Adaptive pressure compensation method and system for underwater thrusters

Abstract

The application relates to the technical field of data processing, and discloses a self-adaptive pressure compensation method and system for an underwater propeller. The method comprises the following steps: establishing a quantitative relationship between a pressure difference and a diaphragm deformation in a ground calibration stage, calculating a reference damping ratio, providing an accurate reference for judging oil deterioration, collecting pressure difference deviation data in a work cycle, and adaptively tracking time-varying drift of oil viscosity through an iterative learning law, so that the defects that a traditional fixed parameter control cannot adapt to working condition changes are overcome, a fractional order control algorithm and a feedforward compensation mechanism are combined to realize active prediction and adjustment, a transient pressure difference peak value is reduced, and high-precision compensation capability is maintained throughout the whole life cycle through continuous learning and reference updating. Through integration of conductivity monitoring and a filter screen self-cleaning mechanism, the problems of an oil pollution detection blind area and unreasonable maintenance period are solved, and the long-term stability and intelligent level of the compensation system are improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to an adaptive pressure compensation method and system for underwater thrusters. Background Technology

[0002] As the core power unit of autonomous underwater vehicles, underwater thrusters need to operate stably for extended periods in marine environments at varying depths. Existing technologies widely employ oil-filled pressure compensation structures to address the pressure difference between the inside and outside of the thruster in deep water. This structure fills the thruster's sealed cavity with oil and incorporates a flexible compensation diaphragm, allowing the internal oil pressure to automatically adjust with changes in external water pressure, thus maintaining a basic balance between internal and external pressures. This significantly reduces the pressure load on the sealing structure. Representative products such as Haoye thrusters have successfully applied magnetic coupling sealing structures and oil-filled self-compensation technology, achieving excellent application results in the field of underwater equipment.

[0003] However, existing oil-filled pressure compensation technology still has significant shortcomings. First, during long-term operation, factors such as motor heating leading to temperature rise, molecular chain breakage causing aging, and water infiltration or particulate contamination can cause time-varying drift in the viscosity and bulk elastic modulus of the compensation oil. However, traditional passive diaphragm compensation structures cannot sense and adapt to the deterioration of oil performance, resulting in a gradual decrease in compensation accuracy over time. Second, during the rapid deepening of the thruster, the passive diaphragm compensation mechanism exhibits a lag in response to changes in external water pressure. Instantaneous pressure differential impacts can still cause fatigue damage to the seals. Existing technologies lack the ability to monitor the oil status in real time and actively adjust the pressure differential. They can only periodically disassemble and inspect the oil quality, resulting in high maintenance costs and delayed fault warnings.

[0004] Further analysis revealed that even with the introduction of an active pressure compensation mechanism, if the time-varying viscoelastic parameters of the oil cannot be accurately identified, the control algorithm will still adjust based on the initial calibration value and will not be able to adapt to changes in actual operating conditions. This necessitates an adaptive parameter identification method to reconstruct the equivalent viscosity and damping ratio of the oil in real time. However, data collected in a single operating cycle is greatly affected by random disturbances and cannot reveal the long-term evolution of oil performance. Therefore, it is necessary to establish a cross-cycle iterative learning mechanism to extract the regular characteristics of oil degradation during repetitive operations and gradually correct the compensation strategy. At the same time, traditional integer-order PI controllers have poor adaptability to the viscoelastic hysteresis response and memory effect of oil-filled systems. It is necessary to design a fractional-order control algorithm specifically for the dynamic characteristics of oil compensation systems to balance fast response and steady-state accuracy. Furthermore, pressure compensation alone cannot solve the problems of oil contamination and filter clogging. It is also necessary to integrate oil quality monitoring and filter self-cleaning functions to form an oil health management system. Summary of the Invention

[0005] This application provides an adaptive pressure compensation method and system for underwater thrusters. It uses an iterative learning observer to identify time-varying viscoelastic parameters of the oil in real time and combines this with a fractional-order control algorithm to achieve active pressure regulation, solving the problems of traditional passive compensation's inability to adapt to oil performance degradation and response lag. Simultaneously, by integrating conductivity monitoring and a filter self-cleaning mechanism, it addresses the issues of blind spots in oil contamination detection and unreasonable maintenance cycles, improving the long-term stability and intelligence level of the compensation system.

[0006] In a first aspect, this application provides an adaptive pressure compensation method for an underwater thruster, the adaptive pressure compensation method for an underwater thruster comprising:

[0007] Step S1: Apply segmented pressure loading to the underwater thruster compensation chamber, record the diaphragm displacement and internal and external pressure difference at each depth point, and calculate the reference damping ratio through the pressure oscillation decay curve;

[0008] Step S2: Collect the measured pressure difference value of the compensation cavity in the operation cycle, calculate the difference between the measured pressure difference value and the theoretical pressure difference value corresponding to the current depth, and obtain the pressure difference deviation dataset by averaging over the depth interval;

[0009] Step S3: Extract the same depth deviation value between the current cycle and the previous cycle from the differential pressure deviation dataset, calculate the viscosity correction term according to the iterative learning law, and calculate the reconstruction damping ratio based on the viscosity correction term;

[0010] Step S4: The degree of oil deterioration is determined by the difference between the reconstructed damping ratio and the reference damping ratio. The feedforward compensation flow rate is calculated based on the reconstructed damping ratio and the external water pressure change rate. The compensation pump flow rate command is generated by combining the fractional integral term of the differential pressure tracking error.

[0011] Step S5: Drive the compensation pump to inject oil according to the compensation pump flow command so that the measured pressure difference value converges to the target value. Store the pressure difference deviation dataset of this cycle and update the theoretical pressure difference value for the next cycle.

[0012] Secondly, this application provides an adaptive pressure compensation system for an underwater thruster, the adaptive pressure compensation system for an underwater thruster comprising:

[0013] The calculation module is used to apply segmented pressure loading to the underwater thruster compensation chamber, record the diaphragm displacement and internal and external pressure difference at each depth point, and calculate the reference damping ratio through the pressure oscillation decay curve.

[0014] The acquisition module is used to acquire the measured pressure difference value of the compensation chamber during the operation cycle, and to obtain the pressure difference deviation dataset by subtracting the measured pressure difference value from the theoretical pressure difference value corresponding to the current depth and averaging it over the depth interval.

[0015] The correction module is used to extract the same depth deviation value between the current cycle and the previous cycle in the differential pressure deviation dataset, calculate the viscosity correction term according to the iterative learning law, and calculate the reconstruction damping ratio based on the viscosity correction term.

[0016] The generation module is used to determine the degree of oil deterioration by subtracting the reconstructed damping ratio from the reference damping ratio, calculate the feedforward compensation flow rate based on the reconstructed damping ratio and the external water pressure change rate, and generate a compensation pump flow command by combining the fractional integral term of the differential pressure tracking error.

[0017] The injection module is used to drive the compensation pump to inject oil according to the compensation pump flow command so that the measured pressure difference value converges to the target value, and to store the pressure difference deviation dataset of the current cycle and update the theoretical pressure difference value for the next cycle.

[0018] Thirdly, an adaptive pressure compensation device for an underwater thruster is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the adaptive pressure compensation device for the underwater thruster to execute the aforementioned adaptive pressure compensation method for the underwater thruster.

[0019] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the aforementioned adaptive pressure compensation method for underwater thrusters.

[0020] The technical solution provided in this application establishes a quantitative correlation between pressure difference and diaphragm deformation by applying segmented pressure loading to the compensation chamber during the ground calibration stage and recording the diaphragm displacement and internal / external pressure difference at each depth point. The benchmark damping ratio is calculated using the pressure oscillation decay curve, providing a quantifiable reference for subsequent assessment of oil degradation. Compared to traditional methods that rely solely on experience to judge oil condition, this method achieves accurate characterization of the oil's viscoelastic properties. In actual operating cycles, the measured pressure difference values ​​of the compensation chamber are continuously collected at high frequency and subtracted from the theoretical pressure difference value corresponding to the current depth. The difference is then averaged across depth intervals to obtain a pressure difference deviation dataset. This dataset not only eliminates the interference of instantaneous random noise but also retains low-frequency trend components reflecting changes in oil performance, providing high-quality input data for iterative learning algorithms. By extracting the same depth deviation value between the current cycle and the previous cycle and calculating the viscosity correction term according to the iterative learning law, this method realizes the function of automatically learning the evolution law of oil performance from repetitive operation. The cumulative update mechanism of the viscosity correction term enables the system to adaptively track the viscosity drift of oil caused by various factors such as temperature rise, aging and contamination. It overcomes the shortcomings of traditional methods that are based on fixed parameter control and cannot adapt to time-varying working conditions. The reconstructed damping ratio calculated according to the viscosity correction term directly reflects the actual viscoelastic state of the current oil. The difference between the reconstructed damping ratio and the reference damping ratio can not only quantitatively judge the degree of oil deterioration, but also serve as a criterion for starting active compensation. This avoids unnecessary active intervention when the oil performance is normal, reduces energy consumption and extends the service life of the compensation pump.

[0021] The core innovation of this method lies in the deep integration of an iterative learning observer and a fractional-order control algorithm into the underwater propulsion oil filling compensation system. The iterative learning observer extracts the regularity characteristics of oil performance degradation by analyzing historical operation cycle data and gradually corrects the control parameters. Compared with traditional adaptive control that only uses current-moment information, the iterative learning mechanism fully explores historical evolution information in the time dimension, enabling the parameter identification accuracy to continuously improve with the number of cycles. The introduction of the fractional-order integral term is specifically optimized for the memory effect and viscoelastic hysteresis characteristics of the oil filling system. The 0.6th-order integral combines the fast response of traditional proportional control with the steady-state accuracy of integral control, while avoiding the over-tuning and oscillation problems that are prone to occur when dealing with viscoelastic systems by integer-order integrals. The feedforward compensation flow is dynamically calculated based on the reconstructed damping ratio and the external water pressure change rate. By intervening in the adjustment before the differential pressure deviates significantly from the target value, the control mode has been transformed from passive response to active prediction. This feedforward mechanism is particularly suitable for handling transient pressure shocks when the thruster rapidly deepens, effectively reducing the peak differential pressure borne by the seals. The driving compensation pump injects oil according to the flow command and calculates the pressure increment based on the oil bulk modulus. After deducting the natural leakage of the diaphragm gap, the predicted differential pressure value is fed back to the controller, forming a closed-loop feedback adjustment mechanism. The current cycle data is stored and the theoretical differential pressure value is updated based on the data from the most recent cycles, so that the system's compensation benchmark can be adaptively adjusted with the long-term evolution of oil performance. Compared with the traditional method that uses a fixed initial calibration curve, which causes the compensation accuracy to decay over time, this method maintains the high-precision operation capability of the compensation system throughout its entire life cycle through continuous learning and benchmark update mechanisms. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of one embodiment of the adaptive pressure compensation method for underwater thrusters in this application.

[0024] Figure 2 This is a schematic diagram of one embodiment of the adaptive pressure compensation system for underwater thrusters in this application.

[0025] Figure 3 This is a schematic block diagram of the structure of the adaptive pressure compensation device for underwater thrusters in an embodiment of the present invention. Detailed Implementation

[0026] This application provides an adaptive pressure compensation method and system for underwater thrusters. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0027] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the adaptive pressure compensation method for underwater thrusters in this application includes:

[0028] Step S1: Apply segmented pressure loading to the underwater thruster compensation chamber, record the diaphragm displacement and internal and external pressure difference at each depth point, and calculate the reference damping ratio through the pressure oscillation decay curve;

[0029] Step S2: Collect the measured pressure difference value of the compensation cavity in the operation cycle, calculate the difference between the measured pressure difference value and the theoretical pressure difference value corresponding to the current depth, and obtain the pressure difference deviation dataset by averaging over the depth interval;

[0030] Step S3: Extract the same depth deviation value between the current cycle and the previous cycle from the differential pressure deviation dataset, calculate the viscosity correction term according to the iterative learning law, and calculate the reconstruction damping ratio based on the viscosity correction term;

[0031] Step S4: The difference between the reconstructed damping ratio and the reference damping ratio is used to determine the degree of oil deterioration. The feedforward compensation flow rate is calculated based on the reconstructed damping ratio and the external water pressure change rate. The compensation pump flow command is generated by combining the fractional integral term of the differential pressure tracking error.

[0032] Step S5: Drive the compensation pump to inject oil according to the compensation pump flow command so that the measured differential pressure value converges to the target value. Store the differential pressure deviation dataset of this cycle and update the theoretical differential pressure value for the next cycle.

[0033] It is understood that the executing entity of this application can be an adaptive pressure compensation system for underwater thrusters, or it can be a terminal or a server; no specific limitation is made here. This application's embodiments use a server as an example for illustration.

[0034] Specifically, an iterative learning observer is used to adaptively identify the viscoelastic parameters of the compensation oil, and a fractional-order control algorithm is combined to achieve active pressure compensation. During the ground calibration phase, silicon-based synthetic oil is injected into the compensation chamber and simulated deep-water pressure is applied. The diaphragm displaces under pressure, and displacement sensors record the displacement and pressure difference at each depth point. The initial elastic modulus is obtained by fitting the stress-strain data using the least squares method. Subsequently, the descent process is simulated by applying pressure at a constant speed, and the pressure in the compensation chamber oscillates. The first and second pressure peaks in the oscillation curve are extracted, and the ratio of the two peaks reflects the system's damping characteristics. The natural logarithm of this ratio is substituted into the characteristic equation to calculate the benchmark damping ratio, which characterizes the standard viscoelastic state of the fresh oil. During actual operation, the adaptive iterative learning observer is activated. The differential pressure sensor collects the measured value of the differential pressure inside and outside the compensation chamber at a frequency of ten times per second. The controller obtains the theoretical differential pressure value corresponding to the current depth from the initial calibration data through linear interpolation. The difference between the measured value and the theoretical value is used to obtain the differential pressure deviation signal. Since the water depth ranges from zero meters to five hundred meters and the oil performance exhibits different deterioration characteristics at different depths, the entire depth range is divided into ten intervals, each spanning fifty meters. The differential pressure deviation signals collected in each interval are processed by time averaging to eliminate the influence of instantaneous fluctuations, obtain representative deviation values ​​for each depth interval, and combine them into a differential pressure deviation dataset. During iterative learning, representative deviation values ​​are extracted between the current k-th cycle and the previous k-th minus one cycle within the same depth range. The difference between the two is the deviation value, which reflects the changing trend of oil performance between the two cycles. Calculating the sensitivity coefficient of oil viscosity to pressure difference response requires the effective force-bearing area of ​​the diaphragm, the equivalent flow area of ​​the oil through the gap around the diaphragm, the flow path length, and the average displacement velocity of the diaphragm. The sensitivity coefficient represents the change in pressure difference response caused by a unit change in viscosity. Taking the reciprocal of the sensitivity coefficient yields the learning gain. The learning gain is multiplied by the deviation value difference and then superimposed with the viscosity correction term stored in the previous cycle. The viscosity correction term for the current cycle is updated according to the iterative learning law. The initial dynamic viscosity value is added to the viscosity correction term to obtain the equivalent viscosity value for the current cycle. This value comprehensively reflects the viscosity change of the oil due to temperature rise and aging. Subsequently, the equivalent viscosity value, initial elastic modulus, and oil density parameters are substituted into the viscoelastic system damping ratio calculation formula to obtain the reconstructed damping ratio.The damping ratio deviation is obtained by subtracting the reconstructed damping ratio from the reference damping ratio. When the deviation is greater than 0.05, it is determined that the oil deterioration exceeds the preset threshold and active compensation needs to be activated. The calculation of the feedforward compensation flow involves the difference between the reconstructed damping ratio and the reference damping ratio, the damping compensation coefficient, the absolute value of the external water pressure change rate, and the directional offset flow. The damping compensation coefficient is pre-calibrated according to the oil type and the structure of the compensation chamber. The external water pressure change rate is converted from the depth change rate of the depth sensor. The directional offset flow is set with positive or negative signs according to the diving or surfacing direction. The feedforward compensation flow is obtained by multiplying the difference between the reconstructed damping ratio and the reference damping ratio by the damping compensation coefficient and then by the absolute value of the external water pressure change rate, and then superimposing the directional offset flow. The differential pressure tracking error is obtained by subtracting the target differential pressure value from the measured differential pressure value. The fractional integral operation is implemented using the discretization defined by Glenn Ward-Letnikov. The control period is set to 0.1 seconds and the memory length is 50 historical windows. The weight coefficient sequence is calculated using fractional binomial coefficients. The fractional integral term is obtained by multiplying the weight coefficient sequence with the historical error sequence term by term and summing the results. The proportional gain is multiplied by the differential pressure tracking error to obtain the proportional control term. The integral gain is multiplied by the fractional integral term to obtain the integral control term. The proportional control term, integral control term, and feedforward compensation flow are added together to generate the compensation pump flow command. After receiving the flow command, the compensation pump calculates the motor speed based on the single-revolution displacement, drives the pump body to draw oil from the bottom of the compensation chamber, filters it through the filter screen, and injects it into the upper part of the compensation chamber. The injection of oil causes a pressure change in the compensation chamber. The pressure increment is calculated by substituting the oil bulk modulus of 1.65 GPa, the injection volume within the control cycle, and the total volume of the compensation chamber into the fluid bulk modulus relationship. The pressure increment is added to the current differential pressure value, and the natural leakage of the diaphragm gap calculated based on the leakage coefficient is subtracted to obtain the predicted differential pressure value. The predicted differential pressure value is fed back to the fractional-order controller to recalculate the flow command for the next cycle, forming a closed-loop regulation. Simultaneously, the local peak values ​​of the pressure curve within each control cycle are extracted to construct a peak sequence. The ratio of adjacent peak values ​​in the peak sequence is calculated. When the ratio of five consecutive peak values ​​is greater than 0.95, it is determined that the system has entered a steady state and the differential pressure no longer oscillates significantly. The compensation pump stops running and switches back to the passive diaphragm compensation mode. After the current cycle is completed, the arithmetic mean of the damping ratios for the ten depth intervals is calculated. The difference between this and the average damping ratio of the previous five cycles is divided by five to obtain the damping ratio evolution rate. This evolution rate reflects the long-term deterioration trend of the oil performance. The percentage of oil degradation is calculated based on the damping ratio evolution rate and the baseline damping ratio. When the degradation exceeds 30%, an oil replacement warning signal is generated to alert maintenance personnel. The differential pressure deviation dataset, viscosity correction term, and reconstructed damping ratio of this cycle are packaged and stored in the iterative learning database. The theoretical differential pressure value is updated based on the depth differential pressure regression surface fitted with the data from the last five cycles.

[0035] In one specific embodiment, step S1 includes:

[0036] Silicon-based synthetic oil of rated volume was injected into the compensation chamber at a standard temperature of 20°C and standard atmospheric pressure to obtain the initial dynamic viscosity value;

[0037] Pressure values ​​corresponding to water depths from 0 to 500 meters are applied to the compensation chamber in a controllable pressure chamber. A pressure acquisition point is set at every 50-meter depth, and the internal oil pressure value and external simulated water pressure value at each acquisition point are recorded. The internal and external pressure difference values ​​at each depth point are calculated.

[0038] The displacement of the diaphragm center point corresponding to each pressure acquisition point is recorded by a displacement sensor installed on the back side of the diaphragm. The initial elastic modulus is calculated by least squares fitting of the internal and external pressure difference value and the displacement of the diaphragm center point.

[0039] The pressure oscillation curve of the compensation chamber was recorded by increasing the pressure from 0 meters to 100 meters at a simulated descent speed of 1 meter per second. The first and second pressure peaks were extracted, and the ratio of the two pressure peaks was divided by the characteristic equation containing the logarithmic term to calculate the reference damping ratio.

[0040] Specifically, 850 ml of silicon-based synthetic oil was injected into the underwater thruster compensation chamber in a calibrated environment on land. The ambient temperature was controlled at 20 degrees Celsius and the standard atmospheric pressure was 0.101 MPa. The initial dynamic viscosity of the oil was measured to be 68 centiliters using a rotational viscometer. This viscosity value characterizes the flow resistance characteristics of the oil under standard conditions. The dynamic viscosity value is subsequently used to calculate the changes in the viscoelastic parameters of the oil under different temperature and pressure conditions. The thruster assembly with the compensation chamber is placed inside a controllable pressure chamber. The pressure chamber is pressurized by injecting water into the chamber using a high-pressure pump to simulate the external pressure corresponding to different water depths. The pressure is gradually increased from the standard atmospheric pressure corresponding to a water depth of 0 meters to 5.1 MPa corresponding to a water depth of 500 meters. A pressure sampling point is set at every 50-meter depth, for a total of 10 sampling points corresponding to depths of 0 meters, 50 meters, 100 meters, and up to 500 meters. At each sampling point, the pressure sensor inside the pressure chamber records the external simulated water pressure value, while the miniature differential pressure sensor installed on the inner wall of the compensation chamber simultaneously records the internal oil pressure value. The pressure difference between the outside and inside at each depth point is calculated by subtracting the internal oil pressure value from the external simulated water pressure value. For example, at a depth of 100 meters, the external pressure is 1.1 MPa while the internal oil pressure is 1.08 MPa, so the pressure difference is 0.02 MPa. A flexible compensating diaphragm is installed at the interface between the compensation chamber and the external water body. One side of the diaphragm is subjected to internal oil pressure, while the other side is subjected to external water pressure. When the external pressure is greater than the internal pressure, the diaphragm indents inward, causing displacement. A displacement sensor, using laser ranging, is installed on the back of the diaphragm to non-contactly measure the displacement of the diaphragm's center point. The displacement of the diaphragm's center point relative to its initial position is recorded at each pressure sampling point. For example, at a depth of 100 meters, the diaphragm's center point displacement is 8 millimeters. The internal and external pressure difference values ​​obtained from 10 sampling points are combined with the corresponding diaphragm center point displacements to form a data pair sequence. The least squares fitting algorithm minimizes... The optimal fitting parameters are determined by the sum of squares of the differences between the actual measured values ​​and the fitted curve values. Specifically, the pressure difference is used as the independent variable and the displacement is used as the dependent variable to establish a linear relationship model. The sum of squares of the residuals of all data pairs is calculated, and the partial derivatives with respect to the fitting parameters are set to zero. The slope between the pressure difference and the displacement is obtained, which is the diaphragm stiffness. The initial elastic modulus is obtained by combining the diaphragm stiffness with the effective force-bearing area of ​​the diaphragm. For example, if the pressure difference displacement slope is 2.5 MPa and the effective force-bearing area of ​​the diaphragm is π multiplied by the square of 0.045 m, then the initial elastic modulus is the slope multiplied by the force-bearing area, which equals 11.94 MPa.During the dynamic response test, the pressurization rate of the pressure chamber was set to a constant increase of 0.1 MPa per second, corresponding to a thruster descent speed of 1 meter per second. Pressurization from 0 meters to 100 meters corresponds to a pressure of 1.1 MPa. During pressurization, the pressure within the compensation chamber did not rise smoothly but exhibited oscillating characteristics. This oscillation was caused by the overshoot and decay process of the viscoelastic system composed of the oil and diaphragm in response to external pressure step changes. A high-speed differential pressure sensor recorded the pressure changes within the compensation chamber at a frequency of 100 times per second, forming a pressure oscillation curve. The local maxima points identified from this curve are the pressure peaks. The first pressure peak occurred at the initial pressurization stage, corresponding to a pressure of 1.12 MPa. The second pressure peak occurred during the decaying oscillation after the first peak, corresponding to a pressure of 1.08 MPa. Dividing the first pressure peak by the second pressure peak yields a ratio of 1. The ratio 0.037 is taken as its natural logarithm, resulting in 0.0363. The characteristic equation is the result of dividing the natural logarithm by twice pi, then dividing by 1, and then taking the square root of the result of dividing the natural logarithm by twice pi. Substituting 0.0363 into the denominator of the characteristic equation, we first calculate that 0.0363 divided by twice pi equals 0.00578. The square of this value is 0.0000334. Adding this square to 1 still results in approximately 1, and taking the square root also results in 1. Therefore, the characteristic equation simplifies to 0.0363 divided by twice pi equals 0.00578. After normalization, the reference damping ratio of 0.162 is obtained. This value characterizes the damping characteristics of fresh oil under standard conditions. When the viscosity of the oil increases due to temperature rise, aging, or contamination, the damping ratio will increase accordingly. The reference damping ratio is stored in the non-volatile memory of the control unit as a reference value for judging oil deterioration.

[0041] In one specific embodiment, step S2 includes:

[0042] The data acquisition module of the adaptive iterative learning observer is activated, and the measured value of the pressure difference between the inside and outside of the compensation cavity is continuously recorded at a sampling frequency of 10Hz during the process of the underwater thruster descending from the water surface to the target operating depth.

[0043] The theoretical pressure difference value corresponding to the current depth is obtained by linear interpolation of the internal and external pressure difference values ​​at each depth point based on the current depth value.

[0044] The difference between the measured differential pressure value and the theoretical differential pressure value is used to obtain the differential pressure deviation signal at each sampling time.

[0045] The water depth from 0 to 500 meters was divided into 10 depth intervals. The pressure difference deviation signal in each depth interval was processed by time averaging to obtain the representative deviation value of each depth interval, and the data were combined to form a pressure difference deviation dataset.

[0046] Specifically, when the underwater thruster begins actual operation, the data acquisition module of the adaptive iterative learning observer is activated. The core function of this module is to monitor the changes in the pressure difference inside and outside the compensation chamber in real time. Throughout the entire process of the thruster diving from the water surface to the target operating depth, the differential pressure sensor works continuously at a sampling frequency of 10 Hz. 10 Hz means that 10 data points are collected per second, that is, the measured value of the pressure difference inside and outside the compensation chamber is recorded every 0.1 seconds. The sampling frequency is set to 10 Hz as a result of balancing data accuracy and storage burden. If the frequency is too low, it will miss the rapidly changing pressure transient process, while if the frequency is too high, it will generate a large amount of redundant data and increase the processing burden. The miniature differential pressure sensor on the inner wall of the compensation chamber directly measures the difference between the internal oil pressure and the external water pressure and outputs the measured value of the differential pressure. This measured value is affected by various factors such as oil temperature changes, diaphragm material creep and oil performance deterioration, and will deviate from the theoretical expected value. The depth sensor synchronously records the current water depth position of the thruster. The controller uses the current depth value to perform linear interpolation calculations on the internal and external pressure difference values ​​obtained from the initial calibration phase at each depth point to obtain the theoretical pressure difference value. The specific processing logic of the linear interpolation algorithm is as follows: First, determine which two adjacent depth points in the calibration data the current depth falls between. For example, if the current depth is 75 meters, it falls between the 50-meter and 100-meter depth points. Then, retrieve the pressure difference values ​​corresponding to these two depth points from memory and record them as the pressure difference value at the 50-meter depth and... The pressure difference at a depth of 100 meters is calculated as follows: the difference between the current depth of 75 meters and the lower limit of 50 meters is 25 meters; the difference between the upper limit of 100 meters and the lower limit of 50 meters is 50 meters. Dividing these two values ​​yields an interpolation weighting coefficient of 0.5. Multiplying this difference by the interpolation weighting coefficient gives the pressure difference increment. Adding this increment to the lower limit pressure difference gives the theoretical pressure difference at the current depth of 75 meters. Linear interpolation assumes a linear relationship between the pressure difference changes between adjacent calibration points, and this assumption has high accuracy within a 50-meter interval. The difference between the measured and theoretical pressure difference is the pressure difference deviation signal. This signal reflects the degree of deviation between the actual compensation system and the ideal state. A positive deviation signal indicates that the actual pressure difference is greater than the theoretical value, suggesting obstructed external pressure transmission or insufficient internal oil compression. A negative deviation signal indicates that the actual pressure difference is less than the theoretical value, suggesting excessive diaphragm deformation or oil leakage. During the continuous descent of the thruster, the pressure difference deviation signal changes continuously over time, forming time-series data.Because the entire operating depth range spans a large area from 0 meters to 500 meters, and the oil properties are affected by temperature and pressure to varying degrees at different depths, the entire depth range is divided into 10 depth intervals, each spanning 50 meters: 0 to 50 meters, 50 to 100 meters, 100 to 150 meters, and so on up to 450 to 500 meters. The pressure difference deviation signals collected within each depth interval are subjected to time averaging. Time averaging involves summing all pressure difference deviation signals collected during the thruster's traversal of that depth interval and dividing by the number of sampling points. For example, the time it takes for the thruster to descend at a speed of 1 meter per second through the 100 to 150 meter depth interval... For 50 seconds, 500 differential pressure deviation signal data points were collected at a frequency of 10 Hz. These 500 data points were summed one by one and divided by 500 to obtain the representative deviation value for that depth interval. Time averaging was used to filter out high-frequency random noise and transient fluctuations in the differential pressure deviation signal, while retaining the low-frequency trend components that reflect changes in oil performance. Ten representative deviation values ​​were calculated for each of the ten depth intervals and arranged in order of depth to form a differential pressure deviation dataset. This dataset served as the input data for subsequent iterative learning algorithms, describing the performance deviation distribution characteristics of the compensation system at different depths during this operation cycle.

[0047] In one specific embodiment, step S3 includes:

[0048] Extract representative deviation values ​​for each depth interval from the differential pressure deviation dataset between the current k-th operation cycle and the previous (k-1)-th operation cycle, and calculate the difference in deviation values ​​for the same depth interval.

[0049] The learning gain is calculated based on the sensitivity coefficient of the oil viscosity to the pressure difference response. The learning gain is multiplied by the difference in deviation value, and the viscosity correction term from the previous cycle is added. The viscosity correction term for the current cycle is obtained according to the iterative learning law.

[0050] The initial dynamic viscosity value is added to the viscosity correction term to obtain the equivalent viscosity value for the current cycle;

[0051] Based on the equivalent viscosity, initial elastic modulus, and oil density parameters, the reconstructed damping ratio is obtained by substituting them into the formula for calculating the damping ratio of a viscoelastic system.

[0052] Specifically, the iterative learning observer extracts the differential pressure deviation dataset between the current k-th operation cycle and the previous k-minus-1 operation cycle from the stored historical database. Each dataset contains representative deviation values ​​for 10 depth intervals. The controller compares the representative deviation values ​​of the same depth interval in the two cycles one by one. For example, for the 100 to 150-meter depth interval, the representative deviation value of this interval is read from the current cycle dataset and recorded as the current deviation, and the representative deviation value of this interval is read from the previous cycle dataset and recorded as the previous deviation. The difference between the current deviation and the previous deviation is calculated to obtain the deviation value difference. The deviation value difference reflects the changing trend of the compensation system performance between the two cycles. A positive difference indicates that the differential pressure deviation is aggravated, indicating that the oil performance is deteriorating faster. A negative difference indicates that the differential pressure deviation is reduced, indicating that the previous correction effect is significant. A difference close to zero indicates that the system state is stable and has not changed significantly. The difference values ​​calculated for the 10 depth intervals constitute a difference sequence. The sensitivity coefficient of oil viscosity to pressure difference response describes the magnitude of the pressure difference response caused by a unit change in oil viscosity. This coefficient is calculated using the effective force-bearing area of ​​the diaphragm, the equivalent flow area of ​​the oil through the gaps around the diaphragm, the flow path length, and the average displacement velocity of the diaphragm. Specifically, the calculation logic is as follows: a small gap exists around the diaphragm, allowing the oil to flow slowly under pressure; the flow resistance is proportional to the oil viscosity; the equivalent flow area characterizes the effective flow cross-section of the gap; the flow path length is the distance the oil travels from one side of the diaphragm to the other; and the average displacement velocity of the diaphragm is obtained by dividing the pressure difference rate by the diaphragm stiffness. The equivalent flow area is then divided by the flow path length and multiplied by the average displacement velocity of the diaphragm. Finally, dividing by the effective force-bearing area of ​​the diaphragm gives the sensitivity coefficient. The reciprocal of this coefficient is defined as the learning gain. The learning gain characterizes the conversion ratio of pressure difference deviation changes to viscosity correction terms. If the learning gain is too large, it will lead to overcorrection and cause system oscillation. If the learning gain is too small, it will lead to undercorrection and slow convergence. The core of the iterative learning law is to multiply the learning gain by the difference in deviation value to obtain the viscosity increment that needs to be superimposed in the current cycle. Then, add this increment to the viscosity correction term stored in the previous cycle to obtain the viscosity correction term for the current cycle. The iterative learning law embodies the idea of ​​extracting patterns from historical data and making gradual corrections. Through the accumulation of multiple cycles, the viscosity correction term is gradually learned to approach the actual change in oil viscosity. The initial dynamic viscosity value is the viscosity benchmark of fresh oil measured at standard temperature during the calibration phase. The viscosity correction term is the viscosity deviation accumulated during the iterative learning process. The two are added together to obtain the equivalent viscosity value of the current cycle. The equivalent viscosity value comprehensively reflects the combined effects of various factors such as the temperature rise caused by motor heating, molecular chain breakage and aging caused by long-term operation, and water infiltration or particulate pollution. An equivalent viscosity value higher than the initial value indicates increased oil flow resistance, while a value lower than the initial value indicates measurement abnormality or algorithm divergence.Calculating the damping ratio of a viscoelastic system requires the equivalent viscosity, initial elastic modulus, and oil density parameters. The damping ratio describes the rate at which the system's oscillations decay. The calculation formula is the square root of the equivalent viscosity divided by twice the product of the initial elastic modulus, oil density, and compensation chamber volume, then divided by the effective force-bearing area of ​​the diaphragm. This formula originates from the definition of the damping ratio of a single-degree-of-freedom mass spring damped system in viscoelastic mechanics. The equivalent viscosity in the numerator characterizes the system's damping characteristics, the initial elastic modulus in the denominator characterizes the system's stiffness characteristics, and the product of oil density and compensation chamber volume characterizes the system's equivalent mass. Substituting these parameters into the formula yields the reconstructed damping ratio. Comparing the reconstructed damping ratio with the initially calibrated reference damping ratio, the difference reflects the degree of deterioration in oil performance. A significantly higher reconstructed damping ratio than the reference damping ratio indicates an abnormal increase in oil viscosity, requiring the initiation of active compensation.

[0053] In one specific embodiment, step S4 includes:

[0054] The damping ratio deviation is obtained by subtracting the reconstructed damping ratio from the reference damping ratio. When the damping ratio deviation is greater than 0.05, it is determined that the oil deterioration exceeds the threshold and active compensation is initiated.

[0055] The difference between the reconstructed damping ratio and the reference damping ratio is multiplied by the damping compensation coefficient, and then multiplied by the absolute value of the external water pressure change rate. The directional offset flow rate is then added to obtain the feedforward compensation flow rate.

[0056] The difference between the set target differential pressure value and the measured differential pressure value is used to obtain the differential pressure tracking error. The differential pressure tracking error is then integrated by fractional integral to obtain the fractional integral term.

[0057] The proportional control term is obtained by multiplying the proportional gain by the differential pressure tracking error, and the integral control term is obtained by multiplying the integral gain by the fractional integral term. The proportional control term, the integral control term, and the feedforward compensation flow are added together to generate the compensation pump flow command.

[0058] Specifically, the controller subtracts the reconstructed damping ratio from the reference damping ratio to obtain the damping ratio deviation value. This deviation value quantifies the degree of oil performance degradation. When the damping ratio deviation value is greater than 0.05, the control logic determines that the degree of oil degradation has exceeded the preset threshold and active compensation needs to be activated. The threshold of 0.05 is set based on a large amount of experimental data showing that when the damping ratio increases by more than one-third of the reference value, passive diaphragm compensation can no longer maintain pressure differential stability. After the flag signal for activating active compensation is triggered, the controller activates the compensation pump drive circuit and begins to calculate the compensation pump flow command. The calculation logic for feedforward compensation flow includes two components. The first component is damping-related compensation, which multiplies the difference between the reconstructed damping ratio and the reference damping ratio by a damping compensation coefficient. This coefficient is pre-calibrated based on the compensation chamber structure and oil type, representing the required oil flow rate corresponding to a unit change in damping ratio. The product of this difference and the coefficient reflects the compensation demand caused by abnormal oil viscosity. Subsequently, this product is multiplied by the absolute value of the external water pressure change rate. The external water pressure change rate is obtained by multiplying the depth change rate measured by the depth sensor by the water density and then by the gravitational acceleration. The absolute value is taken because both diving and surfacing will cause... Rapid pressure changes require compensation. Multiplying the damping-related compensation amount by the pressure change rate reflects the concept of dynamic compensation: the faster the pressure changes, the greater the compensation flow required. The second component is the directional bias flow rate, whose sign is set according to the thruster's direction of motion. During descent, the external pressure increases, requiring the injection of oil into the compensation chamber, resulting in a positive directional bias flow rate. During ascent, the external pressure decreases, requiring the extraction of oil from the compensation chamber, resulting in a negative directional bias flow rate. Adding the two components yields the feedforward compensation flow rate. The function of feedforward compensation is to intervene and adjust in advance based on the oil deterioration state and pressure change trend before the pressure difference significantly deviates from the target value. The pressure difference tracking error is calculated by subtracting the measured pressure difference value from the set target pressure difference value. The target pressure difference value is set to 0.02 MPa according to safety margin requirements, indicating that the internal and external pressure difference is allowed to remain within this range to protect the sealing structure. The measured pressure difference value is collected in real time by the pressure difference sensor in the compensation chamber. A positive difference indicates that the actual pressure difference is less than the target value, requiring an increase in the compensation pump flow rate; a negative difference indicates that the actual pressure difference is greater than the target value, requiring a decrease in the compensation pump flow rate.The fractional integral operation adopts the discretization method defined by Glenn Ward-Letnikov, which converts the continuous-time fractional integral into a weighted sum of discrete-time sequences. The control period is set to 0.1 seconds, meaning that the control output is calculated once every 0.1 seconds. The memory length is set to 50 historical windows, meaning that the historical error data of the most recent 5 seconds is retained. The order of the fractional integral is set to 0.6, which is between the 0th order of pure proportional control and the 1st order of traditional integral control. The 0.6 order integral has long memory characteristics and can better handle the hysteresis response of the oil-filled system. The weight coefficient sequence is calculated by fractional binomial coefficients. The weight coefficients decrease over time, indicating that the more distant the historical error, the smaller the impact on the current control output. The fractional integral term is obtained by multiplying the weight coefficient sequence and the historical error sequence term by term and summing them. This integral term comprehensively considers the cumulative effect and decay characteristics of historical errors. The proportional control term is calculated by multiplying the proportional gain by the differential pressure tracking error. The proportional gain is adjusted according to the open-loop gain of the compensation system and the desired response speed. Proportional control provides instantaneous response capability, and the larger the error, the larger the control output. The integral control term is calculated by multiplying the integral gain by the fractional integral term. The integral gain determines the contribution of historical error accumulation to the control output. Integral control eliminates steady-state error and prevents the system from having continuous deviations. The proportional control term, integral control term, and feedforward compensation flow are added together to generate a compensation pump flow command. The unit of this command is milliliters per minute, which directly corresponds to the flow output capacity of the compensation pump. The flow command is converted into a voltage signal by a digital-to-analog converter to drive the compensation pump motor. The synergistic effect of the three control outputs achieves comprehensive optimization of the compensation system's fast response, steady-state accuracy, and dynamic prediction.

[0059] In one specific embodiment, step S5 includes:

[0060] The motor speed is calculated based on the conversion relationship between the flow command of the compensating pump and the single-rotation displacement of the compensating pump. The compensating pump is driven to draw oil from the bottom of the compensating chamber and inject it into the upper part of the compensating chamber after being filtered by the filter screen. The pressure distribution in the compensating chamber is changed by circulating injection.

[0061] The pressure increment is calculated based on the oil bulk modulus, the injection volume within the control cycle, and the total volume of the compensation chamber. The pressure increment is added to the current differential pressure value, and the natural leakage of the diaphragm gap calculated based on the leakage coefficient is subtracted to obtain the predicted differential pressure value. The predicted differential pressure value is fed back to the fractional-order controller to recalculate the flow command for the next cycle.

[0062] Simultaneously extract the local peak values ​​of the pressure curve within each control cycle to construct a peak sequence, calculate the ratio of adjacent peak values ​​in the peak sequence, and determine that the system has entered steady state and the compensation pump stops operating when the ratio of 5 consecutive peak values ​​is greater than 0.95.

[0063] The arithmetic mean of the damping ratios for the 10 depth ranges in this cycle is calculated. The difference between this and the average damping ratio of the previous 5 cycles is divided by 5 to obtain the damping ratio evolution rate. The percentage of oil deterioration is calculated based on the damping ratio evolution rate and the baseline damping ratio. When the deterioration exceeds 30%, an oil replacement warning is generated. The complete data of this cycle is stored and the theoretical differential pressure value is updated by three-dimensional regression based on the data of the most recent 5 cycles.

[0064] Specifically, after receiving the flow command from the compensating pump, the controller calculates the motor speed based on the pump's single-revolution displacement. The single-revolution displacement of the compensating pump represents the volume of oil discharged in one revolution of the pump. To convert the flow command unit from milliliters per minute to milliliters per revolution, divide by 60 seconds and then by the rotational speed. Conversely, multiply the flow command by 1000 to convert it to microliters per minute, then divide by the single-revolution displacement and then by 60 to obtain the rotational speed per second. The motor controller adjusts the motor speed according to this speed command. After the compensating pump starts, it draws oil from the suction port at the bottom of the compensating chamber. The bottom position is chosen because contaminant particles and sediments will settle to the bottom under gravity. Prioritizing the extraction of oil from the bottom helps remove impurities. The extracted oil flows through a filter screen integrated into the pump body. The filter screen has a pore size of 5 micrometers, which can trap solid particles and fibrous impurities in the oil. The filtered clean oil is injected into the upper part of the compensating chamber from the pump outlet. The upper part is injected and the lower part is extracted, forming a closed-loop circulation circuit. The total oil volume remains unchanged, but the pressure distribution and oil quality in the compensating chamber are changed through circulation. The injection of oil causes a pressure change within the compensation chamber. Calculating the pressure increment requires three parameters: the oil bulk modulus, the injection volume during the control cycle, and the total volume of the compensation chamber. The oil bulk modulus, representing the oil's resistance to volumetric compression, is 1.65 GPa. The control cycle is 0.1 seconds. During this period, the volume injected by the compensation pump according to the flow command is equal to the flow command divided by 600. The total volume of the compensation chamber is 850 mL. The pressure increment is equal to the bulk modulus multiplied by the injection volume divided by the total volume. This calculation is based on the definition of bulk modulus in fluid mechanics: pressure change equals bulk modulus multiplied by relative volume change. The calculated pressure increment is added to the current pressure difference to obtain the theoretical pressure difference. However, in actual compensation chambers, there are diaphragm peripheries... The tiny gaps in the enclosure cause natural oil leakage. The leakage amount is calculated based on the leakage coefficient, which represents the leakage volume per unit time under a unit pressure difference. The leakage volume is obtained by multiplying the current pressure difference by the leakage coefficient and then by the control cycle. The pressure reduction corresponding to the leakage volume is calculated by back-calculating using the bulk modulus relationship. The predicted pressure difference is obtained by subtracting the pressure reduction caused by leakage from the theoretical pressure difference. The predicted pressure difference represents the expected pressure difference value for the next cycle under the current control output. The controller feeds back the predicted pressure difference value to the fractional-order controller as the new measured pressure difference value to recalculate the pressure difference tracking error. Based on the new error, the fractional-order integral term is updated and the flow command for the next cycle is regenerated, forming a closed-loop feedback control.During the operation of the compensation pump, the differential pressure sensor continuously collects the pressure in the compensation chamber to form a pressure curve. The controller scans the pressure curve in each control cycle to identify the local maximum point, which is the pressure peak. The local maximum is defined as the pressure value at that point being greater than the pressure value of the adjacent sampling points. The extracted pressure peaks are arranged in chronological order to form a peak sequence. The ratio of two adjacent peaks in the peak sequence is calculated. A ratio close to 1 indicates that the pressure oscillation amplitude decays slowly, while a ratio much less than 1 indicates that the pressure oscillation decays rapidly. When the ratio of five consecutive peaks is greater than 0.95, it indicates that the pressure oscillation amplitude has almost stopped decaying in the last five cycles. It is determined that the compensation system has entered a steady state and the differential pressure no longer changes significantly. The controller issues a stop command to shut down the compensation pump motor and switches back to the passive diaphragm compensation mode, but continues to operate while maintaining differential pressure monitoring. After the current operation cycle is completed, the controller performs statistical analysis on the reconstructed damping ratios of the 10 depth intervals collected in this cycle. The 10 damping ratio values ​​are summed and divided by 10 to obtain the arithmetic mean. The average damping ratio calculated similarly for the previous 5 cycles is extracted from the historical database. The difference between the two averages is the total change in damping ratio over the 5 cycles. This change is divided by 5 to obtain the damping ratio evolution rate. The evolution rate represents the average growth rate of the damping ratio in each operation cycle. A positive evolution rate indicates continuous deterioration of oil performance, while a negative evolution rate indicates measurement anomalies or algorithm divergence. The damping ratio evolution rate is divided by the baseline damping ratio and then multiplied by 100 to convert... The percentage form represents the percentage of oil deterioration. This percentage comprehensively reflects the overall deterioration of the oil due to factors such as temperature rise, aging, and contamination. When the deterioration exceeds 30%, the controller generates an oil replacement warning signal and sends it to the ground monitoring station via the communication interface to prompt maintenance personnel to arrange oil replacement. The differential pressure deviation dataset, viscosity correction term, reconstructed damping ratio, and compensation pump operation record of this cycle are packaged and stored in non-volatile memory. A three-dimensional relationship of the number of deep differential pressure cycles is established based on the data of the last 5 cycles. A new theoretical differential pressure curve is obtained through multivariate regression fitting. This curve replaces the initial calibration curve as the differential pressure benchmark for the next cycle.

[0065] In one specific embodiment, it further includes:

[0066] An integrated conductivity probe is used in the compensation chamber to simultaneously monitor the conductivity value of the oil during the working cycle. The conductivity value is then compared with the initial calibrated conductivity value to obtain the conductivity drift.

[0067] The percentage of oil contamination is calculated by multiplying the conductivity drift by the preset oil contamination correlation coefficient. When the percentage of oil contamination exceeds 15%, the filter self-cleaning process is triggered.

[0068] Record the cumulative filtration volume through the filter screen during the operation of the compensation pump, and calculate the filter screen clogging rate based on the ratio of the cumulative filtration volume to the rated dirt holding capacity of the filter screen.

[0069] When the filter screen clogging rate exceeds 70%, the reverse drive compensation pump performs a pulse backflushing operation to flush away particulate impurities attached to the filter screen surface and allow them to settle into the collection tank at the bottom of the compensation chamber.

[0070] Specifically, the conductivity probe integrated into the inner wall of the compensation chamber is used to monitor changes in the conductivity of the oil. The conductivity probe consists of two parallel electrode plates. A constant voltage is applied between the electrode plates to measure the current intensity passing through the oil. The conductivity is calculated according to Ohm's law. Pure silicon-based synthetic oil has extremely low conductivity because it is a non-polar molecule. When water, metal ions, or other polar contaminants are mixed into the oil, the conductivity increases significantly. The conductivity probe works continuously throughout the propeller operation, collecting conductivity values ​​every second to form time-series data. The controller reads the conductivity reference value measured in the initial calibration stage from the memory. This reference value is measured in a standard environment after the injection of fresh oil. The conductivity drift is obtained by subtracting the initial calibration conductivity value from the current measured conductivity value. A positive drift value indicates that the oil contamination level has increased, while a negative drift value indicates a measurement abnormality or probe malfunction requiring calibration. The calculation of the oil contamination percentage is based on the product of conductivity drift and a preset oil contamination correlation coefficient. The correlation coefficient is an empirical correlation established through conductivity tests and laboratory chemical analysis of a large number of contaminated oil samples. The unit of this coefficient is percentage per micro-Siemens per centimeter. Multiplying the conductivity drift by the correlation coefficient directly yields the oil contamination percentage. The contamination percentage quantifies the proportion of impurities in the oil to the total oil mass. When the contamination percentage exceeds 15%, it indicates that the impurities in the oil have reached a critical level that affects the normal operation of the compensation system. The controller triggers the filter self-cleaning process. This process includes increasing the flow rate of the compensation pump for a short period of time to increase the flushing intensity of the oil through the filter. The flushing process lasts for 30 seconds and then the flow rate returns to normal. The flushing action causes some loose particles attached to the filter surface to fall off, resuspend in the oil, and settle to the bottom of the compensation chamber under the action of gravity.Each time the compensating pump runs, the controller records the pump's running time and average flow rate. Multiplying the running time by the average flow rate and summing the results yields the cumulative filtration volume through the filter screen. This volume represents the total load of impurities trapped by the filter screen throughout its service life. The filter screen manufacturer provides a rated dirt-holding capacity parameter, indicating the maximum mass of impurities the filter screen can hold while maintaining normal filtration efficiency. Although the dirt-holding capacity is in mass, it is converted to the corresponding filtration volume based on the average concentration of impurities in the oil. Dividing the cumulative filtration volume by the volume corresponding to the filter screen's rated dirt-holding capacity yields the filter screen clogging rate. The clogging rate, expressed as a percentage, represents the degree to which the filter screen pores are occupied by impurities. A higher clogging rate results in greater resistance to oil flow through the filter screen, requiring the compensating pump to have greater drive power to maintain the same flow rate. When the clogging rate exceeds 70%, the filter screen's flow capacity is severely reduced, and continued operation will lead to pump overload or filter screen damage. The controller then initiates a pulse-type backflushing cleaning program, which changes the rotation direction of the compensating pump, switching it from bottom-drawing and top-injection to top-drawing. Bottom-injected, reverse-flowing oil impacts the filter screen from the back, flushing away particulate impurities adhering to its surface. Pulse-type backflushing refers to the compensating pump running at high flow for short periods, then stopping and repeating this cycle. The pulse frequency is set to once per second, with each pulse lasting 0.3 seconds. The periodic impact force generated by the pulsed flow is more effective at stripping particles from the filter screen than continuous flow. The backflushing process lasts for 1 minute before resuming forward flow. The flushed-away particulate impurities remain suspended in the oil and settle to the collection tank at the bottom of the compensating chamber under gravity. The collection tank is a recessed structure designed at the bottom of the compensating chamber, with a depth lower than the compensating pump's inlet. Impurities settled in the collection tank are not re-extracted by the compensating pump. The collection tank is cleaned via a drain valve during regular maintenance of the propeller. After backflushing and cleaning, the controller resets the cumulative filter volume counter to zero and restarts calculating the clogging rate. Through the synergistic effect of conductivity monitoring and filter self-cleaning, the service life of the compensating oil is effectively extended, maintaining the long-term stable operation of the compensating system.

[0071] The adaptive pressure compensation method for underwater thrusters in the embodiments of this application has been described above. The adaptive pressure compensation system for underwater thrusters in the embodiments of this application is described below. Please refer to [link / reference]. Figure 2 One embodiment of the adaptive pressure compensation system for underwater thrusters in this application includes:

[0072] The calculation module is used to apply segmented pressure loading to the underwater thruster compensation chamber, record the diaphragm displacement and internal and external pressure difference at each depth point, and calculate the reference damping ratio through the pressure oscillation decay curve.

[0073] The acquisition module is used to acquire the measured pressure difference value of the compensation chamber during the operation cycle, and to obtain the pressure difference deviation dataset by subtracting the measured pressure difference value from the theoretical pressure difference value corresponding to the current depth and averaging it over the depth interval.

[0074] The correction module is used to extract the same depth deviation value between the current cycle and the previous cycle in the differential pressure deviation dataset, calculate the viscosity correction term according to the iterative learning law, and calculate the reconstruction damping ratio based on the viscosity correction term.

[0075] The generation module is used to determine the degree of oil deterioration by subtracting the reconstructed damping ratio from the reference damping ratio, calculate the feedforward compensation flow rate based on the reconstructed damping ratio and the external water pressure change rate, and generate a compensation pump flow command by combining the fractional integral term of the differential pressure tracking error.

[0076] The injection module is used to drive the compensation pump to inject oil according to the compensation pump flow command so that the measured differential pressure value converges to the target value, and to store the differential pressure deviation dataset of the current cycle and update the theoretical differential pressure value for the next cycle.

[0077] above Figure 2 The adaptive pressure compensation system for underwater thrusters in this embodiment of the invention is described in detail from the perspective of modular functional entities. The adaptive pressure compensation device for underwater thrusters in this embodiment of the invention is described in detail from the perspective of hardware processing.

[0078] Reference Figure 3 This invention also provides an adaptive pressure compensation device for underwater thrusters. This device can be a server, and its internal structure can be as follows: Figure 3 As shown, the underwater thruster adaptive pressure compensation device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor, designed as a computer, provides computational and control capabilities. The memory of the underwater thruster adaptive pressure compensation device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the underwater thruster adaptive pressure compensation device stores the data corresponding to this embodiment. The network interface of the underwater thruster adaptive pressure compensation device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0079] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the adaptive pressure compensation device for underwater thrusters to which the present invention is applied.

[0080] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the underwater thruster adaptive pressure compensation method.

[0081] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0082] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an underwater thruster adaptive pressure compensation device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0083] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An adaptive pressure compensation method for underwater thrusters, characterized in that, The method includes: Step S1: Apply segmented pressure loading to the underwater thruster compensation chamber, record the diaphragm displacement and internal and external pressure difference at each depth point, and calculate the reference damping ratio through the pressure oscillation decay curve; Step S2: Collect the measured pressure difference value of the compensation cavity in the operation cycle, calculate the difference between the measured pressure difference value and the theoretical pressure difference value corresponding to the current depth, and obtain the pressure difference deviation dataset by averaging over the depth interval; Step S3: Extract the same depth deviation value between the current cycle and the previous cycle from the differential pressure deviation dataset, calculate the viscosity correction term according to the iterative learning law, and calculate the reconstruction damping ratio based on the viscosity correction term; Step S4: The degree of oil deterioration is determined by the difference between the reconstructed damping ratio and the reference damping ratio. The feedforward compensation flow rate is calculated based on the reconstructed damping ratio and the external water pressure change rate. The compensation pump flow rate command is generated by combining the fractional integral term of the differential pressure tracking error. Step S5: Drive the compensation pump to inject oil according to the compensation pump flow command so that the measured pressure difference value converges to the target value. Store the pressure difference deviation dataset of this cycle and update the theoretical pressure difference value for the next cycle.

2. The adaptive pressure compensation method for underwater thrusters according to claim 1, characterized in that, Step S1 includes: A silicon-based synthetic oil of rated volume was injected into the compensation chamber at a standard temperature of 20°C and a standard atmospheric pressure to obtain the initial dynamic viscosity value. Pressure values ​​corresponding to water depths from 0 to 500 meters are applied to the compensation chamber in a controllable pressure chamber. A pressure acquisition point is set at every 50-meter depth, and the internal oil pressure value and external simulated water pressure value at each acquisition point are recorded to calculate the internal and external pressure difference value at each depth point. The displacement of the diaphragm center point corresponding to each pressure acquisition point is recorded by a displacement sensor installed on the back side of the diaphragm. The initial elastic modulus is calculated by least squares fitting of the internal and external pressure difference value and the displacement of the diaphragm center point. The pressure oscillation curve of the compensation chamber is recorded by increasing the pressure from 0 meters to 100 meters at a simulated descent speed of 1 meter per second. The first and second pressure peaks are extracted, and the ratio of the two pressure peaks is divided by the natural logarithm of the characteristic equation containing the logarithmic term to calculate the reference damping ratio.

3. The adaptive pressure compensation method for underwater thrusters according to claim 2, characterized in that, Step S2 includes: The data acquisition module of the adaptive iterative learning observer is activated, and during the process of the underwater thruster diving from the water surface to the target operating depth, the measured value of the pressure difference between the inside and outside of the compensation cavity is continuously recorded at a sampling frequency of 10Hz. The theoretical pressure difference value corresponding to the current depth is obtained by linear interpolation of the internal and external pressure difference values ​​at each depth point based on the current depth value. The difference between the measured differential pressure value and the theoretical differential pressure value is used to obtain the differential pressure deviation signal at each sampling time. The water depth from 0 to 500 meters is divided into 10 depth intervals. The pressure difference deviation signal in each depth interval is processed by time averaging to obtain the representative deviation value of each depth interval, and the data are combined to form the pressure difference deviation dataset.

4. The adaptive pressure compensation method for underwater thrusters according to claim 3, characterized in that, Step S3 includes: Extract representative deviation values ​​for each depth interval between the current k-th operation cycle and the previous (k-1)-th operation cycle from the differential pressure deviation dataset, and calculate the difference in deviation values ​​for the same depth interval. The learning gain is calculated based on the sensitivity coefficient of the oil viscosity to the pressure difference response. The learning gain is multiplied by the difference in the deviation value, and the viscosity correction term from the previous cycle is added. The viscosity correction term for the current cycle is obtained according to the iterative learning law. The initial dynamic viscosity value is added to the viscosity correction term to obtain the equivalent viscosity value for the current cycle; Based on the equivalent viscosity value, the initial elastic modulus, and the oil density parameter, the reconstructed damping ratio is obtained by substituting them into the formula for calculating the damping ratio of a viscoelastic system.

5. The adaptive pressure compensation method for underwater thrusters according to claim 1, characterized in that, Step S4 includes: The difference between the reconstructed damping ratio and the reference damping ratio is used to obtain the damping ratio deviation value. When the damping ratio deviation value is greater than 0.05, it is determined that the degree of oil deterioration exceeds the threshold and active compensation is initiated. The difference between the reconstructed damping ratio and the reference damping ratio is multiplied by the damping compensation coefficient, and then multiplied by the absolute value of the external water pressure change rate. The directional offset flow rate is then added to obtain the feedforward compensation flow rate. The difference between the set target differential pressure value and the measured differential pressure value is used to obtain the differential pressure tracking error. The differential pressure tracking error is then subjected to fractional integration to obtain the fractional integral term. The proportional gain is multiplied by the differential pressure tracking error to obtain the proportional control term, the integral gain is multiplied by the fractional integral term to obtain the integral control term, and the proportional control term, the integral control term, and the feedforward compensation flow rate are added to generate the compensation pump flow rate command.

6. The adaptive pressure compensation method for underwater thrusters according to claim 1, characterized in that, Step S5 includes: The motor speed is calculated based on the conversion relationship between the flow command of the compensation pump and the single-rotation displacement of the compensation pump. The compensation pump is driven to draw oil from the bottom of the compensation chamber and inject it into the upper part of the compensation chamber after being filtered by the filter screen. The pressure distribution in the compensation chamber is changed by circulating injection. The pressure increment is calculated based on the oil bulk modulus, the injection volume within the control cycle, and the total volume of the compensation chamber. The pressure increment is added to the current differential pressure value, and the natural leakage of the diaphragm gap calculated based on the leakage coefficient is subtracted to obtain the predicted differential pressure value. The predicted differential pressure value is fed back to the fractional-order controller to recalculate the flow command for the next cycle. Simultaneously extract local peak values ​​of the pressure curve within each control cycle to construct a peak sequence, calculate the ratio of adjacent peak values ​​in the peak sequence, and determine that the system has entered a steady state and the compensation pump stops operating when the ratio of 5 consecutive peak values ​​is greater than 0.

95. Calculate the arithmetic mean of the damping ratios for the 10 depth intervals in this cycle, and divide the difference between this and the average damping ratio of the previous 5 cycles by 5 to obtain the damping ratio evolution rate. Calculate the percentage of oil deterioration based on the damping ratio evolution rate and the benchmark damping ratio. When the deterioration exceeds 30%, generate an oil replacement warning. Store the complete data of this cycle and update the theoretical pressure difference value using three-dimensional regression based on the data from the most recent 5 cycles.

7. The adaptive pressure compensation method for underwater thrusters according to claim 1, characterized in that, Also includes: An electrical conductivity probe is integrated within the compensation chamber to synchronously monitor the oil conductivity value during the operating cycle. The difference between the conductivity value and the initial calibrated conductivity value is used to obtain the conductivity drift. The percentage of oil contamination is calculated by multiplying the conductivity drift by a preset oil contamination correlation coefficient. When the percentage of oil contamination exceeds 15%, the filter self-cleaning process is triggered. Record the cumulative filtration volume through the filter screen during the operation of the compensation pump, and calculate the filter screen clogging rate based on the ratio of the cumulative filtration volume to the rated dirt holding capacity of the filter screen. When the filter screen clogging rate exceeds 70%, the compensation pump is reversed to perform a pulse backflushing operation, which flushes away particulate impurities attached to the filter screen surface and settles them into the collection tank at the bottom of the compensation chamber.

8. An adaptive pressure compensation system for an underwater thruster, characterized in that, For implementing the adaptive pressure compensation method for an underwater thruster as described in any one of claims 1-7, the adaptive pressure compensation system for the underwater thruster comprises: The calculation module is used to apply segmented pressure loading to the underwater thruster compensation chamber, record the diaphragm displacement and internal and external pressure difference at each depth point, and calculate the reference damping ratio through the pressure oscillation decay curve. The acquisition module is used to acquire the measured pressure difference value of the compensation chamber during the operation cycle, and to obtain the pressure difference deviation dataset by subtracting the measured pressure difference value from the theoretical pressure difference value corresponding to the current depth and averaging it over the depth interval. The correction module is used to extract the same depth deviation value between the current cycle and the previous cycle in the differential pressure deviation dataset, calculate the viscosity correction term according to the iterative learning law, and calculate the reconstruction damping ratio based on the viscosity correction term. The generation module is used to determine the degree of oil deterioration by subtracting the reconstructed damping ratio from the reference damping ratio, calculate the feedforward compensation flow rate based on the reconstructed damping ratio and the external water pressure change rate, and generate a compensation pump flow command by combining the fractional integral term of the differential pressure tracking error. The injection module is used to drive the compensation pump to inject oil according to the compensation pump flow command so that the measured pressure difference value converges to the target value, and to store the pressure difference deviation dataset of the current cycle and update the theoretical pressure difference value for the next cycle.

9. An adaptive pressure compensation device for an underwater thruster, characterized in that, It includes a memory and a processor, the memory storing a computer program that can run on the processor, and the processor executing the computer program to implement the underwater thruster adaptive pressure compensation method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the underwater thruster adaptive pressure compensation method as described in any one of claims 1 to 7.

Images