A marine ship hydraulic oil cylinder state monitoring and sealing performance early warning system

Abstract

The application relates to the technical field of ship hydraulic monitoring, and discloses a marine ship hydraulic cylinder state monitoring and sealing performance early warning system. A multi-source sensing data acquisition module of the system synchronously acquires multi-dimensional time sequence operation data of the oil cylinder through vibration, pressure and temperature sensors; a data preprocessing fusion module performs noise suppression, timestamp alignment and multi-modal fusion on the data to generate a standardized oil cylinder state feature sequence; a sealing feature deep extraction module analyzes sealing cavity pressure fluctuation, piston rod displacement oscillation and sealing element deformation trend features from the feature sequence; a multi-scale state evaluation module calculates a sealing performance degradation gradient and a dynamic operation stability coefficient according to the features; and an early warning decision generation module generates a sealing failure risk level and an early warning control instruction based on the calculation results. The system guarantees stable operation of the oil cylinder and reduces maintenance cost and safety risks.

Description

Technical Field

[0001] This invention relates to the field of marine hydraulic monitoring technology, specifically to a marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system. Background Technology

[0002] In marine vessel operations, hydraulic cylinders, as key actuators, are widely used in the drive systems of important equipment such as steering gear, anchor winches, and cargo hoists. Their operational status directly affects the vessel's navigation safety and operational efficiency. The marine environment is characterized by high salt spray, high humidity, strong vibration, and frequent load fluctuations. These harsh conditions significantly impact the sealing performance of hydraulic cylinders. As core components preventing hydraulic oil leakage and the intrusion of external impurities, worn, aged, or deformed seals not only lead to pressure loss and increased energy consumption in the hydraulic system but may also cause corrosion and scratches on internal cylinder components. In severe cases, this can even cause equipment shutdown or safety accidents.

[0003] For condition monitoring of hydraulic cylinders in marine vessels, the industry commonly employs either periodic disassembly and inspection or single-parameter monitoring. Periodic disassembly and inspection requires equipment shutdown, which not only consumes significant operational time and increases maintenance costs, but also risks causing additional damage to cylinder components due to improper handling during disassembly and assembly. Furthermore, this offline monitoring method cannot capture dynamic anomalies during cylinder operation in real time, making it difficult to provide early warnings of sealing performance degradation. Single-parameter monitoring typically focuses only on a single indicator such as hydraulic oil pressure or temperature, ignoring the correlation between other key parameters such as vibration and displacement and sealing performance. For example, monitoring pressure changes within the cylinder cavity solely through pressure sensors cannot accurately distinguish whether pressure fluctuations are caused by seal leakage, load changes, or system pressure regulation, easily leading to misjudgments or missed diagnoses.

[0004] Existing monitoring technologies also have significant limitations in data processing. Factors such as mechanical vibration and electromagnetic interference in the marine vessel operating environment lead to a large amount of noise in the monitoring data collected by sensors. If this noise is not effectively suppressed, it will seriously affect the accuracy and reliability of the data. Furthermore, the sampling frequencies and data formats of different types of sensors vary, making it difficult to effectively fuse multi-source data and form a feature sequence that comprehensively reflects the cylinder's condition. This prevents the monitoring system from extracting key features directly related to sealing performance from multi-dimensional data, such as subtle patterns in sealing cavity pressure fluctuations and changes in the oscillation amplitude of piston rod displacement. Consequently, it cannot accurately assess the degree and trend of sealing performance degradation. When early, slight wear occurs in the seals, existing systems often fail to detect it in time, only discovering it after the sealing performance has severely deteriorated and obvious leakage has occurred, at which point it has already caused economic losses and safety hazards.

[0005] As the marine shipping industry develops towards larger scale and greater intelligence, higher demands are being placed on the reliability and maintenance efficiency of ship equipment. Traditional monitoring methods are no longer sufficient to meet the needs for real-time, accurate, and comprehensive condition monitoring and early warning. Developing a hydraulic cylinder condition monitoring system capable of simultaneous acquisition of multi-source parameters, effective data fusion, precise extraction of sealing features, and early warning of performance degradation has become an urgent problem to be solved in the field of marine ship equipment maintenance. Summary of the Invention

[0006] The purpose of this invention is to provide a marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides a marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system, the system comprising:

[0008] The multi-source sensor data acquisition module is used to simultaneously acquire multi-dimensional time-series operating data of the hydraulic cylinder through vibration sensors, pressure sensors, and temperature sensors.

[0009] The data preprocessing and fusion module is used to perform noise suppression, timestamp alignment and multimodal data fusion on the multidimensional time-series operation data to generate a standardized hydraulic cylinder state feature sequence.

[0010] The sealing feature depth extraction module is used to analyze the sealing cavity pressure fluctuation characteristics, piston rod displacement oscillation characteristics, and seal deformation trend characteristics of the hydraulic cylinder from the standardized cylinder state feature sequence.

[0011] The multi-scale state assessment module is used to calculate the sealing performance degradation gradient and dynamic operation stability coefficient of the hydraulic cylinder based on the pressure fluctuation characteristics of the sealing cavity, the displacement oscillation characteristics of the piston rod, and the deformation trend characteristics of the sealing component.

[0012] The early warning decision generation module is used to generate a seal failure risk level and early warning control instructions based on the seal performance degradation gradient and dynamic operation stability coefficient.

[0013] Preferably, the system further includes:

[0014] The system security management module is used to perform hierarchical verification of user access permissions to the multi-source sensor data acquisition module;

[0015] The real-time health dashboard module is used to visually display the real-time change curves of the sealing performance degradation gradient and dynamic operational stability coefficient.

[0016] Preferably, the multi-source sensor data acquisition module includes:

[0017] The heterogeneous protocol adapter submodule is used to parse different communication protocol data packets of vibration sensors, pressure sensors, and temperature sensors.

[0018] The spatiotemporal calibration engine is used to unify the timestamps of vibration data, pressure data, and temperature data to the ship's master clock reference and compensate for differences in the spatial location of sensors.

[0019] A multi-dimensional data synchronization buffer queue is used to temporarily store aligned multi-sensor data streams and output them to the data preprocessing and fusion module according to a fixed time window.

[0020] Preferably, the sealing feature depth extraction module includes:

[0021] The sealed cavity pressure feature analysis unit is used to extract pressure rise slope features, steady-state pressure pulsation amplitude features, and pressure relief rate features from the standardized cylinder state feature sequence.

[0022] The piston rod displacement feature analysis unit is used to calculate the axial displacement fluctuation frequency feature and radial offset feature of the piston rod from the standardized cylinder state feature sequence.

[0023] The sealing deformation feature modeling unit is used to construct a dynamic stress distribution model of the sealing component based on the pressure rise slope feature, steady-state pressure pulsation amplitude feature, and piston rod axial displacement fluctuation frequency feature.

[0024] Preferably, the multi-scale state assessment module includes:

[0025] The sealing degradation analysis unit is used to calculate the cumulative fatigue of the sealing material based on the pressure relief rate characteristics, piston rod radial offset characteristics, and dynamic stress distribution model of the sealing component.

[0026] The operational stability assessment unit is used to generate the hydraulic cylinder vibration energy entropy value based on the pressure rise slope characteristics, steady-state pressure pulsation amplitude characteristics, and piston rod axial displacement fluctuation frequency characteristics.

[0027] Preferably, the seal degradation analysis unit includes:

[0028] The sealing characteristic degradation gradient calculation subunit is used to calculate the sealing performance degradation rate index based on the time series changes of the cumulative fatigue amount and pressure relief rate characteristics of the sealing material.

[0029] The dynamic stability coefficient generation subunit is used to perform cross-domain coupling analysis on the hydraulic cylinder vibration energy entropy value and the piston rod radial offset characteristics to output the dynamic operation stability coefficient.

[0030] Preferably, the early warning decision generation module includes:

[0031] The risk level mapping unit is used to match a predefined risk level threshold for seal failure based on the combination relationship between the seal performance degradation gradient and the dynamic operating stability coefficient.

[0032] The instruction generation engine is used to generate warning signal triggering logic and maintenance strategy code based on the seal failure risk level threshold.

[0033] Preferably, the system further includes:

[0034] The ship data compression and storage module uses a time-slice compression algorithm to segment and compress the multi-dimensional time-series operational data, and adds ship latitude and longitude location tags.

[0035] The historical data maintenance unit is used to automatically clean up expired data slices according to the preset ship navigation cycle.

[0036] Preferably, the historical data maintenance unit includes:

[0037] The data lifecycle management subunit is used to mark the effective period of the multidimensional time-series operational data based on the ship's navigation log.

[0038] The storage space optimization subunit is used to automatically reduce the storage resolution of data for the corresponding time period when the sealing performance degradation gradient is lower than a set threshold.

[0039] Preferably, the real-time health dashboard module includes:

[0040] An alarm linkage unit is used to trigger the ship's engine room audible and visual alarm device when the risk level of the seal failure exceeds a preset warning value.

[0041] The early warning log generation unit is used to record the timestamp of the early warning control command and the corresponding snapshot of the dynamic operation stability coefficient.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] This marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system utilizes a multi-source sensor data acquisition module, employing vibration, pressure, and temperature sensors to simultaneously acquire multi-dimensional time-series operational data of the hydraulic cylinders. This overcomes the limitations of traditional monitoring methods that rely on single-parameter acquisition. In the complex operating environment of marine vessels, changes in the sealing performance of hydraulic cylinders are often reflected simultaneously in changes in parameters such as vibration, pressure, and temperature. Relying on a single parameter alone cannot comprehensively capture subtle changes in the sealing state. However, the collaborative application of multi-source sensors can collect key data reflecting the cylinder's operating status from different angles, covering various information such as abnormal vibrations caused by seal wear, pressure fluctuations caused by leakage in the sealing cavity, and temperature changes accompanying seal failure. This lays the foundation for accurate subsequent analysis of sealing performance.

[0044] The noise suppression, timestamp alignment, and multimodal data fusion operations performed by the data preprocessing and fusion module effectively solve the problems of severe interference in monitoring data in the marine environment and the difficulty in coordinating the analysis of multi-source data. During the operation of marine vessels, vibrations generated by mechanical operation and interference from electromagnetic equipment can lead to a large amount of noise in the data collected by sensors. Without noise suppression, effective information related to sealing performance in the data will be obscured, affecting the accuracy of subsequent analysis results. Timestamp alignment ensures that operational data collected by different sensors at the same time can be accurately correlated, avoiding data association errors caused by differences in sampling time. Multimodal data fusion can integrate different types of data such as vibration, pressure, and temperature into a standardized cylinder state feature sequence, eliminating differences caused by different data formats and sampling frequencies. This allows the integrated feature sequence to comprehensively and coherently reflect the state changes of the cylinder at different operating stages, providing high-quality data support for the extraction of sealing features.

[0045] The sealing feature depth extraction module can accurately analyze the sealing cavity pressure fluctuation characteristics, piston rod displacement oscillation characteristics, and seal deformation trend characteristics from standardized cylinder state feature sequences, breaking through the bottleneck of traditional monitoring technologies that cannot extract key sealing features from multi-dimensional data. Sealing performance degradation often manifests as subtle changes in the amplitude of sealing cavity pressure fluctuations, abnormal piston rod displacement oscillation frequencies, and a gradual increase in the degree of seal deformation. These features are difficult to detect through intuitive observation or simple data processing. This module, through a professional feature analysis algorithm, can identify these subtle features directly related to sealing performance from complex state feature sequences. For example, it can capture the subtle fluctuation patterns of sealing cavity pressure during early slight wear of the seal, and the differences in piston rod displacement oscillation caused by changes in seal friction, thus providing specific and reliable analytical objects for subsequent assessment of sealing performance degradation.

[0046] The multi-scale condition assessment module calculates the sealing performance degradation gradient and dynamic operational stability coefficient based on the analyzed sealing characteristics, achieving a multi-dimensional and precise assessment of the hydraulic cylinder's sealing condition. Sealing performance degradation is a gradual process, with varying rates and manifestations at different stages; a single indicator cannot comprehensively assess the sealing condition. The sealing performance degradation gradient reflects the rate of change in sealing performance over time, helping to determine the progress of seal wear or aging. The dynamic operational stability coefficient reflects the cylinder's stability under different loads and operating conditions, indirectly reflecting the impact of sealing performance on the overall cylinder's operating state. Through the synergistic analysis of these two indicators, the current status and trends of sealing performance can be comprehensively understood from both time and operational state dimensions, avoiding the one-sidedness of traditional assessment methods due to the reliance on a single indicator.

[0047] The early warning decision generation module generates seal failure risk levels and early warning control commands based on the seal performance degradation gradient and dynamic operational stability coefficient, enabling early warning and precise control of seal failure. When early, slight degradation occurs in the seal, the module can promptly determine the risk level of seal failure based on calculated indicators and generate corresponding early warning control commands. Workers can use the early warning information to develop maintenance plans in advance, performing seal replacement or maintenance without stopping the equipment or with only a short downtime. This avoids the passive situation of waiting for seal failure to occur before repairs are carried out, as is common in traditional methods, reducing operational losses due to equipment downtime and lowering the probability of safety accidents caused by seal failure. Furthermore, the generated early warning control commands provide clear guidance for maintenance work, helping workers accurately grasp the timing and content of maintenance, improving maintenance efficiency, reducing maintenance costs, and ensuring the stable and safe operation of marine vessel hydraulic cylinders and even the entire vessel's equipment. Attached Figure Description

[0048] Figure 1 This is a timing diagram of the marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system described in this invention;

[0049] Figure 2 An architecture diagram of the internal processes of the real-time health dashboard module;

[0050] Figure 3 This is an architecture diagram of the internal process of the multi-source sensor data acquisition module;

[0051] Figure 4 This is an architecture diagram of the internal process of the sealing degradation analysis unit. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see Figure 1 This invention provides a marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system. The system includes: a multi-source sensor data acquisition module, a data preprocessing and fusion module, a sealing feature depth extraction module, a multi-scale condition assessment module, and an early warning decision generation module. The specific implementation steps are as follows:

[0054] The multi-source sensor data acquisition module synchronously acquires multi-dimensional time-series operational data of the hydraulic cylinder through vibration, pressure, and temperature sensors. The data preprocessing and fusion module performs noise suppression, timestamp alignment, and multimodal data fusion on the multi-dimensional time-series operational data to generate a standardized cylinder state feature sequence. The sealing feature depth extraction module analyzes the hydraulic cylinder's sealing cavity pressure fluctuation characteristics, piston rod displacement oscillation characteristics, and seal deformation trend characteristics from the standardized cylinder state feature sequence. The multi-scale state assessment module calculates the hydraulic cylinder's sealing performance degradation gradient and dynamic operational stability coefficient based on the sealing cavity pressure fluctuation characteristics, piston rod displacement oscillation characteristics, and seal deformation trend characteristics. The early warning decision generation module generates a sealing failure risk level and early warning control commands based on the sealing performance degradation gradient and dynamic operational stability coefficient. The entire system is deployed on a marine embedded platform and uses a real-time operating system to process the data stream.

[0055] Example 1: See Figure 2 The system security control module handles hierarchical verification of user access permissions to the multi-source sensor data acquisition module. This module is integrated within the ship's central control system and establishes a communication connection with the ship's user management database. The user management database stores predefined hierarchical role information, including the captain, marine engineer, and maintenance technician roles, each associated with a specific level of operational permissions. When a user initiates an access request through the ship's operating terminal, the system security control module initiates the identity authentication process. The identity authentication process includes a two-factor authentication mechanism: first, verifying the user's unique identification code; second, collecting the user's biometric data. The identification code verification process calls the ship's encrypted service interface for hash value comparison, while biometric verification uses the ship's fingerprint scanner to obtain biological data. The system security control module matches the acquired verification data with records in the user management database using a similarity threshold algorithm. If the similarity between the verification data and the database record is lower than a preset threshold, the system generates an access denial command and activates the audit trail mechanism. The audit trail mechanism records the abnormal access timestamp, terminal device number, and verification failure reason code, writing it to the ship's security log file. The access denial command triggers the ship's operating terminal to display a permission error message.

[0056] The real-time health dashboard module visualizes the continuous changing trends of sealing performance degradation gradient and dynamic operational stability coefficient on the ship's monitoring screen. This module employs a dual-channel data rendering engine: channel one processes the sealing performance degradation gradient data stream, and channel two processes the dynamic operational stability coefficient data stream. The data rendering engine connects to the ship's data bus, acquiring the latest indicator values ​​from the multi-scale condition assessment module every second. The visualization interface includes a main display area and an auxiliary information area. The main display area plots two curves that change over time: curve one represents the sealing performance degradation gradient, and curve two represents the dynamic operational stability coefficient. The curves are plotted using anti-aliasing rendering technology, and the time axis scale is labeled according to the ship's standard time format. The auxiliary information area dynamically displays the indicator values ​​at the current moment, with the reading update frequency synchronized with the data rendering engine. The curve display area supports time range scaling, allowing the ship operator to adjust the displayed time span via the control panel.

[0057] The alarm linkage unit continuously monitors the seal failure risk level data output by the real-time health dashboard module. The monitoring process employs a polling mechanism, reading the risk level value every 200 milliseconds. When the read value exceeds the warning threshold defined in the ship's configuration file, the alarm linkage unit generates a digital control signal. This signal is transmitted to the engine room alarm controller via the ship's internal control network. The alarm controller connects to three sets of audible and visual alarm devices: the first set is installed in the hydraulic station control cabinet, including a red rotating warning light and a 105-decibel buzzer; the second set is located in the engine room, using a dual-color LED alarm panel; and the third set is located on the bridge, equipped with a voice alarm. Upon triggering the control signal, all three sets of devices are activated synchronously. The red rotating warning light enters flashing mode at a frequency of 120 flashes per minute; the buzzer outputs a continuous alarm tone; the dual-color LED alarm panel switches to a solid red state; and the voice alarm plays a pre-recorded hydraulic system warning message. All alarm devices remain active until the ship's operator performs a confirmation operation.

[0058] The early warning log generation unit automatically initiates the log recording process when the alarm linkage unit is activated. The recording process first calls the ship's BeiDou time synchronization module to obtain timestamp data accurate to milliseconds, and then extracts a complete data snapshot of the current dynamic operational stability coefficient from the multi-scale state assessment module. The data snapshot includes the instantaneous value of the coefficient and the fluctuation range value within the previous five minutes. Log recording uses a structured data format, including a timestamp field, a risk level code field, a stability coefficient snapshot field, and an alarm status field. The timestamp field records year, month, day, hour, minute, second, and millisecond information; the risk level code field stores a three-digit code; the stability coefficient snapshot field contains floating-point instantaneous values ​​and fluctuation range data; and the alarm status field records the alarm device trigger status code. The log generation unit writes the structured data to the early warning log partition of the ship's solid-state storage, with each record occupying a fixed-length storage block. The log partition uses a circular storage mechanism, automatically overwriting the oldest record when the storage space reaches its capacity limit.

[0059] The system safety management module and the real-time health dashboard module interact via the ship's high-speed data network. Network communication uses a ship-specific transmission protocol, with data packets containing an encrypted header and a checksum. The system safety management module sends user permission status information to the real-time health dashboard module, which adjusts its interface display based on this information. When a user's permission level is lower than the engineer level, the health dashboard module automatically hides the historical trend curve of the dynamic operational stability coefficient. The alarm linkage unit and the early warning log generation unit share an event trigger, which generates a trigger pulse when the seal failure risk level exceeds a threshold. The trigger pulse simultaneously activates the alarm control circuit and the log recording circuit, achieving millisecond-level synchronization between alarm actions and log recording. The ship's electrical system provides dual redundant power supplies to all modules, automatically switching to emergency power in case of main power failure. Communication lines between modules are configured with physically isolated interfaces to prevent electromagnetic interference from affecting signal transmission.

[0060] Example 2: See Figure 3The heterogeneous protocol adaptation submodule is deployed inside the ship's sensor gateway device, and its processor runs multiple communication protocol parsing threads. These parsing threads include RS-485 protocol parsing, CAN bus protocol parsing, and Ethernet protocol parsing. When a vibration sensor sends data via the RS-485 interface, the RS-485 protocol parsing thread receives the binary data stream, identifies the start and stop bits, and extracts the payload after stripping the check byte. When a pressure sensor transmits via the CAN bus, the CAN bus protocol parsing thread captures standard format data frames, filters the arbitration identifier to match the preset sensor ID, and converts the data bytes into engineering unit values. Data packets transmitted by a temperature sensor via Ethernet are processed by the Ethernet protocol parsing thread, which parses the TCP / IP protocol stack, verifies the integrity of the data packets, and reads the valid temperature value. All parsing results are converted into a unified internal data structure containing a sensor ID field, a timestamp field, and a numerical field, which is then passed to the spatiotemporal calibration engine via a shared memory area.

[0061] The spatiotemporal calibration engine runs on the ship's main control computer, and its core functions include time reference synchronization and spatial position compensation. The time reference synchronization module connects to the ship's GPS timing system, acquiring Coordinated Universal Time (UTC) signals as the master clock source. After receiving the data structure from the heterogeneous protocol adaptation submodule, the engine first compares the difference between the internal timestamp of the data packet and the master clock. Data packets with a time difference exceeding 100 milliseconds trigger a time correction procedure, which uses a linear interpolation algorithm to recalculate the time series of the data points. The spatial position compensation function calls a preset sensor position parameter table, which records the three-dimensional coordinates of each sensor in the ship's coordinate system. The compensation algorithm calculates the spatial vectors of the hydraulic cylinder body and each sensor, generating a position compensation factor matrix. Vibration sensor data is multiplied by an axial compensation factor, pressure sensor data by a radial compensation factor, and temperature sensor data by a spherical compensation factor. The compensated data stream is appended with a calibration mark and written to a multidimensional data synchronization buffer queue.

[0062] The multidimensional data synchronization buffer queue is implemented as a dual-buffer architecture with a fixed 100-millisecond time window. The front buffer receives calibration data output from the spatiotemporal calibration engine, while the rear buffer supplies data to the data preprocessing and fusion module. Buffer management employs a ping-pong switching mechanism: when the 100-millisecond time window is reached, the front buffer automatically locks and switches to the rear buffer, while the newly activated empty buffer begins receiving data. The queue output interface pushes data packet groups in the order of the time window, with each data packet group containing all calibrated sensor data within that time window. The queue overflow protection mechanism initiates data sampling when the buffer reaches 90% capacity, with the sampling rate dynamically adjusted according to the load.

[0063] The sealing cavity pressure feature analysis unit extracts three types of feature parameters from the standardized feature sequence output by the data preprocessing and fusion module. The pressure rise slope feature is achieved by analyzing the initial segment of the pressure-time curve; the algorithm selects the interval from the static baseline to the steady-state value and calculates the pressure change per unit time. The steady-state pressure pulsation amplitude feature uses a peak detection method to identify the maximum deviation of pressure fluctuations during the steady-state pressure duration. The pressure relief rate feature analysis focuses on the pressure decrease phase, recording the time difference between the working pressure and the ambient pressure, and calculating the pressure decay per unit time. These feature parameters are updated in real-time to a shared feature register for use by the sealing component deformation feature modeling unit.

[0064] The piston rod displacement feature analysis unit processes displacement data from a standardized feature sequence. Axial displacement fluctuation frequency characteristics are calculated using real-time spectrum analysis, windowing the continuous signals acquired by the displacement sensors and applying a Fast Fourier Transform to obtain the dominant frequency component. Radial offset feature analysis fuses displacement sensor array data, constructing a spatial vector model based on three sets of orthogonal displacement sensor readings to calculate the spatial angle and offset distance of the piston rod axis from the theoretical baseline. The displacement feature parameter output includes frequency values ​​and offset vectors, which are then transmitted to the seal degradation analysis unit.

[0065] A dynamic stress distribution model is constructed using a sealing deformation feature modeling unit. The model input interface receives characteristics of the pressure rise slope, steady-state pressure pulsation amplitude, and axial displacement fluctuation frequency. The modeling process utilizes a finite element analysis library to discretize the hydraulic cylinder sealing system into a three-dimensional mesh model. Pressure characteristics are mapped to model boundary load conditions, and displacement frequency characteristics are converted into periodic load functions. The stress calculation engine solves for the nodal stress distribution, outputting the maximum principal stress value and stress concentration factor of the sealing contact area. The model update frequency is synchronized with the sensor data acquisition rate, recalculating the stress distribution map after each data update. The stress distribution data is stored as a structured array and transmitted to the multi-scale condition assessment module via the ship's data bus.

[0066] The ship's sensor network adopts a star topology. Vibration sensors are mounted on the surface of the hydraulic cylinder body along three orthogonal axes, pressure sensors are integrated into the pressure interface of the sealed cavity, and temperature sensors are embedded in the piston rod guide sleeve. All sensors are connected to the ship's sensor gateway via waterproof connectors. The gateway device is equipped with an anti-interference shielding layer, and its internal circuit board is divided into analog signal processing and digital signal conversion areas. The parsing logic of the heterogeneous protocol adaptation submodule is embedded in the gateway's programmable logic device, and the spatiotemporal calibration engine runs on the real-time operating system partition of the ship's main control computer. The multidimensional data synchronization buffer queue is implemented as a dual-port memory, physically connected to the main control computer and the data preprocessing fusion module processor. The three units of the sealing feature depth extraction module are deployed on a dedicated digital signal processor, which is equipped with a high-speed cache to store intermediate values ​​of feature parameters. The feature data transmission path adopts a parallel bus architecture, and the bus clock frequency is synchronized with an integer multiple of the sensor sampling rate. During system operation, the sensor data stream enters the feature parsing process after protocol parsing, spatiotemporal calibration, and buffer relay. The final generated set of sealing feature parameters is transmitted to downstream modules through a memory-mapped interface.

[0067] Example 3: See Figure 4 The seal degradation analysis unit processes three types of input features from the seal feature depth extraction module: a time series of pressure relief rate features, a dataset of piston rod radial offset features, and the output of the seal dynamic stress distribution model. This unit is deployed on a ship-specific digital signal processor, which is equipped with a high-speed cache to store real-time input data. The time series of pressure relief rate features contains the pressure relief rate value per minute, the dataset of piston rod radial offset features records the three-dimensional spatial offset vector, and the output of the seal dynamic stress distribution model provides the stress distribution matrix of the sealing ring. The core function of the unit is to calculate the cumulative fatigue of the seal material, integrating the three types of input features for joint analysis. The pressure relief rate feature reflects the leakage state of the sealing system, and its numerical change is related to the degree of seal wear. The piston rod radial offset feature generates additional stress load parameters through a geometric transformation algorithm, which represent the additional stress increment caused by the offset. The stress matrix output by the seal dynamic stress distribution model serves as the basic stress field data. The calculation of the cumulative fatigue uses an improved linear damage accumulation model, the formula of which is:

[0068] in: This represents the total fatigue damage. This represents the stress amplitude in the i-th time interval, which is obtained by superimposing additional stress load parameters on the basic stress field data; k is a material constant stored in the ship material database; C is the fatigue strength coefficient, which is retrieved from the database according to the material type of the seal. is the duration of the i-th time interval, fixed at 10 milliseconds; n is the total number of stress cycles. The calculation is performed every 10 milliseconds, and the result is written to the ship's memory fatigue accumulation register.

[0069] The operational stability assessment unit simultaneously processes the pressure rise slope characteristic, steady-state pressure pulsation amplitude characteristic, and piston rod axial displacement fluctuation frequency characteristic. The unit runs on the parallel computing core of a digital signal processor, with the three characteristic parameters input through a dedicated data channel. The pressure rise slope characteristic reflects the hydraulic system's response speed, and its value is taken from the latest sampling point of the characteristic sequence. The steady-state pressure pulsation amplitude characteristic includes pulsation peak and trough data pairs. The piston rod axial displacement fluctuation frequency characteristic provides the amplitude spectrum of the dominant frequency component. The hydraulic cylinder vibration energy entropy value generation process employs the information entropy algorithm: first, an energy distribution histogram containing three types of characteristics is constructed, divided into 100 intervals at 0.1 unit intervals; then, the frequency of occurrence of characteristic values ​​in each interval is counted, and the probability distribution is calculated; finally, the Shannon entropy formula is applied to obtain the dimensionless entropy value. The entropy value calculation cycle is once every 100 milliseconds, and the calculation result is transmitted to the dynamic stability coefficient generation subunit.

[0070] The sealing characteristic degradation gradient calculation subunit monitors the temporal evolution trend of fatigue accumulation in the sealing material. This subunit reads historical data from the fatigue accumulation register every 60 seconds, extracting the fatigue accumulation sequence over the past ten minutes. Simultaneously, it collects the characteristic change in pressure relief rate during the same time period, calculated through differential calculation of adjacent sampling points. The sealing performance degradation rate index is calculated using a two-parameter gradient algorithm: a two-dimensional feature space is constructed with the time axis as the independent variable and fatigue accumulation and pressure relief rate change as dependent variables; a least-squares straight line is fitted in the feature space, and the slope of the line is the degradation rate index. The index value is output once per minute and sent to the early warning decision generation module via the ship's data bus.

[0071] The dynamic stability coefficient generation subunit performs cross-domain coupling analysis. Input sources include the hydraulic cylinder vibration energy entropy and piston rod radial offset characteristics. Entropy data comes from the operational stability assessment unit's update stream, updated 10 times per second, while the piston rod radial offset characteristics are input at a frequency of 100 times per second. The coupling analysis employs the covariance matrix method: first, the two types of feature data are aligned along the time axis with a time alignment accuracy of 1 millisecond; then, a two-dimensional data matrix is ​​constructed, with column vectors representing the entropy sequence and offset amplitude sequence, respectively; next, the covariance eigenvalues ​​of the matrix are calculated, and the largest eigenvalue is used as the basic coupling parameter; finally, the basic coupling parameter is mapped to a 0-1 normalized interval, outputting the dynamic operational stability coefficients. The coefficient update frequency is synchronized with the entropy input, outputting 10 values ​​per second. The coefficient values ​​are formatted as single-precision floating-point numbers and written to the ship's shared memory area for use by the real-time health dashboard module.

[0072] The calculation logic for the cumulative fatigue of the sealing material is synchronized with the monitoring of changes in the pressure relief rate characteristics. When a sudden change occurs in the pressure relief rate characteristics, the fatigue accumulation calculation automatically switches to high-speed mode, and the sampling interval is shortened from 10 milliseconds to 1 millisecond. The radial offset characteristics of the piston rod are preprocessed during the generation of the dynamic stability coefficient: the offset vector is converted to the Euclidean norm to eliminate directional influences. The calculation of the hydraulic cylinder vibration energy entropy uses dynamic histogram technology, and the histogram interval width is automatically adjusted according to the distribution range of feature values. The sealing performance degradation gradient calculation subunit starts a robust regression algorithm when data is abnormal to reduce the impact of outliers. The covariance matrix calculation of the dynamic stability coefficient generation subunit adopts a recursive update strategy, incrementally updating matrix elements each time new data arrives to avoid full recalculation.

[0073] The ship-specific digital signal processor allocates two hardware threads to the seal degradation analysis unit: thread one handles fatigue accumulation calculation, and thread two handles vibration energy entropy generation. The processor is equipped with a mathematical co-accelerator to speed up exponential and matrix calculations. The dynamic stability coefficient generation subunit uses a dedicated vector register to store intermediate results of the covariance matrix. All feature parameter transmission paths employ error checking mechanisms, and cyclic redundancy check codes are added to data transmission. The time series of seal material fatigue accumulation is stored in the ship's non-volatile memory in timestamp-value pairs format. The entropy calculation results of the running stability assessment unit are synchronously written to the ship's real-time database. The seal performance degradation rate index generation process is logged, including input data checksums and calculation timestamps. The output channel of the dynamic running stability coefficient is configured with a double-buffered structure to ensure continuous data transmission. The processor temperature monitoring module automatically reduces the calculation frequency when the chip temperature exceeds a threshold to maintain stable system operation.

[0074] Example 4: The risk level mapping unit maintains a seal failure risk level mapping table, which is stored in the non-volatile memory of the ship's embedded system. The mapping table adopts a three-dimensional data structure, with the three dimensions corresponding to the seal performance degradation gradient range, the dynamic operational stability coefficient range, and the risk level code, respectively. During the operation of the ship's hydraulic system, this unit continuously receives real-time parameter streams from the multi-scale condition assessment module. When a new data packet arrives, the unit executes the following matching process: first, it reads the seal performance degradation gradient value and searches for the interval containing that value in dimension one of the mapping table; then, it reads the dynamic operational stability coefficient and determines the corresponding interval in dimension two; finally, it cross-locates to obtain the risk level code. The mapping table update mechanism is implemented through the ship's maintenance port, and the engine engineer can connect to dedicated equipment to modify the threshold range.

[0075] Table 1: Sealing Failure Risk Level Mapping Table

[0076]

[0077] Referring to Table 1, the instruction generation engine is activated after the risk level mapping unit outputs the level code. This engine contains a pre-compiled instruction template library, with each template corresponding to a specific level code. Taking level code 301 (warning status) as an example: the engine retrieves the ALERT2 instruction set from the template library, which contains three-level response logic. The primary response generates a yellow warning signal code, in the format of a 16-bit binary instruction "1100101011001010", which is sent to the bridge alarm via the ship's CAN bus. The secondary response generates the maintenance strategy code "MTC202", which contains four maintenance instructions: check the torque of the sealing cavity pressure interface bolts (standard value 350 N·m), clean the piston rod surface (using a hydraulic-specific cleaner), record the hydraulic oil temperature (for 30 minutes), and back up the system operation log. The tertiary response generates a hydraulic system load limiting instruction, reducing the maximum working pressure setting of the cylinder to 75% of the rated value.

[0078] Example of implementation process: During the operation of the hydraulic steering gear system of an ocean-going cargo ship, the multi-scale condition assessment module continuously outputs monitoring parameters. At a certain moment, the data packet shows a sealing performance degradation gradient of 0.82% and a dynamic operating stability coefficient of 0.63. The risk level mapping unit searches the mapping table and matches the intersection of the gradient range of 0.5-1.0% and the stability coefficient of 0.50-0.70, determining risk level code 301. This code triggers the ALERT2 response procedure of the command generation engine. The yellow warning light on the bridge control panel flashes at a frequency of 1Hz, and the LCD display pops up the text warning "Hydraulic sealing system warning level 301". The maintenance strategy code "MTC202" is transmitted to the engineer's handheld terminal, and the terminal interface displays a list of four maintenance tasks and operation instructions. The ship control bus synchronously sends a pressure limit command to the hydraulic controller, and the upper limit of the steering gear hydraulic system pressure is automatically adjusted to 21MPa (originally rated at 28MPa).

[0079] The execution paths for instructions differ depending on the risk level. When risk level code 401 (dangerous state) is triggered, the instruction generation engine executes the ALERT3 response procedure. This procedure first plays a pre-recorded emergency voice announcement via the ship's public address system, containing specific equipment numbers and risk codes. Simultaneously, a red warning signal code "1110001110001110" is generated, triggering three parallel operations: the first signal activates the engine room main alarm (105-decibel rotating siren); the second signal sends a shutdown command to the hydraulic power unit (20-second soft shutdown sequence); and the third signal is encrypted and transmitted to the ship's satellite communication equipment, automatically generating a distress message containing a snapshot of the equipment status. The maintenance strategy code is upgraded to "MTC401," containing emergency sealing and isolation operation guidelines and spare parts replacement procedures, which are pushed to the chief engineer's workstation via the ship's local area network.

[0080] The ship's network architecture ensures reliable command transmission. Warning signal codes are transmitted via a dedicated alarm control bus, employing a dual-ring redundant topology with a transmission rate of 1 Mbps. Maintenance strategy codes are transmitted via the ship's Ethernet, with network switches configured with VLANs to isolate alarm and maintenance traffic. Shutdown commands are hardwired to the hydraulic controller, using shielded twisted-pair cables. The execution status of all commands is monitored: the alarm control bus provides feedback on relay engagement status signals, Ethernet transmission receives TCP acknowledgments, and the hardwired circuit detects current continuity. Status monitoring data is returned to the command generation engine, which records command execution logs containing three key elements: timestamp, command code, and device response code.

[0081] The risk level mapping unit has an exception handling mechanism. When input parameters exceed the mapping table range, the unit automatically switches to a conservative strategy: the maximum value range is used when gradient parameters exceed limits, and the minimum value range is used when stability coefficients exceed limits. The mapping table version management function records the operator ID and timestamp for each modification. The instruction generation engine's template library implements a write protection mechanism, allowing only devices with encryption keys to update template content. During ship navigation, the system automatically verifies the integrity of the mapping table every hour, using a cyclic redundancy check algorithm to detect data corruption. If verification fails, a backup and recovery process is triggered, restoring the initial mapping table from a read-only partition of the ship's memory.

[0082] The command generation engine supports dynamic strategy adjustments. The ship's maintenance system feedback interface receives maintenance task completion status codes. When a torque check completion code for task "MTC202" is received, the engine automatically releases the hydraulic system pressure limit command. If a piston rod cleaning completion code is not received within two hours, the engine upgrade command is "MTC202+", which increases the piston rod surface temperature monitoring frequency to once per minute. Historical command records are stored in the ship's black box equipment, using the aerospace industry standard ARINC429. Each record contains 64 bytes of command data and 8 bytes of status check code.

[0083] Example 5: The ship data compression and storage module operates a multi-dimensional time-series data stream, which continuously originates from a multi-source sensor data acquisition module. The module's core processor runs a time-slice compression algorithm with a fixed time window of 30 seconds. At the end of each time window, the data collector extracts all raw sensor data packets for that period, including triaxial data from vibration sensors, millisecond-level sampling values ​​from pressure sensors, and temperature sensor readings. The compression engine initiates a Huffman coding program: first, it scans the data packets to construct a character frequency table, which includes statistics on the frequency of data values; then, it generates an optimal binary tree, assigning short codes to high-frequency data and long codes to low-frequency data; finally, it outputs compressed data blocks with an attached check sequence. The compression process is accompanied by ship positioning operations; the module calls the ship's GPS receiver to obtain the current latitude and longitude coordinates. The coordinate data is formatted, with longitude values ​​converted to the ISO6709 standard format and latitude values ​​padded to second-level accuracy. The converted latitude and longitude strings are embedded in the header of the compressed data block, generating a compressed unit with a location tag. The compressed unit is written to the ship's solid-state memory via a direct memory access channel, with the storage path organized by a year-month-day hierarchical directory.

[0084] The historical data maintenance unit performs automatic data management based on the ship's sailing cycle. The unit's internal clock synchronizes with the ship's master timer system and reads the cycle parameters defined in the ship's sailing configuration file. The standard configuration defines three sailing cycle phases: departure phase (48 hours after anchoring), cruising phase (48 hours after departure to 72 hours before arrival), and berthing phase (72 hours before arrival). The unit continuously monitors the ship's sailing state machine, triggering a data cleanup procedure when the state machine switches between cycle phases. The cleanup procedure scans the compressed storage area and identifies the timestamp attributes of expired data slices. The expiration determination logic is as follows: for the departure phase, the most recent 5 days of data are retained; for the cruising phase, the current data and the previous 10 days of data are retained; for the berthing phase, only the current task data is retained. Expired slices are marked with a deletion flag, and a background process batch erases the storage blocks. The erasure operation performs physical data overwriting, using a three-pass random data rewriting strategy.

[0085] The historical data maintenance unit comprises a data lifecycle management subunit and a storage space optimization subunit. The data lifecycle management subunit interfaces with the ship navigation log database, establishing a mapping between compressed data slices and navigation logs. The mapping process parses navigation log entries, extracting log timestamps and ship status codes. When a new compression unit is generated, the management subunit retrieves the navigation log time window and matches the log status code to mark the lifecycle attributes of the data slice. Attribute marking uses color coding: red indicates critical mission data (such as port entry / exit operation periods), yellow indicates routine navigation data, and green indicates low-value data. The marking information is written to the compression unit's metadata area, which is stored independently of the compressed data body.

[0086] The storage space optimization subunit monitors the sealing performance degradation gradient parameters output by the multi-scale state assessment module. This subunit is equipped with a gradient threshold comparator, with the threshold set at 0.3% / hour. When the real-time gradient value remains below the threshold for 30 minutes, the optimization subunit activates a resolution adjustment program. The program locks the compressed data slices for the corresponding time period and restores the original data stream using the decompression engine. The restored data is sent to the downsampling processor, which applies different strategies according to data type: vibration data retains peak sampling points, keeping one maximum value for every 10 original data points; pressure data uses an arithmetic mean method, merging every 5 points into one mean point; temperature data directly deletes intermediate values, retaining only the first and last data points of each minute. The downsampled data stream is then recompressed using time slices to generate a low-resolution compressed unit that overwrites the original storage area.

[0087] The ship's solid-state storage employs a multi-tiered storage architecture. The primary storage area stores high-resolution compressed data, configured with a RAID1 disk array for redundancy protection. The secondary storage area stores downsampled, low-resolution data, using a single-disk storage mode. A storage space optimization sub-unit maintains a storage mapping table, recording the physical address and resolution identifier of data slices. When the system reads historical data, the access interface automatically selects the optimal resolution dataset based on the requested time range. The ship's navigation log database integrates a compressed storage index module, allowing engineers to retrieve historical data using a combination of time and geographic location criteria.

[0088] The data lifecycle management subunit implements a marker propagation mechanism. When a compressed data slice is downsampled, the original color-coded attributes are inherited by the new data unit. The storage space optimization subunit is configured with an anomaly handling procedure: if a sudden change in the sealing performance degradation gradient is detected after downsampling (a change exceeding 0.5% between adjacent sampling points), the procedure automatically restores the high-resolution data for that period. The restoration process calls the ship's backup storage area, which retains a snapshot of the original data from the most recent 72 hours. The ship's communication system provides a remote interface for data maintenance. In ship-to-shore data transmission mode, low-resolution data is sent first, and high-resolution data is transmitted on demand.

[0089] The compressed storage module's processor is configured with a dedicated buffer. The front-end input buffer is designed with a capacity of 32MB to hold 30 seconds of uncompressed raw data stream. An independent memory pool is allocated to the compression engine's workspace to avoid disk read / write blocking the processing flow. The task scheduler of the historical data maintenance unit manages concurrent operations: data cleaning tasks are scheduled during system idle periods, and resolution adjustment tasks have real-time preemption priority. The ship's power management system provides the storage module with an independent power supply line with voltage regulation, automatically switching to backup power when voltage fluctuations exceed 10%. All storage operations are logged in an audit log, which includes the operation type code, data slice ID, and execution result code. The audit file is separately encrypted and stored in the ship's black box device.

[0090] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0091] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A marine vessel hydraulic cylinder condition monitoring and sealing performance early warning system, characterized in that, The method comprises the following steps: A multi-source sensing data acquisition module is used to synchronously acquire multi-dimensional time sequence operation data of the hydraulic oil cylinder through a vibration sensor, a pressure sensor and a temperature sensor; A data preprocessing and fusion module is used to perform noise suppression, timestamp alignment and multi-modal data fusion on the multi-dimensional time sequence operation data to generate a standardized oil cylinder state feature sequence; A sealing feature deep extraction module is used to analyze sealing cavity pressure fluctuation features, piston rod displacement oscillation features and sealing element deformation trend features of the hydraulic oil cylinder from the standardized oil cylinder state feature sequence; A multi-scale state evaluation module is used to calculate a sealing performance degradation gradient and a dynamic operation stability coefficient of the hydraulic oil cylinder according to the sealing cavity pressure fluctuation features, the piston rod displacement oscillation features and the sealing element deformation trend features; An early warning decision generation module is used to generate a sealing failure risk level and a pre-warning control instruction based on the sealing performance degradation gradient and the dynamic operation stability coefficient. The sealing feature deep extraction module comprises: A sealing cavity pressure feature analysis unit is used to extract a pressure rise slope feature, a steady-state pressure pulsation amplitude feature and a pressure relief rate feature from the standardized oil cylinder state feature sequence; A piston rod displacement feature analysis unit is used to calculate a piston rod axial displacement fluctuation frequency feature and a radial offset feature from the standardized oil cylinder state feature sequence; A sealing element deformation feature modeling unit is used to construct a sealing element dynamic stress distribution model according to the pressure rise slope feature, the steady-state pressure pulsation amplitude feature and the piston rod axial displacement fluctuation frequency feature.

2. The marine vessel hydraulic cylinder condition monitoring and seal performance early warning system of claim 1, wherein, Further comprising: A system safety management module is used to perform hierarchical verification on the operation permission of a user accessing the multi-source sensing data acquisition module; A real-time health board module is used to visually display real-time change curves of the sealing performance degradation gradient and the dynamic operation stability coefficient.

3. The marine vessel hydraulic cylinder condition monitoring and seal performance early warning system of claim 1, wherein, The multi-source sensing data acquisition module comprises: A heterogeneous protocol adaptation sub-module is used to analyze different communication protocol data packets of the vibration sensor, the pressure sensor and the temperature sensor; A space-time calibration engine is used to unify the time stamps of vibration data, pressure data and temperature data to a ship main clock reference and compensate for the spatial position differences of the sensors; A multi-dimensional data synchronous buffer queue is used to temporarily store the aligned multi-sensor data stream and output it to the data preprocessing and fusion module in a fixed time window.

4. The marine hydraulic cylinder condition monitoring and seal performance warning system of claim 1, wherein, The multi-scale state evaluation module comprises: A sealing degradation analysis unit is used to calculate a sealing element material fatigue accumulation according to the pressure relief rate feature, the piston rod radial offset feature and the sealing element dynamic stress distribution model; An operation stability evaluation unit is used to generate a hydraulic oil cylinder vibration energy entropy value based on the pressure rise slope feature, the steady-state pressure pulsation amplitude feature and the piston rod axial displacement fluctuation frequency feature.

5. The marine vessel hydraulic cylinder condition monitoring and seal performance early warning system of claim 4, wherein, The sealing degradation analysis unit comprises: A sealing feature degradation gradient calculation sub-unit is used to calculate a sealing performance degradation rate index according to the time sequence change amount of the sealing element material fatigue accumulation and the pressure relief rate feature. A dynamic stability coefficient generating subunit is configured to cross-domain coupling analyze the hydraulic cylinder vibration energy entropy value and the piston rod radial offset value, and output a dynamic operation stability coefficient.

6. The marine hydraulic cylinder condition monitoring and seal performance warning system of claim 1, wherein, The early warning decision generating module comprises: A risk level mapping unit is configured to match a predefined sealing failure risk level threshold according to a combined relationship between the sealing performance degradation gradient and the dynamic operation stability coefficient; An instruction generating engine is configured to generate an early warning signal triggering logic and a maintenance strategy code based on the sealing failure risk level threshold.

7. The marine hydraulic cylinder condition monitoring and seal performance warning system of claim 1, wherein, Further comprising: A ship data compression storage module is configured to segmentally compress the multi-dimensional time-series operation data by using a time-slice compression algorithm, and add a ship latitude and longitude position label; A historical data maintenance unit is configured to automatically clean up expired data slices according to a preset ship navigation cycle.

8. The marine vessel hydraulic cylinder condition monitoring and seal performance early warning system of claim 7, wherein, The historical data maintenance unit comprises: A data life cycle management subunit is configured to mark an effective period of the multi-dimensional time-series operation data based on a ship navigation log; A storage space optimization subunit is configured to automatically reduce a storage resolution of data in a corresponding time period when the sealing performance degradation gradient is lower than a set threshold.

9. The marine hydraulic cylinder condition monitoring and seal performance warning system of claim 2, wherein, The real-time health board module comprises: An alarm linkage unit is configured to trigger a ship engine room audible and visual alarm device when the sealing failure risk level exceeds a preset warning value; An early warning log generating unit is configured to record a timestamp of the early warning control instruction and a corresponding dynamic operation stability coefficient snapshot.

Images