Self-adaptive control system of pipeline cleaning robot

Abstract

The invention relates to the technical field of adaptive control, in particular to a pipeline cleaning robot adaptive control system which comprises a fluid sensing module, a path optimization module, a position calibration module, a pressure adjusting module and an effect evaluation module. According to the method, the laminar flow index, the turbulence intensity coefficient and the backflow probability value of the fluid are obtained through feature comparison, the flowing mode in the pipeline is mastered, the movement path and cleaning parameters of the robot are dynamically adjusted, the working requirements under different pipeline conditions are met, and the movement stability and efficiency of the robot in the complex pipeline environment are ensured; the actual motion posture and position of the robot are analyzed in real time and compared with a preset track, the positioning precision of the robot is improved, the pressure parameter of the cleaning nozzle and the torque compensation value of the rotating motor are dynamically adjusted, and accurate control over the cleaning pressure is ensured; and the insufficient cleaning position is recognized, the cleaning time is adjusted, and the stable and good cleaning effect is kept.

Description

Technical Field

[0001] The present invention relates to the technical field of adaptive control, and particularly to an adaptive control system for a pipeline cleaning robot. Background Art

[0002] The technical field of adaptive control includes control technologies with self-adjusting capabilities in automatic control systems. In a dynamic system, when system parameters or the external environment change, the parameters of the controller are automatically adjusted to maintain the stability of the system performance. It is applied to scenarios that need to handle system uncertainties and changes, including aerospace, robotics, automotive, and industrial automation. The core content is to obtain the state information of the system in real time and make corresponding adjustments to the controller to ensure that the system can maintain good control effects under changing working conditions.

[0003] Among them, the adaptive control system for a pipeline cleaning robot refers to a control system for robot operation that addresses possible environmental changes and task requirement changes in the actual application of the pipeline cleaning robot. The technical matters solved by this patent theme mainly include the specific application of adaptive control algorithms in robot cleaning tasks, especially how to cope with complex fluid environment changes inside the pipeline through adaptive control technology. The system uses sensors to monitor the changes inside the pipeline in real time and dynamically adjusts the robot's behavior and cleaning parameters through adaptive algorithms to ensure the stability of the cleaning effect. Specifically, an adaptive control algorithm based on a feedback mechanism is used to precisely control the robot to adapt to the working requirements under different pipeline conditions, and the best cleaning effect is achieved by adjusting parameters such as the movement path, cleaning pressure, and speed.

[0004] The current pipeline cleaning robot is usually composed of a cleaning nozzle, a rotating motor, a drive system, a sensor, etc. When dealing with changes in the complex fluid environment inside the pipeline, the traditional pipeline cleaning robot control technology uses fixed preset parameters for path control, and cannot be dynamically adjusted according to factors such as the fluid state of the pipeline, the radius of the bend, and the change in the pipe diameter. It is difficult to adapt to the significant changes in the fluid state in the pipeline due to changes in flow rate, pressure, and temperature, resulting in unstable cleaning effects. The cleaning effect is poor in complex environments, and the control accuracy of the posture and position is low. In the bends or narrow areas of the pipeline, it is easy to deviate from the predetermined path, affecting the cleaning quality. In a pipeline environment with high turbulence intensity or backflow, the robot's motion path and cleaning pressure cannot be adjusted in time, resulting in uneven cleaning or substandard cleaning effects. There are deficiencies in robot position calibration and posture control, and the lack of real-time feedback and precise adjustment capabilities for the robot's actual motion posture and position leads to positioning deviations and motion instability of the robot in complex pipelines. There is a lack of an effective cleaning effect evaluation mechanism, and the cleaning strategy cannot be adjusted in real time, resulting in some areas not being thoroughly cleaned. When dealing with complex pipeline environments, it is difficult to ensure the stability and consistency of the cleaning effect, limiting its application in a wider range of scenarios. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an adaptive control system for a pipeline cleaning robot.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: The pipeline cleaning robot adaptive control system comprises:

[0007] The fluid sensing module monitors the state of the pipeline fluid, obtains the instantaneous flow velocity sequence, dynamic pressure waveform, and pipe wall temperature value, calculates the flow velocity fluctuation value, pressure gradient value, and temperature change value, and calculates the fluid laminar index, turbulence intensity coefficient, and backflow probability value at multiple locations in the pipeline through feature comparison to obtain the flow pattern coefficient;

[0008] The path optimization module calls the flow pattern coefficient, uses the turbulence intensity coefficient and the backflow probability value, combines the curvature radius of the curves at multiple locations of the pipeline, and the pipe diameter contraction ratio parameters, calculates the path curvature adjustment amount and the movement speed, obtains the curve deceleration ratio and the straight pipe section acceleration threshold, and generates the path trajectory parameters;

[0009] The position calibration module calls the path trajectory parameters, uses the inertial measurement unit and the gyroscope to analyze the actual motion posture and position of the robot, calculates the position offset and angle deviation by comparing with the preset trajectory, obtains the horizontal offset correction and the posture compensation angle, and generates the position compensation parameters;

[0010] The pressure regulation module calls the position compensation parameter, uses the horizontal offset correction amount, combines with the wall contact pressure value collected by the pressure sensor in real time, calculates the difference between the expected value and the measured value of the contact pressure, adjusts the pressure parameter of the cleaning nozzle and the torque compensation value of the rotating motor, and generates the cleaning pressure parameter.

[0011] Based on the cleaning pressure parameter, the effect evaluation module uses the image acquisition device, by comparing the image sequences before and after the pipe wall cleaning, extracts the residue coverage area ratio and the texture clarity difference value, identifies the insufficient cleaning positions and adjusts the cleaning duration, and generates the pipeline cleaning operation parameter.

[0012] As a further solution of the present invention, the flow pattern coefficient includes a fluid laminar flow index, a turbulence intensity coefficient, and a recirculation probability value, the path trajectory parameter includes the robot working path, the turning driving speed, and the straight pipe forward speed threshold, the position compensation parameter includes a position offset correction amount, an angle offset correction amount, and position offset data, the cleaning pressure parameter includes the cleaning nozzle control parameter and the rotating motor control parameter, and the pipeline cleaning operation parameter includes the residue coverage area ratio, the texture clarity difference value, and the cleaning time adjustment value.

[0013] As a further solution of the present invention, the fluid perception module includes:

[0014] The feature extraction sub-module monitors the pipeline fluid state, obtains the instantaneous flow rate sequence by using the flow rate sensor, obtains the dynamic pressure waveform by using the pressure sensor, obtains the wall temperature value by using the temperature sensor, calculates the flow rate fluctuation value, the pressure gradient value, and the temperature change value under the current fluid state, and establishes the fluid state feature vector.

[0015] The feature comparison sub-module calls the fluid state feature vector, extracts the feature vector data of multiple known fluid states, and adopts the formula:

[0016]

[0017] Judges the closeness between the current fluid state and the known state, and obtains the feature similarity score.

[0018] Among them, S represents the feature similarity score, w i represents the weight of each feature item, X i represents the i-th feature value of the current state, represents the i-th feature value under the known fluid state, n is the total number of features, and i represents the i-th feature item in the feature vector.

[0019] Based on the feature similarity score, the state calculation sub-module identifies the flow states at multiple positions in the pipeline by comparing with various known fluid states, extracts the corresponding laminar flow index, turbulent intensity coefficient, and recirculation probability value of the fluid, and obtains the flow pattern coefficient.

[0020] As a further solution of the present invention, the path optimization module includes:

[0021] The geometric feature extraction sub-module calls the flow pattern coefficient, and uses the geometric feature information of the pipeline to extract the bend curvature radius and pipe diameter contraction ratio parameters at multiple positions in the pipeline, and obtains the geometric feature analysis result;

[0022] Based on the geometric feature analysis result, the path adjustment calculation sub-module calls the turbulent intensity coefficient and recirculation probability value, combines the bend curvature radius and pipe diameter contraction ratio, and uses the formula:

[0023]

[0024] Performs path curvature adjustment calculation, obtains the path curvature adjustment amount, and generates path adjustment parameters;

[0025] Among them, C adjust is the path curvature adjustment amount, T curvature is the bend curvature radius, R is the response coefficient of the flow, P shrink is the pipe diameter contraction ratio, and L is the length of the flow path;

[0026] The motion trajectory planning sub-module calculates the motion speed of the robot at multiple positions in the pipeline according to the path adjustment parameters and combines the flow characteristics, obtains the bend deceleration ratio and the straight pipe section acceleration threshold, and obtains the path trajectory parameters.

[0027] As a further solution of the present invention, the position calibration module includes:

[0028] The attitude and position analysis sub-module calls the path trajectory parameters, and by analyzing the measurement data of the inertial measurement unit and the gyroscope, identifies the actual motion position and attitude of the robot, records the current coordinates and angles of the robot, and obtains the real-time position data of the robot;

[0029] The horizontal offset correction quantum module uses the real-time position data of the robot, according to the actual coordinates of the robot, by comparing with the preset trajectory, uses the formula:

[0030]

[0031] Obtains the horizontal offset correction amount;

[0032] Among them, D offset is the horizontal offset correction amount, X actis the actual X coordinate of the robot, X ide is the preset X coordinate, Y act is the actual Y coordinate of the robot, Y ide is the preset Y coordinate;

[0033] Based on the horizontal offset correction amount, the attitude compensation angle sub-module uses the actual attitude of the robot, compares it with the preset trajectory, calculates the attitude compensation angle of the robot, adjusts the angular difference between the robot and the preset trajectory, and generates position compensation parameters.

[0034] As a further solution of the present invention, the pressure adjustment module includes:

[0035] The pressure difference calculation sub-module adjusts the deviation between the robot and the preset trajectory according to the position compensation parameter, combines the real-time horizontal offset correction amount of the robot, combines the contact pressure value of the pipe wall obtained by the pressure sensor, calculates the difference between the measured pressure and the preset expected value, and obtains the pressure difference data;

[0036] The cleaning pressure adjustment sub-module is based on the pressure difference data and uses the formula:

[0037]

[0038] Calculates the adjusted cleaning nozzle pressure, adjusts the pressure parameter of the cleaning nozzle, and obtains the pressure parameter of the cleaning nozzle;

[0039] where P adj is the adjusted cleaning nozzle pressure, P des is the preset target contact pressure, ΔP' is the pressure difference, that is, the difference between the measured pressure and the target pressure, α is the correction coefficient used to adjust the influence of the pressure difference on the nozzle pressure adjustment, μ is the fluid viscosity coefficient, and D is the nozzle diameter;

[0040] The rotary motor torque compensation sub-module further calculates the torque compensation value of the rotary motor according to the pressure parameter of the cleaning nozzle, considering the required pressure change of the cleaning nozzle, combines the load response characteristics of the motor, adjusts the motor output, and generates the cleaning pressure parameter.

[0041] As a further solution of the present invention, the effect evaluation module includes:

[0042] The image comparison sub-module is based on the cleaning pressure parameter, obtains the pipe wall images before and after cleaning through the image acquisition device, uses image analysis to extract the residue coverage area and texture clarity of multiple regions in the image, and generates the pipe wall image analysis result;

[0043] Based on the analysis result of the pipe wall image, the cleaning effect analysis sub-module calculates the residue coverage ratio and texture clarity difference value of multiple regions by comparing the residue coverage area and texture clarity in the images before and after cleaning, using the formula:

[0044]

[0045] Evaluate the cleaning effect at multiple positions and calculate the cleanliness score;

[0046] Among them, C tot is the cleanliness score, which measures the cleaning effect, A bf,i' is the residue coverage area of region i' before cleaning, A af,i' is the residue coverage area of region i' after cleaning, T bf,i' is the texture clarity of region i' before cleaning, T af,i' is the texture clarity of region i' after cleaning, w 1 is the weight coefficient of the residue coverage area, w 2 is the weight coefficient of the texture clarity, N' is the total number of regions divided in the image, and i' is the index of the region in the image;

[0047] The cleaning duration adjustment sub-module calls the cleanliness score, identifies the positions with insufficient cleaning, and adjusts the cleaning duration of the robot to generate pipeline cleaning operation parameters.

[0048] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0049] In the present invention, the laminar flow index, turbulent intensity coefficient, and recirculation probability value of the fluid are obtained through feature comparison, the flow pattern inside the pipeline is mastered, the movement path and cleaning parameters of the robot are dynamically adjusted to adapt to the working requirements under different pipeline conditions, ensuring that the robot can adaptively adjust its movement trajectory according to the flow characteristics inside the pipeline, solving the problem of path deviation of traditional robots, analyzing the actual movement posture and position of the robot in real time and comparing with the preset trajectory to improve the positioning accuracy of the robot, ensuring that the robot always maintains an accurate movement trajectory in a complex pipeline environment, combining the wall contact pressure value collected by the pressure sensor in real time, calculating the difference between the expected contact pressure value and the measured value, dynamically adjusting the pressure parameter of the cleaning nozzle and the torque compensation value of the rotating motor to ensure the precise control of the cleaning pressure, avoiding problems of incomplete cleaning or damage caused by too high or too low pressure, and the comparison of the image sequences before and after the pipe wall cleaning realizes the identification of the positions with insufficient cleaning and the adjustment of the cleaning duration, maintaining a stable cleaning effect in a complex and changeable pipeline environment, ensuring that each region can be cleaned with high quality, improving the efficiency and effect of the pipeline cleaning operation, and improving the cleaning efficiency and accuracy. Brief Description of the Drawings

[0050] Figure 1 It is the system flowchart of the present invention;

[0051] Figure 2 It is the flowchart of the fluid sensing module of the present invention;

[0052] Figure 3 It is the flowchart of the path optimization module of the present invention;

[0053] Figure 4 It is the flowchart of the position calibration module of the present invention;

[0054] Figure 5 It is the flowchart of the pressure regulation module of the present invention;

[0055] Figure 6 It is the flowchart of the effect evaluation module of the present invention. Detailed implementation manners

[0056] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0057] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. In addition, in the description of the present invention, the meaning of "plural" is two or more unless otherwise specifically defined.

[0058] Please refer to Figure 1 , the adaptive control system of the pipeline cleaning robot includes:

[0059] The fluid sensing module monitors the pipeline fluid state, obtains the instantaneous flow rate sequence, dynamic pressure waveform, and pipe wall temperature value, calculates the flow rate fluctuation value, pressure gradient value, and temperature change value, and calculates the laminar flow index, turbulent intensity coefficient, and recirculation probability value at multiple positions in the pipeline through feature comparison, and obtains the flow pattern coefficient;

[0060] The path optimization module calls the flow pattern coefficient, uses the turbulent intensity coefficient and recirculation probability value, combines the bend curvature radius and pipe diameter contraction ratio parameters at multiple positions of the pipeline, calculates the path curvature adjustment amount and movement speed, obtains the bend deceleration ratio and straight pipe section acceleration threshold, and generates the path trajectory parameters;

[0061] The position calibration module calls the path trajectory parameters, uses the inertial measurement unit and gyroscope to analyze the actual motion posture and position of the robot, calculates the position offset and angle deviation by comparing with the preset trajectory, obtains the horizontal offset correction amount and attitude compensation angle, and generates position compensation parameters;

[0062] The pressure regulation module calls the position compensation parameters, uses the horizontal offset correction amount, combines with the pipe wall contact pressure value collected by the pressure sensor in real time, calculates the difference between the expected value and the measured value of the contact pressure, adjusts the pressure parameters of the cleaning nozzle and the torque compensation value of the rotating motor, and generates cleaning pressure parameters;

[0063] The effect evaluation module is based on the cleaning pressure parameters, uses the image acquisition device, by comparing the image sequences before and after the pipe wall cleaning, extracts the residue coverage area ratio and texture clarity difference value, identifies the positions with insufficient cleaning and adjusts the cleaning duration, and generates pipeline cleaning operation parameters.

[0064] The flow pattern coefficients include the fluid laminar flow index, turbulent intensity coefficient, and recirculation probability value. The path trajectory parameters include the robot working path, corner driving speed, and straight pipe forward speed threshold. The position compensation parameters include the position offset correction amount, angle offset correction amount, and position offset data. The cleaning pressure parameters include the cleaning nozzle control parameters and rotating motor control parameters. The pipeline cleaning operation parameters include the residue coverage area ratio, texture clarity difference value, and cleaning time adjustment value.

[0065] Please refer to Figure 2 , the fluid sensing module includes:

[0066] The feature extraction sub-module monitors the pipeline fluid state, obtains the instantaneous flow rate sequence using the flow rate sensor, obtains the dynamic pressure waveform using the pressure sensor, obtains the pipe wall temperature value using the temperature sensor, calculates the flow rate fluctuation value, pressure gradient value, and temperature change value under the current fluid state, and establishes a fluid state feature vector;

[0067] In the fluid state monitoring, first obtain the relevant data of the fluid through the flow rate sensor, pressure sensor, and temperature sensor. Assume that the instantaneous flow rate sequence recorded by the flow rate sensor is [1.2, 1.3, 1.4, 1.2, 1.3] m / s, the dynamic pressure waveform recorded by the pressure sensor is [100, 102, 104, 103, 101] Pa, and the pipe wall temperature of the temperature sensor is [30.1, 30.2, 30.3, 30.1, 30.0] °C. The flow rate fluctuation value is calculated as, and the flow rate fluctuation value can be obtained by calculating the standard deviation of the flow rate sequence. The formula for the standard deviation is:

[0068]

[0069] where, V iis the i-th flow velocity value, is the mean value of the flow velocity sequence, and n is the length of the flow velocity sequence.

[0070] For the flow velocity sequence [1.2, 1.3, 1.4, 1.2, 1.3], first calculate the mean flow velocity

[0071]

[0072] Therefore, the flow velocity fluctuation value is 0.1 m / s.

[0073] For the calculation of the pressure gradient value, the pressure gradient value is obtained by calculating the pressure difference between adjacent measurement points. Assuming the distance between measurement points is Δx = 1 m, the pressure gradient formula is:

[0074]

[0075] where ΔP is the pressure difference between two adjacent measurement points, and Δx is the distance between measurement points.

[0076] Assume the pressure sequence is [100, 102, 104, 103, 101] (unit: Pa), and the distance between measurement points is assumed to be 1 meter. Then the pressure gradient value is calculated as follows:

[0077]

[0078] Therefore, the pressure gradient value is 2 Pa / m.

[0079] For the calculation of the temperature change value, the temperature change value is obtained by calculating the maximum change in temperature. The formula for the temperature change value is:

[0080] TC = max(T) - min(T);

[0081] where T is the pipe wall temperature sequence.

[0082] The temperature sequence is [30.1, 30.2, 30.3, 30.1, 30.0] (unit: °C), the maximum value is 30.3 °C, and the minimum value is 30.0 °C. So the temperature change value is:

[0083] TC = 30.3 - 30.0 = 0.3 °C;

[0084] Therefore, the temperature change value is 0.3 °C.

[0085] The feature comparison sub-module calls the fluid state feature vector, extracts the feature vector data of multiple known fluid states, and uses the formula:

[0086]

[0087] Judge the proximity between the current fluid state and the known state, and obtain the feature similarity score;

[0088] where S represents the feature similarity score, w i represents the weight of each feature item, X i represents the i-th eigenvalue of the current state, represents the i-th eigenvalue under the known fluid state, n is the total number of features, and i represents the i-th feature item in the feature vector;

[0089] In the process of feature comparison, first call the fluid state feature vector and compare it with the known fluid state feature vector. Suppose the known fluid state feature vectors are: known state 1: [0.12, 0.025, 0.09], known state 2: [0.08, 0.03, 0.12], known state 3: [0.1, 0.02, 0.1]. Suppose the weights of each feature are w 1 = 0.4, w 2 = 0.3, w 3 = 0.3, and the feature vector of the current fluid state is [0.1, 0.02, 0.1]. Substitute it into the formula to calculate the feature similarity score:

[0090] Calculate the similarity score of known state 1:

[0091] S 1 = 0.4·|0.1 - 0.12| + 0.3·|0.02 - 0.025| + 0.3·|0.1 - 0.09| = 0.0125;

[0092] For the calculation of known state 2:

[0093] S 2 = 0.4·|0.1 - 0.08| + 0.3·|0.02 - 0.03| + 0.3·|0.1 - 0.12|;

[0094] S 2 = 0.4·0.02 + 0.3·0.01 + 0.3·0.02 = 0.008 + 0.003 + 0.006 = 0.018;

[0095] For the calculation of known state 3:

[0096] S 3 = 0.4·|0.1 - 0.1| + 0.3·|0.02 - 0.02| + 0.3·|0.1 - 0.1|;

[0097] S 3 = 0.4·0 + 0.3·0 + 0.3·0 = 0;

[0098] As shown in Table 1, the example data shows the comparison of different fluid state characteristics and the calculated similarity scores.

[0099]

[0100] As shown in Table 1, the calculated feature similarity scores help identify the current fluid state.

[0101] The final calculation results show that the current fluid state is closest to the known state 3 because its feature similarity score is 0, meaning that the feature vector of the current fluid state exactly matches the feature vector of the known state 3. For the known state 1 and the known state 2, the similarity scores are 0.0125 and 0.018 respectively, indicating that their matching degrees with the current state are relatively low. Based on this, the system determines that the current fluid state is closest to the known state 3 and extracts the corresponding fluid laminar flow index, turbulent intensity coefficient, and recirculation probability value accordingly.

[0102] The state calculation sub-module, based on the feature similarity scores, identifies the flow states at multiple positions in the pipeline by comparing with multiple known fluid states, extracts the corresponding fluid laminar flow index, turbulent intensity coefficient, and recirculation probability value, and obtains the flow pattern coefficient;

[0103] According to the obtained feature similarity scores, the system will select the closest known fluid state and extract its corresponding fluid laminar flow index, turbulent intensity coefficient, and recirculation probability value. Suppose that in the closest known state, the fluid laminar flow index is 0.85, the turbulent intensity coefficient is 0.12, and the recirculation probability value is 0.07. The system will identify the current fluid state of the pipeline based on the results of the similarity scores. In this way, the system determines that the flow characteristics of the current fluid state are: fluid laminar flow index 0.85, turbulent intensity coefficient 0.12, and recirculation probability value 0.07. The final results provide the basis for determining the flow pattern through these fluid state parameters.

[0104] Please refer to Figure 3 , the path optimization module includes:

[0105] The geometric feature extraction sub-module calls the flow pattern coefficient and uses the geometric feature information of the pipeline to extract the bend curvature radius and pipe diameter contraction ratio parameters at multiple positions in the pipeline to obtain the geometric feature analysis results;

[0106] Call the flow pattern coefficient, utilize the geometric feature information of the pipeline, extract the bend curvature radius and pipe diameter shrinkage ratio parameters at multiple positions inside the pipeline. In practical applications, obtaining geometric feature data usually requires sensor measurement and calculation. First, arrange multiple curvature sensors inside the pipeline. Each sensor monitors the curvature information of a measurement point and records the curvature radius data. For example, at a certain pipeline position, the measured curvature radius is 2.5 m, and at another position, the curvature radius is 3.1 m. Meanwhile, utilize the pipe diameter sensors to obtain the pipe diameter data at different positions of the pipeline to calculate the pipe diameter shrinkage ratio parameter. Assume that the initial pipe diameter at a certain position is 0.8 m and the shrunk pipe diameter is 0.6 m. Then the pipe diameter shrinkage ratio is calculated as follows:

[0107]

[0108] Among them, P shrink represents the pipe diameter shrinkage ratio, D init represents the original pipe diameter, D new represents the current pipe diameter.

[0109] Substitute the values for calculation:

[0110]

[0111] It indicates that the pipe diameter at this position has shrunk by 25%. For the measurement point data at multiple positions, record the curvature radius and pipe diameter shrinkage ratio parameters of all measurement points and perform normalization processing to ensure that the data can be used for subsequent path adjustment calculations, and finally obtain the geometric feature analysis result.

[0112] Based on the geometric feature analysis result, the path adjustment calculation sub-module calls the turbulence intensity coefficient and the recirculation probability value, combines the bend curvature radius and the pipe diameter shrinkage ratio, and uses the formula:

[0113]

[0114] Perform path curvature adjustment calculation to obtain the path curvature adjustment amount and generate path adjustment parameters;

[0115] Among them, C adjust is the path curvature adjustment amount, T curvature is the bend curvature radius, R is the response coefficient of the flow, P shrink is the pipe diameter shrinkage ratio, and L is the length of the flow path;

[0116] Based on the geometric feature analysis results, the turbulence intensity coefficient and the recirculation probability value are called, and combined with the bend radius of curvature and the pipe diameter contraction ratio, the path curvature adjustment amount is calculated. During the pipeline flow optimization process, the flow characteristics at different positions determine the curvature change of the path. If the turbulence intensity is large and the recirculation probability is high at a certain position, the path adjustment amount should be increased to reduce flow instability. The path curvature adjustment amount is calculated using the formula:

[0117]

[0118] where C adjust is the path curvature adjustment amount, T curvature is the bend radius of curvature, R is the flow response coefficient, P shrink is the pipe diameter contraction ratio, and L is the length of the flow path.

[0119] Suppose in a certain pipeline section, the measured radius of curvature T curvature = 2.5 m, the pipe diameter contraction ratio P shrink = 0.25, the flow path length L = 5 m, and the flow response coefficient R = 1.8. Substitute into the formula for calculation:

[0120]

[0121] The calculated path curvature adjustment amount is 4.49. According to this adjustment amount, the path is corrected and the path parameters are optimized, and finally the path adjustment parameters are generated.

[0122] The motion trajectory planning sub-module calculates the motion speeds of the robot at multiple positions in the pipeline according to the path adjustment parameters and combines the flow characteristics, obtains the bend deceleration ratio and the straight pipe section acceleration threshold, and gets the path trajectory parameters;

[0123] According to the path adjustment parameters and combined with the flow characteristics in the pipeline, calculate the motion speeds of the robots at each position and adjust their speeds to match the flow state of the pipeline. For the bend area, due to the large curvature change, a deceleration ratio needs to be set to ensure that the robot can pass smoothly. Suppose the path curvature adjustment amount is 4.49, then the bend deceleration ratio is calculated as follows:

[0124]

[0125] where S bend is the bend deceleration ratio, and C adjust is the path curvature adjustment amount. Substitute into the calculation:

[0126]

[0127] That is, the bend deceleration ratio at this position is 18.2%, indicating that in the bend area, the robot should reduce its speed by 18.2%.

[0128] For the straight pipe section, an acceleration threshold needs to be set to increase the moving speed with lower resistance. Assuming the current flow velocity is 2 m / s and the turbulence intensity coefficient is 0.12, the acceleration threshold for the straight pipe section is calculated as follows:

[0129] A thresh = V flow × (1 + T intensity );

[0130] Where A thresh represents the acceleration threshold for the straight pipe section, V flow represents the flow velocity, and T intensity represents the turbulence intensity coefficient. Substitute the values for calculation:

[0131] A thresh = 2 × (1 + 0.12) = 2 × 1.12 = 2.24 m / s;

[0132] That is, the acceleration threshold for the straight pipe section is 2.24 m / s. Finally, combining the bend deceleration ratio and the straight pipe section acceleration threshold, calculate the trajectory adjustment at each position and optimize the movement mode of the robot in the pipeline to obtain the path trajectory parameters.

[0133] Table 2: Data related to path adjustment

[0134]

[0135] As shown in Table 2, the path curvature adjustment amount, bend deceleration ratio, and straight pipe section acceleration threshold at each position are calculated and used to optimize the robot's movement trajectory to ensure that the path adjustment is reasonable and matches the fluid characteristics.

[0136] Please refer to Figure 4 , the position calibration module includes:

[0137] The attitude and position analysis sub-module calls the path trajectory parameters, analyzes the measurement data of the inertial measurement unit and the gyroscope, identifies the actual movement position and attitude of the robot, records the current coordinates and angle of the robot, and obtains the real-time position data of the robot;

[0138] In this sub-module, first use the inertial measurement unit (IMU) and the gyroscope to obtain the real-time movement data of the robot, including the coordinates (X and Y axis coordinates) of the robot and its angle (attitude). The sensors continuously collect the dynamic data of the robot. By analyzing these data, the current movement state and position of the robot can be accurately judged. Through the inertial measurement unit, we can obtain the acceleration and angular velocity of the robot, while the gyroscope provides the angle information. Assume the current coordinates of the robot are X act = 5.2 m and Y act = 3.5 m, and the current angle is θ act= 45°, and these data are measured and obtained in real time by sensors. Next, the coordinates and angles at each moment recorded by the sensors will be stored and used for subsequent trajectory comparison. At this time, the robot compares its real-time coordinates and angles with the preset trajectory data to determine whether there is a deviation in its position on the current trajectory. Assume that the preset trajectory position of the robot is X ideal = 5.0 m and Y ideal = 3.0 m, while the actual position is X act = 5.2 m and Y act = 3.5 m. By comparison, it can be identified whether the robot deviates from the predetermined trajectory. The ultimate goal of this sub-module is to obtain the real-time position data of the robot and compare it with the preset trajectory to provide a data basis for subsequent calculation of the horizontal offset correction amount. Assume that the robot starts from the starting point and travels along a certain preset trajectory, and the path is formed by connecting multiple points. We need to monitor and record the position of the robot in real time so as to correct its traveling trajectory at any time. In this way, the robot can adjust its path at any time according to the real-time feedback provided by devices such as inertial sensors and gyroscopes to ensure that it can accurately travel to the target position. If the robot deviates at a certain moment, the position deviation can be timely identified through the analysis function of this sub-module, and the required data can be prepared for subsequent path adjustment and attitude compensation.

[0139] The sub-module for calculating the horizontal offset correction amount utilizes the real-time position data of the robot. According to the actual coordinates of the robot, by comparing with the preset trajectory, using the formula:

[0140]

[0141] the horizontal offset correction amount is obtained;

[0142] where D offset is the horizontal offset correction amount, X act is the actual X coordinate of the robot, X ide is the preset X coordinate, Y act is the actual Y coordinate of the robot, Y ide is the preset Y coordinate;

[0143] The function of this sub-module is to calculate the horizontal offset correction amount between the actual position data of the robot and the preset trajectory. The calculation formula is:

[0144]

[0145] where D offset is the horizontal offset correction amount, X act is the actual X coordinate of the robot, X ideal is the preset X coordinate, Y act is the actual Y coordinate of the robot, Y idealis the preset Y coordinate. The horizontal offset correction calculates the Euclidean distance between the current position of the robot and the predetermined position, reflecting the degree to which the robot deviates from the preset trajectory.

[0146] Assume the actual coordinates of the robot are X act = 5.2 m and Y act = 3.5 m, while the preset trajectory coordinates are X ideal = 5.0 m and Y ideal = 3.0 m. Then substitute these values into the formula for calculation:

[0147]

[0148] Therefore, the horizontal offset correction is 0.54 m, which means there is a deviation of 0.54 meters between the current position of the robot and the preset trajectory. The robot needs to make appropriate adjustments based on this deviation. This correction provides a basis for subsequent attitude compensation to ensure that the robot can travel accurately along the preset trajectory. If similar deviations continuously occur during the operation of the robot, by calculating the horizontal offset correction each time, the robot's path can be gradually adjusted to minimize the deviation from the preset trajectory. This adjustment process is dynamic, correcting the position of the robot in real time to ensure that it can finally reach the target position without errors or unstable operation due to continuous deviation.

[0149] The sub-module for calculating the attitude compensation angle, based on the horizontal offset correction and using the actual attitude of the robot, compares it with the preset trajectory to calculate the attitude compensation angle of the robot, adjusts the angular difference between the robot and the preset trajectory, and generates position compensation parameters;

[0150] This sub-module further corrects the movement direction of the robot by calculating the attitude compensation angle to ensure that it travels correctly along the preset trajectory. The calculation formula for the attitude compensation angle is:

[0151] θ comp = θ act - θ ideal ;

[0152] where θ comp is the attitude compensation angle, θ act is the actual angle of the robot, and θ ideal is the preset angle. The calculation of the attitude compensation angle is to correct the turning of the robot so that it can move forward in the correct direction.

[0153] Assume the actual angle of the robot is θ act = 45°, and the angle of the preset trajectory is θ ideal = 40°. Then the process of calculating the attitude compensation angle is as follows:

[0154] θ comp = 45° - 40° = 5°;

[0155] The obtained attitude compensation angle is 5°, indicating that the robot needs to adjust its angle to the right by 5° to be consistent with the preset trajectory. The calculation of this compensation angle is based on the difference between the angle of the robot's current position and the angle of the preset trajectory, ensuring that the robot moves forward in the correct direction. Assume that the attitude compensation process of the robot is dynamic, and the actual angle and target angle of the robot are used for real-time calculation at each step, and its angle is adjusted. Through this mechanism, the robot can gradually correct the angle difference from the preset trajectory during operation and finally travel along the predetermined route.

[0156] Please refer to Figure 5 , the pressure regulation module includes:

[0157] The pressure difference calculation sub-module adjusts the deviation between the robot and the preset trajectory according to the position compensation parameter and combines the real-time horizontal offset correction amount of the robot, obtains the contact pressure value of the pipe wall through the pressure sensor, calculates the difference between the measured pressure and the preset expected value, and obtains the pressure difference data;

[0158] In this sub-module, first, according to the position compensation parameter and combining the real-time horizontal offset correction amount of the robot, the deviation between the robot and the preset trajectory is adjusted. By adjusting the robot's deviation in real time, this step ensures that the robot can accurately travel along the predetermined trajectory. The real-time offset of the robot (including horizontal and vertical offsets) is detected by the sensors of the robot body (such as IMU sensors and gyroscopes) and is dynamically updated in combination with the position compensation algorithm. For this step, the real-time deviation is calculated based on the difference between the actual movement and the expected trajectory. For example, the errors of the X-axis and Y-axis are measured between the actual position and the target position of the robot, and these errors are integrated and corrected to achieve the desired accuracy. The pressure difference calculation is the second part of this sub-module. By obtaining the contact pressure value of the pipe wall, the difference between the measured pressure and the preset expected value is calculated. Specifically, the contact pressure value comes from the pressure sensor, which measures and feeds back the data of the pressure inside the pipe in real time. This sub-module compares these data with the preset target pressure and calculates the pressure difference between the measured pressure and the target pressure. The calculation formula:

[0159] ΔP' = P measured - P desired ;

[0160] where, ΔP' is the pressure difference, P measured is the measured pressure, P desired is the target pressure.

[0161] Suppose the wall contact pressure read by the pressure sensor is 4.8 MPa, while the target set pressure value is 5.0 MPa. The pressure difference is calculated as follows:

[0162] ΔP' = 4.8 MPa - 5.0 MPa = -0.2 MPa;

[0163] The obtained pressure difference is -0.2 MPa, and this data will provide a basis for subsequent pressure regulation.

[0164] The cleaning pressure adjustment sub-module is based on the pressure difference data and uses the formula:

[0165]

[0166] Calculate the adjusted cleaning nozzle pressure, adjust the pressure parameters of the cleaning nozzle, and obtain the pressure parameters of the cleaning nozzle;

[0167] where P adj is the adjusted cleaning nozzle pressure, P des is the preset target contact pressure, ΔP' is the pressure difference, that is, the difference between the measured pressure and the target pressure, α is the correction coefficient used to adjust the influence of the pressure difference on the nozzle pressure adjustment, μ is the fluid viscosity coefficient, and D is the nozzle diameter;

[0168] The function of this sub-module is to adjust the pressure parameters of the cleaning nozzle according to the pressure difference data. First, based on the pressure difference data obtained from the pressure difference calculation sub-module, this data is combined with the target pressure and adjusted through the formula. Specifically, the pressure of the cleaning nozzle needs to be corrected according to the difference between the measured pressure and the target pressure in order to achieve the set cleaning effect. The formula is as follows:

[0169]

[0170] where P adj is the adjusted cleaning nozzle pressure, P des is the preset target contact pressure, ΔP' is the pressure difference, that is, the difference between the measured pressure and the target pressure, α is the correction coefficient used to adjust the influence of the pressure difference on the nozzle pressure adjustment, μ is the fluid viscosity coefficient, which affects the resistance of fluid flow, and D is the nozzle diameter, which affects the flow rate and the magnitude of the nozzle pressure.

[0171] Suppose in a certain cleaning task, the target contact pressure is 5.0 MPa, the measured pressure is 4.8 MPa, the viscosity coefficient μ = 0.001 Pa·s, the nozzle diameter D = 0.05 m, and the correction coefficient α = 1.2. Then the pressure difference ΔP' = -0.2 MPa. Substituting into the formula for calculation, we get:

[0172]

[0173] The finally obtained adjusted cleaning nozzle pressure is 4.9952 MPa. This adjustment ensures that the output pressure of the nozzle is very close to the target pressure value, thus optimizing the cleaning effect and avoiding the influence of excessive pressure fluctuations on the cleaning quality.

[0174] The rotation motor torque compensation sub-module calculates the torque compensation value of the rotation motor according to the pressure parameters of the cleaning nozzle, considering the required pressure change of the cleaning nozzle and combining with the load response characteristics of the motor, adjusts the motor output, and generates the cleaning pressure parameters;

[0175] The main task of this sub-module is to calculate the torque compensation value of the rotation motor according to the pressure parameters of the cleaning nozzle. The pressure change of the cleaning nozzle will affect the load of the motor, and the torque of the motor must be compensated according to the adjustment of the nozzle pressure to ensure that the motor can provide sufficient power. According to the pressure difference data, the motor load will change accordingly, which will in turn affect its output torque. In this sub-module, the pressure adjustment requirements of the cleaning nozzle need to be identified first, and the corresponding torque compensation value is calculated based on these requirements. Specifically, the torque compensation value of the motor is calculated by considering the relationship between the pressure change and the motor load. If the pressure adjustment of the cleaning nozzle is large, the torque compensation value of the motor needs to be increased accordingly to adapt to the changed pressure requirements. Calculation formula:

[0176]

[0177] Where, T adj is the adjusted torque compensation value of the rotation motor, T base is the reference torque, that is, the motor output torque without pressure adjustment, β is the motor load response coefficient, used to adjust the relationship between the motor torque response and the pressure change, ΔP' is the pressure difference, that is, the difference between the actually measured pressure and the target pressure, P des is the target pressure, that is, the target contact pressure required by the cleaning nozzle.

[0178] Suppose, the reference torque T base = 10 Nm, the motor load response coefficient β = 0.05, the pressure difference ΔP' = -0.2 MPa, the target pressure P des = 5.0 MPa, substitute into the formula for calculation:

[0179]

[0180] Through calculation, the obtained adjusted torque compensation value of the rotation motor is 9.98 Nm. This torque value is obtained by adjusting the load response of the motor based on the measured pressure difference, ensuring that the motor can provide appropriate power output during the cleaning process and maintaining the stable working state of the cleaning nozzle.

[0181] During this process, the torque compensation value of the motor is proportional to the pressure difference, and is further optimized by adjusting the motor load response characteristics. This sub-module will finally calculate the required rotational motor torque compensation value, so as to provide appropriate adjustment for the motor and ensure the stability and accuracy of the motor output during the cleaning process.

[0182] Please refer to Figure 6 , the effect evaluation module includes:

[0183] The image comparison sub-module, based on the cleaning pressure parameters, obtains the inner wall images of the pipe before and after cleaning through an image acquisition device, and uses image analysis to extract the residue coverage area and texture clarity of multiple regions in the image, generating the analysis result of the inner wall image of the pipe;

[0184] The main task of this sub-module is to obtain the inner wall images of the pipe before and after cleaning through an image acquisition device, and use image analysis technology to extract the residue coverage area of multiple regions in each image. Specifically, the image processing system first performs high-definition image acquisition on the inner wall of the pipe to obtain the image data before and after cleaning. Then, through image analysis algorithms, especially edge detection and binarization processing, the system identifies the residue regions in the image and calculates the coverage area of these regions. The key to this step is to separate the regions before and after cleaning through threshold processing and calculate the proportion of residues in each region. Finally, the system outputs the residue coverage area data of each region for further analysis of the cleaning effect. The formula for calculating the residue coverage area in the image is as follows:

[0185]

[0186] where A area,i' is the residue coverage area of region i', W and H are the width and height of the image respectively, I raw (x, y) is the pixel value at position (x, y) in the image, Mask(x, y) is the binarization mask, with a value of 1 in the residue region and 0 in other positions, and ∑ represents the summation over all pixels in the image.

[0187] Suppose there is an image before cleaning, with an image size of W = 10 pixels and H = 10 pixels. The residues in the image are represented by a binarization mask (Mask), with a mask value of 1 in the residue region and 0 in other regions. Assume that the pixel value of the image is I raw (x, y) = 1 in the residue region and I raw (x, y) = 0 in other regions, and there are 25 pixels in the region where the mask is 1.

[0188] Calculate the coverage area of the residues:

[0189]

[0190] Since the pixel value in the residue-covered area of the image is 1 and the mask value is also 1, the covered area is the number of pixels with a mask value of 1, that is:

[0191] A area,1 = 25 pixels;

[0192] The result shows that the residue-covered area is 25 pixels.

[0193] Based on the analysis result of the pipe wall image, the cleaning effect analysis sub-module calculates the residue-covered area ratio and texture clarity difference value of multiple regions by comparing the residue-covered area and texture clarity in the images before and after cleaning, and uses the formula:

[0194]

[0195] to evaluate the cleaning effect at multiple positions and calculate the cleanliness score;

[0196] where C tot is the cleanliness score, measuring the cleaning effect, A bf,i' is the residue-covered area of region i' before cleaning, A af,i' is the residue-covered area of region i' after cleaning, T bf,i' is the texture clarity of region i' before cleaning, T af,i' is the texture clarity of region i' after cleaning, w 1 is the weight coefficient of the residue-covered area, w 2 is the weight coefficient of the texture clarity, N' is the total number of regions divided in the image, and i' is the index of the region in the image;

[0197] Based on the image analysis result provided by the image comparison sub-module, this sub-module calculates the residue-covered area ratio and texture clarity difference value of multiple regions in the images before and after cleaning. By comparing the images before and after cleaning, the system calculates the cleaning effect of each region and finally generates an overall cleanliness score. To evaluate the cleanliness, the following formula is used to calculate the improvement effect of the residue-covered area and texture clarity by weighting:

[0198]

[0199] where A bf,i' is the residue-covered area of region i' before cleaning, A af,i' is the residue-covered area of region i' after cleaning, T bf,i′ is the texture clarity of region i' before cleaning, T af,i′ is the texture clarity of region i' after cleaning, w 1 and w 2They are the weight coefficients of the residue coverage area and the texture clarity respectively, N′ is the total number of regions into which the image is divided, and C tot is the overall cleanliness score.

[0200] For example, assume there are three regions, Region 1: A bf,1 = 50%, A af,1 = 10%, T bf,1 = 0.8, T af,1 = 0.9, Region 2: A bf,2 = 60%, A af,2 = 20%, T bf,2 = 0.7, T af,2 = 0.85, Region 3: A bf,3 = 40%, A af,3 = 15%, T bf,3 = 0.75, T af,3 = 0.88, the weight coefficients w 1 = 0.6 and w 2 = 0.4, calculate the cleanliness score:

[0201] Calculate the contribution of each region respectively:

[0202] Region 1:

[0203]

[0204] C 1 = 0.6·0.8 + 0.4·0.125;

[0205] C 1 = 0.48 + 0.05;

[0206] C 1 = 0.53;

[0207] Region 2:

[0208]

[0209] C 2 = 0.6·0.6667 + 0.4·0.2143; C 2 = 0.4 + 0.0857;

[0210] C 2 = 0.4857;

[0211] Region 3:

[0212]

[0213] C 3 = 0.6·0.625 + 0.4·0.1733;

[0214] C 3 = 0.375 + 0.0693;

[0215] C 3 = 0.4443;

[0216] Overall cleanliness score:

[0217] C tot = C 1 + C 2 + C 3 ;

[0218] C tot = 0.53 + 0.4857 + 0.4443;

[0219] C tot = 1.46;

[0220] Therefore, the overall cleanliness score is 1.46.

[0221] The cleaning duration adjustment sub-module calls the cleanliness score, identifies the positions with insufficient cleaning, and adjusts the cleaning duration of the robot to generate pipeline cleaning operation parameters;

[0222] The cleaning duration adjustment sub-module automatically identifies the areas with insufficient cleaning based on the cleanliness score of the cleaning effect analysis sub-module and adjusts the cleaning duration. Specifically, the system determines which areas have insufficient cleaning effects based on the calculated cleanliness score. If the cleanliness score of a certain area is low, it means that the cleaning of that area is not thorough. To ensure complete cleaning of that area, the system will correspondingly extend the cleaning duration, especially in areas where the proportion of the residue coverage area is large or the texture clarity difference value is small. In addition, the cleaning duration adjustment sub-module also considers the moving speed and cleaning efficiency of the robot in different areas and comprehensively adjusts the duration to optimize the cleaning process. According to the cleanliness score C tot and the cleaning effects of each area, the formula for adjusting the cleaning duration T adj is:

[0223]

[0224] where T adj,i′ is the adjusted cleaning duration of area i′, T base is the basic cleaning duration, α is the adjustment factor that controls the intensity of the duration adjustment, C tot,i′ is the cleanliness score of area i', and 1 - C tot,i′ is used to represent the degree of insufficient cleanliness.

[0225] For example, assume that the basic cleaning duration of area 1 is T base = 5 minutes, and the cleanliness score of this area is Ctot,1 = 0.8. Assuming α = 0.2, the adjusted cleaning duration is as follows:

[0226] T adj,1 = 5×(1 + 0.2·(1 - 0.8)) = 5×(1 + 0.2×0.2) = 5×1.04 =

[0227] 5.2 minutes;

[0228] Through this formula, the cleaning duration can be reasonably adjusted according to the cleaning effect to ensure that each area can be sufficiently cleaned.

[0229] The above are only the preferred embodiments of the present invention and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. The adaptive control system of the pipeline cleaning robot is characterized by: The system comprises: The fluid sensing module monitors the state of the pipeline fluid, obtains flow rate, pressure, and temperature data, calculates the fluid state at multiple locations in the pipeline through feature comparison, and obtains the flow pattern coefficient; The path optimization module calls the flow pattern coefficient, uses the turbulence intensity coefficient and the backflow probability value, combines the curvature radius of the curve of the pipeline at multiple locations, and the pipe diameter contraction ratio, calculates the path curvature adjustment amount and the movement speed, and generates the path trajectory parameters; The position calibration module calls the path trajectory parameters, uses the inertial measurement unit and the gyroscope to analyze the motion posture and position of the robot and compares them with the preset trajectory, obtains the horizontal offset correction amount and the posture compensation angle, and generates the position compensation parameters; The pressure regulating module calls the position compensation parameter, uses the horizontal offset correction amount, and combines the pipe wall contact pressure value to calculate the difference between the expected contact pressure value and the measured contact pressure value, adjusts the pressure parameter of the cleaning nozzle and the torque compensation value of the rotary motor, and generates the cleaning pressure parameter; Based on the cleaning pressure parameters, the effect evaluation module uses image acquisition equipment to compare the image sequences before and after the pipe wall cleaning, identify the insufficiently cleaned positions and adjust the cleaning time to generate pipeline cleaning operation parameters.

2. The adaptive control system of the pipeline cleaning robot according to claim 1 is characterized in that: The flow pattern coefficient includes the fluid laminar index, the turbulence intensity coefficient, and the backflow probability value; the path trajectory parameters include the robot working path, the curve driving speed, and the straight pipe forward speed threshold; the position compensation parameters include the position offset correction amount, the angle offset correction amount, and the position offset data; the cleaning pressure parameters include the cleaning nozzle control parameters and the rotating motor control parameters; the pipeline cleaning operation parameters include the residue coverage area ratio, the texture clarity difference value, and the cleaning time adjustment value.

3. The adaptive control system of the pipeline cleaning robot according to claim 1, characterized in that: The fluid sensing module comprises: The feature extraction submodule monitors the state of the pipeline fluid, obtains the instantaneous flow velocity sequence using the flow velocity sensor, obtains the dynamic pressure waveform using the pressure sensor, obtains the pipe wall temperature value using the temperature sensor, calculates the flow velocity fluctuation value, pressure gradient value and temperature change value under the current fluid state, and establishes the fluid state feature vector; The feature comparison submodule calls the fluid state feature vector, extracts feature vector data of multiple known fluid states, and uses the formula: Determine the closeness between the current fluid state and the known state and obtain the feature similarity score; Among them, S represents the feature similarity score, w i Represents the weight of each feature item, X i represents the i-th eigenvalue of the current state, represents the i-th eigenvalue under the known fluid state, n is the total number of features, and i represents the i-th eigenterm in the eigenvector; The state calculation submodule identifies the flow states at multiple locations in the pipeline based on the feature similarity score and compares it with multiple known fluid states, extracts the corresponding fluid laminar flow index, turbulence intensity coefficient and backflow probability value, and obtains the flow pattern coefficient.

4. The adaptive control system of the pipeline cleaning robot according to claim 1, characterized in that: The path optimization module includes: The geometric feature extraction submodule calls the flow pattern coefficient and uses the geometric feature information of the pipeline to extract the curvature radius of the bend and the pipe diameter contraction ratio parameters at multiple positions in the pipeline to obtain the geometric feature analysis results; The path adjustment calculation submodule uses the turbulence intensity coefficient and the backflow probability value based on the geometric feature analysis results, combined with the curve curvature radius and the pipe diameter contraction ratio, and uses the formula: Perform path curvature adjustment calculation, obtain path curvature adjustment amount, and generate path adjustment parameters; Among them, C adjust is the path curvature adjustment, T curvature is the curvature radius of the curve, R is the response coefficient of the flow, P shrink is the tube diameter contraction ratio, L is the length of the flow path; The motion trajectory planning submodule calculates the motion speed of the robot at multiple positions in the pipeline according to the path adjustment parameters and the flow characteristics, obtains the deceleration ratio of the curve and the acceleration threshold of the straight pipe section, and obtains the path trajectory parameters.

5. The adaptive control system of the pipeline cleaning robot according to claim 1, characterized in that: The position calibration module comprises: The posture and position analysis submodule calls the path trajectory parameters, identifies the actual motion position and posture of the robot by analyzing the measurement data of the inertial measurement unit and the gyroscope, records the current coordinates and angle of the robot, and obtains the real-time position data of the robot; The module for calculating the horizontal offset correction uses the real-time position data of the robot and compares the actual coordinates of the robot with the preset trajectory using the formula: Get the horizontal offset correction; Among them, D offset is the horizontal offset correction, X act is the actual X coordinate of the robot, X ide is the preset X coordinate, Y act is the actual Y coordinate of the robot, Y ide is the preset Y coordinate; The posture compensation angle calculation submodule calculates the posture compensation angle of the robot based on the horizontal offset correction amount and the actual posture of the robot by comparing it with the preset trajectory, adjusts the angle difference between the robot and the preset trajectory, and generates position compensation parameters.

6. The adaptive control system of the pipeline cleaning robot according to claim 1, characterized in that: The pressure regulating module comprises: The pressure difference calculation submodule adjusts the deviation between the robot and the preset trajectory according to the position compensation parameter and the real-time horizontal offset correction of the robot, obtains the contact pressure value of the pipe wall in combination with the pressure sensor, calculates the difference between the measured pressure and the preset expected value, and obtains the pressure difference data; The cleaning pressure adjustment submodule uses the formula based on the pressure difference data: Calculating the adjusted cleaning nozzle pressure, adjusting the pressure parameters of the cleaning nozzle, and obtaining the pressure parameters of the cleaning nozzle; Among them, P adj is the adjusted cleaning nozzle pressure, P des is the preset target contact pressure, ΔP′ is the pressure difference, i.e., the difference between the measured pressure and the target pressure, α is the correction coefficient used to adjust the effect of the pressure difference on the nozzle pressure adjustment, μ is the fluid viscosity coefficient, and D is the nozzle diameter; The rotating motor torque compensation submodule further calculates the torque compensation value of the rotating motor according to the pressure parameters of the cleaning nozzle, considers the pressure change required for the cleaning nozzle, and combines the load response characteristics of the motor to adjust the motor output and generate the cleaning pressure parameters.

7. The adaptive control system of the pipeline cleaning robot according to claim 1, characterized in that: The effect evaluation module includes: The image comparison submodule obtains the pipe wall images before and after cleaning through the image acquisition device based on the cleaning pressure parameters, and uses image analysis to extract the residue coverage area and texture clarity of multiple areas in the image to generate pipe wall image analysis results; The cleaning effect analysis submodule calculates the residue coverage area ratio and texture clarity difference value of multiple regions based on the pipe wall image analysis results by comparing the residue coverage area and texture clarity in the images before and after cleaning, using the formula: Evaluate cleaning performance at multiple locations and calculate cleanliness scores; Among them, C tot is the cleanliness score, which measures the cleaning effect, A bf,i′ is the residue coverage area of ​​region i′ before cleaning, A af,i′ is the residue coverage area of ​​region i′ after cleaning, T bf,i′ is the texture clarity of region i′ before cleaning, T af,i′ is the texture clarity of region i′ after cleaning, w1 is the weight coefficient of the residue coverage area, w2 is the weight coefficient of texture clarity, N′ is the total number of regions divided into images, and i′ is the index of the region in the image; The cleaning time adjustment submodule calls the cleanliness score, identifies the insufficiently cleaned position, adjusts the cleaning time of the robot, and generates pipeline cleaning operation parameters.

Images