Pipeline dewatering robot failure prediction maintenance system

Abstract

This invention provides a fault prediction and maintenance system for pipeline dredging robots, relating to the field of pipeline dredging technology. It includes a partitioning module, used to construct a ring-shaped analysis region based on the physical relationship of four measurement points on the hydraulic power station and the dredging robot body. This ring-shaped analysis region is divided into three operational status analysis rings: inner, middle, and outer. The inner ring corresponds to the core components of the hydraulic power station, the middle ring to the key execution components of the dredging robot, and the outer ring to auxiliary components. This invention acquires real-time operating parameters from multiple measurement points of the pipeline dredging robot and analyzes their characteristics according to ring classification. Combined with a pre-trained fault prediction model, it generates component health status indicators. When these indicators fall below a preset threshold, corresponding maintenance is performed, achieving early fault prediction and timely maintenance, ensuring stable equipment operation.

Description

Technical Field

[0001] This invention relates to the field of pipeline dredging technology, and in particular to a pipeline dredging robot fault prediction and maintenance system. Background Technology

[0002] In the field of pipeline maintenance, the control system of dredging robots plays an important role in ensuring the stability and efficiency of dredging operations. Existing dredging robot control systems can usually achieve basic walking, dredging and video monitoring functions, but there is still room for improvement in their adaptive adjustment capabilities under actual complex working conditions.

[0003] For example, while existing control systems can display the operating parameters of the hydraulic power station and the real-time motion status of the robot, they mainly focus on status monitoring and command issuance. They have certain limitations in dynamically adjusting the hydraulic power output according to the actual workload. When the dredging device encounters changes in resistance or changes in the working conditions inside the pipeline, the system often cannot automatically adjust the output characteristics of the hydraulic power station in a timely manner to adapt to the load requirements. This may affect the robot's working efficiency and may even cause abnormalities in components during long-term operation. This control method, which is based on fixed parameter settings or relies on frequent manual adjustments by operators, may have its response speed and control accuracy affected when facing complex and ever-changing dredging environments. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a fault prediction and maintenance system for pipeline dredging robots, thereby improving the adaptability and work efficiency of dredging robots in actual operations.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] The first aspect is the pipeline dredging robot fault prediction and maintenance system, which includes:

[0007] The acquisition module is used to acquire operating parameters in real time through four measuring points set on the hydraulic power station and the body of the pipeline dredging robot. The operating parameters include vibration data, pressure data and temperature data, and the operating parameters are arranged in chronological order to form a time-series data sequence.

[0008] The partitioning module is used to construct a ring-shaped analysis area based on the physical relationship between the four measurement points on the hydraulic power station and the dredging robot body. The ring-shaped analysis area is divided into three operating state analysis rings: inner, middle and outer. The inner ring corresponds to the core components of the hydraulic power station, the middle ring corresponds to the key execution components of the dredging robot, and the outer ring corresponds to the auxiliary components.

[0009] The classification module is used to classify the four measurement points in the time series data according to the operating status analysis ring, and calculate the characteristic parameters of the inner, middle and outer operating status analysis rings, including vibration amplitude, pressure fluctuation and temperature change rate.

[0010] The adjustment module is used to analyze the characteristic parameters of the ring belt based on the internal, middle and external operating status, generate component operating status adjustment values, and input the operating status adjustment values ​​into the pre-trained fault prediction model to obtain component health status indicators.

[0011] The execution module is used to perform corresponding maintenance operations when any health status indicator falls below a preset threshold.

[0012] In a second aspect, a computing device includes:

[0013] One or more processors;

[0014] A storage device for storing one or more programs that, when executed by one or more processors, enable the one or more processors to implement the system.

[0015] Thirdly, a computer-readable storage medium storing a program that, when executed by a processor, implements the system.

[0016] The above-described solution of the present invention has at least the following beneficial effects:

[0017] By collecting real-time core parameters of the hydraulic power station and robot body, including vibration, pressure, and temperature, the status of critical components can be remotely monitored without manual entry into the pipeline. Automatic maintenance is triggered by thresholds, such as adjusting pressure when the hydraulic power station is overloaded or stopping actions when a component malfunctions, reducing the frequency of personnel contact with high-risk environments and mitigating safety risks at their source. Through ring-zone division and characteristic parameter calculation, such as monitoring hydraulic pump pressure fluctuations in the inner ring and analyzing the vibration amplitude of the walking / auger motor in the middle ring, the type of faulty component can be identified. Combined with quantitative health indicators output by a pre-trained fault prediction model, potential faults, such as pipeline blockage and decreased oil tank cooling, can be identified in advance, preventing the fault from escalating into a complete machine shutdown, shortening unplanned downtime, and ensuring the continuity of dredging operations. Through parameter classification and targeted maintenance, on-demand maintenance is achieved. For example, when the health indicators of the hydraulic power station are low, only the output pressure is adjusted; when auxiliary components malfunction, alarms are precisely triggered. Fault warnings also reduce secondary damage to components caused by sudden faults, such as gear wear caused by untreated abnormal motor vibration, reducing spare parts replacement and manual repair costs. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a pipeline dredging robot fault prediction and maintenance system provided in an embodiment of the present invention.

[0019] Figure 2 This is a flowchart illustrating the process of performing corresponding maintenance operations when any health status indicator is lower than a preset threshold, as provided by an embodiment of the present invention. Detailed Implementation

[0020] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0021] like Figure 1 As shown, an embodiment of the present invention proposes a pipeline dredging robot fault prediction and maintenance system, comprising:

[0022] The acquisition module is used to acquire operating parameters in real time through four measuring points set on the hydraulic power station and the body of the pipeline dredging robot. The operating parameters include vibration data, pressure data and temperature data, and the operating parameters are arranged in chronological order to form a time-series data sequence.

[0023] The partitioning module is used to construct a ring-shaped analysis area based on the physical relationship between the four measurement points on the hydraulic power station and the dredging robot body. The ring-shaped analysis area is divided into three operating state analysis rings: inner, middle and outer. The inner ring corresponds to the core components of the hydraulic power station, the middle ring corresponds to the key execution components of the dredging robot, and the outer ring corresponds to the auxiliary components.

[0024] The classification module is used to classify the four measurement points in the time series data according to the operating status analysis ring, and calculate the characteristic parameters of the inner, middle and outer operating status analysis rings, including vibration amplitude, pressure fluctuation and temperature change rate.

[0025] The adjustment module is used to analyze the characteristic parameters of the ring belt based on the internal, middle and external operating status, generate component operating status adjustment values, and input the operating status adjustment values ​​into the pre-trained fault prediction model to obtain component health status indicators.

[0026] The execution module is used to perform corresponding maintenance operations when any health status indicator falls below a preset threshold.

[0027] In this embodiment of the invention, by collecting core parameters of the hydraulic power station and robot body in real time, including vibration, pressure, and temperature, the status of key components can be remotely monitored without manual entry into the pipeline. Automatic maintenance is triggered by thresholds, such as adjusting pressure when the hydraulic power station is overloaded or stopping actions when a component malfunctions, reducing the frequency of personnel contact with high-risk environments and mitigating safety risks at their source. Through annular zone division and characteristic parameter calculation, such as monitoring hydraulic pump pressure fluctuations in the inner zone and analyzing the vibration amplitude of the walking / auger motor in the middle zone, the type of component to which the fault belongs can be located. Combined with the quantitative health indicators output by the pre-trained fault prediction model, potential faults can be identified in advance, such as pipeline blockage and decreased oil tank heat dissipation, preventing the fault from escalating into a complete machine shutdown, shortening unplanned downtime, and ensuring the continuity of dredging operations. Through parameter classification and targeted maintenance, on-demand maintenance is achieved. For example, when the health indicators of the hydraulic power station are low, only the output pressure is adjusted; when auxiliary components malfunction, alarms are precisely triggered. Simultaneously, fault warnings reduce secondary damage to components caused by sudden faults, such as gear wear caused by untreated abnormal motor vibration, reducing spare parts replacement and manual repair costs.

[0028] In a preferred embodiment of the present invention, operating parameters are acquired in real time through four measuring points set on the hydraulic power station and the main body of the pipeline dredging robot. These operating parameters include vibration data, pressure data, and temperature data. The operating parameters are arranged in chronological order to form a time-series data sequence, including:

[0029] Four measuring points were set at the hydraulic pump outlet of the hydraulic power station, the main control valve group outlet, the walking hydraulic motor inlet of the dredging robot body, and the auger hydraulic motor inlet of the dredging device. Specifically, based on the structural characteristics of the hydraulic power station and body of the pipeline dredging robot (the hydraulic power station integrates a hydraulic pump and a main control valve group, and the robot body is equipped with a walking hydraulic motor and an auger hydraulic motor for the dredging device), the four measuring points were precisely deployed. The first measuring point was installed at the hydraulic pump outlet of the hydraulic power station, fixed with a corrosion-resistant metal clamp, and an oil-resistant rubber pad was added to the inside of the clamp to ensure that the sensor probe was in close contact with the outer wall of the hydraulic pump outlet pipe, with the contact gap controlled within 0.5mm; the second measuring point was installed at the hydraulic pump outlet of the hydraulic power station. At the outlet end of the main control valve group of the pressure power station, corresponding to the main oil outlet interface of the main control valve group, the sensor is fixed to the side of the interface flange by a threaded connection. The sensor detection surface is perpendicular to the oil flow direction to ensure the accuracy of pressure data acquisition. The third measurement point is installed at the inlet end of the walking hydraulic motor of the dredging robot body. A clamp-type fixing structure is used to fasten the sensor to the middle section of the oil inlet pipe of the walking hydraulic motor. The contact parts between the pipe and the sensor are treated with rust removal and coated with rust inhibitor to adapt to the humid environment inside the pipeline. The fourth measurement point is installed at the inlet end of the auger hydraulic motor of the dredging device, fitting the straight pipe section of the auger hydraulic motor oil inlet interface. The sensor shell is made of 316 stainless steel to resist the corrosion of mud and sewage inside the pipeline.

[0030] Collect the operating parameters of four measurement points. The operating parameters include vibration data, pressure data, and temperature data. Timestamp the operating parameters of each measurement point respectively, and arrange and integrate the operating parameters of the four measurement points in the order of the timestamps, forming a time-series data sequence that includes spatial position attributes and time attributes. Specifically include: Start the multi-parameter sensing acquisition system supporting the four measurement points. The system includes a piezoelectric vibration sensor with a range of 0 to 20 g, a strain gauge pressure sensor with a range of 0 to 30 MPa, and a K-type thermocouple temperature sensor with a range of -50 to 200 °C. Synchronously collect the three types of operating parameters of each measurement point. The first is vibration data, which reflects the running smoothness of rotating components such as hydraulic pumps and motors. The acquisition frequency is set to 10 Hz, and the vibration acceleration values along the axial and radial directions of the pipeline are recorded. The second is pressure data, which reflects the pressure loss and stability during the hydraulic power transmission process. The acquisition frequency is set to 5 Hz, and the real-time pressure value of the hydraulic oil in the pipeline is recorded. The third is temperature data, which reflects the heat generation situation of components due to friction and power loss. The acquisition frequency is set to 1 Hz, and the temperature value of the outer wall of the pipeline is recorded. For each set of parameters collected at each measurement point, mark the acquisition time through the internal clock of the system, unify the timestamp format, ensure that the parameter time bases of the four measurement points are completely consistent, and arrange the parameters of the four measurement points in order of the timestamps, forming a structured data table. Each row of data includes the timestamp, the position of the measurement point, such as the outlet of the hydraulic pump, the inlet of the walking hydraulic motor, the vibration acceleration value (axial / radial), the pressure value, and the temperature value. Thus, construct a time-series data sequence that includes both spatial position attributes (the component positions corresponding to the measurement points) and time attributes (the acquisition timestamps). Each piece of data in the sequence completely associates the three-dimensional information of time, position, and the three types of parameter values.

[0031] In this embodiment, by directly setting measurement points on the core components of the hydraulic power station and the robot body, remote real-time acquisition of operating parameters is achieved, eliminating the need for personnel to enter the pipeline for monitoring, and fundamentally reducing the safety risks of personnel coming into contact with toxic gases and deep water environments, thus solving the safety pain points of manual monitoring; The three types of parameters collected, namely vibration, pressure, and temperature, respectively correspond to the mechanical running smoothness of components, the hydraulic power transmission state, and the heat generation situation of components, comprehensively reflecting the running state of the core components, avoiding misjudgment of faults caused by single-parameter monitoring, and improving the integrity of parameter monitoring; By unifying the timestamps and integrating spatial position attributes, a time-series data sequence including time, position, and parameters is formed, solving the problems of fragmented and uncorrelated robot data, and improving the accuracy of fault analysis.

[0032] In a preferred embodiment of the present invention, based on the physical relationship between four measurement points on the hydraulic power station and the dredging robot body, a ring-shaped analysis region is constructed, and the ring-shaped analysis region is divided into three operating state analysis rings: inner, middle, and outer. The inner ring corresponds to the core components of the hydraulic power station, the middle ring corresponds to the key execution components of the dredging robot, and the outer ring corresponds to auxiliary components, including:

[0033] Using the hydraulic pump of the hydraulic power station as the power source, a circular analysis region from the power source to the execution end is constructed based on the flow path of the hydraulic oil and the direction of power transmission. Specifically, this includes: taking the hydraulic pump of the hydraulic power station as the power source, this hydraulic pump is the core power output component of the hydraulic system, installed on the left side of the rear frame of the robot, and fixed to the frame with four M8 bolts. Its output pressure directly determines the power supply intensity of all downstream hydraulic components. If the hydraulic pump experiences abnormalities such as piston wear or variable displacement mechanism jamming, it will directly lead to insufficient power of the entire machine (output pressure below 8MPa) or a sudden pressure drop (instantaneous drop exceeding 5MPa). Based on this, the complete flow path and pipeline parameters of the hydraulic oil are first systematically analyzed. Specifically, the hydraulic pump outputs high-pressure oil through a 16mm inner diameter high-pressure oil pipe at the outlet end, and the normal operating pressure range is 10 to 15MPa. The oil is then delivered to the main control valve assembly, which is fitted with a 16mm inner diameter oil pipe at the inlet and two separate 12mm inner diameter oil pipes at the outlet. Its core function is to adjust the pressure and flow rate of the oil according to the work instructions, with an adjustment range of 0 to 50 L / min. The adjusted oil is then delivered through two 12mm high-pressure oil pipes. One pipe flows to the walking hydraulic motors on both sides of the dredging robot (one on the left and one on the right, with the motor inlet fitted with a 12mm inner diameter oil pipe), and the other flows to the auger hydraulic motor of the front-end dredging device (also fitted with a 12mm inner diameter oil pipe at the inlet). The oil drives the two types of motors to complete their work (the walking motor drives the tracks to rotate, and the auger motor drives the auger to break up the sludge). The oil is then collected through an 18mm inner diameter return oil pipe and finally returned to the oil tank, which has a capacity of 50L. The tank is installed on the right rear of the robot and fixed to the frame with three M6 bolts.

[0034] Based on the power transmission direction, power is transmitted unidirectionally from the hydraulic pump to the actuators. Hydraulic energy is gradually converted into mechanical energy during this process. The hydraulic pump converts the input 380V three-phase AC power into hydraulic energy. The main control valve group adjusts the parameters of the hydraulic energy. The travel hydraulic motor and the auger hydraulic motor then convert the adjusted hydraulic energy into rotational mechanical energy. An XY plane coordinate system is established with the robot's front-end center as the origin (X-axis forward along the robot's axis, Y-axis perpendicular to the X-axis and upward). The spatial distribution coordinates of each component are analyzed. Specifically, the core power components are concentrated at the robot's rear end (area X≥450mm). The hydraulic pump's center point coordinates are X=500mm, Y=0mm, and the main control valve group's center point coordinates are X=550mm, Y=0mm, with a 50mm distance between them to ensure space for oil pipe connection. The actuators are distributed in the middle and front of the robot. The center point coordinates of the walking hydraulic motors (left and right sides) are X=300mm, Y=200mm and X=300mm, Y=-200mm, respectively, and they are symmetrically arranged to ensure walking stability. The auger hydraulic motor's center point coordinates are X=100mm, Y=0mm, located at the front center to ensure that the crushing range covers the pipe cross-section.Auxiliary components are distributed around the core and actuator components. The center point of the oil tank is at coordinates X=520mm, Y=150mm. Piping accessories, including high-pressure oil pipes, quick-connect couplings, and pipe clamps, are arranged along the edge of the frame. Most of the couplings are concentrated in the X=350-500mm area, which is the connection path between the core and actuator components. The lighting device consists of two LED high-intensity light modules, located at X=50mm, Y=50mm and X=50mm, Y=-50mm respectively at the front end. The imaging device is a full HD camera located at X=30mm, Y=0mm at the front end, with the lens facing the working direction to capture forward images. Based on the above hydraulic oil flow path and component spatial distribution characteristics, a ring-shaped analysis area is drawn with the geometric center of the hydraulic pump (coordinates X=500mm, Y=0mm) as the center. The inner boundary of this ring-shaped analysis area is the outermost point of the hydraulic pump's outer perimeter (coordinates X=500mm±150mm). Using X=0mm±100mm as a baseline, a circular boundary with a radius of 180mm is determined using the distance formula between two points. This ensures the inner boundary completely covers the entire hydraulic pump, including the pump body flange, inlet and outlet ports, etc. The outer boundary extends to the outermost auxiliary components of the robot, namely the front-end imaging device (X=30mm, Y=0mm) and the right-side oil tank (X=520mm, Y=250mm). The distance from the center point of each component to the center of the circle is calculated, and the maximum value, assumed to be 550mm, is taken as the outer boundary radius, forming a circular boundary. The annular area must meet strict coverage requirements; the center point of all components must fall between the inner boundary (radius 180mm) and the outer boundary (radius 550mm), and the relative positions of each component within the annular area must be completely consistent with the actual physical layout of the robot, such as the oil pipe routing, component spacing, and fixing method. Furthermore, this annular area must be marked on the robot's CAD assembly drawing, with both the inner and outer boundaries marked with solid blue lines.

[0035] The annular analysis area is radially divided into three operational state analysis rings. The inner ring corresponds to the hydraulic power generation area, including the hydraulic pump and main control valve assembly; the middle ring corresponds to the power transmission and core execution area, including the travel hydraulic motor and auger hydraulic motor; and the outer ring corresponds to the functional auxiliary area, including the oil tank, pipeline accessories, lighting devices, and imaging devices. Specifically, along the radial direction of the annular analysis area (extending from the center X=500mm, Y=0mm to the outer boundary radius of 550mm), based on the principles of component functional correlation and power transmission sequence, the annular area is divided into three independent operational state analysis rings. The boundary definition, component composition, functional positioning, and labeling requirements for each ring are as follows: The radial range of the inner ring (hydraulic power generation area) is from 180mm to 330mm, where the inner boundary is completely consistent with the inner boundary of the annular analysis area (radius 180mm). The outer boundary is a virtual boundary with a radius of 330mm, which is determined based on the outermost point of the main control valve assembly (coordinate X=650m). The straight-line distance from the center of the circle (x=500mm, y=0mm) is calculated and assumed to be 330mm. The core components and fixing method of the inner ring are as follows: The center point coordinates of the hydraulic pump are x=500mm, y=0mm, and it is fixed to the frame with four M8 bolts. Its outlet end is connected to the inlet end of the main control valve group through a 16mm inner diameter high-pressure oil pipe. The oil pipe is 50mm long, and both ends of the oil pipe are sealed with compression fittings with a pressure rating of 25MPa. The center point coordinates of the main control valve group are x=550mm, y=0mm, y=0mm, and y=0mm. Y=0mm, fixed to the frame extension plate on the right side of the hydraulic pump by two M6 bolts, with a bolt tightening torque of 12N・m; the valve group integrates two pilot-operated pressure valves (adjustment range 0 to 25MPa) and two solenoid flow valves, with a flow range of 0 to 50L / min. Its main function is to receive commands from the control system and adjust the pressure and flow of the hydraulic pump output oil to the range required by the actuators, such as a travel hydraulic motor with an appropriate pressure of 8 to 12MPa and an auger hydraulic motor with an appropriate pressure of 12 to 15MPa;The inner ring belt functions as the core of hydraulic energy generation and control. The hydraulic pump converts externally input electrical energy into hydraulic energy, outputting high-pressure oil with a pressure range of 10 to 15 MPa. The main control valve group precisely adjusts the oil parameters according to the real-time needs of walking and dredging operations, ensuring that downstream actuators receive a matching power supply. Its adjustment accuracy directly affects the action response speed and operational stability of the actuators. The virtual outer boundary of this ring belt needs to be marked with a red dashed line in the CAD drawing. The boundary coordinate expressions are X=500±330cosθ, Y=0±330sinθ, where θ is the polar angle in the polar coordinate system. Specifically, it is defined as the angle formed by rotating counterclockwise to a point on the boundary, with the center of the annular analysis area (X=500mm, Y=0mm) as the origin and the positive X-axis direction (towards the robot's front end) as the starting edge. The value range is 0 to 360°. By continuously varying θ from 0° to 360°, the following parameters can be calculated: The coordinates of all points on the boundary are calculated as follows: (e.g., when θ=0°, the coordinates are X=500+330×1=830mm, Y=0+330×0=0mm; when θ=90°, the coordinates are X=500+330×0=500mm, Y=0+330×1=330mm). This forms a complete ring boundary, ensuring full coverage of the circumference and no component is omitted. During the labeling process, the positional relationship between the boundary and the hydraulic pump and main control valve assembly must be verified one by one. For example, the hydraulic pump needs to be checked. Check whether the flange edge (coordinates X=500+150=650mm, Y=0) falls inside the boundary (X=830mm is the rightmost point of the boundary, 650mm<830mm, which meets the requirements), and whether the solenoid valve of the main control valve group (coordinates X=550+80=630mm, Y=0) exceeds the boundary. If there is interference between the boundary and the structural components, such as if the edge coordinates of a component exceed the boundary coordinates, the boundary radius can be finely adjusted, with an adjustment range not exceeding 5mm, until the requirement of no interference is met.

[0036] The radial range of the middle ring belt (power transmission and core execution area) is 330mm to 480mm. Its inner boundary seamlessly connects with the outer boundary (330mm radius) of the inner ring belt. The outer boundary is a virtual boundary with a radius of 480mm, which is determined by the straight-line distance from the outermost point of the travel hydraulic motor (coordinates X=300mm, Y=200mm) to the center of the circle, ensuring that the outer boundary can completely cover all execution components. The components and connection methods included in the middle ring belt are as follows: travel hydraulic motors (left and right sides), where the center point coordinates of the left motor are X=300mm, Y=200mm, and the center point coordinates of the right motor are X=300mm, Y=-200mm. Both are connected to the robot track drive shaft via flanges. The flange is fitted with four M10 bolts to ensure slippage-free transmission. The motor inlet is connected to the first outlet of the main control valve group via a 12mm inner diameter high-pressure oil pipe, 250mm long. A 25MPa pressure-resistant check valve is connected in series on the pipe to prevent backflow and motor reversal. The auger hydraulic motor's center point coordinates are X=100mm, Y=0mm, and it is connected to the auger shaft via an XL3 type flexible coupling. The coupling can compensate for installation errors (radial deviation ≤0.1mm). The motor inlet is connected to the second outlet of the main control valve group via a 12mm inner diameter high-pressure oil pipe, 450mm long. The outer layer of the oil pipe is wrapped with a flame-retardant and wear-resistant sleeve with a temperature resistance range of -40 to 120℃, capable of withstanding pipe... The inner ring belt is designed to withstand wear from silt and oil corrosion within the dredging channel. It functions as the power transmission terminal and the core of the operation execution. It receives regulated hydraulic energy from the inner ring belt via high-pressure oil pipes. The travel hydraulic motor converts this hydraulic energy into the mechanical energy of the track rotation, driving the robot to move forward, backward (speed 0.3 to 1 m / min), and turn (turning radius ≤ 500 mm). The auger hydraulic motor converts the hydraulic energy into the mechanical energy of the auger rotation, driving it to break up viscous silt (70% to 90% water content) or sandy silt (≥ 50% sand content) within the pipeline at a rated speed of 150 rpm, directly achieving the core function of dredging operations (dredging efficiency ≥ 200 m³ / h). The virtual outer boundary of this ring belt needs to be defined in the CAD drawing. The yellow dashed lines mark the boundary coordinates, which are X=500±480cosθ and Y=0±480sinθ. The definition of θ is consistent with the inner ring. This expression generates a complete circle for the outer boundary. For example, when θ=180°, the coordinates are X=500-480=20mm and Y=0, corresponding to the robot's front end; when θ=270°, the coordinates are X=500 and Y=0-480=-480mm, corresponding to the bottom of the robot. This ensures the boundary covers the locations of all actuators. After boundary division, verification is required. The specific process is to first calculate the distance from the center point of all motors to the center of the circle (X=500mm, Y=0mm), where the distance from the center point of the left-side walking hydraulic motor to the center is assumed to be 282.The distance is 282.8 mm, which is less than 330 mm, since the inner boundary of the middle ring belt has a radius of 330 mm. This does not meet the requirement of 330 mm < distance < outer boundary radius. Therefore, the outer boundary radius of the middle ring belt needs to be slightly adjusted to 485 mm so that the distance of the left travel hydraulic motor (282.8 mm) is less than 485 mm, thus meeting the range requirement of the middle ring belt. The right travel hydraulic motor is symmetrically distributed with the left travel hydraulic motor, and the distance from its center point to the center of the circle is the same as that on the left, approximately 282.8 mm. After slightly adjusting the outer boundary radius to 485 mm, this also meets the requirement. The required range for the middle ring belt is 330 mm < distance < 485 mm. Assuming the distance from the center point of the auger hydraulic motor to the center of the circle is 400 mm, this distance is greater than the inner boundary of the middle ring belt (330 mm) and less than the adjusted outer boundary radius (485 mm), thus meeting the middle ring belt range requirement. If subsequent inspections reveal that the distance between the center points of other motors exceeds the specified range of 330 mm < distance < outer boundary radius, the outer boundary radius can be fine-tuned, with the adjustment range controlled within 8 mm. Ultimately, this ensures that all motors are completely within the middle ring belt, with no components crossing the belt.

[0037] The outer ring (functional auxiliary area) has a radial range of 480mm to 550mm. Its inner boundary connects with the outer boundary of the middle ring (radius 480mm), and its outer boundary perfectly matches the outer boundary of the annular analysis area (radius 550mm). Calculations show that the robot's outermost auxiliary components, the front-end imaging device (X=30mm, Y=0mm), are assumed to be 470mm from the center; the right-side fuel tank (X=520mm, Y=250mm), is assumed to be 250.8mm from the center, all falling within this range, ensuring coverage of all auxiliary components. The components included in the outer ring and... The layout is as follows: the center point of the oil tank is at X=520mm, Y=150mm, with a capacity of 50L. It is constructed from welded steel plate with a thickness of 3mm and a weld strength ≥350MPa. The top of the tank has a filler port with an air filter (20μm filtration accuracy), and the bottom has a DN20 drain valve. The oil tank is connected to the oil outlets of all motors via an 18mm inner diameter return oil pipe. A 10μm precision return oil filter (0.1m² filtration area, flow rate ≥60L / min) is connected in series on the return oil pipe. The tank integrates two cooling fans with an airflow of 150... The operating voltage is 24V. The fan starts at an oil temperature ≥55℃ and stops at ≤30℃, ensuring stable control of the circulating oil temperature within the normal range of 30 to 55℃. The piping accessories include three types of components: high-pressure oil pipes, male and female quick couplings, and pipe clamps. The high-pressure oil pipes have an inner diameter of 12 to 16mm, are made of steel wire braided rubber tubing (inner layer of oil-resistant rubber, outer layer of steel wire braid), have a temperature range of -40 to 120℃, and a pressure rating of 30MPa. They are used to connect the hydraulic components of the inner ring belt (main control valve assembly) and the middle ring belt (motor). The male and female quick couplings are made of brass (chrome-plated for rust prevention), have a pressure rating of 20MPa, and a insertion / extraction life of ≥500 cycles. Insertion and extraction force ≤50N, used for quick disassembly and maintenance of oil pipes; pipe clamps are made of nylon (temperature resistant from -30 to 80℃), with a fixing spacing of 50mm, and are fixed along the edge of the frame with M4 bolts to keep the oil pipe close to the frame and prevent wear caused by vibration; the lighting device consists of two LED high-intensity modules, each with a power of 30W (input voltage 24V), a luminous flux of 2800lm (color temperature 6000K, close to natural light), and a waterproof rating of IP68 (can be immersed in 1m deep water for 30 minutes without leakage); the two modules are installed at X=50mm, Y=50mm and X=50mm, Y=-50mm respectively at the front end of the robot, connected by 2×1.5 The control system is connected using standard copper core wires, supporting brightness adjustment from 0 to 100% (PWM dimming). In dimly lit environments within pipelines (illuminance ≤ 5 lux), it can provide an illumination intensity of ≥ 200 lux, eliminating blind spots. The imaging device consists of a single full HD camera with a resolution of 1920×1080 (pixel density 300 dpi), a frame rate of 30fps (no ghosting), and a 90° field of view, covering over 80% of the pipeline cross-section. The camera is mounted at the robot's front end at X=30mm and Y=0mm, with a sapphire glass lens (scratch-resistant) and an IP68 protection rating. Transmission is achieved via a 75Ω coaxial cable (transmission distance ≥ 10... The system connects to the control system (0m), with a minimum illumination of 0.01 lux (starlight-level night vision). Even in low-light environments (such as when there is no lighting inside the pipeline), it can still clearly capture the state of sludge accumulation and the condition of the pipeline wall. The outer ring belt is positioned to provide auxiliary support for the core components. The oil tank is responsible for storing hydraulic oil (the oil level is maintained at 1 / 2 to 2 / 3 of its capacity) and maintaining a stable oil temperature through a cooling fan. This prevents the oil from being affected by high-temperature oxidation (the oxidation rate doubles when the temperature is >60℃) or low-temperature viscosity increase (the viscosity is >1000cSt when the temperature is <10℃), which would affect power transmission. The pipeline accessories ensure that the hydraulic oil transmission path is unobstructed and reduce pressure loss (pressure loss along the oil pipe ≤0).(5MPa / 10m); The lighting device provides sufficient light source for operations inside the pipeline, ensuring that the imaging device can clearly collect environmental information; the imaging device provides a visual basis for remote control of the working environment, and operators can adjust the dredging path through real-time images. Although these four types of components do not directly participate in power generation or operation execution, they jointly ensure the stable operation of the core components of the inner ring and middle ring. For example, stable oil temperature can extend motor life by 30% and improve operational safety. Visualization can prevent the robot from colliding with the inner wall of the pipeline. The virtual inner boundary of this ring needs to be marked with a green dashed line in the CAD drawing. The boundary coordinate expression is X=500±480cosθ, Y=0±480sinθ, where θ is still the polar angle in the polar coordinate system (defined as before, with a value of 0 to 360°). The inner boundary generated by this expression can accurately connect to the outer boundary of the middle ring, ensuring that there is no overlap or gap between the rings. After the boundary is divided, a verification is required to check whether the mounting bases of all auxiliary components fall completely within the radius of 480mm to 550mm, with the oil tank as the reference. Taking the fixed bracket as an example, based on the center point of the fuel tank (X=520mm, Y=150mm), the distance from the center point of the fuel tank to the center of the circle (X=500mm, Y=0mm) is assumed to be 541mm. Combining this with the outer boundary radius of the outer ring of 550mm, the outermost edge of the fuel tank fixed bracket should not exceed 529mm. That is, 520mm plus the difference between 550mm and 541mm. After confirmation, the distance from the outermost edge of the adjusted bracket to the center of the circle is approximately 550mm, which falls exactly within a radius of 480mm to 5... Within a 50mm range, the requirements are met. If, during inspection, the mounting bases of other auxiliary components are found to exceed the above range, such as the distance from the edge of the lighting device's base to the center being less than 480mm or greater than 550mm, this can be corrected by fine-tuning the component's mounting position. For example, the lighting device can be moved 5mm towards the center. It is important to note that all fine-tuning adjustments must not exceed 10mm. Ultimately, ensure that the mounting bases of all auxiliary components fall entirely within a radius of 480mm to 550mm, with no auxiliary components crossing the ring.

[0038] After the three ring zones are divided, a final verification needs to be performed in conjunction with the robot's CAD assembly drawing. The specific process involves extracting the center point coordinates of all components, including the hydraulic pump, main control valve group, travel motor, auger motor, oil tank, lighting device, imaging device, and key connectors of pipeline accessories. Then, using the center of the annular analysis area (X=500mm, Y=0mm) as a reference, the distance from the center point of each component to that center is calculated. The core objective is to confirm that the center point of each component falls within only one of the three ring zones, without any cross-ring distribution. It is necessary to explain the ring zone assignment verification for the hydraulic pump. The center point coordinates of the hydraulic pump are X=500mm, Y=0mm, and its distance to the center is 0mm. The inner boundary radius of the inner ring zone is 180mm. Since the inner boundary of the inner ring zone is defined as completely covering the outer perimeter of the hydraulic pump, even if the center point of the hydraulic pump is inside the 180mm inner boundary, its entire... The pump body, including the pump body, inlet and outlet ports, and fixed flanges, all fall within the inner ring, with a radial range of 180mm to 330mm. Therefore, the hydraulic pump's ring affiliation meets the requirements and is clearly classified as belonging to the inner ring. If, during verification, it is found that a component is distributed across rings, for example, the distance from the center point of a critical joint of a pipeline accessory to the center of the circle is 320mm, while the outer boundary radius of the inner ring is 330mm and the inner boundary radius of the middle ring is also 330mm. The center point of this joint is located both within the 330mm range of the inner ring and outside the 330mm range of the middle ring, which is considered a cross-ring distribution. In this case, the boundary radius of the corresponding ring needs to be fine-tuned. For example, the outer boundary radius of the inner ring can be adjusted from 330mm to 325mm, so that the 320mm distance of the joint's center point falls only within the adjusted inner ring. The fine-tuning range of all ring boundaries must be controlled within 10mm, ultimately strictly meeting the verification standard of one component per ring.

[0039] Based on the physical component affiliation of the four measurement points in the time-series data sequence, they are mapped to either the inner loop or the middle loop, respectively. Specifically, this involves: first, clarifying the basic information and component affiliation of the four measurement points. The deployment locations and corresponding components of the four measurement points have been determined previously. Measurement point MP1 corresponds to the hydraulic pump outlet (the hydraulic pump belongs to the inner loop), measurement point MP2 corresponds to the main control valve group outlet (the main control valve group belongs to the inner loop), measurement point MP3 corresponds to the travel hydraulic motor inlet (the travel hydraulic motor belongs to the middle loop), and measurement point MP4 corresponds to the auger hydraulic motor inlet (the auger hydraulic motor belongs to the middle loop); extracting from the previously collected time-series data sequence... Complete data for each measurement point, including timestamp, vibration axial value, vibration radial value, pressure value, temperature value, data quality code, and acquisition device ID, is tagged according to component affiliation. All time-series data for MP1 and MP2 are tagged as inner loop data, and all time-series data for MP3 and MP4 are tagged as middle loop data. Data storage uses an SQLite relational database, constructing two dedicated data tables: an inner loop data table and a middle loop data table. The data table fields are completely consistent with the time-series data format, and both use the timestamp and measurement point code as primary keys to avoid duplicate data for the same measurement point at the same timestamp; data write frequency... Synchronized with the sensor acquisition frequency (vibration data 10Hz, pressure data 5Hz, temperature data 1Hz), each set of data is written to the corresponding data table in real time after acquisition, with a write delay controlled within 100ms to ensure data real-time performance. Finally, a mapping accuracy check is performed to ensure no confusion between the inner and middle loop data. The first step is code comparison, exporting the measurement point code fields from the inner and middle loop data tables, confirming that the inner loop data table contains only MP1 and MP2, and the middle loop data table contains only MP3 and MP4. If other codes appear, the attribution relationship is rechecked. The second step is random sampling, randomly selecting 100 data entries from the inner loop data table. The data, accounting for approximately 5% of the hourly sample size, verifies the matching between the measurement point code and the parameter value. For example, the pressure value of MP1 should be between 10 and 15 MPa. If it exceeds this range and the data quality code is 0, it is determined to be data written incorrectly. The third step is spatial association. Combining the robot's CAD assembly drawing, the measurement point code is associated with the center point coordinates of the corresponding component. For example, MP1 corresponds to the center point of the hydraulic pump X=500mm and Y=0mm, confirming it as an inner ring component. This ensures that the data marking and spatial assignment are consistent. The verification must meet three criteria: no abnormal coding, sampling error rate ≤0.1%, and 100% spatial association matching. If these criteria are not met, the mapping process is re-executed.

[0040] For the auxiliary components of the outer ring belt, the operating status of the auxiliary components of the outer ring belt is inferred by analyzing the pressure fluctuations and temperature change rates of the inner and middle ring belts, thus obtaining the inference result of the operating status of the outer ring belt. Specifically, based on the pressure data of the measuring points of the hydraulic pump and main control valve group in the inner ring belt, the fluctuation range of the hydraulic power station's output pressure is calculated; based on the temperature data of the measuring points of the traveling hydraulic motor and auger hydraulic motor in the middle ring belt, the temperature change rate of the key actuators is calculated. This includes: because sensors cannot be directly deployed in the outer ring belt auxiliary components, oil tank, pipeline accessories, etc., due to installation space limitations, key indicators need to be indirectly calculated through the pressure and temperature parameters of the inner and middle ring belts to infer their operating status. This is specifically divided into two parts: calculating the fluctuation range of the hydraulic power station's output pressure and calculating the temperature change rate of the key actuators. The calculation of the fluctuation range of the hydraulic power station's output pressure first involves screening stable operating periods. A continuous 5-minute stable period is selected from the inner ring belt data table. The screening criteria must simultaneously meet three requirements: vibration stability (the axial standard deviation of the vibration at the hydraulic pump outlet measuring point does not exceed 0.05g), vibration radial standard deviation does not exceed 0.08g; main control... The axial standard deviation of vibration at the valve assembly outlet measurement point should not exceed 0.06g, and the radial standard deviation should not exceed 0.09g. By calculating the standard deviation, interference from abnormal component vibration is eliminated. Pressure stability is ensured (pressure fluctuations at the hydraulic pump outlet measurement point and the main control valve assembly outlet measurement point should not exceed 0.2MPa to avoid sudden pressure changes due to load increases). Data quality is also assessed (no severe interference records with a data quality code of 2 are found within the time period to ensure data reliability). If multiple time periods meet the requirements, the latest time period is selected. For example, if the current time is 15:30, then select... The time period from 15:25 to 15:30 was selected to ensure that the data matched the current equipment status. Next, pressure data extraction and preprocessing were performed. Pressure data from the hydraulic pump outlet and main control valve group outlet during the stable period were extracted from the inner loop data table. Since the pressure acquisition frequency was 5Hz, 1500 samples were collected for each component within 5 minutes. Outliers were removed using the 3σ principle. The mean and standard deviation of the main control valve group outlet pressure data were calculated using the same method, and outliers exceeding the corresponding range were removed. If data loss occurred after outlier removal, the loss rate should not exceed 0%.5% is used to replace the outlier by interpolating the arithmetic mean of two adjacent normal data points. For example, if the hydraulic pump outlet pressure data at 15:25:02 is an outlier, the mean of the normal pressure values ​​at 15:25:01 and 15:25:03 is used to calculate the outlier value, ensuring data continuity. For the preprocessed hydraulic pump outlet pressure data, the maximum and minimum pressure values ​​within that time period are identified. The pressure fluctuation value of the hydraulic pump is obtained by subtracting the minimum pressure value from the maximum pressure value. The same method is used to process the main control valve group outlet pressure data to obtain the pressure fluctuation value of the main control valve group. The arithmetic mean of the hydraulic pump pressure fluctuation value and the main control valve group pressure fluctuation value is taken as the fluctuation range of the hydraulic power station output pressure. The calculation result is retained to two decimal places. For example, the fluctuation range of the hydraulic power station output pressure is 0.23 MPa. The smaller this value, the more stable the hydraulic power station output pressure.

[0041] The calculation of the temperature change rate of key execution components begins with the extraction of temperature data. From the central ring data table, temperature data completely consistent with the pressure stabilization period is extracted, specifically the temperature data at the inlet of the travel hydraulic motor and the auger hydraulic motor. The temperature sampling frequency is 1Hz, with 300 samples collected from each motor within 5 minutes. After extraction, the data integrity and quality are checked, requiring a data missing rate of no more than 0.5%. If missing data exists, it is filled with the normal temperature value from the previous moment. All data have a quality code of 0 to avoid interference affecting the calculation results. Next, the temperature change rate of a single motor is calculated. Considering the slow temperature change of the hydraulic system oil (the rate of change does not exceed 0.5℃ / min under normal operating conditions), an interval of 10 seconds (equivalent to 1 / 6 of a minute) is chosen between two adjacent timestamps. This interval accurately reflects the temperature change trend while avoiding numerical fluctuations caused by excessively short intervals, such as instantaneous temperature jumps caused by electromagnetic interference. For the temperature data at the inlet of the travel hydraulic motor, the temperature values ​​corresponding to the i-th and (i+1)-th timestamps are taken. The temperature change rate of the travel hydraulic motor per interval is obtained by subtracting the temperature value of the previous time from the temperature value of the next time interval and then dividing by the time interval. The arithmetic mean of the temperature change rates of all single intervals is taken to obtain the average temperature change rate of the travel hydraulic motor. The same method is used to process the temperature data at the inlet of the auger hydraulic motor to obtain the average temperature change rate of the auger hydraulic motor. The calculation results of both average temperature change rates are retained to three decimal places. Finally, the average temperature change rate of the key actuator is calculated, that is, the arithmetic mean of the average temperature change rate of the travel hydraulic motor and the average temperature change rate of the auger hydraulic motor is taken as the average temperature change rate of the key actuator, with the unit being ℃ / min. When the average temperature change rate of the key actuator is positive, it indicates that the temperature of the actuator is rising. The larger the value, the more obvious the heating of the actuator.

[0042] Correlation analysis was performed between pressure fluctuation range and temperature change rate. When pressure fluctuation increases and temperature change rate rises abnormally, it is determined that there is a risk of blockage in the outer ring pipeline accessories. When pressure fluctuation is stable but temperature change rate rises slowly, it is determined that the heat dissipation performance of the outer ring oil tank has decreased. The inference results of the outer ring operation status are obtained, which include: the historical normal operation range is based on the statistical analysis of data from the factory stage and field fault-free operation data. It is divided into three steps. The first step is data collection at the factory stage. Before leaving the factory, three working conditions of the actual operation scenario are simulated, namely no-load condition (the robot remains stationary and does not perform dredging operation), half-load condition, and half-load condition. The first step involved collecting stable data for 5 consecutive minutes under two conditions: a loaded condition (the robot only performs walking motions while the auger idles) and a full-load condition (the robot simultaneously performs walking motions and auger crushing motions, crushing viscous sludge with a moisture content controlled at 70%). For each condition, 10 sets of stable data were collected over 5 months. For each set of data, the fluctuation range of the hydraulic power station's output pressure and the average temperature change rate of key actuators were calculated. The second step involved collecting fault-free data from the field, gathering 50 sets of fault-free data from the robot's 3-month on-site dredging operations. The fault-free criteria were no maintenance records and no fault alarms during the operation. Information output and dredging efficiency remain stable and not less than 200 m³ / h. For each set of field data, the fluctuation range of the hydraulic power station output pressure and the average temperature change rate of key actuators are calculated. The third step is statistical analysis, integrating all valid data, namely 30 sets from the factory stage and 50 sets from the field without faults, for a total of 80 sets. The mean and standard deviation of the hydraulic power station output pressure fluctuation range, as well as the mean and standard deviation of the average temperature change rate of key actuators, are calculated. The mean of the hydraulic power station output pressure fluctuation range minus 2 standard deviations is taken as the mean of the hydraulic power station output pressure fluctuation range plus 2. The range of standard deviations is taken as the historical normal range of the hydraulic power station's output pressure fluctuation range. The range from the mean of the average temperature change rate of the key actuators minus 2 standard deviations to the mean of the average temperature change rate of the key actuators plus 2 standard deviations is taken as the historical normal range of the average temperature change rate of the key actuators. This range can cover 95.4% of the normal operating data, effectively avoiding misjudgment. For example, the normal reference range of the hydraulic power station's output pressure fluctuation range is 0.15 to 0.35 MPa, and the normal reference range of the average temperature change rate of the key actuators is 0.02 to 0.08 °C / min.

[0043] The criteria for assessing the risk of blockage in pipeline fittings require the simultaneous fulfillment of two indicators: the fluctuation range of the hydraulic power unit's output pressure must increase by more than 20% compared to the upper limit of the historical normal range, and the average temperature change rate of the critical actuators must abnormally increase by more than 50% compared to the upper limit of the historical normal range. For example, if the normal upper limit of the hydraulic power unit's output pressure fluctuation range is 0.35 MPa, a 20% increase would result in a threshold of 0.42 MPa; and the normal upper limit of the average temperature change rate of the critical actuators is 0.08℃ / min, a 50% increase would result in a threshold of 0.12℃ / min. When the actual calculated fluctuation range of the hydraulic power unit's output pressure is 0.45 MPa and the average temperature change rate of the critical actuators is... The judgment condition is met when the flow rate is 0.13℃ / min. The logical derivation process is as follows: when the pipeline accessories in the outer ring are blocked, such as hoses and joints, the cross-sectional area of ​​the oil flow will decrease, which will directly lead to a significant increase in the oil flow resistance. The increase in resistance will disrupt the stability of the output pressure of the inner ring, which is manifested as an increase in the fluctuation range of the hydraulic power station output pressure. At the same time, the friction between the oil and the blockage, such as sludge clumps and impurities, will be significantly enhanced, and the additional heat generated will increase accordingly. This heat will be transferred to the middle ring during the oil circulation process, causing the inlet temperature of the actuator to rise. Ultimately, this will manifest as an increase in the average temperature change rate of the key actuator. Therefore, it can be inferred that there is a risk of blockage in the pipeline accessories of the outer ring.

[0044] Based on the increase in the range of hydraulic power unit output pressure fluctuation and the increase in the average temperature change rate of key actuators, the blockage risk is divided into three levels: Mild blockage (20% to 30% increase in hydraulic power unit output pressure fluctuation and 50% to 80% increase in the average temperature change rate of key actuators), where the blockage is minor and has little impact on operation, requiring inspection of pipeline accessories during the next routine maintenance, such as weekly maintenance; Moderate blockage (30% to 50% increase in hydraulic power unit output pressure fluctuation and 80% to 120% increase in the average temperature change rate of key actuators), where the blockage has already affected normal operation, such as robot movement. If the pumping speed decreases and dredging efficiency drops, the current operation must be suspended and maintenance arranged within 24 hours. For severe blockages, where the hydraulic power station output pressure fluctuation range increases by more than 50% and the average temperature change rate of key actuators increases by more than 120%, the blockage is severe, oil flow resistance is extremely high, and there is a risk of pipeline rupture. Immediate shutdown and maintenance are required to prevent further accidents. To infer a decrease in the oil tank's heat dissipation performance, two indicators must be met simultaneously: the hydraulic power station output pressure fluctuation range must remain within the historical normal range, and the difference between this fluctuation range and the historical normal range average must not exceed 0.01 MPa, i.e., the fluctuation amplitude must not exceed 5%. For key actuators… The average temperature change rate of the actuator slowly increases compared to the upper limit of the historical normal range, with an increase exceeding 10% per hour. For example, if the normal upper limit of the average temperature change rate of the critical actuator is 0.08℃ / min, after an increase of 10% per hour, the threshold after 1 hour is 0.088℃ / min, and the threshold after 2 hours is 0.096℃ / min. When the actual calculated output pressure fluctuation range of the hydraulic power station is 0.25MPa (within the normal range) and the average temperature change rate of the critical actuator is 0.09℃ / min (an increase of 12.5% ​​within 1 hour), the judgment condition is met. (Logical derivation process) When the heat dissipation performance of the oil tank in the outer ring decreases, such as due to blockage of the oil tank radiator or failure of the cooling fan, the normal heat generated by the oil during circulation cannot be effectively dissipated through the oil tank, resulting in continuous heat accumulation. Since the oil flow path is not obstructed (the pipeline is not blocked), the oil flow resistance does not change significantly, so the pressure fluctuation in the inner ring remains stable, specifically, the fluctuation range of the hydraulic power station output pressure is within the normal range. However, the accumulated heat will be transferred to the middle ring with the oil circulation, causing the inlet temperature of the actuator to rise slowly, which ultimately manifests as an increase in the average temperature change rate of the key actuator. Therefore, it can be inferred that the heat dissipation performance of the oil tank in the outer ring decreases.

[0045] Based on the hourly increase in the average temperature change rate of critical actuators, the degree of decline in oil tank heat dissipation performance is divided into three levels: mild decline, where the average temperature change rate of critical actuators increases by 10% to 20% per hour, resulting in a slight decrease in heat dissipation efficiency and a slow rise in oil temperature; the oil tank radiator needs to be cleaned, such as removing dust and impurities from the surface of the heat sink fins, to restore heat dissipation capacity; moderate decline, where the average temperature change rate of critical actuators increases by 20% to 30% per hour, resulting in a significant decrease in heat dissipation efficiency and a continuous rise in oil temperature; the oil tank cooling fan needs to be replaced to improve heat dissipation; and severe decline, where the average temperature change rate of critical actuators increases by more than 30% per hour, indicating that the heat dissipation system has completely failed, and the oil temperature rises rapidly (potentially exceeding 60°C, at which point the oil oxidation rate doubles). The system must be shut down and the oil tank heat dissipation system replaced, and the radiator and cooling fan should be replaced simultaneously to prevent oil deterioration due to high temperatures from affecting the lifespan of the hydraulic system.

[0046] After the correlation analysis is completed, a status report for the outer ring auxiliary components is generated. The report must include five core components: the analysis period, calculated indicators (specific values ​​for the hydraulic power station output pressure fluctuation range and the average temperature change rate of key actuators, along with corresponding historical normal ranges; for example, the hydraulic power station output pressure fluctuation range is 0.45 MPa, with a historical normal range of 0.15 to 0.35 MPa; the average temperature change rate of key actuators is 0.13℃ / min, with a historical normal range of 0.02 to 0.08℃ / min); status inferences (e.g., moderate risk of blockage in pipeline accessories, normal oil tank cooling); and judgment criteria (e.g., hydraulic power station output pressure fluctuation range). The operating range is 0.45 MPa, which is 42.9% higher than the historical upper limit of 0.35 MPa; the average temperature change rate of key actuators is 0.13℃ / min, which is 62.5% higher than the historical upper limit of 0.08℃ / min, meeting the judgment criteria for moderate blockage risk; maintenance recommendations include suspending the current dredging operation, disassembling and inspecting the pipeline joints from the main control valve group to the travel hydraulic motor within 12 hours, and clearing any blockages; after maintenance, data needs to be collected again to verify whether the blockage risk has been eliminated. After the report is generated, it is transmitted in real time to the display terminal of the robot control system via industrial Ethernet, and the report is also stored in the local database in PDF format.

[0047] This embodiment divides the analysis area into three rings—inner, middle, and outer—based on the hydraulic oil flow path and power transmission direction. This clarifies the attribution of hydraulic power generation, execution, and auxiliary components, resolving the problem of ambiguous fault location and improving fault location efficiency. By mapping measurement points to the inner and middle rings, direct parameter monitoring of core components such as hydraulic pumps, main control valve groups, travel hydraulic motors, and auger hydraulic motors is achieved, ensuring that the operating status of core components is monitored without omission and avoiding major failures caused by unmonitored core components. Addressing the lack of direct measurement points for auxiliary components in the outer ring, the status is indirectly inferred through the correlation analysis of pressure and temperature parameters in the inner and middle rings, filling the monitoring gap for auxiliary components and avoiding missed faults in auxiliary components such as pipeline blockage and reduced oil tank heat dissipation, thus reducing the risk of system downtime due to auxiliary component failures. The ring division and status inference logic avoids blind disassembly and repair, reducing maintenance costs, improving maintenance efficiency, and solving the pain points of blind and costly robot maintenance.

[0048] In a preferred embodiment of the present invention, the four measurement points in the time-series data sequence are classified according to the operational status analysis rings, and the characteristic parameters of the inner, middle, and outer operational status analysis rings are calculated respectively, including vibration amplitude, pressure fluctuation, and temperature change rate, including:

[0049] The hydraulic pump and main control valve assembly measurement point data in the time-series data sequence are categorized into inner loop data, while the travel hydraulic motor and auger hydraulic motor measurement point data are categorized into middle loop data. Specifically, this involves: first, clarifying the source of the time-series data sequence, which is generated based on data collected from four previously deployed measurement points (hydraulic pump outlet measurement point, main control valve assembly outlet measurement point, travel hydraulic motor inlet measurement point, and auger hydraulic motor inlet measurement point), including fields such as timestamp, vibration axial value, vibration radial value, pressure value, temperature value, data quality code, and acquisition device ID; then, classifying the data according to the loop to which the component belongs. The inner loop data classification involves categorizing all time-series data from the hydraulic pump outlet measurement point and all time-series data from the main control valve assembly outlet measurement point into inner loop data, and then using the database loop... The identification field is marked as the inner loop and stored in the inner loop data table, ensuring that the table only contains data from the two measurement points mentioned above, without any other loop data mixed in. The middle loop data classification involves classifying all time-series data from the travel hydraulic motor inlet measurement point and all time-series data from the auger hydraulic motor inlet measurement point into middle loop data. This is marked as the middle loop by the loop identification field in the database and stored in the middle loop data table, ensuring that this table only contains data from the two measurement points mentioned above and is completely separated from the inner loop data. After classification, a data integrity check must be performed to check whether the data acquisition frequency of each measurement point in the two data tables meets the requirements (vibration data 10Hz, pressure data 5Hz, temperature data 1Hz), and the missing rate must not exceed 0.5%. If missing data exists, it is filled with normal data from the previous moment.

[0050] Based on the inner ring band data, the maximum amplitude of the vibration data at the hydraulic pump measuring point is calculated as the inner ring band vibration amplitude. The standard deviation of the pressure data at the main control valve group measuring point is calculated as the inner ring band pressure fluctuation. The minute-by-minute change in the temperature data at the hydraulic pump and main control valve group measuring points is calculated as the inner ring band temperature change rate. Specifically, based on the categorized inner ring band data, within a continuous 5-minute stable operating period (the selection criteria for the stable period are the same as before: stable vibration, stable pressure, and qualified data quality), the following three key parameters are calculated sequentially. First, the inner ring band vibration amplitude is calculated. The calculation object is the vibration data at the hydraulic pump outlet measuring point, including the axial and radial vibration values. Data preprocessing is performed first to extract the stable time... For all vibration axial and radial values ​​at the hydraulic pump outlet measurement point within the section, the absolute value of the vibration value at each time stamp is taken to eliminate the influence of positive and negative directions. Then, the maximum amplitude is calculated. All pre-processed vibration axial and radial absolute values ​​are traversed, and the maximum value is selected. This maximum value is the vibration amplitude of the inner ring belt, used to reflect the operational stability of the core power component (hydraulic pump) of the inner ring belt. Next, the pressure fluctuation of the inner ring belt is calculated. The calculation object is the pressure data of the main control valve group outlet measurement point. First, the pressure average is calculated, and all pressure values ​​of the main control valve group outlet measurement point within the stable period are extracted. The sum of each pressure value is divided by the sample size (since the pressure acquisition frequency is 5Hz, the sample size for 5 minutes is calculated). The pressure mean is obtained (for a value of 1500). Then, the sum of squared differences is calculated. For each pressure value, the square of the difference between that value and the pressure mean is calculated, and then all squared differences are summed. Finally, the standard deviation is calculated by dividing the sum of squared differences by the square root of the result (sample size minus 1). This standard deviation is the pressure fluctuation of the inner ring, reflecting the stability of the inner ring pressure parameters. Finally, the temperature change rate of the inner ring is calculated. The calculation objects are the temperature data at the hydraulic pump outlet measurement point and the temperature data at the main control valve group outlet measurement point. First, the time interval is determined, selecting 1 minute as the time interval. The stable period (5 minutes) is divided into 4 consecutive 1-minute sub-periods (each sub-period is 60 seconds). Next, calculate the temperature change per minute at a single measurement point. For the hydraulic pump outlet temperature data, calculate the temperature change for each sub-period (sub-period end temperature minus sub-period start temperature) to obtain four hydraulic pump temperature changes. Use the same method to calculate four minute temperature changes for the main control valve group outlet temperature data. Finally, calculate the inner ring temperature change rate. Take the arithmetic mean of the four temperature changes for the hydraulic pump (to obtain the average temperature change of the hydraulic pump) and the arithmetic mean of the four temperature changes for the main control valve group (to obtain the average temperature change of the main control valve group). Then take the arithmetic mean of these two averages, which is the inner ring temperature change rate, used to reflect the heating trend of the core components of the inner ring.

[0051] Based on the middle ring belt data, the average vibration amplitude of the measuring points of the travel hydraulic motor and the auger hydraulic motor is calculated as the middle ring belt vibration amplitude. The range of the pressure data at the measuring points of the travel hydraulic motor and the auger hydraulic motor is calculated as the middle ring belt pressure fluctuation. The maximum minute change of the temperature data at the measuring points of the travel hydraulic motor and the auger hydraulic motor is calculated as the middle ring belt temperature change rate. Specifically, based on the categorized middle ring belt data, during a 5-minute stable operating period, the following three key parameters are calculated sequentially: First, the middle ring belt vibration amplitude is calculated, taking the vibration data at the inlet measuring point of the travel hydraulic motor and the vibration data at the inlet measuring point of the auger hydraulic motor as the calculation objects. First, the vibration amplitude of a single motor is calculated. For the travel hydraulic motor, its axial and radial vibration values ​​are extracted. The absolute value of the vibration value at each time stamp is taken, and the maximum value is selected as the vibration amplitude of the travel hydraulic motor. The vibration amplitude of the auger hydraulic motor is calculated using the same method. Next, the vibration amplitude of the middle belt is calculated by dividing the sum of the vibration amplitudes of the travel hydraulic motor and the auger hydraulic motor by 2 and taking the arithmetic mean of the two motor vibration amplitudes. This arithmetic mean reflects the overall operational stability of the middle belt actuator. Finally, the pressure fluctuation of the middle belt is calculated, using the pressure data at the inlet measurement point of the travel hydraulic motor and the auger... The pressure data at the hydraulic motor inlet measuring point is analyzed as follows: First, the maximum and minimum pressure values ​​within a stable period are identified for the travel hydraulic motor pressure data. The pressure range of the travel hydraulic motor is calculated by subtracting the minimum pressure value from the maximum. The same method is used to calculate the pressure range of the auger hydraulic motor. Next, the pressure fluctuation of the middle ring belt is calculated by dividing the sum of the travel hydraulic motor pressure range and the auger hydraulic motor pressure range by 2 and taking the arithmetic mean of the two motor pressure ranges. This arithmetic means the pressure fluctuation of the middle ring belt, which reflects the pressure stability of the middle ring belt actuator. Finally, the temperature change rate of the middle ring belt is calculated, with the temperature at the travel hydraulic motor inlet measuring point as the calculation object. Based on the temperature data at the inlet measurement point of the auger hydraulic motor, the temperature change per minute of a single motor is first calculated. For the temperature data of the travel hydraulic motor, the time interval is divided into sub-periods of 1 minute, and the temperature change of each sub-period (end temperature minus start temperature) is calculated to obtain the temperature change of 4 travel motors. The same method is used to calculate the 4 temperature changes per minute of the auger hydraulic motor. Then, the temperature change rate of the middle ring belt is calculated. From the 4 temperature changes of the travel motor and the 4 temperature changes of the auger motor, the largest value is selected. This maximum value is the temperature change rate of the middle ring belt, which is used to reflect the maximum heating trend of the middle ring belt actuator.

[0052] Based on the inference results of the outer ring zone operating status, the risk level of pipeline accessory blockage is quantified as an outer ring zone pressure fluctuation characteristic parameter, and the degree of decline in oil tank heat dissipation performance is quantified as an outer ring zone temperature change rate characteristic parameter. Specifically, based on the inference results of the outer ring zone operating status, namely the risk level of pipeline accessory blockage and the degree of decline in oil tank heat dissipation performance, the two states are quantified as corresponding outer ring zone characteristic parameters. The quantification criteria are consistent with the previous risk / performance level classification standards. For the risk level of pipeline accessory blockage, if it is inferred to be mild blockage (the output pressure fluctuation range of the hydraulic power station increases by 20% to 30%, and the average temperature change rate of key actuators increases by 50% to 80%), the outer ring zone pressure fluctuation characteristic parameter is quantified to 0.2 to 0.3; if it is inferred to be moderate blockage (the output pressure fluctuation range of the hydraulic power station increases by 30% to 50%, and the average temperature change rate of key actuators increases by 80% to 120%), the outer ring zone pressure fluctuation characteristic parameter is quantified to 0.3-0.5; if it is inferred to be severe ...5; if it is inferred to be severe blockage (the output pressure fluctuation range of the hydraulic power station increases by 30% to 50%, and the average temperature change rate of key actuators increases by 80% to 120%), the outer ring zone pressure If the pressure fluctuation of the outer ring band increases by more than 50% (or the average temperature change rate of critical actuators increases by more than 120%), the outer ring band pressure fluctuation characteristic parameter is quantized to be greater than 0.5. If no blockage risk is inferred, the outer ring band pressure fluctuation characteristic parameter is quantized to 0 to 0.2. Regarding the degree of decline in the tank's heat dissipation performance, if it is inferred to be a slight decline (the average temperature change rate of critical actuators increases by 10% to 20% per hour), the outer ring band temperature change rate characteristic parameter is quantized to 0.1 to 0.2. If it is inferred to be a moderate decline (the average temperature change rate of critical actuators increases by 20% to 30% per hour), the outer ring band temperature change rate characteristic parameter is quantized to 0.2 to 0.3. If it is inferred to be a severe decline (the average temperature change rate of critical actuators increases by more than 30% per hour), the outer ring band temperature change rate characteristic parameter is quantized to be greater than 0.3. If the heat dissipation performance is inferred to be normal, the outer ring band temperature change rate characteristic parameter is quantized to 0 to 0.1. The two quantized characteristic parameters need to be associated and stored with the inner ring band and middle ring band parameters to form a complete ring band state parameter dataset.

[0053] This embodiment achieves precise binding of time-series data with the ring belts through clear ring belt data classification, avoiding confusion between different ring belt data. For the inner ring belt, three key parameters—vibration amplitude, pressure fluctuation, and temperature change rate—are extracted to quantitatively characterize the operating status of the hydraulic pump and main control valve group. This reflects the stability, pressure stability, and heating trend of the power core, providing direct evidence for determining whether the power core is operating normally. For the middle ring belt, vibration amplitude, pressure fluctuation, and temperature change rate are calculated. Through calculation methods such as average, range, and maximum, the overall operating status of the two actuators is comprehensively reflected. This allows for rapid identification of abnormal vibration, sudden pressure changes, or excessive heating of the actuators, providing quantitative indicators for actuator status monitoring. The qualitative state inference results of the outer ring belt are transformed into quantitative feature parameters, eliminating the ambiguity of the qualitative description. This allows the outer ring belt status to be directly correlated with the parameters of the inner and middle ring belts for analysis. Simultaneously, it provides standardized parameter support for overall system status assessment and maintenance priority determination, improving the accuracy and efficiency of maintenance decisions.

[0054] In a preferred embodiment of the present invention, based on the characteristic parameters of the inner, middle, and outer operating status analysis ring, component operating status adjustment values ​​are generated, and these operating status adjustment values ​​are input into a pre-trained fault prediction model to obtain component health status indicators, including:

[0055] The three characteristic parameters of vibration amplitude, pressure fluctuation, and temperature change rate of the inner ring belt are weighted and fused to generate the operating status adjustment value of the hydraulic power station. Similarly, the three characteristic parameters of vibration amplitude, pressure fluctuation, and temperature change rate of the middle ring belt are weighted and fused to generate the operating status adjustment value of the dredging robot's actuators, including a walking hydraulic motor and an auger hydraulic motor. Finally, the two characteristic parameters of pressure fluctuation and temperature change rate of the outer ring belt are weighted and fused to generate the operating status adjustment value of the auxiliary components. Specifically, this involves: firstly, based on historical fault statistics covering the fault types, frequencies, and impact degrees of the hydraulic power station, actuators, and auxiliary components, determining the weighting coefficients for each characteristic parameter. The weight allocation follows the principle that the higher the sensitivity of the component's state characterization, the greater the impact of the fault. Following the principle that parameters with greater loudness have higher weights, the vibration amplitude in the inner ring is 0.3, pressure fluctuation is 0.4, and temperature change rate is 0.3; in the middle ring, the vibration amplitude is 0.3, pressure fluctuation is 0.35, and temperature change rate is 0.35; and in the outer ring, the pressure fluctuation characteristic parameter and temperature change rate characteristic parameter are both 0.5, with the total weight of each ring being 1. Based on these weights, the adjustment value of the hydraulic power station's operating status corresponding to the inner ring is calculated. Because the vibration amplitude, pressure fluctuation, and temperature change rate of the inner ring have different dimensions, each parameter needs to be normalized to the range of 0 to 1. The normalization formula is: Normalized value = (Actual parameter value - Minimum value of the parameter's historical normal range) ÷ (Maximum value of the parameter's historical normal range - Parameter's historical normal range) The minimum value is calculated based on the historical normal range of parameters, which is derived from previous factory tests and field fault-free data. If the actual value of the parameter exceeds the normal range, the maximum value is normalized to 1, and the minimum value is normalized to 0. Then, the hydraulic power station operating status adjustment value is calculated through weighted fusion. The formula is: Hydraulic power station operating status adjustment value = normalized inner ring belt vibration amplitude × 0.3 + normalized inner ring belt pressure fluctuation × 0.4 + normalized inner ring belt temperature change rate × 0.3. This value ranges from 0 to 1. The closer the value is to 0, the closer the operation is to the normal state. The closer the value is to 1, the more obvious the deviation from the normal state. Next, the operating status adjustment value of the actuator corresponding to the middle ring belt is calculated. The actuator, namely the characteristic parameters of the travel hydraulic motor and the auger hydraulic motor (middle ring belt) The vibration amplitude, pressure fluctuation, and temperature change rate of the outer ring belt are normalized using the same formula as the inner ring belt. Then, a weighted fusion calculation is performed to determine the operating status adjustment value of the actuator. The formula is: Actuator operating status adjustment value = Normalized middle ring belt vibration amplitude × 0.3 + Normalized middle ring belt pressure fluctuation × 0.35 + Normalized middle ring belt temperature change rate × 0.35. This value ranges from 0 to 1; a smaller value indicates more stable actuator operation. Finally, the operating status adjustment value of the auxiliary components corresponding to the outer ring belt is calculated. The characteristic parameters of the outer ring belt (pressure fluctuation characteristic parameter, temperature change rate characteristic parameter) are already in the range of 0 to 1. Because they have been quantified, they are directly calculated using a weighted fusion calculation. The formula is: Auxiliary component operating status adjustment value = Outer ring belt pressure fluctuation characteristic parameter × 0.35.5 + outer ring temperature change rate characteristic parameter × 0.5, this value ranges from 0 to 1, the closer the value is to 0, the more normal the condition of the auxiliary component.

[0056] The operating status adjustment values ​​of the hydraulic power station, the actuators, and the auxiliary components are used as component operating status adjustment values ​​and input into a pre-trained fault prediction model to obtain component health status indicators, including the health status indicators of the hydraulic power station, the actuators, and the auxiliary components. Specifically, the hydraulic power station operating status adjustment value is compared with the historical normal operating data range of the hydraulic power station, and the hydraulic power station health status indicator is calculated using the multilayer perceptron in the pre-trained fault prediction model. The actuator operating status adjustment value is compared with the historical normal operating data range of the actuator, and the actuator health status indicator is calculated using the time series analysis algorithm in the pre-trained fault prediction model. The auxiliary component operating status adjustment value is compared with the historical normal operating data range of the auxiliary component, and the auxiliary component health status indicator is calculated using the association rule mining algorithm in the pre-trained fault prediction model. Finally, three independent component health status indicators ranging from 0 to 1 are obtained, where 1 represents a completely healthy state and 0 represents a completely faulty state. Specifically, the pre-trained fault prediction model is first constructed and trained. The training data includes two types: one is three months of field data... The dataset consists of two sets of data: 1) Fault-free operating data, covering various working conditions, including loop parameters, operating status adjustment values, and corresponding health labels of 1 (indicating complete health); and 2) Simulated data for 12 typical faults, such as hydraulic pump piston wear, motor jamming, and pipeline blockage, including loop parameters, operating status adjustment values, and corresponding health labels from 0 to 0.8 (smaller values ​​indicate more severe faults) before and after the fault. These two sets of data are divided into training and validation sets in a 7:3 ratio. Based on the training data, three sub-models are designed for the different characteristics of the hydraulic power station, actuators, and auxiliary components. The structure and parameters of each sub-model are detailed below. The calculation and training verification process is as follows: Since the health status of the hydraulic power station needs to be judged comprehensively by combining the operating status adjustment value and the degree of deviation from the normal range, a multilayer perceptron with a fixed structure is designed. The input layer has 2 neurons, corresponding to the hydraulic power station operating status adjustment value and the degree of deviation of the adjustment value, respectively; the hidden layer has 2 layers, with 16 neurons in the first layer and 8 neurons in the second layer, and the activation function is ReLU, which enhances the nonlinear fitting ability and ensures the capture of complex correlations between parameters; the output layer has 1 neuron, with the activation function being Sigmoid, which strictly maps the output to the range of 0 to 1, directly representing the health index.

[0057] The status of the actuators (walking hydraulic motor, auger hydraulic motor) changes dynamically over time, and the trend needs to be captured through time series analysis. An ARIMA model is used, and the model parameters are determined using the AIC criterion (Akaike Information Criterion, which measures the balance between goodness of fit and complexity; smaller values ​​indicate better overall model performance) as p=2, d=1, and q=2. Here, p is the number of autoregressive terms, indicating that the first p lagged observations in the sequence are used for prediction; here, p=2 means using the first two lagged observations. d is the difference order, indicating that the non-stationary sequence is transformed into a stationary one. The number of differencing operations required for the sequence is d=1, meaning that stationarity is achieved through one differencing operation. q is the number of moving average terms, indicating that the prediction result is corrected using the first q lag prediction errors. Here, q=2 means that the first 2 lag prediction errors are used. The specific input, output, and calculation process are as follows: The input is the adjustment value of the running status of the execution component within one hour (collected once every 5 minutes, for a total of 12 times, recorded in chronological order as the 1st to 12th adjustment values); the output is two adjustment values ​​for the next 10 minutes (adjustment value for the next 5 minutes and adjustment value for the next 10 minutes).

[0058] The extrapolation and calculation of the model proceeded step by step according to the following logic: First, the series was stabilized. Since d=1, the 12 adjusted values ​​were first-order differencing to obtain 11 stationary series data (2nd adjusted value - 1st adjusted value, 3rd adjusted value - 2nd adjusted value, ..., 12th adjusted value - 11th adjusted value). All stationary series data needed to pass the ADF test (p-value < 0.05) to ensure that the ARIMA model's requirement for series stationarity was met. Next, the autoregressive term was calculated. Since p=2, the autoregressive term for a stationary series data was 0.6 × the value of that data. The first stationary value is calculated as follows: + 0.3 × the first two stationary values ​​of the data, where 0.6 and 0.3 are typical autoregressive coefficients obtained from the training set fitting. The specific values ​​fluctuate slightly with the training data, but must meet the model stability requirements. For example, the autoregressive term for the third stationary value = 0.6 × the second stationary value + 0.3 × the first stationary value. Then, a moving average term is calculated. Since q = 2, the moving average term for a stationary sequence of data = (-0.2) × the first lag error of the data + (-0.1) × the first two lag errors of the data (-0.2 and -0.1 are typical autoregressive coefficients obtained from the training set fitting). The typical moving average coefficient is obtained, where the lag error = actual stationary value - predicted stationary value, and the predicted stationary value = 0.6 × the previous stationary value + 0.3 × the previous two stationary values ​​+ (-0.2) × the previous lag error + (-0.1) × the previous two lag errors + constant term 0.02; finally, short-term forecast extrapolation is performed. The first point is the prediction of the future value of the stationary sequence. The stationary value in the next 5 minutes = 0.6 × the 12th stationary value + 0.3 × the 11th stationary value + (-0.2) × the 12th lag error + (-0.1)×11th lag error+0.02; 10-minute stable value = 0.6×5-minute stable value+0.3×12th stable value+(-0.2)×5-minute lag error (5-minute lag error = actual 5-minute stable value - predicted 5-minute stable value)+(-0.1)×12th lag error+0.02; The second point is the restoration of the original adjustment value: 5-minute adjustment value = 12th adjustment value + 5-minute stable value; 10-minute adjustment value = 5-minute adjustment value + 10-minute stable value.

[0059] The health status of auxiliary components (pipeline fittings, oil tank) is synergistically correlated with the parameters of the inner and middle annular zones. Based on the Apriori algorithm, association rules between the adjustment values ​​of auxiliary components and the parameters of the inner and middle annular zones are mined. The rule structure is fixed as follows: from the first term (adjustment value of auxiliary component operating status, pressure fluctuation of the inner annular zone, and temperature change rate of the middle annular zone) to the second term (health label). The parameter settings and process of the mining are as follows: First, frequent itemsets are selected based on a minimum support of 85%. Frequent itemsets are combinations of samples containing both the first and second terms. For example, if the adjustment value of the auxiliary component operating status is ≤0.2, and the pressure fluctuation of the inner annular zone is ≤0.3MPa, and the temperature change rate of the middle annular zone is ≤0.3MPa... The change rate of temperature is ≤0.1℃ / min, and the health label is 1. This step eliminates low-frequency and meaningless parameter combinations. Then, association rules are extracted from the selected frequent items. Further filtering is performed on rules with support ≥85% and confidence ≥90% as the final judgment criteria. Support = number of samples covered by the rule / total number of samples (reflecting the universality of the rule), and confidence = number of samples that satisfy both the preceding condition and the subsequent health label / number of samples that satisfy the preceding condition (reflecting the reliability of the rule). The final rule base should contain ≥20 rules and cover different parameter combinations to avoid judgment bias caused by a single rule.

[0060] After completing the model structure design, training and validation were carried out separately for the three types of sub-models to ensure that the performance of each model met the standards. The specific process is as follows: For multilayer perceptron training and validation, the Adam optimizer was used with a fixed learning rate of 0.001 to balance training speed and stability. The number of iterations was set to 500. The cross-entropy loss function was minimized through the training set, and the final training set loss value must be ≤0.08. The validation set accuracy was calculated as the number of correctly predicted samples / the total number of validation samples. A correct prediction was defined as the deviation between the model's output health indicator and the actual health label of the sample ≤0.1, requiring a validation set accuracy ≥95%. The model learned the evolution of the execution component's state over time by fitting the time series trend to the training set. The validation set prediction error (MAE) was calculated as the sum of the absolute values ​​of the differences between the predicted and actual values ​​of all validation samples / the number of validation samples, requiring a validation set prediction error ≤0.05 to ensure short-term prediction accuracy. After generating a rule base based on the training set... The rule matching accuracy is verified using a validation set. The rule matching accuracy is calculated as the number of correctly matched samples divided by the total number of validation samples. A correct match is defined as a sample parameter that conforms to the description of the first term of the rule and the sample health label that is consistent with the second term of the rule. The rule matching accuracy is required to be ≥92%. The pre-trained fault prediction model (including three sub-models) is subjected to 5-fold cross-validation. All training data is divided into 5 groups, and 4 groups are used as the training set and 1 group as the validation set in turn. The training and validation are repeated 5 times. The final overall accuracy is calculated as the total number of correctly predicted samples in all folds divided by the total number of samples. The accuracy is required to be ≥94%. The recall is calculated as the number of correctly predicted fault samples divided by the total number of actual fault samples. Fault samples refer to samples with a health label of 0 to 0.8. Correctly predicted fault samples refer to samples whose health index output by the model is 0 to 0.8 and whose deviation from the actual health label is ≤0.1. The recall is required to be ≥91% to ensure the stability of the model on different data subsets.

[0061] After model training and validation, based on the operational status adjustment values, health status indicators for the hydraulic power station, actuators, and auxiliary components are calculated. All indicators range from 0 to 1, where 1 indicates a completely healthy component and 0 indicates a completely faulty component. The specific calculation process is as follows: Calculating the health status indicators of the hydraulic power station requires first obtaining the degree to which the operational status adjustment value deviates from the normal range, and then inputting this deviation into the model in conjunction with the adjustment value itself. The specific steps are as follows: First, calculate the deviation of the adjustment value. The formula for calculating the deviation is: Deviation = |Hydraulic power station operational status adjustment value – Historical normal range mean| ÷ Historical normal range standard deviation. Here, the historical normal range mean is the sum of all hydraulic power station operational status adjustment values ​​in the fault-free data ÷ the total number of fault-free samples, and the historical normal range standard deviation is the sum of the squares of the differences between each fault-free adjustment value and the historical normal range mean ÷ (total number of fault-free samples – 1). Then, take the square root of the result. After obtaining these two key parameters—the hydraulic power station operational status adjustment value and the deviation of the adjustment value—both are input into the pre-training process. The completed multilayer perceptron has two neurons in its input layer, corresponding to the two parameters respectively, which can accurately receive and transmit parameter information. Its hidden layer contains two layers (16 neurons in the first layer and 8 neurons in the second layer). The ReLU activation function is used to process the nonlinear relationship between parameters. It can capture the operating status adjustment value reflecting the current load level. For example, excessive load will accelerate component wear. The deviation of the adjustment value reflects the degree of deviation from the normal benchmark. If the deviation exceeds the limit, it may indicate the synergistic effect of potential abnormalities. This law is learned by the model during training by learning the parameter combination and corresponding health label under different working conditions in the training set. 1 represents complete health, and 0 to 0.8 represent different fault degrees. Finally, the information processed by the hidden layer is transmitted to the output layer (1 neuron, Sigmoid activation function). This layer strictly maps the calculation result to the range of 0 to 1. The directly output value is the health status index of the hydraulic power station, realizing the quantitative representation of the health status of the components.

[0062] The operating status of the actuators (walking hydraulic motor, auger hydraulic motor) changes dynamically over time. Health assessment needs to consider both whether the current trend is normal and whether the short-term predicted status deviates from the baseline. Therefore, the calculation process revolves around the sequence characteristics and predicted values ​​output by the ARIMA model. The specific steps are as follows: First, a time series is constructed. The ARIMA model needs to capture the trend of state evolution through continuous time data. Therefore, data is collected at a frequency of 12 times (5 minutes / time × 12 times = 1 hour) to collect the adjustment values ​​of the actuator's operating status every 5 minutes for 1 hour. The original sequence is then formed in chronological order. Following the aforementioned ARIMA model processing flow (first, the original sequence is first differencing to achieve stabilization, then the parameters are fitted using the least squares method), a stationary sequence and fixed fitting parameters are obtained, namely, autoregressive coefficients of 0.6 and 0.3, moving average coefficients of -0.2 and -0.1, and a constant term of 0.02. Stationary sequences can eliminate trend interference from the original sequence. Next, sequence similarity is calculated using a dynamic time warping algorithm. First, the comparison object is defined: the current stationary sequence of the executing component. After ARIMA model sequence stabilization, a similarity is obtained compared to a historical normal stationary sequence (constructed based on fault-free data, fully recording the sequence fluctuation characteristics of the executing component under healthy conditions, serving as a health benchmark for similarity comparison). For example, when the executing component is running fault-free, the adjusted values ​​collected every 5 minutes for one consecutive hour are 0.12, 0.13, 0.12, 0.14, 0.13, 0.15, 0.14, 0.13, 0.15, 0.14, 0.16, 0.15. After performing a first-order difference on this original sequence, the resulting stationary sequence is 0.01, -0.01, 0.02, -0.01, 0.02, -0.01, -0.01, 0.02, -0.01, 0.02, -0.01, 0.02, -0.01, 0.02, -0.01, 0.02, -0.01, 0.02, -0.01, 0.02, -0.01.01. This sequence is a historical normal stationary sequence, with small data fluctuations and no abnormal trends, reflecting the dynamic characteristics of the execution component under healthy conditions. Next, the dynamic time warping algorithm is initiated. The core function of this algorithm is to flexibly adjust the node matching relationship between the two sequences. For example, it non-strictly aligns a fluctuation node in the current sequence with a similar fluctuation node in the historical normal sequence, eliminating minor differences in the timing rhythm of the two sequences and ensuring the fairness of the comparison. Then, after the optimal matching path is determined, the distance between each pair of matching nodes (the node of the current stationary sequence and the node of the historical normal stationary sequence) on the path is calculated, and the distances of all node pairs are summed to obtain the cumulative distance reflecting the overall difference between the two sequences. Finally, the cumulative distance is normalized, mapping the result to the range of 0 to 1. This normalized result is the sequence similarity. The smaller the cumulative distance, the smaller the overall difference between the two sequences. The closer the normalized sequence similarity is to 1, the more consistent the current dynamic trend of the execution component is with the dynamic trend under healthy conditions. Next, the predicted value deviation is calculated, which is used to evaluate... The deviation between the future 10-minute adjustment value output by the ARIMA model and the normal baseline is calculated using the formula: Predicted Value Deviation = |Future 10-minute Adjustment Value - Historical Normal Range Mean of Executor Component| ÷ Historical Normal Range Standard Deviation of Executor Component. Here, the historical normal range mean and standard deviation of the execution component are based on fault-free data statistics, corresponding to the average adjustment value and fluctuation range during normal operation of the execution component, respectively. This indicator quantifies whether there is an abnormal tendency in the short-term predicted state. Finally, a health indicator is calculated. Considering that the current state trend has a higher weight in influencing real-time health, and the short-term predicted state is used as an auxiliary reference, the execution component health status indicator = sequence similarity × 0.6 + (1 - Predicted Value Deviation) × 0.4. Here, (1 - Predicted Value Deviation) converts the deviation into a positive health contribution value. The smaller the deviation, the closer this value is to 1, and the greater the positive contribution to the health indicator. Through the weighting of 0.6 and 0.4, a balance between the current state and the short-term trend is achieved, ensuring that the indicator comprehensively reflects the health status of the execution component.

[0063] The health status of auxiliary components is synergistically correlated with the parameters of the inner and middle ring bands. Therefore, it is necessary to first match the corresponding rules in the rule base, and then calculate the indicators based on the rules. The specific steps are as follows: First, perform rule matching. Match the three parameters related to the status of the auxiliary components—the adjustment value of the auxiliary component's operating status, the current pressure fluctuation of the inner ring band, and the current temperature change rate of the middle ring band—with the normal rules and fault rules in the rule base. The normal rule refers to the rule with a health label of 1. Its upper limit is the maximum allowable value of the adjustment value of the auxiliary component's operating status in the preceding item of the rule. For example, if the rule is that the adjustment value of the auxiliary component's operating status is ≤0.2, the pressure fluctuation of the inner ring band is ≤0.3MPa, and the temperature change rate of the middle ring band is ≤0.1℃ / min, the upper limit of the rule with health label 1 is 0.2, which represents the maximum allowable range of the adjustment value when the auxiliary component is in a healthy state. The fault rule refers to the rule with a health label of 0 to 0.8. Its lower limit is the minimum allowable value of the adjustment value of the auxiliary component's operating status in the preceding item of the rule. For example, if the rule is that the adjustment value of the auxiliary component's operating status is ≥0.6, the upper limit of the rule with health label 0 is 0.2. The lower limit of rule .5 is 0.6, representing the minimum threshold for adjustment values ​​when auxiliary components show signs of failure. Then, health indicators are calculated based on the matching results. If the match is a normal rule, the health status indicator = (rule upper limit - auxiliary component operating status adjustment value) / rule upper limit × 0.7 + rule matching degree × 0.3, where the rule matching degree is the degree of agreement between the sample parameters and the preceding rule description. A perfect match results in a rule matching degree of 1. For example, if the sample auxiliary component operating status adjustment value is 0.15 and the inner ring pressure fluctuation is 0.25 MPa... If all conditions are met, the rule matching degree is 1. If the actual pressure fluctuation of the inner ring is 0.32 MPa, exceeding the 0.3 MPa limit of the rule, the rule matching degree is calculated as 1 - (0.32 - 0.3) / 0.3 to quantify the degree of conformity between the parameters and the rule. If the matching rule is a fault rule, the auxiliary component health status index = 1 - (auxiliary component operating status adjustment value - rule lower limit) / (1 - rule lower limit). This formula can be used to quantify the health status in reverse according to the degree to which the adjustment value exceeds the fault threshold.

[0064] After all the health status indicators of the hydraulic power station, actuators, and auxiliary components are calculated, the final output of these three independent component health status indicators is as follows: hydraulic power station health status indicator, actuator health status indicator, and auxiliary component health status indicator. All indicators are in the range of 0 to 1, where 1 indicates that the component is completely healthy and 0 indicates that the component is completely faulty, which helps to accurately determine the components that need maintenance and the maintenance priority.

[0065] This embodiment addresses the issue of the one-sidedness of single-parameter representation by comprehensively considering the impact of multiple feature parameters on the component status through weighted fusion of multiple feature parameters; normalization processing eliminates dimensional differences, ensuring the comparability of fusion results; and ring-zone generation of operating status adjustment values ​​realizes independent state quantification of hydraulic power station, actuators, and auxiliary components, laying the foundation for accurate assessment. A pre-trained fault prediction model selects an appropriate algorithm for different component characteristics, improving the accuracy of health status assessment. The model undergoes rigorous training and validation to ensure the reliability of the output health indicators in the 0-1 range, which can intuitively reflect the component's health level. Three independent indicators accurately locate faulty components, providing a clear basis for targeted maintenance.

[0066] like Figure 2 As shown, in another preferred embodiment of the present invention, when any health status indicator is lower than a preset threshold, a corresponding maintenance operation is performed, including:

[0067] When the health status index of the hydraulic power station falls below the first preset threshold, the output pressure of the hydraulic power station is adjusted to reduce the workload of the hydraulic pump. When the health status index of the actuator falls below the second preset threshold, the walking hydraulic motor's walking action or the auger hydraulic motor's sludge removal action is stopped. When the health status index of the auxiliary component falls below the third preset threshold, a maintenance alarm is issued, instructing the oil tank, pipeline accessories, lighting devices, or imaging devices to be inspected. Specifically, this includes: defining the preset thresholds for the health status indexes of each component. These thresholds are determined based on the distribution of health indicators in fault-free operation data and the equipment's safety redundancy requirements. Specifically, the first preset threshold (hydraulic power station) is 90% of the minimum value of the fault-free health index of the hydraulic power station. The calculation method is: first preset threshold = minimum value of fault-free health index of hydraulic power station × 0.9. This calculation ensures that the threshold is below the fault-free state limit and retains safety redundancy. For example, hydraulic... When the minimum health indicator value in the fault-free data of the power station is 0.67, the first preset threshold is 0.67 × 0.9 ≈ 0.6. The second preset threshold (for the execution component) is 83% of the minimum health indicator value in the fault-free data of the execution component, calculated as: second preset threshold = minimum health indicator value of the execution component × 0.83. This setting can prevent unnecessary shutdowns caused by slight fluctuations in the indicators. For example, when the minimum health indicator value in the fault-free data of the execution component is 0.6, the second preset threshold is 0.6 × 0.83 ≈ 0.5. The third preset threshold (for auxiliary components) is 80% of the minimum health indicator value in the fault-free data of the auxiliary component, calculated as: third preset threshold = minimum health indicator value of the auxiliary component × 0.8. This ensures that abnormalities in the auxiliary components can be identified early. For example, when the minimum health indicator value in the fault-free data of the auxiliary component is 0.5, the third preset threshold is 0.5 × 0.8 = 0.4.

[0068] When the health status index of the hydraulic power station is less than the first preset threshold, the output pressure adjustment mechanism is activated. The output pressure adjustment value is calculated as: Output pressure adjustment value = Current output pressure of the hydraulic power station × (Health status index of the hydraulic power station / First preset threshold). This formula ensures that the output pressure is positively correlated with the health status, preventing the maintenance of a high load when the health index is too low. For example, when the current output pressure is 15MPa, the health status index is 0.5, and the first preset threshold is 0.6, the output pressure adjustment value = 15 × (0.5 / 0.6) = 12.5MPa. Subsequently, a load reduction operation is performed, controlling the displacement adjustment mechanism of the hydraulic pump to reduce the output pressure to the calculated adjustment value. At the same time, the adjustment time, the pressure before and after the adjustment, and the health index are recorded to form an emergency handling log.

[0069] When the health status index of the actuator is less than the second preset threshold, a shutdown action is initiated according to the type of actuator. If the actuator is a walking hydraulic motor, a power-off command is triggered to stop the walking hydraulic motor's walking action, including forward, backward, and turning actions. At the same time, the walking control handle is locked to prevent misoperation. If the actuator is an auger hydraulic motor, a shutdown command is triggered to stop the sludge removal system, stopping the auger hydraulic motor's rotational sludge removal action. At the same time, the feed valve of the auger conveying channel is closed to prevent material accumulation.

[0070] When the health status index of the auxiliary component is less than the third preset threshold, the maintenance alarm system is activated. The alarm triggering logic is as follows: the type of component associated with the abnormality is determined according to the association rules corresponding to the health status index of the auxiliary component. If the abnormality is associated with the oil tank parameters, such as excessive oil temperature fluctuation, the alarm indicates that the oil tank needs maintenance; if it is associated with the pipeline parameters, such as abnormal pressure loss, the alarm indicates that the pipeline accessories need maintenance; if it is associated with the monitoring device parameters, such as signal transmission interruption, the alarm indicates that the lighting device or video device needs maintenance. The alarm output forms include audible and visual alarms through the equipment control panel, including flashing red lights and buzzers, and pop-up prompts on the remote monitoring platform. At the same time, maintenance prompts are pushed to the maintenance personnel terminal, specifying the components to be maintained and their priorities.

[0071] After maintenance operations are performed, the health status indicators of the hydraulic power station, actuators, and auxiliary components are continuously monitored until all health status indicators recover to above their corresponding preset thresholds. Specifically, this includes: after maintenance personnel complete repair operations such as adjusting hydraulic pump pressure, replacing faulty motors, and cleaning the oil tank, continuous health status monitoring is initiated, with a monitoring frequency of collecting the operational status adjustment values ​​and loop parameters of each component every 5 minutes; data processing is performed according to model calculation logic (multilayer sensor, ARIMA, association rule mining) to calculate the health status indicators of the hydraulic power station, actuators, and auxiliary components in real time; during continuous monitoring, the following conditions are used to determine whether the health status of each component has recovered: the recovery condition for a single component is that the health status indicator of a certain component is not lower than its corresponding preset threshold for three consecutive times (each time at 5-minute intervals). The values ​​(hydraulic power station not less than the first preset threshold, actuator not less than the second preset threshold, and auxiliary component not less than the third preset threshold) are set to ensure that the indicators recover to a stable state rather than fluctuating instantaneously. The overall recovery condition is that all three components meet the individual component recovery condition, that is, the health status indicators of the hydraulic power station, actuator, and auxiliary component are not less than the corresponding preset threshold for three consecutive times. When the overall recovery condition is met, the emergency treatment state is lifted. For the hydraulic power station, the output pressure is gradually restored to the normal working pressure at a recovery rate of 5MPa / minute to avoid sudden pressure rise impacting the system. For the actuator, the travel control handle is unlocked or the auger feed valve is opened to allow the actuator to resume normal operation. For the auxiliary component, the audible and visual alarms and remote prompts are turned off. The maintenance recovery time and health indicators before and after recovery are recorded, and the equipment maintenance file is updated.

[0072] This embodiment, by monitoring health status indicators in real time and promptly initiating emergency procedures when indicators fall below preset thresholds, can prevent minor abnormalities such as hydraulic power station overload damage and actuator jamming and wear from developing into serious malfunctions, thus reducing equipment maintenance costs. Precise shutdown actions for actuators can avoid safety accidents caused by malfunctions of actuators in faulty states. Targeted alarms for auxiliary components ensure that maintenance personnel can quickly locate the objects to be repaired, reducing the risk of main system shutdowns due to auxiliary system anomalies. Continuous monitoring and frequent recovery assessments after maintenance can avoid repeated shutdowns caused by incomplete maintenance, ensuring that the equipment is restored to a stable and healthy state before being put back into use. At the same time, emergency procedures based on health indicators, rather than blind shutdowns, can reduce unnecessary production interruptions and improve the overall operating efficiency of the equipment.

[0073] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the system as described above. All implementations in the above system embodiments are applicable to this embodiment and can achieve the same technical effects.

[0074] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the system as described above. All implementations in the above system embodiments are applicable to this embodiment and can achieve the same technical effects.

[0075] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A pipeline dredging robot fault prediction and maintenance system, characterized in that, include: The acquisition module is used to acquire operating parameters in real time through four measuring points set on the hydraulic power station and the body of the pipeline dredging robot. The operating parameters include vibration data, pressure data and temperature data, and the operating parameters are arranged in chronological order to form a time-series data sequence. The module is divided into sections. Using the hydraulic pump of the hydraulic power station as the power source, and based on the flow path of the hydraulic oil and the direction of power transmission, a ring-shaped analysis region is constructed from the power source to the execution end. This ring-shaped analysis region is then radially divided into three operational state analysis rings: the inner ring corresponds to the hydraulic power generation area, including the hydraulic pump and main control valve group; the middle ring corresponds to the power transmission and core execution area, including the travel hydraulic motor and the auger hydraulic motor; and the outer ring corresponds to the functional auxiliary area, including the oil tank, pipeline accessories, lighting devices, and imaging devices. The physical components corresponding to the four measurement points in the time-series data are then analyzed. The components are assigned to either the inner or middle ring belt. Based on the pressure data from the measuring points of the hydraulic pump and main control valve group in the inner ring belt, the fluctuation range of the hydraulic power station's output pressure is calculated. Based on the temperature data from the measuring points of the traveling hydraulic motor and auger hydraulic motor in the middle ring belt, the temperature change rate of key actuators is calculated. The pressure fluctuation range and temperature change rate are correlated. When the pressure fluctuation increases and the temperature change rate rises abnormally, it is determined that there is a risk of blockage in the outer ring belt pipeline accessories. When the pressure fluctuation is stable but the temperature change rate rises slowly, it is determined that the heat dissipation performance of the outer ring belt oil tank has decreased, thus obtaining the inference result of the outer ring belt's operating status. The classification module categorizes the hydraulic pump and main control valve assembly measurement point data in the time-series data into inner loop data, and the travel hydraulic motor and auger hydraulic motor measurement point data into middle loop data. Based on the inner loop data, it calculates the maximum amplitude of the vibration data at the hydraulic pump measurement point as the inner loop vibration amplitude, the standard deviation of the pressure data at the main control valve assembly measurement point as the inner loop pressure fluctuation, and the minute-by-minute change in the temperature data at the hydraulic pump and main control valve assembly measurement points as the inner loop temperature change rate. Based on the middle loop data, it calculates the average vibration amplitude at the travel hydraulic motor and auger hydraulic motor measurement points as the middle loop vibration amplitude, the range of the pressure data at the travel hydraulic motor and auger hydraulic motor measurement points as the middle loop pressure fluctuation, and the maximum minute-by-minute change in the temperature data at the travel hydraulic motor and auger hydraulic motor measurement points as the middle loop temperature change rate. Based on the outer loop operating status inference results, it quantifies the pipeline accessory blockage risk level as an outer loop pressure fluctuation characteristic parameter and the degree of deterioration in the oil tank heat dissipation performance as an outer loop temperature change rate characteristic parameter. The adjustment module is used to analyze the characteristic parameters of the ring belt based on the internal, middle and external operating status, generate component operating status adjustment values, and input the operating status adjustment values ​​into the pre-trained fault prediction model to obtain component health status indicators. The execution module is used to perform corresponding maintenance operations when any health status indicator falls below a preset threshold.

2. The pipeline dredging robot fault prediction and maintenance system according to claim 1, characterized in that, Operating parameters are acquired in real time through four measuring points set on the hydraulic power station and the main body of the pipeline dredging robot. These operating parameters include vibration data, pressure data, and temperature data. The operating parameters are arranged in chronological order to form a time-series data sequence, including: Four measuring points were set at the hydraulic pump outlet of the hydraulic power station, the main control valve group outlet, the walking hydraulic motor inlet of the dredging robot body, and the auger hydraulic motor inlet of the dredging device. Operational parameters were collected from four measurement points, including vibration data, pressure data, and temperature data. The operational parameters of each measurement point were assigned a unique timestamp. The operational parameters of the four measurement points were then arranged and integrated according to the order of the timestamps to form a time-series data sequence containing spatial location attributes and temporal attributes.

3. The pipeline dredging robot fault prediction and maintenance system according to claim 2, characterized in that, Based on the characteristic parameters of the internal, middle, and external operating status analysis ring, component operating status adjustment values ​​are generated. These adjustment values ​​are then input into a pre-trained fault prediction model to obtain component health status indicators, including: The three characteristic parameters of vibration amplitude, pressure fluctuation and temperature change rate of the inner ring belt are weighted and fused to generate the operating status adjustment value of the hydraulic power station. The three characteristic parameters of vibration amplitude, pressure fluctuation and temperature change rate of the middle ring belt are weighted and fused to generate the operating status adjustment value of the dredging robot's execution components, including the walking hydraulic motor and the auger hydraulic motor. The two characteristic parameters of pressure fluctuation and temperature change rate of the outer ring belt are weighted and fused to generate the operating status adjustment value of the auxiliary components. The operating status adjustment values ​​of the hydraulic power station, the execution component, and the auxiliary component are used as component operating status adjustment values ​​and input into the pre-trained fault prediction model to obtain component health status indicators, including the health status indicators of the hydraulic power station, the execution component, and the auxiliary component.

4. The pipeline dredging robot fault prediction and maintenance system according to claim 3, characterized in that, The operating status adjustment values ​​of the hydraulic power station, the actuators, and the auxiliary components are used as component operating status adjustment values ​​and input into a pre-trained fault prediction model to obtain component health status indices, including hydraulic power station health status indices, actuator health status indices, and auxiliary component health status indices, including: The hydraulic power station's operating status adjustment value is compared with the historical normal operating data range of the hydraulic power station, and the health status index of the hydraulic power station is calculated using the multilayer perceptron in the pre-trained fault prediction model. The operating status adjustment value of the actuator is compared with the historical normal operating data range of the actuator, and the health status index of the actuator is calculated using the time series analysis algorithm in the pre-trained fault prediction model. The operating status adjustment value of the auxiliary component is compared with the historical normal operating data range of the auxiliary component, and the health status index of the auxiliary component is calculated using the association rule mining algorithm in the pre-trained fault prediction model. Finally, three independent component health status indices ranging from 0 to 1 are obtained, where 1 represents a completely healthy state and 0 represents a completely faulty state.

5. The pipeline dredging robot fault prediction and maintenance system according to claim 4, characterized in that, When any health status indicator falls below a preset threshold, the corresponding maintenance operation is performed, including: When the health status index of the hydraulic power station is lower than the first preset threshold, the output pressure of the hydraulic power station is adjusted to reduce the working load of the hydraulic pump. When the health status index of the actuator is lower than the second preset threshold, the walking hydraulic motor is stopped or the sludge removal action of the auger hydraulic motor is stopped. When the health status index of the auxiliary component is lower than the third preset threshold, a maintenance alarm is issued, indicating that the oil tank, pipeline accessories, lighting device or imaging device should be inspected. After the maintenance operation is performed, continue to monitor the changes in the health status indicators of the hydraulic power station, the execution components, and the auxiliary components until all health status indicators recover to above the corresponding preset threshold.

6. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the system as described in any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the system as described in any one of claims 1 to 5.

Images