Submarine pipeline mud landing point intelligent monitoring system control device and implementation method thereof

Abstract

The invention relates to the field of underwater robots, in particular to an intelligent monitoring system control device for a subsea pipeline mud landing point and an implementation method of the intelligent monitoring system control device. The method comprises the following steps: linearizing a nonlinear model of the underwater robot through Taylor expansion, and inputting a linear part of the nonlinear model into a model prediction controller; the linearized model error and ocean current disturbance are combined into composite disturbance, and a nonlinear disturbance observer is used for estimating the composite disturbance; and inputting the obtained estimator and the linearized model into a model prediction controller, introducing the robot pose acquired by the sensor and the actual robot pose obtained by calculation, establishing a cost function, and obtaining the optimal solution of the cost function by adopting a # imgabs0 # fish optimization algorithm to obtain the generalized thrust. A large amount of computing resources are saved, the composite interference is effectively estimated, the robustness of the implementation method of the control device is improved, the pose error in the task process of the robot is reduced, and the intelligent level of the system is improved.

Description

Technical Field

[0001] The present invention relates to the field of radio detection technology, and in particular to a control device for an intelligent monitoring system for a mud point of a submarine pipeline and an implementation method thereof. Background Art

[0002] Currently, the most common approach for submarine pipeline laying operations in water depths less than 200 meters is to combine an unmanned, autonomous, remotely controlled underwater robot (UAV) with a multi-purpose mother vessel. However, in deep water (greater than 200 meters), this approach not only increases operational risks but also increases the negative impact of current interference and the cross-operation of other vessels, making it more difficult to operate. Therefore, combining an UAV with an unmanned surface vessel (USV) not only reduces operational costs and risks, but also increases the system's intelligence and simplifies mud point monitoring.

[0003] The interference of ocean currents will still cause unnecessary position changes of unmanned autonomous underwater robots when conducting mud spot monitoring, affecting the monitoring of mud spots under the sea pipe.

[0004] Therefore, a control device for an intelligent monitoring system for mud points at offshore pipelines and an implementation method thereof are provided. Summary of the Invention

[0005] The object of the present invention is to provide a control device for an intelligent monitoring system for mud points of submarine pipelines and an implementation method thereof, so as to solve the problem proposed in the above-mentioned background technology that the interference of ocean currents still causes unnecessary position changes of unmanned autonomous underwater robots when performing mud point monitoring, thereby affecting the monitoring of mud points of submarine pipelines.

[0006] To achieve the above object, the present invention provides a method for implementing an intelligent monitoring system for mud points of submarine pipelines, comprising the following steps:

[0007] S1. Linearize the nonlinear model of the underwater robot through Taylor expansion and input its linear part into the model predictive controller;

[0008] S2, combining the linearized nonlinear model error of the underwater robot and the ocean current disturbance into a composite disturbance, and using a nonlinear disturbance observer to estimate the composite disturbance to obtain an estimated compensation amount;

[0009] S3, input the estimated compensation amount obtained in S2 and the nonlinear model of the underwater robot after linearization in S1 into the model predictive controller, and bring in the robot posture collected by the sensor and the actual posture of the robot calculated, establish the cost function, and use The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust;

[0010] S4. To address the phenomenon of thrust delay caused by differences in communication structure and equipment, an improved grasshopper optimization algorithm combined with model prediction is used to calculate the thrust of each thruster.

[0011] As a further improvement of this technical solution, the specific steps of linearizing the model in S1 are:

[0012] Enter the robot's dynamic model and enter it in the following form:

[0013] c=f(a,b);

[0014] Among them, a = (x, y, z, φ, θ, ψ); b = (u, v, w, p, q, r);

[0015] Where c represents the rate of change of state; f(a, b) represents the relationship between the state change and the current state and control input in the robot dynamics model; a represents the robot's six-degree-of-freedom state variable, x represents the robot's position along the X-axis in the global coordinate system, y represents the robot's position along the Y-axis in the global coordinate system, z represents the robot's position along the Z-axis in the global coordinate system, φ represents the robot's roll angle in the global coordinate system, θ represents the robot's pitch angle in the global coordinate system, and ψ represents the robot's yaw angle in the global coordinate system; b represents the robot's six-degree-of-freedom velocity variable, u represents the robot's velocity component along the X-axis in the body coordinate system, v represents the robot's velocity component along the Y-axis in the body coordinate system, w represents the robot's velocity component along the Z-axis in the body coordinate system, p represents the robot's angular velocity component around the X-axis of the body coordinate system, q represents the robot's velocity component along the Y-axis in the body coordinate system, and r represents the robot's velocity component along the Z-axis in the body coordinate system;

[0016] Set c=f(a,b) in (a d ,b d ) and ignore the second-order and higher terms, then:

[0017]

[0018] Where a d represents the value of the state variable at the operating point; b d Represents the value of the speed variable at the operating point.

[0019] As a further improvement of the present technical solution, in S2, the linearized nonlinear model error of the underwater robot and the ocean current disturbance are combined into a composite disturbance, and the composite disturbance is estimated using a nonlinear disturbance observer.

[0020] As a further improvement of this technical solution, in S3, the compensation amount obtained in S2 and the linearized model in S1 are input into the model prediction controller, and the robot posture collected by the sensor and the actual posture of the robot calculated are brought in to establish a cost function, and the cost function is adopted. The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust, where the mathematical expression of the cost function is:

[0021]

[0022] Where N represents the prediction time domain, that is, the system behavior is predicted for the next N steps; e(k+i|k) represents the difference between the expected posture at the future time k+i predicted at the current time k and the actual posture; Q1 is used to assign different weights to different components of the error; T represents the input variable of the thruster; and Q2 is used to assign different weights to different components of the control input.

[0023] As a further improvement of this technical solution, in S3, the specific steps of optimizing and solving the cost function are as follows:

[0024] S3.1, yes Initialize the fish population size, number of iterations, individual initial positions and adsorption hosts;

[0025] S3.2. Generate a random number for each crucian carp to determine the value of H(i). the type of host to which the fish attach;

[0026] S3.3, Fish individuals update their positions based on the type of host they attach to;

[0027] S3.4, Fish according to the previous generation The fish's position relative to the current host's location, with small movements around the host;

[0028] S3.5, Fish decide whether to switch hosts based on a cost function;

[0029] S3.6. Repeat step S3.2 until the maximum number of iterations is reached to obtain the specific thrust value F of the robot propeller.

[0030] As a further improvement of the present technical solution, in S4, in order to address the phenomenon of thrust delay of thrusters caused by differences in communication structure and equipment, an algorithm combining an improved grasshopper optimization algorithm and a model prediction is used to calculate the thrust of each thruster, wherein the thrust distribution includes normal thrust distribution of thrusters based on the improved grasshopper optimization algorithm and delayed thrust distribution of thrusters based on the model prediction algorithm.

[0031] As a further improvement of this technical solution, in the normal thrust distribution of the thrust of the thruster based on the improved grasshopper optimization algorithm, the specific steps of allocating the thrust of the thruster to the normal thruster are as follows:

[0032] S4.1.1. Initialize the grasshopper population and randomly generate the grasshopper position X i , maximum number of iterations L, maximum adjustable coefficient c max and the minimum value c min ;

[0033] S4.1.2. Define the fitness function:

[0034] J=k1|F-(X i +X e )|+k2P(X i );

[0035] Where J represents the fitness function; k1 represents the weight coefficient of the control deviation term; F represents the total thrust expected by the underwater robot; X i Indicates the grasshopper position; X e represents the effect of ocean current disturbance on the propeller; P(X i ) represents the candidate solution X i Related penalty terms; k2 represents the weight coefficient of the control penalty term;

[0036] S4.1.3. Calculate the fitness of each individual at the current iteration number;

[0037] S4.1.4. Determine whether the current number of iterations, l, has reached the maximum number of iterations, l. If so, exit the loop and select the individual with the highest fitness function as the optimal solution for normal thruster thrust. If not, divide the grasshopper population into elite grasshoppers, spectator grasshoppers, and scout grasshoppers, and calculate the adjustable coefficient c for each.

[0038] S4.1.5, update the position of each grasshopper;

[0039] S4.1.6. Sort the grasshoppers according to their fitness function values, and the grasshoppers in the first half produce offspring according to the invasive weed optimization algorithm;

[0040] S4.1.7. Update the fitness function values of all individuals, iteration number l+1, and repeat step S4.1.4 until the maximum number of iterations is reached.

[0041] As a further improvement of this technical solution, in S4.1.4, the grasshopper population is divided into elite grasshoppers, bystander grasshoppers, and scout grasshoppers, wherein the mathematical expression involved in the elite grasshoppers is:

[0042]

[0043] Where c i Used to adjust the movement intensity of individual grasshoppers; c max represents the maximum value of the initial exploration intensity; c min Indicates the minimum value of the final development intensity; l indicates the current number of iterations; L indicates the maximum number of iterations; ζ is used to control c i The decay rate

[0044] The math involved in watching the grasshopper is:

[0045]

[0046] If the number of iterations l is less than Randomly generate adjustable coefficients; if the number of iterations l is greater than or equal to Then the mathematical expression involved in detecting grasshoppers is:

[0047] As a further improvement of the present technical solution, in the thrust distribution of the delayed thruster based on the model prediction algorithm, the specific steps of distributing thrust to the delayed thruster are:

[0048] S4.2.1. The mathematical expression for improving and discretizing the system model is:

[0049]

[0050] Where, X(k)=[x(k)v(k-1)], u(k)=v(k)-v(k-1), Y(k)=x(k), C(k)=[I 3×3 O 3×9 ];

[0051] Where X(k) represents the system state at the current time k; x(k) represents the position of the thruster; v(k-1) represents the thrust change; u(k) represents the control input at the current time k; v(k) represents the control input at the current time; Y(k) represents the observable state at the current time k; A(k) represents the coefficient matrix related to the control input; J(k) represents the Jacobian matrix; t represents the sampling time interval; I 6×6 represents the 6×6 unit matrix; B(k) represents the state transition matrix; O 6×6 represents a 6×6 zero matrix; C(k) represents the observation matrix; I 3×3 represents the 3×3 identity matrix; O 3×9 represents a 3×9 zero matrix;

[0052] S4.2.2. Define the sequence τ(k) of future control inputs and the sequence x(k) of future state variables;

[0053] S4.2.3. Define the relationship between the predicted state and the predicted input:

[0054] Z(k)=D(k)τ(k)+E(k)z(k);

[0055] in,

[0056]

[0057] E(k)=[B(k)B(k) 2 B(k) 3 B(k) n ];

[0058] Where Z(k) represents an element in the sequence z(k) of future state variables; D(k) represents the lower triangular matrix; E(k) represents the influence matrix of the current state on the predicted state;

[0059] S4.2.4. Define the discretized cost function:

[0060]

[0061] Where M(k) represents the discretized cost function; e(k+i|k) represents the state error at the future i-th sampling moment predicted at the current moment k; k3 represents the coefficient of the square term of the weighted prediction error; k4 represents the coefficient of the square term of the weighted control input;

[0062] S4.2.5. Perform rolling optimization based on the state variable constraints, control input constraints, and state change rate constraints to obtain the optimal thrust solution for the delayed thruster at the next moment.

[0063] In another aspect, the present invention provides a control device for an intelligent monitoring system for a mud point on a submarine pipeline, comprising a posture acquisition unit, a model linearization unit, a composite disturbance merging unit, a generalized thrust calculation unit, and a thruster thrust calculation unit. The control device for an intelligent monitoring system for a mud point on a submarine pipeline implements any one of the above-described methods for implementing an intelligent monitoring system for a mud point on a submarine pipeline:

[0064] Wherein, the posture acquisition unit acquires the posture information of the robot through sensors;

[0065] The model linearization unit linearizes the nonlinear model of the underwater robot through Taylor expansion while ignoring disturbances, and inputs the linear part into the model predictive controller;

[0066] The composite disturbance merging unit merges the linearized model error and the ocean current disturbance into a composite disturbance, and estimates the composite disturbance using a nonlinear disturbance observer;

[0067] The generalized thrust calculation unit inputs the estimated compensation amount and the linearized model into the model predictive controller, and brings in the robot posture collected by the sensor and the calculated actual posture of the robot to establish a cost function, and adopts The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust;

[0068] The thrust calculation unit calculates the thrust of each thruster by using an improved grasshopper optimization algorithm combined with model prediction to address the phenomenon of thrust delay caused by communication structure and equipment differences.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] Aiming at the goal of intelligent monitoring of pipeline mud impact points under current interference, the multi-thruster underwater robot control device is combined with advanced implementation methods. This greatly improves the maneuverability of the unmanned autonomous underwater robot in the deep sea, increases its ability to maintain the target posture at the target position, and improves the efficiency of video information acquisition, which has great engineering practical significance. It saves a lot of computing resources, effectively estimates complex interference, improves the robustness of the control device implementation method, reduces the posture error during the robot mission, and improves the intelligence level of the system. It overcomes the interference of a large number of complex seabed currents, can achieve precise control of multiple thrusters under multi-angle and complex current conditions, and realizes the difficult and multi-pose video capture target of pipeline mud impact points, expanding the applicable scenarios of underwater robot mud impact point detection and improving the operational efficiency of mud impact point detection. A thrust distribution strategy combining an improved grasshopper optimization algorithm with model predictive control is adopted to effectively deal with the uncertain thrust delay caused by the internal communication structure of the equipment, and reasonably distributes thrust considering the thrust limit of the thruster, thereby improving the success rate of the underwater robot mud impact point detection, extending the life of the thruster, reducing the failure rate, and increasing the unmanned intelligence level of the detection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The figure is a flow chart of the overall method of the present invention. DETAILED DESCRIPTION

[0072] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0073] Example 1: Please refer to Figure 1 As shown, this embodiment provides a method for implementing an intelligent monitoring system for marine pipeline mud points, including the following steps:

[0074] S1. Linearize the nonlinear model of the underwater robot through Taylor expansion and input its linear part into the model predictive controller;

[0075] S2, combining the linearized nonlinear model error of the underwater robot and the ocean current disturbance into a composite disturbance, and using a nonlinear disturbance observer to estimate the composite disturbance to obtain an estimated compensation amount;

[0076] S3, input the estimated compensation amount obtained in S2 and the nonlinear model of the underwater robot after linearization in S1 into the model predictive controller, and bring in the robot posture collected by the sensor and the actual posture of the robot calculated, establish the cost function, and use The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust;

[0077] S4. To address the phenomenon of thrust delay caused by differences in communication structure and equipment, an improved grasshopper optimization algorithm combined with model prediction is used to calculate the thrust of each thruster.

[0078] In this embodiment S1, the specific steps of model linearization are:

[0079] Enter the robot's dynamic model and enter it in the following form:

[0080] c=f(a,b);

[0081] Among them, a = (x, y, z, φ, θ, ψ); b = (u, v, w, p, q, r);

[0082] Where c represents the rate of change of state; f(a, b) represents the relationship between the state change and the current state and control input in the robot dynamics model; a represents the robot's six-degree-of-freedom state variable, x represents the robot's position along the X-axis in the global coordinate system, y represents the robot's position along the Y-axis in the global coordinate system, z represents the robot's position along the Z-axis in the global coordinate system, φ represents the robot's roll angle in the global coordinate system, θ represents the robot's pitch angle in the global coordinate system, and ψ represents the robot's yaw angle in the global coordinate system; b represents the robot's six-degree-of-freedom velocity variable, u represents the robot's velocity component along the X-axis in the body coordinate system, v represents the robot's velocity component along the Y-axis in the body coordinate system, w represents the robot's velocity component along the Z-axis in the body coordinate system, p represents the robot's angular velocity component around the X-axis of the body coordinate system, q represents the robot's velocity component along the Y-axis in the body coordinate system, and r represents the robot's velocity component along the Z-axis in the body coordinate system;

[0083] Set c=f(a,b) in (a d ,b d) and ignore the second-order and higher terms, then:

[0084]

[0085] Where a d represents the value of the state variable at the operating point; b d Represents the value of the speed variable at the operating point.

[0086] The underwater robot dynamics model takes into account linear motion and angular motion, and includes factors such as inertial force, Coriolis force, centripetal force, damping force, buoyancy and external force. The dynamics model is expressed as:

[0087]

[0088] In the formula, η = [x, y, z, φ, θ, ψ] T It represents the robot's posture vector, that is, the robot's six-degree-of-freedom state variable; Represents the first-order derivative (velocity) of the pose vector; represents the second derivative (acceleration) of the pose vector; M represents the mass matrix, which depends on the mass and inertia tensor of the robot and may also be a function of the pose in some cases; represents the Coriolis and centrifugal force matrices, which depend on the current velocity and attitude; D represents the damping term, which is usually a linear or nonlinear matrix proportional to the velocity; g(η) represents the generalized force vector due to gravity and buoyancy; τ represents the control input applied to the robot;

[0089] In this embodiment S2, the linearized nonlinear model error of the underwater robot and the ocean current disturbance are combined into a composite disturbance, and the composite disturbance is estimated using a nonlinear disturbance observer, wherein the disturbance in the X direction of the nonlinear disturbance observer is estimated as follows:

[0090]

[0091] in,

[0092] g x =1 / (mX u );

[0093] Where, represents the estimated compensation amount of disturbance in the X direction; γ x1 Represents the intermediate variable of the disturbance observer in the X direction, which is used to calculate the disturbance estimate in the X direction; represents the gain coefficient of the disturbance observer in the X direction, which is used to control γ x1 Dynamic response speed; u rrepresents the velocity component of the robot along the X axis in the body coordinate system; f1(u r ,v r ,r) represents the dynamic characteristics of the robot; g x T represents the coefficient related to the gravity and buoyancy of the disturbance observer in the X direction; x represents the component of thrust in the X direction; d x Representation based on internal state variables and observer gain The calculated disturbance estimate; γ x2 Represents the state variable of the disturbance observer in the X direction, which reflects the state of the system through a dynamic update mechanism and is used to assist in calculating the disturbance estimate d x ; represents the design parameters of the disturbance observer in the X direction, which is used to control The dynamic response speed; m represents the mass of the robot; v r represents the velocity component of the robot along the Y axis in the body coordinate system; r represents the angular velocity component of the robot around the Z axis in the body coordinate system; X u|u| represents the hydrodynamic resistance related to the square of velocity; represents the hydrodynamic resistance proportional to the speed; Indicates that due to the lateral velocity v r The hydrodynamic resistance caused by the coupling effect of the angular velocity r; X rr represents the hydrodynamic drag due to the angular velocity r around the Z axis; x g Indicates the offset of the center of gravity relative to the center of buoyancy in the X-axis direction; y g Indicates the offset of the center of gravity relative to the center of buoyancy in the Y-axis direction; X u represents the additional mass coefficient associated with the robot's motion along the X-axis;

[0094] The disturbance in the Y direction of the nonlinear disturbance observer is estimated as;

[0095]

[0096] Among them, f2(u r ,v r ,r)=(-mu r r+my g r 2 =mx g r+Y m v r ∣v r ∣+Y r|r| r|r|+Y r v r +Y r rY ur u r r+Ymm u r v r ) / (mY r );

[0097] g y =1 / (mY v );

[0098] Where, represents the estimated compensation amount of disturbance in the Y direction; γ y1 Represents the intermediate variable of the disturbance observer in the Y direction, which is used to calculate the disturbance estimate in the Y direction; represents the gain coefficient of the disturbance observer in the Y direction, which is used to control γ y1 Dynamic response speed; v r represents the velocity component of the robot along the Y axis in the body coordinate system; T y represents the component of thrust in the Y direction; d y Representation based on internal state variables and observer gain The calculated disturbance estimate; Represents the state variable of the disturbance observer in the Y direction, which is used to assist in calculating the disturbance estimate d y ; represents the design parameters of the disturbance observer in the Y direction, which is used to control γ y2 Dynamic response speed; Y m represents the hydrodynamic drag coefficient related to the square of velocity; Y r|r| represents the hydrodynamic drag coefficient related to the square of angular velocity; Y r represents the hydrodynamic resistance proportional to the speed; Y ur represents the hydrodynamic resistance caused by the coupling effect of the X-axis velocity and the angular velocity around the Z-axis; Y mm represents the hydrodynamic resistance caused by the coupling of X-axis velocity and Y-axis velocity; g y It represents the coefficient related to the gravity and buoyancy of the disturbance observer in the Y direction, which is used to convert the control input into the form of acceleration; Yv represents the additional mass coefficient related to the motion of the robot along the Y axis.

[0099] In this embodiment S3, the compensation amount obtained in S2 and the linearized model in S1 are input into the model predictive controller, and the robot posture collected by the sensor and the actual posture of the robot calculated are brought in to establish the cost function, and the cost function is adopted. The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust, where the mathematical expression of the cost function is:

[0100]

[0101] Where N represents the prediction time domain, that is, the system behavior is predicted for the next N steps; e(k+i|k) represents the difference between the expected posture at the future time k+i predicted at the current time k and the actual posture; Q1 is used to assign different weights to different components of the error; T represents the input variable of the thruster; and Q2 is used to assign different weights to different components of the control input.

[0102] In this embodiment S3, the specific steps of optimizing the cost function are as follows:

[0103] S3.1, yes Initialize the fish population size, number of iterations, individual initial positions and adsorption hosts;

[0104] S3.2. Generate a random number for each crucian carp to determine the value of H(i). The host type that the fish adsorbs, if H(i)=0, The fish is attached to the whale; if H(i)=1, Fish cling to sailfish;

[0105] S3.3, The fish individual updates its position according to the host type it is attached to; if Fish attached to the swordfish, then:

[0106]

[0107] Where, represents the new position of the i-th individual at the t+1-th iteration; represents the position of the best position found so far at iteration t; represents the position of an individual randomly selected from the population at the tth iteration;

[0108] like The fish clings to the whale, then:

[0109]

[0110] in,

[0111] Where, represents the next position of the i-th individual after the t-th iteration; D represents the distance between the current individual and the global optimal solution; s represents the scaling factor, which is used to control the change of the search range over time; k represents the exponential coefficient, which is used to adjust the influence of the scaling factor c on the search range; represents the current position of the i-th individual at the t-th iteration; a X Represents the iterative process of the system, which is used to optimize the changing trend of the control behavior in the algorithm; titeramax Indicates the maximum number of iterations;

[0112] S3.4, Fish according to the previous generation The fish's position and the current host's position move in a small range around the host. The fish's positions are:

[0113]

[0114] Where, Represents the new position of the current individual in the search space; Indicates the location of the current candidate solution; X p Indicates the target position; rand represents a random variable, which is used to control the randomness of movement;

[0115] S3.5, The fish determines whether to switch hosts based on the cost function. Fish change hosts; if The fish continued to feed around the host;

[0116] S3.6. Repeat step S3.2 until the maximum number of iterations is reached to obtain the specific thrust value F of the robot propeller.

[0117] In this embodiment S4, in order to address the phenomenon of thrust delay of thrusters caused by differences in communication structure and equipment, an algorithm combining the improved grasshopper optimization algorithm and model prediction is used to calculate the thrust of each thruster, wherein the thrust distribution includes normal thrust distribution of thrusters based on the improved grasshopper optimization algorithm and delayed thrust distribution of thrusters based on the model prediction algorithm.

[0118] In the normal thrust distribution of the thrusters based on the improved grasshopper optimization algorithm in this embodiment, the specific steps of distributing the thrust to the normal thrusters are as follows:

[0119] S4.1.1. Initialize the grasshopper population and randomly generate the grasshopper position X i , maximum number of iterations L, maximum adjustable coefficient c max and the minimum value c min ;

[0120] S4.1.2. Define the fitness function:

[0121] J=k1|F-(X i +X e )|+k2P(X i );

[0122] Where J represents the fitness function; k1 represents the weight coefficient of the control deviation term; F represents the total thrust expected by the underwater robot; X i represents the position of the grasshopper, that is, the current thrust distribution scheme of the underwater robot; Xe represents the impact of ocean current disturbance on the propeller; P(X i ) represents the candidate solution X i The relevant penalty term is used to constrain the feasibility of the thrust distribution scheme; k2 represents the weight coefficient of the control penalty term;

[0123] S4.1.3. Calculate the fitness of each individual at the current iteration number;

[0124] S4.1.4. Determine whether the current number of iterations, l, has reached the maximum number of iterations, l. If so, exit the loop and select the individual with the highest fitness function as the optimal solution for normal thruster thrust. If not, divide the grasshopper population into elite grasshoppers, spectator grasshoppers, and scout grasshoppers, and calculate the adjustable coefficient c for each.

[0125] S4.1.5. Update the position of each grasshopper:

[0126]

[0127] Where, represents the new position of the i-th grasshopper in the d-dimensional space; N represents the total number of grasshoppers in the population; ub d represents the upper bound of the d-dimensional space; lb d The lower bound of the d-th dimension; It represents the simulation of attraction and repulsion behavior between grasshoppers; represents the optimal solution of the objective function in dimension d, which is used to guide the grasshopper to the target position and ensure that the optimization process converges towards the desired solution. d represents the target value;

[0128] S4.1.6. Sort the grasshoppers by their fitness function values. The first half of the grasshoppers produce offspring using the invasive weed optimization algorithm. The number of offspring is:

[0129]

[0130] Where s represents the number of offspring; f i Indicates the fitness value of the current individual; f best Indicates the fitness value of the individual with the highest fitness value in the population; f worst Indicates the fitness value of the individual with the lowest fitness value in the population; s max Indicates the maximum number of offspring allowed; s min Indicates the minimum number of offspring allowed;

[0131] Then the initial position of the offspring is normally distributed as follows:

[0132]

[0133] Where, σ iter Represents the standard deviation of the normal distribution under the current iteration number l; q is used to control the decay rate of the standard deviation; σ init represents the initial standard deviation; σ final represents the final standard deviation;

[0134] Update the position of the new offspring based on a random walk strategy:

[0135] X new =randX fbest +randX sbest ;

[0136] Where, X new Indicates the position of the new offspring; X fbest Indicates the position of the individual with the highest fitness value in the current population; X sbest Indicates the position of the individual with the second highest fitness value in the current population;

[0137] S4.1.7. Update the fitness function values of all individuals, iteration number l+1, and repeat step S4.1.4 until the maximum number of iterations is reached.

[0138] In this embodiment S4.1.4, the grasshopper population is divided into elite grasshoppers, bystander grasshoppers, and scout grasshoppers, where the mathematical expression involved in the elite grasshoppers is:

[0139]

[0140] Where c i represents the social behavior coefficient of the i-th grasshopper, which is used to adjust the movement intensity of the individual grasshopper; c max represents the maximum value of the initial exploration intensity; c min represents the minimum value of the final development intensity; l represents the current number of iterations; L represents the maximum number of iterations; ζ represents a constant used to control c i The decay rate

[0141] The math involved in watching the grasshopper is:

[0142]

[0143] If the number of iterations l is less than Randomly generate adjustable coefficients; if the number of iterations l is greater than or equal to Then the mathematical expression involved in detecting grasshoppers is:

[0144] In the thrust distribution of the delayed thruster based on the model prediction algorithm in this embodiment, the specific steps of distributing thrust to the delayed thruster are:

[0145] S4.2.1. The mathematical expression for improving and discretizing the system model is:

[0146]

[0147] Where, X(k)=[x(k)v(k-1)], u(k)=v(k)-v(k-1), Y(k)=x(k), C(k)=[I 3×3 O 3×9 ];

[0148] Where X(k) represents the system state at the current time k; x(k) represents the position of the thruster; v(k-1) represents the thrust change; u(k) represents the control input at the current time k; v(k) represents the control input at the current time; Y(k) represents the observable state at the current time k; A(k) represents the coefficient matrix related to the control input, which is used to describe the influence of the control input on the system state; J(k) represents the Jacobian matrix; t represents the sampling time interval; I 6×6 represents the 6×6 unit matrix; B(k) represents the state transfer matrix, which is used to describe the relationship between the system states; O 6×6 represents a 6×6 zero matrix; C(k) represents the observation matrix, which is used to map the system state to the observable output; I 3×3 represents the 3×3 identity matrix; O 3×9 represents a 3×9 zero matrix;

[0149] S4.2.2. Define the sequence τ(k) of future control inputs and the sequence x(k) of future state variables:

[0150]

[0151] Where u(k+i|k) represents the control input at the future i-th sampling time predicted at the current time k, n represents the prediction time domain length; z(k) represents the sequence of future state variables; X(k+i|k) represents the system state at the future i-th sampling time predicted at the current time k;

[0152] S4.2.3. Define the relationship between the predicted state and the predicted input:

[0153] Z(k)=D(k)τ(k)+E(k)z(k);

[0154] in,

[0155]

[0156] E(k)=[B(k)B(k) 2 B(k) 3 B(k) n ];

[0157] Where Z(k) represents an element in the sequence z(k) of future state variables; D(k) represents the lower triangular matrix, which is used to control the influence of input on the predicted state; E(k) represents the influence matrix of the current state on the predicted state;

[0158] S4.2.4. Define the discretized cost function:

[0159]

[0160] Where M(k) represents the discretized cost function; e(k+i|k) represents the state error at the future i-th sampling moment predicted at the current moment k; k3 represents the coefficient of the weighted prediction error square term, which is used to penalize the degree to which the state deviates from the reference trajectory; k4 represents the coefficient of the weighted control input square term, which is used to penalize the energy consumption of the control input;

[0161] S4.2.5. Perform rolling optimization based on the state variable constraints, control input constraints, and state change rate constraints to obtain the optimal thrust solution for the delayed thruster at the next moment. The expressions for the state variable constraints, control input constraints, and state change rate constraints are:

[0162] The state variable constraints are:

[0163] x min ≤X(k+i|k)≤x max ;

[0164] Where x min Indicates the minimum value of the state variable; x max Indicates the maximum value of the state variable;

[0165] The control input constraints are:

[0166] u min ≤u(k+i|k)≤u max ;

[0167] Where x min Indicates the minimum value of the control input variable; x max Indicates the maximum value of the control input variable;

[0168] The state change rate constraint is:

[0169] Δx min ≤X(k+i+1|k)-X(k+i|k)≤Δxmax ;

[0170] Where X(k+i+1|k) represents the system state at the future i+1 sampling time predicted by the current time k; Δx min Indicates the minimum allowable value of state change; Δx max Indicates the maximum allowed value of the state change.

[0171] Example 2:

[0172] This embodiment provides a control device for an intelligent monitoring system for a mud point on a subsea pipeline, including a posture acquisition unit, a model linearization unit, a compound disturbance merging unit, a generalized thrust calculation unit, and a thruster thrust calculation unit. A method for implementing an intelligent monitoring system for a mud point on a subsea pipeline based on any one of the above-described control devices for the intelligent monitoring system for a mud point on a subsea pipeline is provided:

[0173] Among them, the posture acquisition unit collects the robot's posture information through sensors; the sensors include inertial navigation, DVL, depth gauge, altimeter, USBL and image processing computer; after the various parts of the system enter the working area, the unmanned autonomous underwater robot collects its own posture information through inertial navigation, DVL and depth gauge and other equipment, and uses the USBL base array installed on the surface unmanned ship for positioning correction to obtain accurate posture information, and sends its own streamlined status information to the pipe-laying ship in the form of underwater acoustic communication and receives control instructions; the unmanned autonomous underwater robot uses the assembled cameras and lights and other equipment to collect video information of the mud points, and transmits the video information to the repeater through underwater wireless optical communication; the unmanned autonomous underwater robot uses the assembled cameras and lights and other equipment to collect video information of the mud points, and transmits the video information to the repeater through underwater wireless optical communication; the repeater is equipped with optical fiber compass, depth gauge and other equipment to collect its own posture information, and uses the USBL base array installed on the surface unmanned ship for positioning correction to obtain accurate position information After receiving the video information using wireless optical communication, the repeater transmits the video information to the surface unmanned vessel through the optical fiber umbilical cable; the surface unmanned vessel is equipped with GPS, optical fiber compass and other equipment, which can collect its own precise positioning information, and correct the rough positioning information of the repeater and the unmanned autonomous underwater robot through USBL; the surface unmanned vessel receives the video information and status information transmitted by the repeater through the optical fiber umbilical cable, and at the same time, an antenna is installed on the top of the unmanned vessel, which sends the information to the pipe-laying vessel in the form of a wireless network, and receives the unmanned vessel operation instructions and repeater retraction and extension instructions sent by the pipe-laying vessel; the pipe-laying vessel receives the video information transmitted by the surface unmanned vessel through the radio station, and the main control computer in the pipe-laying vessel's control room, the unmanned vessel control computer, the repeater and the underwater robot control computer calculates it, and sends the next operation instructions to the surface unmanned vessel, the repeater and the unmanned autonomous underwater robot through a combination of hydroacoustic communication and wireless communication.

[0174] The model linearization unit linearizes the nonlinear model of the underwater robot through Taylor expansion and inputs its linear part into the model predictive controller;

[0175] The composite disturbance merging unit combines the linearized nonlinear model error of the underwater robot with the ocean current disturbance into a composite disturbance. The composite disturbance is estimated using a nonlinear disturbance observer to obtain an estimated compensation. The robot monitors the pipeline in the deep sea, so only the longitudinal and lateral water currents are considered, ignoring the influence of the vertical water current.

[0176] The generalized thrust calculation unit inputs the estimated compensation and the linearized nonlinear model of the underwater robot into the model predictive controller, and takes the robot posture collected by the sensor and the calculated actual posture of the robot into it to establish a cost function and adopt The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust;

[0177] The thrust calculation unit uses an improved grasshopper optimization algorithm combined with model prediction to calculate the thrust of each thruster to address the thrust delay caused by communication structure and equipment differences.

[0178] The underwater robot is equipped with eight thrusters, including four vertical thrusters, two main thrusters and two side thrusters. The thruster layout is as follows:

[0179] The distance between the two front vertical pushers and the front end of the robot body is a, the distance between the front vertical pusher and the rear vertical pusher on the same side is b, the distance between the rear vertical pusher and the rear end of the robot body is c, the spacing between the left and right vertical pushers is d, the spacing between the left vertical pusher and the left side of the robot body is e, and the same applies to the right side; the distance between the side pusher and the upper end of the robot body is p, the distance between the front side pusher and the front end of the robot body is h, the spacing between the two vertical pushers is i, and the distance between the rear vertical pusher and the rear end of the robot is j; the distance between the left main pusher and the left side of the robot is n, and the same applies to the right side, the spacing between the two main pushers is o, the angle between the main pusher and the XOZ plane is f, the distance between the main pusher and the upper side of the robot is q, and the same applies to the distance to the lower side; the distance between the upper side pusher and the upper side of the robot body is l, and the same applies to the distance to the lower side, the spacing between the two side pushers is m, the distance between the side pusher and the left side of the robot body is r, and the same applies to the distance to the right side.

[0180] The vertical thrusters and side thrusters use slotted thrusters, which can change the complexity of the flow field of the propeller inlet flow and increase the propulsion efficiency of the thrusters; the main thrusters use ducted thrusters, which can reduce the energy loss caused by the propeller tail vortex and the blade loss of the propeller blades, thereby improving the propulsion efficiency of the thrusters.

[0181] The control devices of the intelligent detection system all use forward and reverse thrusters, which significantly improve the movement ability of unmanned autonomous underwater robots and the monitoring level of offshore pipeline mud points, thereby improving the monitoring efficiency of offshore pipeline mud points.

[0182] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for implementing an intelligent monitoring system for mud points on offshore pipelines, characterized in that: The following steps are involved: S1. Linearize the nonlinear model of the underwater robot through Taylor expansion and input its linear part into the model predictive controller; S2, combining the linearized nonlinear model error of the underwater robot and the ocean current disturbance into a composite disturbance, and using a nonlinear disturbance observer to estimate the composite disturbance to obtain an estimated compensation amount; S3, input the estimated compensation amount obtained in S2 and the nonlinear model of the underwater robot after linearization in S1 into the model predictive controller, and bring in the robot posture collected by the sensor and the actual posture of the robot calculated, establish the cost function, and use The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust; S4. To address the phenomenon of thrust delay caused by differences in communication structure and equipment, an improved grasshopper optimization algorithm combined with model prediction is used to calculate the thrust of each thruster.

2. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 1 is characterized in that: The specific steps of model linearization in S1 are: Enter the robot's dynamic model and enter it in the following form: c=f(a,b); Among them, a = (x, y, z, φ, θ, ψ); b = (u, v, w, p, q, r); Where c represents the rate of change of state; f(a, b) represents the relationship between the state change and the current state and control input in the robot dynamics model; a represents the robot's six-degree-of-freedom state variable, x represents the robot's position along the X-axis in the global coordinate system, y represents the robot's position along the Y-axis in the global coordinate system, z represents the robot's position along the Z-axis in the global coordinate system, φ represents the robot's roll angle in the global coordinate system, θ represents the robot's pitch angle in the global coordinate system, and ψ represents the robot's yaw angle in the global coordinate system; b represents the robot's six-degree-of-freedom velocity variable, u represents the robot's velocity component along the X-axis in the body coordinate system, v represents the robot's velocity component along the Y-axis in the body coordinate system, w represents the robot's velocity component along the Z-axis in the body coordinate system, p represents the robot's angular velocity component around the X-axis of the body coordinate system, q represents the robot's velocity component along the Y-axis in the body coordinate system, and r represents the robot's velocity component along the Z-axis in the body coordinate system; Set c=f(a,b) in (a d ,b d ) and ignore the second-order and higher terms, then: Where a d represents the value of the state variable at the operating point; b d Represents the value of the speed variable at the operating point.

3. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 1 is characterized in that: In S2, the linearized nonlinear model error of the underwater robot and the ocean current disturbance are combined into a composite disturbance, and the composite disturbance is estimated using a nonlinear disturbance observer.

4. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 1 is characterized in that: In S3, the compensation amount obtained in S2 and the linearized model in S1 are input into the model predictive controller, and the robot posture collected by the sensor and the actual posture of the robot calculated are brought in to establish the cost function, and the cost function is used. The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust, where the mathematical expression of the cost function is: Where N represents the prediction time domain, that is, the system behavior is predicted for the next N steps; e(k+i|k) represents the difference between the expected posture at the future time k+i predicted at the current time k and the actual posture; Q1 is used to assign different weights to different components of the error; T represents the input variable of the thruster; and Q2 is used to assign different weights to different components of the control input.

5. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 1 is characterized in that: In S3, the specific steps for optimizing the cost function are as follows: S3.1, yes Initialize the fish population size, number of iterations, individual initial positions and adsorption hosts; S3.

2. Generate a random number for each crucian carp to determine the value of H(i). the type of host to which the fish attach; S3.3, Fish individuals update their positions based on the type of host they attach to; S3.4, Fish according to the previous generation The fish's position relative to the current host's location, with small movements around the host; S3.5, Fish decide whether to switch hosts based on a cost function; S3.

6. Repeat step S3.2 until the maximum number of iterations is reached to obtain the specific thrust value F of the robot propeller.

6. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 1 is characterized in that: In S4, in order to address the phenomenon of thrust delay of thrusters caused by differences in communication structure and equipment, an algorithm combining an improved grasshopper optimization algorithm and a model prediction is used to calculate the thrust of each thruster, wherein the thrust distribution includes normal thrust distribution of thrusters based on the improved grasshopper optimization algorithm and delayed thrust distribution of thrusters based on the model prediction algorithm.

7. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 6 is characterized in that: In the normal thrust distribution of the thrust of the thruster based on the improved grasshopper optimization algorithm, the specific steps of allocating the thrust of the thrust of the thruster are as follows: S4.1.

1. Initialize the grasshopper population and randomly generate the grasshopper position X i , maximum number of iterations L, maximum adjustable coefficient c max and the minimum value c min ; S4.1.

2. Define the fitness function: J=k1|F-(X i +X e )|+k2P(X i ); Where J represents the fitness function; k1 represents the weight coefficient of the control deviation term; F represents the total thrust expected by the underwater robot; X i Indicates the grasshopper's position; X e represents the effect of ocean current disturbance on the propeller; P(X i ) represents the candidate solution X i Related penalty terms; k2 represents the weight coefficient of the control penalty term; S4.1.

3. Calculate the fitness of each individual at the current iteration number; S4.1.

4. Determine whether the current number of iterations l has reached the maximum number of iterations l. If so, exit the loop and select the individual with the highest fitness function as the optimal solution for normal thruster thrust; If not exceeded, the grasshopper population is divided into elite grasshoppers, bystander grasshoppers and scout grasshoppers, and their respective adjustable coefficients c are calculated; S4.1.5, update the position of each grasshopper; S4.1.

6. Sort the grasshoppers according to their fitness function values, and the grasshoppers in the first half produce offspring according to the invasive weed optimization algorithm; S4.1.

7. Update the fitness function values of all individuals, iteration number l+1, and repeat step S4.1.4 until the maximum number of iterations is reached.

8. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 7 is characterized in that: In S4.1.4, the grasshopper population is divided into elite grasshoppers, bystander grasshoppers, and scout grasshoppers, where the mathematical expression involved in the elite grasshoppers is: Where c i represents the social behavior coefficient of the i-th grasshopper; c max represents the maximum value of the initial exploration intensity; c min Indicates the minimum value of the final development intensity; l indicates the current number of iterations; L indicates the maximum number of iterations; ζ represents a constant used to control c i The decay rate The math involved in watching the grasshopper is: If the number of iterations l is less than Randomly generate adjustable coefficients; If the number of iterations l is greater than or equal to Then the mathematical expression involved in detecting grasshoppers is:

9. The method for implementing the intelligent monitoring system for mud points of submarine pipelines according to claim 6 is characterized in that: In the thrust distribution of the delayed thruster based on the model prediction algorithm, the specific steps of distributing thrust to the delayed thruster are: S4.2.

1. The mathematical expression for improving and discretizing the system model is: where X(k) = [x(k) v(k - 1)], u(k) = v(k) - v(k - 1), Y(k) = x(k), C(k) = [I 3×3 O 3×9 ; Where X(k) represents the system state at the current time k; x(k) represents the position of the thruster; v(k-1) represents the thrust change; u(k) represents the control input at the current time k; v(k) represents the control input at the current time; Y(k) represents the observable state at the current time k; A(k) represents the coefficient matrix related to the control input; J(k) represents the Jacobian matrix; t represents the sampling time interval; I 6×6 represents the 6×6 unit matrix; B(k) represents the state transition matrix; O 6×6 represents a 6×6 zero matrix; C(k) represents the observation matrix; I 3×3 represents the 3×3 identity matrix; O 3×9 represents a 3×9 zero matrix; S4.2.

2. Define the sequence τ(k) of future control inputs and the sequence x(k) of future state variables; S4.2.

3. Define the relationship between the predicted state and the predicted input: Z(k)=D(k)τ(k)+E(k)z(k); in, E(k)=[B(k)B(k) 2 B(k) 3 B(k) n ]; Where Z(k) represents an element in the sequence z(k) of future state variables; D(k) represents the lower triangular matrix; E(k) represents the influence matrix of the current state on the predicted state; S4.2.

4. Define the discretized cost function: Where M(k) represents the discretized cost function; e(k+i|k) represents the state error at the future i-th sampling moment predicted at the current moment k; k3 represents the coefficient of the square term of the weighted prediction error; k4 represents the coefficient of the square term of the weighted control input; S4.2.

5. Perform rolling optimization based on the state variable constraints, control input constraints, and state change rate constraints to obtain the optimal thrust solution for the delayed thruster at the next moment.

10. A control device for an intelligent monitoring system for a mud point on a submarine pipeline, comprising a posture acquisition unit, a model linearization unit, a composite disturbance merging unit, a generalized thrust calculation unit, and a propeller thrust calculation unit, characterized in that: The control device of the intelligent monitoring system for mud points of submarine pipelines is used to implement the method for implementing the intelligent monitoring system for mud points of submarine pipelines as claimed in any one of claims 1 to 9: Wherein, the posture acquisition unit acquires the posture information of the robot through sensors; The model linearization unit linearizes the nonlinear model of the underwater robot through Taylor expansion and inputs its linear part into the model predictive controller; The composite disturbance merging unit merges the linearized nonlinear model error of the underwater robot and the ocean current disturbance into a composite disturbance, and uses the nonlinear disturbance observer to estimate the composite disturbance to obtain the estimated compensation amount; The generalized thrust calculation unit inputs the estimated compensation and the linearized nonlinear model of the underwater robot into the model predictive controller, and takes the robot posture collected by the sensor and the calculated actual posture of the robot into it to establish a cost function and adopt The fish optimization algorithm obtains the optimal solution of the cost function and obtains the generalized thrust; The thrust calculation unit calculates the thrust of each thruster by using an improved grasshopper optimization algorithm combined with model prediction to address the phenomenon of thrust delay caused by communication structure and equipment differences.

Images