions to each kinematic pair, drive a single joint each time, and record the linear velocity and angular velocity in the x, y, and z directions of the end position. Figure 5 shown.

[0031] Repeat the above steps six times, recording the linear and angular velocities in the x, y, and z directions for each joint. This matrix is ​​then assembled into a 6x6 matrix, which is the Jacobian matrix of the robotic arm. Compare this matrix with the matrix generated in MATLAB. If the error is within the acceptable range, the Jacobian matrix is ​​correctly constructed.

[0032] S2.4: Establishing the Theoretical Kinetic Equations Using the Lagrangian Method The essence of the Lagrangian method is to use the system energy to differentiate the variables in the system to obtain the net external force or net external torque.

[0033] S2.4.1: Calculate total kinetic energy For the i-th connecting rod, its kinetic energy consists of two parts: translational kinetic energy and rotational kinetic energy: (2-8) Where, ---The mass of the i-th connecting rod; ---The center of mass position vector of the i-th link; ---Center of mass velocity; ---Connecting rod angular velocity vector; ---Inertia tensor in the mass center coordinate system; The total kinetic energy is the sum of the kinetic energies of all connecting rods: (2-9) S2.4.2: Calculate total potential energy The potential energy comes from gravity, assuming that the acceleration due to gravity is , the potential energy of the i-th connecting rod is: (2-10) The total potential energy is: (2-11) S2.4.3: Subtracting total kinetic energy from total potential energy The Lagrangian function of the system is defined as (2-12) Where, L---Lagrange function; T---system kinetic energy; V---system potential energy; S2.4.4: Calculate the resultant external moment By taking the derivative of equation (2-13), we can get the total external torque of the system, which is: (2-13) The system dynamics equation can be obtained: (2-14) Where, ---Inertia matrix; ---Coriolis force and centrifugal force terms; --- Gravity term; ---Driving force / torque vector; S3: Using a hierarchical identification strategy to deal with the nonlinearity of impact-deformation coupling Given the significant "impact-coupling effect" in drilling operations of drilling rigs, that is, the end drill bit maintains high-speed rotation while performing high-frequency impact, this coupling effect causes parameters such as friction, damping and contact stiffness in the dynamic equations to exhibit obvious nonlinear and time-varying characteristics, making conventional linear parameter identification methods difficult to directly apply.

[0034] To this end, the present invention adopts a hierarchical identification strategy, which divides system dynamics modeling and identification into two layers: S3.1: Perform offline identification of static structural parameters S3.1.1: Determine identification objectives Connecting rod quality , center of mass position , moment of inertia .

[0035] S3.1.2: Set operating conditions Collect data in a no-load, no external impact, and low-speed steady motion state to ensure stable dynamic response and minimal noise impact.

[0036] S3.1.3: Build an identification model Based on the Lagrangian dynamics modeling results, a standard regression model is constructed: (3-1) Where, ---Structural parameter regression matrix; ---Generalized coordinate vector, generalized velocity, generalized acceleration; ---The structural parameter vector to be identified is in the following form: (3-2) S3.1.4: Perform parameter solution Standard linear identification techniques such as least squares or QR decomposition are used to identify Estimation is performed and deterministic structural parameters are output, which can be used to initialize the dynamic model .

[0037] S3.2: Implementing online updates of time-varying disturbance parameters S3.2.1: Clarify update objectives Estimate parameters that evolve dynamically with operating conditions, including friction parameters (Coulomb friction , viscous friction ), equivalent impact stiffness , elastic structural damping Other unmodeled disturbance terms.

[0038] S3.2.2: Set up identification conditions Online data collection is performed during the rock drilling operation, which is a working condition with strong impact-rotation coupling and frequent external force disturbances.

[0039] S3.2.3: Reconstruct the identification model Using sliding window + recursive least squares (RLS) with known structural parameters, the model is reconstructed as follows: (3-3) Where, ---Disturbance term regression matrix (including velocity, acceleration, historical shock and other characteristic terms); ---Time-varying parameter vector, in the following form: (3-4) S3.2.4: Implement online model correction The real-time estimated Feedback into the dynamic model can correct the model in real time and improve the adaptability to impact and elastic deformation.

[0040] The final model form is: (3-5) Where, ---Basic dynamics terms; ---Time-varying nonlinear terms.

[0041] Through this hierarchical structure, basic structural parameters can be estimated with high precision under static conditions, while time-varying parameters are continuously tracked and corrected under impact conditions, significantly improving the model's dynamic adaptability and anti-interference modeling capabilities. It is particularly suitable for the identification and control of drilling operations of rock drill arms in complex rock environments.

[0042] S4: Reconstructing the linear regression model Linearizing the above kinetic equation into parameter form, we get the kinetic regression model: (4-1) Where, --- Dynamic regression matrix (consisting of generalized coordinates and their derivatives); --- Dynamic parameter vector (including and other common parameters); S5: Data collection and working condition classification The present invention classifies working conditions and performs stage-by-stage control on the data acquisition process according to the physical characteristics of different stages under rock drilling working conditions.

[0043] S5.1: Divide the working conditions into stages S5.1.1: Perform no-load stretch sampling When the robot arm is not in contact with the rock surface, the actuator is controlled to move according to a preset structured excitation trajectory (such as multi-frequency sine, PRBS, etc.) to obtain dynamic response data in a low-interference, no-load environment; S5.1.2: Conduct sampling in stable drill sections While the drill bit of the main drilling arm continues to drill into the rock, the other robotic arm repeats the movement along a predetermined trajectory, and the excitation signal is in the form of a superposition of "drilling amplitude + trajectory control".

[0044] This stage is used to identify the coupled friction parameters, equivalent stiffness terms, etc., and to collect medium-frequency disturbance characteristics in an environment with dynamic contact and stable impact input.

[0045] S5.1.3: Sampling of mutation segments When the system performs direction switching, suddenly encounters different rock formations or high-frequency impact interference, the data segment is marked as a high-disturbance state, corresponding to strong interference, which is suitable for online disturbance modeling.

[0046] S5.2: Implement data collection strategy S5.2.1: Perform multi-channel simultaneous acquisition Synchronously record multi-dimensional signals such as joint driving torque, current, voltage, angle, speed, etc. to ensure that all variables required for dynamic modeling are complete; S5.2.2: Set up a high-frequency sampling mechanism A high-speed acquisition system with a sampling frequency of no less than 500Hz is used to capture the instantaneous dynamic response within the impact cycle and retain the high-frequency information of the signal.

[0047] S5.2.3: Implement automatic labeling of working conditions By setting a threshold function (such as acceleration RMS or impact amplitude), the working condition label is automatically marked for subsequent model switching.

[0048] S6: Lumped Perturbation Modeling and Joint Identification The present invention further considers the non-modeled lumped disturbance terms existing in the system during rock drilling operations, including system friction, impact excitation and structural elasticity.

[0049] S6.1: Constructing a nonlinear friction model A joint friction model containing multiple typical friction components is constructed to describe the nonlinear resistance characteristics of joints under low speed, reversing or impact conditions. The friction force is expressed as: (6-1) Where, ---Coulomb friction coefficient; --- is the viscous friction coefficient; ---Joint angular velocity.

[0050] S6.2: Establishing the impact equivalent stiffness model In the rock drilling impact stage, the equivalent stiffness term is introduced Describe the periodic response excited by the impact frequency and the rock mass reaction force: (6-2) S6.3: Establishing Structural Elastic Deformation Terms Considering the elastic response of the flexible structure of the manipulator under impact, the additional displacement is estimated online. , and introduced into the correction term of the dynamic inverse model: (6-3) Where, ---Equivalent compliance stiffness coefficient.

[0051] Perturbation parameters and structural parameters Together they constitute a joint parameter set for system identification, and online identification is achieved through sliding window recursive least squares (SW-RLS) or extended Kalman filter (EKF).

[0052] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A high-precision identification method for the dynamic parameters of a long boom of a drilling rig based on hierarchical identification and nonlinear coupling modeling, characterized in that: The steps include: S1: pre-identification processing; S2: Establish the dynamic equations of the robotic arm; S3: Adopting a hierarchical identification strategy to deal with the nonlinearity of impact-deformation coupling; S4: Reconstructed linear regression model; S5: data collection and working condition classification; S6: Lumped disturbance modeling and joint identification.

2. The high-precision identification method for the long boom dynamic parameters of a drilling rig based on hierarchical identification and nonlinear coupling modeling according to claim 1 is characterized in that: Step S1 is specifically as follows: S1.1: Build a 3D model of the long cantilever; S1.2: Simplified long cantilever model; S1.3: Determine DH parameters and coordinate system.

3. The high-precision identification method for the dynamic parameters of a long boom of a drilling rig based on hierarchical identification and nonlinear coupling modeling according to claim 1 is characterized in that: Step S2 is specifically as follows: S2.1: Get the coordinate transformation matrix; S2.2: Establish pose matrix; S2.3: Compute Jacobian matrix expressions; S2.4: Use the Lagrangian method to establish the theoretical dynamic equations.

4. The high-precision identification method for the long boom dynamic parameters of a drilling rig based on hierarchical identification and nonlinear coupling modeling according to claim 1 is characterized in that: Step S3 is specifically as follows: S3.1: Perform offline identification of static structural parameters; S3.2: Implement online updating of time-varying disturbance parameters.

5. The high-precision identification method for the long boom dynamic parameters of a drilling rig based on hierarchical identification and nonlinear coupling modeling according to claim 1, characterized in that: Step S5 is specifically as follows: S5.1: Divide the working conditions into stages; S5.2: Implement data collection strategy.

6. The high-precision identification method for the dynamic parameters of a long boom of a drilling rig based on hierarchical identification and nonlinear coupling modeling according to claim 1, characterized in that: Step S6 is specifically as follows: S6.1: Construct a nonlinear friction model; S6.2: Establish an impact equivalent stiffness model; S6.3: Establish structural elastic deformation terms.

Images