A multi-objective optimization method for drift rotary drilling based on drill string state estimation

Abstract

This invention discloses a multi-objective optimization method for tunnel rotary drilling based on drill string state estimation, comprising: establishing a drill string state observer: constructing a nonlinear state observer to reconstruct the state of the drill string dynamics model and obtain the bottom hole drilling speed; establishing a multi-objective optimization model for drilling parameters: namely, a mechanical drilling speed model and a drill bit wear model; determining the parameters of the mechanical drilling speed model using regression analysis: the parameters in the mechanical drilling speed model are obtained by regression analysis of the borehole speed, borehole pressure, and bottom hole drilling speed, including the coal and rock strata coefficient, threshold drilling pressure, and speed index; and optimizing the parameters of the multi-objective optimization model for drilling parameters using a non-dominated sorting genetic algorithm. This invention conducts multi-objective optimization research on near-horizontal rotary drilling in tunnels based on drill string state estimation, which has guiding significance for the actual rotary drilling process in coal mine tunnels and plays an important role in improving drilling efficiency and ensuring drilling performance.

Description

Technical Field

[0001] This invention belongs to the field of underground tunnel drilling technology in coal mines, and relates to the optimization of operating parameters in rotary drilling, specifically to a multi-objective optimization method for rotary tunnel drilling based on drill string state estimation. Background Technology

[0002] Coal, as a vital basic energy source, ensures the energy needs of economic and social development and people's livelihoods. Coal mine geology differs from oil mines, characterized by high in-situ gas pressure, high gas content, soft coal seams, and low permeability, making it highly susceptible to underground collapses and stuck drill bits during drilling. Complex rock structures and unpredictable lithological variations affect drill bit wear, drilling speed, and borehole quality. With the increasing depth and intensity of coal mining, the probability and difficulty of managing various coal mine disasters are also constantly increasing. Currently, underground tunnel drilling in coal mines is still in the transition from mechanization to automation, with many key technologies yet to be solved. The availability of high-performance and high-reliability explosion-proof components is still very limited, and the level of informatization, intelligence, and practicality needs further improvement.

[0003] Near-horizontal rotary drilling in tunnels is a crucial means of ensuring the green and efficient mining of coal resources. During drilling, drillers often rely on their own experience to adjust operating parameters, making it difficult to simultaneously improve drilling efficiency and save costs. The complex geological conditions of coal mine tunnels and the uneven coal seam structure make it difficult to predict coal seam conditions during drilling; the friction and resistance between the drill bit and the coal seam constantly change, leading to unreasonable feed pressure and power head rotation speeds. To ensure the safety and efficiency of coal mine tunnel drilling, it is necessary to monitor and adjust equipment parameters to maintain a steady drilling process. However, manual drilling requires a significant amount of time and effort, increasing the difficulty of drilling work and affecting drilling efficiency and borehole quality. The presence of faults, folds, and other geological structures in coal seams significantly affects the measurement results of measurement-while-drilling tools, making it difficult to obtain in-hole data and resulting in large errors, failing to provide effective information to guide the adjustment of operating parameters during the drilling process. To balance multiple indicators such as drilling efficiency, drilling cost, and drilling safety, it is crucial to study multi-objective optimization of the drilling process and provide optimal recommended values ​​for operating parameters to ensure drilling performance. Summary of the Invention

[0004] This invention addresses the problems of difficulty in obtaining in-hole data and lack of basis for optimizing operating parameters in rotary drilling of underground coal mine tunnels. It studies a multi-objective optimization method for rotary drilling of tunnels based on drill string state estimation, and obtains reasonable optimized operating parameters, which can improve drilling efficiency and quality and reduce costs.

[0005] The technical solution adopted in this invention is as follows:

[0006] A multi-objective optimization method for tunnel rotary drilling based on drill string state estimation includes: establishing a drill string state observer: establishing the state-space equation of the drill string dynamics model based on the axial and torsional dimensions of the drill string dynamics model, constructing a nonlinear state observer to reconstruct the state of the drill string dynamics model, and obtaining the bottom hole drilling speed; establishing a multi-objective optimization model for drilling parameters: establishing a multi-objective optimization model for drilling parameters, namely, a mechanical drilling speed model and a drill bit wear model, with the maximum drilling speed and minimum drill bit wear as optimization objectives; determining the parameters of the mechanical drilling speed model using regression analysis: the parameters in the mechanical drilling speed model are obtained by regression analysis of the borehole head rotation speed, borehole head drilling pressure, and bottom hole drilling speed, including the coal and rock strata coefficient, threshold drilling pressure, and rotation speed exponent; and optimizing the parameters of the multi-objective optimization model for drilling parameters using a non-dominated sorting genetic algorithm.

[0007] Optionally, the state-space equations of the drill string dynamics model are:

[0008]

[0009]

[0010]

[0011] Where i = a represents the axial drill string dynamics model, i = t represents the torsional drill string dynamics model, n = 3 represents the degrees of freedom of the model, and the dynamic coefficient matrix J i ,K i C i Determined by the coefficients of the established axial and torsional dynamics model, u i For each dimension, w i (x i ) represents the drill bit-coal-rock interaction in the drill string dynamics model corresponding to each dimension, ||w i (x i )|| <M i M i The maximum value generated by the drill bit-coal-rock interaction; where S1 = [1, 0] 1×(n-1) ] T S2 = [0 1×(n-1) [,-1] T .

[0012] Optionally, the nonlinear state observer is:

[0013]

[0014] in, To input the estimated state variables, To output the estimated state variables, L is the gain matrix. Estimate the drill bit-coal-rock interaction for each dimension of the corresponding drill string dynamics model; solve the linear matrix inequalities using the LMI toolbox in MATLAB. The positive definite matrix P and the real number μ can be obtained, and the observer gain L = μP can be obtained. -1 C i T .

[0015] Optionally, the mechanical drilling rate model is:

[0016]

[0017] Where the mechanical drilling rate ROP is mm / s, and A1 = k d A2 = M is the coal and rock strata coefficient, A3 = λ is the threshold drilling pressure, and N is the rotation speed index (r / min).

[0018] Optionally, the drill bit wear model is as follows:

[0019]

[0020] Where D1 and D2 are the drilling pressure influence coefficients; Q1 and Q2 are the rotation speed influence coefficients; A f C1 is the formation abrasiveness coefficient; N is the tooth wear reduction coefficient; W is the rotational speed (r / min); h is the drilling pressure (MPa); and h is the relative wear of the drill bit cutting teeth.

[0021] Optionally, the parameter optimization of the multi-objective optimization model for drilling parameters using a non-dominated sorting genetic algorithm includes:

[0022] Using drill bit rotation speed and drilling pressure as decision variables, and comprehensively considering the mechanical drilling speed model and the drill bit wear model, a multi-objective optimization model is established. The parameter optimization can be expressed by the mathematical model of equation (24):

[0023]

[0024] Where h is the relative wear of the drill bit cutting teeth, ROP is the mechanical drilling speed (mm / s), and the tooth wear rate is v. t h represents the relative wear of the drill bit cutting teeth; drill bit rotation speed N, r / min; drilling pressure W, MPa.

[0025] Invention benefits:

[0026] This invention addresses the challenges of obtaining in-hole data and lacking a basis for decision optimization in rotary drilling of underground coal mine tunnels. It conducts a multi-objective optimization study on near-horizontal rotary drilling in tunnels based on drill string state estimation. This research provides guidance for actual rotary drilling processes in coal mine tunnels and plays a crucial role in improving drilling efficiency and ensuring drilling performance. Attached Figure Description

[0027] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings:

[0028] Figure 1 This is a roadmap for the multi-objective optimization technique for drilling parameters in this invention;

[0029] Figure 2 This is a comparison chart of the drilling speed calculated by the model of the present invention and the estimated bottom hole drilling speed;

[0030] Figure 3 This is the initial population distribution diagram of the NSGA-II algorithm of this invention;

[0031] Figure 4 This is a population distribution diagram after 60 iterations of the NSGA-II algorithm of this invention;

[0032] Figure 5 This is a population distribution diagram for the NSGA-II algorithm of this invention after 120 iterations. Detailed Implementation

[0033] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] This invention focuses on near-horizontal rotary drilling in tunnels, establishing drill string dynamics models in both axial and torsional dimensions, designing a nonlinear state observer, and establishing a mapping relationship between the drill string motion states at the borehole opening and bottom. It proposes an optimization target evaluation model for mechanical drilling speed and drill bit wear, and based on borehole opening data and bottom state estimation, optimizes the two operating parameters of rotational speed and feed pressure using the NSGA-II algorithm. Finally, it verifies and analyzes the proposed method using actual drilling data from a coal mine in Huainan, demonstrating its effectiveness and practicality. Specifically, it includes:

[0035] 1. Establish a drill string status observer:

[0036] To address the difficulty in obtaining in-hole data during rotary drilling in underground coal mine tunnels, a nonlinear state observer is constructed to reconstruct the state of the drill string system, based on the drill string dynamics model considering both axial and torsional dimensions. According to Lyapunov stability theory, stability analysis of the estimation error system is conducted, and a real-time design and update method for the state observer parameters is proposed. This enables in-hole drill bit state estimation based on the nonlinear state observer. Specifically, the state-space equations of the drill string dynamics model are established based on the axial and torsional dimensions; a nonlinear state observer is constructed to reconstruct the state of the drill string dynamics model, thereby obtaining the bottom hole drilling speed.

[0037] 2. Establish a multi-objective optimization model for drilling parameters:

[0038] Drilling parameters are a crucial factor directly restricting and influencing drilling efficiency and costs. Optimizing drilling parameters effectively reduces drill bit wear and improves drilling speed and efficiency. Due to the complexity of the underground working environment in coal mines, numerous factors affect drilling quality; therefore, multiple factors need to be considered comprehensively when establishing a drilling model. Using maximum drilling speed and minimum drill bit wear as optimization objectives, a multi-objective optimization model is established that can quantitatively assess drilling efficiency and reflect the impact of drilling parameters on the drilling process. The established mechanical drilling speed model and drill bit wear model have been validated with extensive actual field data and can objectively reflect drilling patterns. Specifically, a multi-objective optimization model for drilling parameters, namely the mechanical drilling speed model and the drill bit wear model, is established with maximum drilling speed and minimum drill bit wear as optimization objectives.

[0039] 3. Determine the parameters of the mechanical drilling rate model using regression analysis:

[0040] Multiple regression analysis is a statistical method used to study the relationship between multiple independent variables and a dependent variable. It can determine the influence of independent variables on the dependent variable and the interactions among these independent variables. Since relevant parameters in the mechanical drilling rate model are difficult to measure based on actual field conditions, regression analysis can be performed using easily measurable borehole data, including borehole rotation speed, borehole pressure, and bottom hole drilling speed estimated by a state observer. This allows the deriving of the coal and rock strata coefficient, threshold drilling pressure, and rotation speed exponent from the mechanical drilling rate model, resulting in a practically valuable drilling rate equation.

[0041] 4. Multi-objective optimization of drilling parameters based on NSGA-II

[0042] A mechanical drilling rate model and a drill bit wear model have been established. In practical engineering, it is not only necessary to pursue the optimality of a single objective, but also to consider both mechanical drilling rate and drill bit wear comprehensively. These two are mutually restrictive; improving one objective will come at the cost of reducing the effectiveness of the other. Therefore, an effective algorithm is needed to find the optimal solution. In recent years, multi-objective optimization algorithms have achieved good results in solving practical problems. Multi-objective genetic algorithms are optimization algorithms that can continuously optimize a specific problem without a dominated front. Non-dominated sorting genetic algorithms (NSGA-II) are a typical multi-objective genetic algorithm, especially effective for two objectives, and have been widely used in some complex engineering problems. Using the NSGA-II algorithm to solve the proposed optimization problem and obtain the Pareto optimal solution provides decision-makers with an effective method for selecting the optimal solution in practical engineering.

[0043] Multi-objective optimization of drilling parameters plays a crucial role in rotary drilling of coal mine tunnels. Parameter optimization can help the coal mining industry improve drilling efficiency, reduce drilling costs, and improve borehole quality. Based on the research content of this invention, the designed technical roadmap is as follows: Figure 1 As shown. The borehole information is input into the drill string dynamics model to obtain state information, and then the borehole state estimation information, i.e., the bottom hole drilling rate, is obtained through a nonlinear state observer. Using the borehole state estimation information, nonlinear multiple regression is used to identify the parameters of the mechanical drilling rate model, obtaining parameters including the coal and rock strata coefficient, threshold drilling pressure, and rotational speed exponent. Multi-objective optimization is then performed to solve the operational parameters, and finally, the optimization results are analyzed. Specifically, this includes:

[0044] Step 1: To address the challenges of obtaining in-hole data and lacking a basis for optimizing operating parameters during near-horizontal rotary drilling in coal mine tunnels, a nonlinear state observer based on axial and torsional drill string dynamics models is designed.

[0045] Step 1.1: Establish the drill string dynamics model

[0046] The axial drill string dynamics model is expressed by the following equation.

[0047]

[0048] Where, j an (n = 1, 2, 3) represents the quality coefficient of each connecting part, h an and k an These represent the axial stiffness and damping coefficient of each connection part, respectively. n =[z1,z2,z3] T For the axial displacement of each part, mm, u a Represents the input force, N, w a For the axial drill bit-coal rock interaction force, N; Equation (1) can form the coefficient matrix J of the dynamic model. a H a K a ;

[0049] in

[0050] Axial drill bit-coal rock interaction force w a The model is shown below:

[0051]

[0052] In the formula, w a The force represented by N is the reaction force from the interaction between the drill bit and the coal and rock at the drill bit location. This force acts at the drill bit location. C1 is the axial drilling speed of the drill bit, in mm / s; C2 is a constant related to the drill bit speed, which determines the magnitude of the drill bit speed; C2 is the nonlinear characteristic coefficient between the coal / rock and the PDC drill bit.

[0053] Based on the axial dynamic coefficient matrix and the drill bit-coal-rock interaction, the axial dynamic model can be represented by the following matrix form.

[0054]

[0055] Where S1 = [1, 0] 1×(n-1) ] T S2 = [0 1×(n-1) [,-1] T .

[0056] The dynamic model of the torsional drill string is expressed by the following equation.

[0057]

[0058] Where, j tn h tn k tn These represent the inertia coefficient, damping coefficient, and stiffness coefficient of each part of the model, respectively, θ n =[θ1,θ2,θ3] T The angle offset of each turntable section is expressed in rad. t The input torque to the power head is N·m, w t Let J be the load torque of the drill bit-coal-rock interaction, in N·m. Equation (4) can form the dynamic coefficient matrix J. t H t K t ;

[0059] in

[0060] In this invention, the Karnopp friction model with an exponentially decaying friction term is used to describe the torsional dimension of the drill bit-coal-rock interaction, and its formula is shown below:

[0061]

[0062] In the formula, w t The torque represented by the drill bit-coal-rock interaction load is N·m; T eb It is the contact torque, N·m; D is the drill bit rotation speed, in r / min; v T is a very small positive number; sb It is the static moment, N·m; R b W is the drill bit radius, in meters. ob Represents drilling pressure, MPa; μ b It is the coefficient of dry friction.

[0063] Based on the torsional dynamics coefficient matrix and the drill bit-coal-rock interaction, the torsional dynamics model can be represented by the following matrix form:

[0064]

[0065] Step 1.2: Establish the state-space equations of the drill string dynamics model

[0066] Based on the axial and torsional drill string dynamics models, state variables are introduced to construct the drill string state-space equations for subsequent state variable observation. x, as the state variable in the axial drill string dynamics model a .choose x is the state variable of the torsional drill string dynamics model. t ,in Since the axial drill string dynamics model (3) and the torsional drill string dynamics model (6) have the same structure, and by combining the set state vectors and through matrix substitution, the following general formula (7) of the drill string dynamics model can be used to describe the state space equations of the axial and torsional drill string dynamics models respectively.

[0067]

[0068]

[0069]

[0070] Where i = a represents the axial drill string dynamics model, i = t represents the torsional drill string dynamics model, n = 3 represents the degrees of freedom of the model, and the dynamic coefficient matrix J i ,K i C i Determined by the coefficients of the established axial and torsional dynamics model, u i For each dimension, the corresponding input (units correspond to the model), w i (x i ) represents the drill bit-coal-rock interaction in the drill string dynamics model corresponding to each dimension (units correspond to the model), ||w i (x i )|| <M i M i This represents the maximum value generated by the drill bit-coal-rock interaction (units correspond to the model).

[0071] Step 1.3: Stability Analysis of Nonlinear State Observer

[0072] The following lemmas are used when performing Lyapunov stability analysis on a state observer.

[0073] Lemma 1: When V(t) is a continuous function and w(t) is an exponentially bounded function, if there exist scalars α>0 and b>0, then the following inequality holds for all t>t0:

[0074]

[0075] Further, it can be seen that

[0076]

[0077] Lemma 2: When a system has an exponential decay rate α > 0 and is exponentially eventually bounded, if a constant β ≥ 1, then

[0078]

[0079] in

[0080]

[0081] κ is called ||x(t)|| 2 The ultimate upper limit.

[0082] A state observer is a tool used to estimate the state of a system. Its function is to estimate the system's state variables, including those not directly measured, by measuring the system's outputs and inputs, and using the system's dynamic model. Based on known orifice information, the state observer estimates the information inside the orifice.

[0083] In equation (5), a feedback matrix L is introduced, and the error between the observer estimate and the estimated quantity of the system is added. Construct the following state observer:

[0084]

[0085] in, To input the estimated state variables, To output estimated state variables.

[0086] Introducing state error quantity To analyze the convergence and convergence rate of the observer, subtracting (7) from (12) yields:

[0087]

[0088] Introducing the state error r into (13) yields:

[0089]

[0090] Given ||w i (x i )|| <M i , We set but Matrix A i B i and C i The system state estimation error is determined by the established model parameters. Therefore, whether the system state estimation error can be reduced to zero mainly depends on the feedback matrix L. If a suitable feedback matrix L can be selected, the system state estimation error can have a good decay rate.

[0091] For the estimation error r, construct the Lyapunov function:

[0092] V = r i T Pr i (15);

[0093] Where P is a positive definite symmetric matrix. Differentiating equation (13) and adding 2αV to both sides of the equation, we get...

[0094]

[0095] Where α > 0 and is a real number.

[0096] Based on the above conditions, it can be concluded that... Substituting into (16), we get

[0097]

[0098] set up The feedback gain matrix L = μP is set based on the property of linear matrix inequalities. -1 C i T μ is a constant to be replaced. Substituting it into equation (13) yields...

[0099]

[0100] in

[0101] When it was established, it meant V is an exponentially bounded state, and by Lemma 1, there exists...

[0102]

[0103] in It is not less than λ max The constant of (P) can be obtained from (19).

[0104]

[0105] According to Lemma 2, the system is actually exponentially stable, and ||r i (t)||2 The ultimate upper bound is

[0106] Solving linear matrix inequalities using the LMI toolbox in MATLAB The positive definite matrix P and the real number μ can be obtained, and the observer gain L = μP can be obtained. -1 C i T Substituting into equation (12) completes the design of the nonlinear observer. The feed force or torque is used as input to realize the in-hole state estimation based on the axial or torsional drill string system model, providing in-hole information for the next step of multi-objective optimization problem.

[0107] Step 2: In actual underground drilling in coal mines, drilling speed and drill bit wear have a significant impact on coal mine production efficiency, safety and environmental protection. Taking maximum drilling speed and minimum drill bit wear as optimization objectives, a multi-objective optimization model is established that can quantitatively evaluate drilling efficiency and reflect the impact of drilling parameters on the drilling process.

[0108] Step 2.1: Establish the drilling rate model

[0109] The modified Young's mechanical drilling rate model is a typical physical model that comprehensively considers the influence of multiple factors on the drilling rate:

[0110]

[0111] Where ROP is the mechanical drilling rate, mm / s; k d C is the formation drillability coefficient. p C is the pressure difference influence coefficient. H λ is the hydraulic purification coefficient; W is the drilling pressure (MPa); M is the threshold drilling pressure (MPa); N is the rotational speed (r / min); λ is the rotational speed index; C2 is the tooth wear coefficient; h is the relative wear of the drill bit teeth (0 represents no wear, 1 represents complete wear). Where C... P and C H The value of k is set to 1 in this invention. d The values ​​of M and λ are related to the actual drilling conditions at the site. These values ​​are obtained through regression analysis. To facilitate the determination of model parameters through regression analysis, the drilling speed model is simplified:

[0112]

[0113] Where A1 = k dLet A1 be the coal and rock strata coefficient, A2 = M be the threshold drilling pressure, and A3 = λ be the rotational speed index. Regression analysis yields A1, A2, and A3. In the drilling speed model, 1 / 1 + C2h represents the effect of tooth wear on the drilling speed. This effect is already reflected in the drill bit wear model below, so it is ignored in the drilling speed model. This model has good interpretability and has been validated by a large amount of actual field data, showing good consistency with actual drilling conditions.

[0114] Step 2.2: Establish a drill bit wear model

[0115] The drill bit wear in this invention mainly considers tooth wear, and the speed model of tooth wear is as follows:

[0116]

[0117] Where D1 and D2 are the drilling pressure influence coefficients; Q1 and Q2 are the rotation speed influence coefficients; A f C1 is the formation abrasiveness coefficient; N is the rotational speed (r / min); W is the drilling pressure (MPa); h is the relative wear of the drill bit cutting teeth (0 represents no wear at all, 1 represents complete wear). The specific values ​​of the correlation coefficients are given in the experimental section by referring to relevant materials.

[0118] Step 2.3: Establish a multi-objective optimization model

[0119] Drill bit rotation speed and drill pressure are two important factors affecting drilling efficiency and quality, and are also two easily controllable variables during the drilling process. ROP is the mechanical drilling speed, in mm / s, and the tooth wear rate is v. t (Percentage of drill bit wear per hour), where h is the relative wear of drill bit cutting teeth. Taking drill bit rotation speed N, r / min and drilling pressure W, MPa as decision variables, and h as the relative wear of drill bit cutting teeth, a multi-objective optimization model is established by comprehensively considering mechanical drilling speed and drill bit wear. The optimization problem can be represented by the mathematical model of equation (24).

[0120]

[0121] Step 3: Regression analysis is a statistical method used to study the relationship between independent and dependent variables. By using borehole data combined with bottom hole information estimated by the state observer, the coefficients A1, A2 and A3 in the simplified form (20) of the mechanical drilling rate model can be obtained through regression analysis.

[0122] Step 3.1: Data collection: The collected data includes the bottom hole drilling speed, borehole rotation speed, and drilling pressure observed by the status observer.

[0123] Step 3.2: Data processing: Preprocess the initial data using mean filtering.

[0124] Step 3.3: Estimate parameters: Estimate the parameters in the drilling rate model using nonlinear multiple regression.

[0125] Using the rotational speed, drilling pressure measured at the borehole opening, and the bottom hole drilling rate estimated by the state observer, a nonlinear multiple regression model was fitted to the mechanical drilling rate model. The fitted parameters were A1 = 12.9392, A2 = 8.4479, and A3 = -0.3277. The results were analyzed using common evaluation indicators of regression analysis, as shown in Table 1.

[0126] Table 1 Evaluation parameters for regression analysis

[0127]

[0128] Step 3.4: Test the model. Judge the goodness of fit of the model and test whether the regression model is suitable, including testing the significance and goodness of fit of the model.

[0129] Mean squared error (MSE) is the ratio of the sum of squares of the deviations between observed and true values ​​to the number of observations. A smaller MSE indicates better accuracy in describing the experimental data using the prediction model. Mean absolute error (MAO) better reflects the actual error in the predicted values. R 2 The value of R is between 0 and 1, and can be understood as the goodness of fit. 2 A value of 0.9881 indicates that the model can explain 98% of the output variation.

[0130] After determining the coefficients of the drilling speed model through nonlinear multiple regression, the drilling speed obtained from the mechanical drilling speed model and the bottom hole drilling speed estimated by the state observer are compared, for example... Figure 2 The drilling rate calculated by the mechanical drilling rate model and the drilling rate estimated by the state observer are relatively close, indicating that the drilling rate equation fitted by nonlinear multiple regression has good performance and practical value.

[0131] The optimization problem (22) above proposes two optimization objectives, requiring an effective algorithm to find the optimal solution. The NSGA-II algorithm has shown good performance in solving bi-objective optimization problems and has been widely used in practical engineering problems. This method is applicable to the present invention.

[0132] Step 4: NSGA-II mainly overcomes some shortcomings of the NSGA algorithm. In NSGA-II, a fast non-dominated sorting algorithm is used. Since multiple optimization objectives cannot be simultaneously optimized, the population is divided into different levels of Pareto solutions based on their dominance relationships. The algorithm aims to find a set of solutions that are as close as possible to the Pareto optimal domain, requiring both convergence and diversity of solutions. NSGA-II also employs an elite retention strategy, significantly increasing the probability of retaining superior individuals from the parent generation. Furthermore, the NSGA-II algorithm uses a crowding comparison operator as the comparison standard among individuals in the population, overcoming the NSGA algorithm's requirement for manually specifying shared parameters.

[0133] The population size, number of iterations, and probabilities of crossover and mutation in the NSGA-II algorithm are set according to the actual problem. From the obtained Pareto solutions, a satisfactory solution is selected as the optimized value of the drilling parameters based on the actual site conditions and drilling requirements, providing a useful reference for decision-makers in practical engineering problems.

[0134] Step 4.1: Initialize the population

[0135] The correlation coefficients in the drill bit wear equation can be determined based on the actual working conditions and the type of drill bit used. D1 = 0.0148, D2 = 6.38, Q1 = 1.5, Q2 = 6.53 × 10⁻⁶ -5 C1 = 5, A f =0.00228. The relevant parameter settings in the optimization algorithm were determined based on experience, with a population size of 100 and a generation count of 120. Figure 3 This indicates the distribution of solutions across different genetic generations.

[0136] Step 4.2: Pareto Optimal Solution Analysis

[0137] from Figure 4-5 It can be seen that initially, the population is scattered in the solution space. When the population evolves to the 60th generation, the individuals in the population are distributed more evenly. When the final evolution is completed, the individuals in the population are evenly distributed in the solution space, and the obtained Pareto curve is relatively smooth. The distribution of the final solution confirms the effectiveness of the NSGA-II algorithm. Three sets of Pareto optimal solutions are selected for analysis and comparison.

[0138] Table 2 Comparison of the three groups of Pareto optimal solutions

[0139]

[0140] Table 2 shows that as the drilling speed increases, the drill bit wear rate accelerates, indicating a mutually restrictive relationship between the two; improving one performance objective will lead to a decrease in the other. In actual coal mine drilling operations, there are certain requirements for the drilling speed. First, the drilling speed should be kept within a certain range, and then the drill bit wear rate should be considered. The selection of a satisfactory solution still needs to be determined based on the actual working conditions on site.

[0141] Step 5: To further verify the improvement of drill string state estimation on optimization results, experimental analysis was conducted using measured data from a coal mine in Huainan.

[0142] Three drilling experiments were conducted at the coal mine drilling site, using drill rods with a length of 0.75m. The resulting data was substantial and highly reliable. In the experiments described below, the correlation coefficient of the drill bit wear model was consistent with the simulation data used above, and the parameters in the mechanical drilling rate model were obtained through regression fitting of measured data. This paper uses bottom hole information observed by a state observer and optimizes the data using the NSGA-II algorithm, while comparing the results with actual field operating parameters. Due to the increase in mechanical drilling rate, another optimization objective, drill bit wear rate, will also increase. The optimized target value selection scheme in Table 2 aims to keep the change in drill bit wear rate relatively small while allowing a larger increase in drilling rate. The comparison is shown in Table 3.

[0143] Table 3 Comparison of Actual Coal Mine Drilling Data and Optimized Data

[0144]

[0145] Analysis of Table 3 leads to the following conclusions: After optimization using the method of the present invention based on the borehole bottom information estimated by the drill string condition, the drilling speed is 26.07 mm / s, which is 32.47% higher than the actual data collected.

[0146] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0147] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0148] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

Claims

1. A multi-objective optimization method for tunnel rotary drilling based on drill string state estimation, characterized in that, include: Establish a drill string state observer: Based on the drill string dynamics model in the axial and torsional dimensions, establish the state space equation of the drill string dynamics model, construct a nonlinear state observer to reconstruct the state of the drill string dynamics model, and obtain the bottom hole drilling speed; Establish a multi-objective optimization model for drilling parameters: With the maximum drilling speed and minimum drill bit wear as optimization objectives, establish a multi-objective optimization model for drilling parameters, namely the mechanical drilling speed model and the drill bit wear model; Regression analysis method to determine the parameters of mechanical drilling rate model: The parameters in the mechanical drilling rate model are obtained by regression analysis of the borehole speed, borehole pressure and bottom hole speed, including the coal and rock strata coefficient, threshold pressure and speed index. Parameter optimization of a multi-objective optimization model for drilling parameters is performed using a non-dominated sorting genetic algorithm. The state-space equations of the drill string dynamics model are as follows: (7)； ； in i = a For axial drill string dynamics model, i = t For the torsional drill string dynamics model, n= 3 represents the degrees of freedom of the model, and the dynamic coefficient matrix. Determined by the coefficients of the established axial and torsional dynamics model. u i For the corresponding inputs of each dimension, w i ( x i ) represents the drill bit-coal-rock interaction in the drill string dynamics model corresponding to each dimension. |w i ( x i )| |<M i , M i The maximum value generated by the drill bit-coal-rock interaction; where .

2. The multi-objective optimization method for tunnel rotary drilling based on drill string state estimation according to claim 1, characterized in that, The nonlinear state observer is: (12)； in, To input the estimated state variables, To output estimated state variables, L Here is the gain matrix. Estimate the drill bit-coal-rock interaction for each dimension of the corresponding drill string dynamics model; solve the linear matrix inequalities using the LMI toolbox in MATLAB. Positive definite matrices can be obtained. P and real numbers Obtain observer gain .

3. The multi-objective optimization method for tunnel rotary drilling based on drill string state estimation according to claim 1 or 2, characterized in that, The mechanical drilling speed model is as follows: (22)； Among them, mechanical drilling rate ROP mm / s A 1 =k d This is the coal and rock strata coefficient. A 2 =M Threshold drilling pressure, A 3 =λ The speed index; N , where is the rotational speed, in r / min.

4. The multi-objective optimization method for tunnel rotary drilling based on drill string state estimation according to claim 1 or 2, characterized in that, The drill bit wear model is as follows: (23)； Where D1 and D2 are the drilling pressure influence coefficients; Q1 and Q2 are the rotation speed influence coefficients; A f C1 is the abrasiveness coefficient of the formation; C2 is the tooth wear reduction coefficient. N Rotational speed, r / min W Drilling pressure, MPa h This represents the relative wear of the drill bit's cutting teeth.

5. The multi-objective optimization method for tunnel rotary drilling based on drill string state estimation according to claim 1 or 2, characterized in that, The parameter optimization of the drilling parameter multi-objective optimization model using the non-dominated sorting genetic algorithm includes: Using drill bit rotation speed and drilling pressure as decision variables, and comprehensively considering the mechanical drilling speed model and the drill bit wear model, a multi-objective optimization model is established. The parameter optimization can be expressed by the mathematical model of equation (24): (24)； in, h This represents the relative wear of the drill bit's cutting teeth. ROP The mechanical drilling speed is ____ mm / s; the tooth wear rate is ____. ; h The relative wear of the drill bit cutting teeth; drill bit rotation speed. N r / min; drilling pressure W , MPa.

Images