A compensation method for dynamic error of marine winch metering
By synchronously sampling the winch drum angular displacement and rope acceleration signals, constructing a set of sine and cosine basis functions for projection reconstruction, identifying high-frequency vibrations and compensating for low-frequency trends, the problem of large metering errors in marine winches was solved, achieving high-precision and stable metering results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUNAN TIANJIAN OFFSHORE ENG EQUIP CO LTD
- Filing Date
- 2026-03-05
- Publication Date
- 2026-05-15
AI Technical Summary
The high-frequency vibrations and reciprocating elastic expansion and contraction generated by ocean winches under wave and hull motion result in large dynamic errors in meter counting. Existing technologies cannot adaptively adjust filtering parameters and cannot effectively distinguish between high-frequency vibrations and actual rope length changes, leading to a decrease in meter counting accuracy and stability.
By synchronously sampling the winch drum angular displacement and rope acceleration signals, a set of sine and cosine basis functions is constructed for projection reconstruction. The dominant frequency of high-frequency vibration is identified, the low-frequency trend is accumulated and the high-frequency error is subtracted to achieve dynamic error compensation.
It achieves high-precision metering in harsh marine environments, eliminates vibration and noise, adapts to different sea conditions, simplifies controller design, and improves the real-time performance and robustness of metering.
Smart Images
Figure CN121786334B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of marine winch measurement and control technology in marine operation equipment, specifically a method for compensating for dynamic errors in marine winch meter counting. Background Technology
[0002] Marine winches, as key actuators in ships and marine engineering equipment, are widely used in applications such as acoustic towed hull deployment and recovery, marine observation mooring deployment, sampler lifting and lowering, and underwater robot mooring operations. The winch's rope length or lowering depth is typically obtained through a mechanical counting wheel or drum connected to an encoder, serving as a fundamental measurement parameter for the upper-level control system to implement depth control, safety collision avoidance, and task determination. In calm or gentle sea conditions, existing meter-counting methods are generally sufficient for engineering requirements.
[0003] Due to the significant wave, swell, and six-degree-of-freedom motion characteristics of the marine environment, winches often operate in a highly disturbed environment below the undulating deck. The rope, under the coupling effect of wave excitation and hull motion, generates significant high-frequency vibrations and reciprocating elastic stretching, causing jitter and high-frequency oscillations in the encoder angular displacement signal connected to the drum or counting wheel. This results in large dynamic errors in the metering results over a short timescale. Especially in deep water, with large fluctuations, and rapid cable deployment, the metering curve exhibits severe sawtooth fluctuations, making it difficult to reflect the true trend of rope length change. Common improvement approaches in existing technologies include: increasing encoder resolution, optimizing the mechanical transmission structure to reduce backlash, or performing simple time-moving averaging or low-pass filtering with a fixed cutoff frequency on the metering signal in the host computer; some solutions also attempt to introduce experience-based fixed transfer function models or universal filters to perform frequency domain filtering or state estimation of the measurement signal. However, the above methods generally suffer from the following common problems: First, the filtering parameters are usually pre-tuned and fixed, making it difficult to adaptively adjust the suppression level of different wave frequency bands according to sea state changes. During periods of transition from stable to drastic sea state changes, insufficient filtering or excessive smoothing can easily occur, leading to excessive residual vibration errors or suppression of the actual rope length change, respectively. Second, most methods do not jointly analyze the rope length measurement signal with synchronously acquired dynamic quantities (such as rope end acceleration and structural vibration response), making it impossible to distinguish between "high-frequency vibration components excited by sea state" and "actual rope length change components caused by operating conditions" from an energy or spectral characteristic perspective. This makes it difficult to achieve targeted dynamic error compensation when frequency bands overlap or operating conditions change abruptly. Third, many methods rely on fixed mathematical models established through experience, which have limited applicability to different ship types, winch structures, and cable types. Once the operating conditions exceed the model's assumptions, the metering accuracy and stability significantly decrease, requiring frequent manual parameter adjustments on-site. There is a lack of a unified spectral analysis and differentiation mechanism closely coupled with the actual measurement signal.
[0004] Therefore, this case aims to propose a method for compensating dynamic errors in meter counting of marine winches. First, the winch drum angular displacement and rope acceleration signals are converted into rope length and acceleration time series through synchronous sampling. Then, a set of sine and cosine basis functions is constructed using discrete angular frequencies within a single sliding window. The original rope length signal is then projected and reconstructed. Next, the upper limit of the dominant high-frequency vibration frequency is identified through the acceleration power spectrum. The components not exceeding this upper limit are accumulated to extract the low-frequency main trend. Finally, the low-frequency trend and high-frequency error components are subtracted from the original signal to achieve dual compensation output for meter counting errors. Summary of the Invention
[0005] This invention provides a method for compensating for dynamic errors in meter counting of marine winches, thereby helping to solve the problems mentioned in the background art.
[0006] This invention provides the following technical solution: a method for compensating dynamic errors in meter counting of marine winches, comprising:
[0007] By synchronously sampling the angular displacement acquisition mechanism and the acceleration acquisition mechanism along the rope direction, the time series of the drum angular displacement and the time series of acceleration are obtained, and the time series of the drum angular displacement is converted into the original rope time series.
[0008] Within a single data window, a one-dimensional rope time series vector and a one-dimensional acceleration time series vector are constructed according to the sampling order.
[0009] Based on the total number of sampling points and the sampling period, a discrete angular frequency is set to generate sinusoidal basis function vectors and cosine basis function vectors, which constitute the orthogonal basis function set of harmonic decomposition.
[0010] Projection operations are performed on the original long-time rope sequence at each discrete angular frequency to obtain the sine component coefficients and cosine component coefficients. The reconstructed and accumulated rope length measurement values are obtained by the harmonic reconstruction.
[0011] Perform spectral analysis on the acceleration time series, calculate the acceleration power spectrum intensity and the average power spectrum intensity, identify the set of vibration-dominant discrete angular frequency indices, and select the discrete angular frequency with the largest angular frequency value as the upper limit of the high-frequency disturbance frequency.
[0012] By accumulating harmonic components within a discrete angular frequency range not exceeding the upper limit of the high-frequency disturbance frequency, a sequence of low-frequency main trend rope length measurements is obtained.
[0013] Subtract the corresponding low-frequency main trend rope length measurement value from the original rope length measurement value to obtain the high-frequency dynamic error component, and form an error vector according to the sampling order;
[0014] The original rope length measurement value is subtracted from the corresponding high-frequency dynamic error component to obtain the compensated rope length measurement value. The compensated rope length measurement value sequence is formed according to the sampling order and used as the output result of dynamic error compensation for ocean winch meter counting.
[0015] Optionally, the step of synchronously sampling through the angular displacement acquisition mechanism and the acceleration acquisition mechanism along the rope direction to obtain the drum angular displacement time series and acceleration time series, and converting the drum angular displacement time series into the original rope time series, specifically includes:
[0016] An angular displacement acquisition mechanism is set up along the winch rope exit path. The angular displacement acquisition mechanism is implemented by an angular displacement encoder. An equal interval sampling period is selected, the total number of sampling points in a single data window is determined, and multiple sampling times are obtained sequentially according to the sampling period at the start of sampling. The drum angular displacement measurement value is recorded at each sampling time.
[0017] Obtain the radius parameter of the winch drum, multiply the drum angular displacement measurement value at each sampling time with the drum radius parameter, and generate the original rope length measurement value corresponding to each sampling time in chronological order;
[0018] An acceleration acquisition mechanism along the rope direction is set up on the winch rope exit path. The acceleration acquisition mechanism uses an acceleration sensor to keep the acceleration sampling and rope length sampling synchronized in time. At each sampling moment, the acceleration sampling value along the rope direction is recorded and formed into an acceleration time series according to the sampling order.
[0019] Optionally, the step of constructing a one-dimensional rope time series vector and a one-dimensional acceleration time series vector in the sampling order within a single data window specifically includes:
[0020] Arrange all the original rope length measurements obtained in a single data window in the order of sampling to form a one-dimensional rope length time sequence vector containing only the original rope length measurements.
[0021] All acceleration sample values obtained within a single data window are arranged sequentially according to the sampling order to form a one-dimensional acceleration time series vector containing only acceleration sample values.
[0022] Optionally, the step of setting discrete angular frequencies based on the total number of sampling points and the sampling period, generating sine basis function vectors and cosine basis function vectors to form a harmonic decomposition orthogonal basis function set, specifically includes:
[0023] Based on the total number of sampling points within a single data window, the number of discrete angular frequencies used in harmonic decomposition is set to one-quarter of the total number of sampling points;
[0024] Based on the total number of sampling points and the sampling period, perform an angular frequency calculation operation for each discrete angular frequency index. Multiply the angular constant representing a complete period with the discrete angular frequency index, and divide the result with the product of the total number of sampling points and the sampling period to obtain the discrete angular frequency corresponding to the discrete angular frequency index.
[0025] For each discrete angular frequency, the corresponding sinusoidal basis function values are calculated at all sampling times and arranged into a sinusoidal basis function vector according to the sampling order;
[0026] For each discrete angular frequency, the corresponding cosine basis function values are calculated at all sampling times and arranged into a cosine basis function vector according to the sampling order, so that the sine basis function vector and the cosine basis function vector constitute a harmonic decomposition orthogonal basis function set within a finite time window.
[0027] Optionally, the step of performing projection operations on the original long-time rope sequence at each discrete angular frequency to obtain the sine component coefficients and cosine component coefficients, reconstructing and accumulating them to obtain the harmonic reconstructed rope length measurement value sequence specifically includes:
[0028] For each discrete angular frequency, at all sampling times, the original rope length measurement value is multiplied by the corresponding sine basis function value at each sampling time. The product result of all sampling points within a single data window is summed. The sum is multiplied by a constant two and then divided by the total number of sampling points to obtain the sine component coefficients at the current discrete angular frequency.
[0029] For each discrete angular frequency, at all sampling times, the original rope length measurement value is multiplied by the cosine basis function value at the corresponding sampling time. The product result of all sampling points within a single data window is summed. The sum is multiplied by a constant two and then divided by the total number of sampling points to obtain the cosine component coefficients at the current discrete angular frequency.
[0030] At each sampling time, the sine basis function value and the sine component coefficients at the corresponding sampling time are multiplied, and the cosine basis function value and the cosine component coefficients at the corresponding sampling time are multiplied. The two product results are added together to obtain the harmonic reconstruction component at the current discrete angular frequency.
[0031] At each sampling time, the harmonic reconstruction components obtained at all discrete angular frequencies are accumulated to obtain a sequence of harmonic reconstruction rope length measurements.
[0032] Optionally, the step of performing spectral analysis on the acceleration time series, calculating the acceleration power spectrum intensity and the average power spectrum intensity, identifying the set of dominant discrete angular frequencies, and selecting the discrete angular frequency with the largest angular frequency value as the upper limit of the high-frequency disturbance frequency specifically includes:
[0033] For each discrete angular frequency, at all sampling times, the acceleration sample value is multiplied by the corresponding sine basis function value at each sampling time. The product results at all sampling points are summed to obtain the cumulative value of the sine component, and the cumulative value of the sine component is squared. At the same time, the acceleration sample value is multiplied by the corresponding cosine basis function value at each sampling time. The product results at all sampling points are summed to obtain the cumulative value of the cosine component, and the cumulative value of the cosine component is squared. The two squared results are then added together to obtain the acceleration power spectrum intensity at the current discrete angular frequency.
[0034] The acceleration power spectrum intensity obtained at all discrete angular frequencies is averaged. The sum of all acceleration power spectrum intensities is then divided by the number of discrete angular frequencies to obtain the average power spectrum intensity.
[0035] Among all discrete angular frequencies, the discrete angular frequency indices with acceleration power spectrum intensity greater than average power spectrum intensity are selected to form a set of vibration-dominant discrete angular frequency indices.
[0036] When the set of vibration-dominant discrete angular frequencies is not empty, the discrete angular frequency with the largest angular frequency value is selected from the set of vibration-dominant discrete angular frequencies as the upper limit of the vibration-dominant frequency; when the set of vibration-dominant discrete angular frequencies is empty, the discrete angular frequency with the largest angular frequency value is selected from all discrete angular frequencies as the upper limit of the vibration-dominant frequency, and the upper limit of the vibration-dominant frequency is used as the upper limit of the high-frequency disturbance frequency.
[0037] Optionally, the step of accumulating harmonic components within a discrete angular frequency range not exceeding the upper limit of the high-frequency disturbance frequency to obtain a sequence of low-frequency main trend rope length measurements specifically includes:
[0038] In all discrete angular frequency indices, select those whose corresponding angular frequencies are not higher than the upper limit of the high-frequency disturbance frequency to form a low-frequency index set.
[0039] At each sampling time, for each discrete angular frequency in the low-frequency index set, the sine component coefficient corresponding to each discrete angular frequency is multiplied with the sine basis function value at each sampling time, and the cosine component coefficient corresponding to each discrete angular frequency is multiplied with the cosine basis function value at each sampling time. The two product results are added together to obtain the low-frequency harmonic components at each sampling time and each discrete angular frequency.
[0040] At each sampling moment, the low-frequency harmonic components obtained at all discrete angular frequencies in the low-frequency index set are accumulated to form a sequence of low-frequency main trend rope length measurements representing the low-frequency main trend of winch rope length change over time.
[0041] Optionally, the step of subtracting the corresponding low-frequency main trend rope length measurement value from the original rope length measurement value to obtain the high-frequency dynamic error component, and forming an error vector according to the sampling order, specifically includes:
[0042] At each sampling moment, the low-frequency main trend rope length measurement value corresponding to each sampling moment is subtracted from the original rope length measurement value to obtain the high-frequency dynamic error component corresponding to each sampling moment.
[0043] According to the sampling time sequence, the high-frequency dynamic error components of all sampling times are arranged sequentially to form a high-frequency dynamic error vector.
[0044] Optionally, the step of subtracting the corresponding high-frequency dynamic error component from the original rope length measurement value to obtain the compensated rope length measurement value, and forming a sequence of compensated rope length measurement values according to the sampling order, as the output result of the dynamic error compensation for the marine winch meter counting, specifically includes:
[0045] At each sampling moment, the original rope length measurement value is subtracted from the high-frequency dynamic error component corresponding to each sampling moment to obtain the compensated rope length measurement value corresponding to each sampling moment.
[0046] According to the sampling time sequence, the compensated rope length measurement values of all sampling times are arranged sequentially to form a sequence of compensated rope length measurement values, which serves as the output result of dynamic error compensation for ocean winch meter counting.
[0047] The present invention has the following beneficial effects:
[0048] 1. The collected raw data of drum angular displacement and acceleration are organized into one-dimensional rope time-series vectors and one-dimensional acceleration time-series vectors respectively, according to the sampling order, within a fixed-length data window. A sliding window combined with vectorization processing allows subsequent basis function projection and spectral analysis to be performed in parallel on the same time scale, avoiding the computational redundancy and real-time bottlenecks of traditional point-by-point calculations. Firstly, by determining the window length and sampling frequency, the signal resolution can be dynamically adjusted according to the actual operating speed of different winches. Secondly, vectorized storage facilitates acceleration through matrix operations, making it suitable for online real-time processing. Thirdly, it lays a unified data structure foundation for subsequent discrete frequency basis function generation and projection reconstruction. Compared with existing methods that only use single-time-domain signal processing or direct calculation using discrete Fourier transform, this method can flexibly control the balance between frequency resolution and time response delay based on window parameters.
[0049] 2. By calculating the discrete angular frequencies based on the total number of sampling points and the sampling period, and generating corresponding sine and cosine basis function vectors for each angular frequency, a complete set of orthogonal basis functions for harmonic decomposition is formed. The generation of harmonic decomposition basis functions is synchronized with the time window, ensuring that the value vector of each basis function precisely corresponds to each sampling moment within the current data window. Furthermore, sine and cosine basis functions are used to project the signal, taking into account both amplitude and phase information. This approach achieves several advantages: first, it preserves the time-domain representation of the signal at each discrete frequency component, facilitating subsequent individual processing of specific frequency components; second, the basis function set mathematically constitutes a complete orthogonal system, ensuring minimal reconstruction error; and third, it avoids the global processing required by traditional Fourier transforms, reducing boundary effects. Compared to existing techniques that use fixed filter banks to extract frequency band components, this method can achieve accurate projection and reconstruction of any selected frequency component, providing flexibility for personalized error compensation.
[0050] 3. Perform spectral analysis on the acceleration time series at each discrete angular frequency, calculate the power spectral intensity at each frequency, and compare it with the average intensity at all frequency points to identify the dominant discrete angular frequency index set. First, calculate the power spectrum based on the sum of squares of the projection coefficients of the sine and cosine basis functions of the acceleration signal, replacing conventional Fourier spectrum analysis and directly utilizing the already generated basis function projection results. Second, use the average power spectral intensity as a baseline to dynamically filter out abnormally high vibration components, avoiding insufficient sensitivity or false detection problems caused by fixed threshold settings. Quickly determine the upper limit of the dominant vibration frequency, thus providing accurate boundaries for low-frequency trend extraction and high-frequency error filtering; at the same time, reduce the computational complexity of frequency domain peak detection. Compared with the traditional process that requires performing FFT on the entire signal and then peak detection, this method directly uses the projection coefficients for power calculation, resulting in less computation and higher real-time performance.
[0051] 4. After identifying the upper limit of the high-frequency disturbance frequency, this method selects all discrete angular frequencies not exceeding this upper limit, and accumulates the corresponding harmonic reconstruction components at each sampling time to obtain the low-frequency main trend signal of the winch rope length changing over time. The projected reconstruction values of multiple frequency components are accumulated at once to obtain the entire low-frequency trend, eliminating the need for additional filter design; and the low-frequency trend bandwidth can be dynamically adjusted according to the upper limit of the dominant frequency. On the one hand, this trend signal accurately reflects the actual rope length change during the rope pulling process, removing high-frequency vibration noise; on the other hand, because it is based on projection reconstruction, no secondary convolution or IIR filter is required, ensuring system stability and real-time performance. Compared with existing methods using fixed low-pass filtering or moving averages, this method can adaptively adjust the bandwidth according to the vibration characteristics during operation, balancing noise suppression and trend preservation.
[0052] 5. Subtract the low-frequency main trend rope length measurement from the original rope length measurement to obtain the high-frequency dynamic error component, and construct the error vector according to the sampling order. The error is extracted directly using the difference between the reconstructed low-frequency trend signal and the original signal, eliminating the need for a dedicated high-pass filter or differential; furthermore, the error vector remains synchronized with the original data, facilitating subsequent compensation. Firstly, the high-frequency error component accurately corresponds to short-term deviations caused by vibration, requiring no additional parameter adjustments; secondly, the error vector is directly used for compensation, simplifying controller design and allowing for a more compact connection with the feedback loop of the mechanical system. Compared to traditional methods based on differentiation or high-pass filtering for error extraction, this method is more intuitive, easier to adjust, and has lower computational overhead.
[0053] 6. Finally, by subtracting the high-frequency dynamic error component from the original rope length value sequentially, a sequence of compensated rope length measurements is generated as the final output. This combines the low-frequency trend and high-frequency error after dual separation and compensation, preserving the static and slowly changing characteristics of the winch while eliminating errors caused by vibration, achieving high-precision meter counting. Firstly, the compensated output signal corresponds strictly one-to-one with the original sample, facilitating integration with downstream control and display systems. Secondly, this compensation method exhibits high robustness against noise and mechanical vibration in harsh marine environments. Thirdly, it can operate online in real time, making it suitable for embedded controllers. Compared to existing technologies that only perform single filtering or do not separate trends, this method can eliminate noise while preserving trend information, balancing accuracy and stability. Attached Figure Description
[0054] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation
[0055] 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.
[0056] Example, refer to Figure 1 A method for compensating dynamic errors in meter counting of marine winches, comprising:
[0057] By synchronously sampling the angular displacement acquisition mechanism and the acceleration acquisition mechanism along the rope direction, the time series of the drum angular displacement and the time series of acceleration are obtained, and the time series of the drum angular displacement is converted into the original rope time series.
[0058] Within a single data window, a one-dimensional rope time series vector and a one-dimensional acceleration time series vector are constructed according to the sampling order.
[0059] Based on the total number of sampling points and the sampling period, a discrete angular frequency is set to generate sinusoidal basis function vectors and cosine basis function vectors, which constitute the orthogonal basis function set of harmonic decomposition.
[0060] Projection operations are performed on the original long-time rope sequence at each discrete angular frequency to obtain the sine component coefficients and cosine component coefficients. The reconstructed and accumulated rope length measurement values are obtained by the harmonic reconstruction.
[0061] Perform spectral analysis on the acceleration time series, calculate the acceleration power spectrum intensity and the average power spectrum intensity, identify the set of vibration-dominant discrete angular frequency indices, and select the discrete angular frequency with the largest angular frequency value as the upper limit of the high-frequency disturbance frequency.
[0062] By accumulating harmonic components within a discrete angular frequency range not exceeding the upper limit of the high-frequency disturbance frequency, a sequence of low-frequency main trend rope length measurements is obtained.
[0063] Subtract the corresponding low-frequency main trend rope length measurement value from the original rope length measurement value to obtain the high-frequency dynamic error component, and form an error vector according to the sampling order;
[0064] The original rope length measurement value is subtracted from the corresponding high-frequency dynamic error component to obtain the compensated rope length measurement value. The compensated rope length measurement value sequence is formed according to the sampling order and used as the output result of dynamic error compensation for ocean winch meter counting.
[0065] First, by setting up an angular displacement acquisition mechanism and an acceleration acquisition mechanism along the rope direction, the angular displacement of the drum and the acceleration of the rope are sampled synchronously. Then, the angular displacement data is converted into a long time series of the original rope, and a time series vector of rope length and acceleration is constructed within a data window of fixed length to ensure that the subsequent processing has a unified sample structure in both the time domain and the frequency domain. Next, based on the number of sampling points and the sampling period within the window, sine and cosine basis function vectors are constructed to form a set of orthogonal basis functions for harmonic decomposition, ensuring orthogonal decomposition of each frequency component of the signal. Subsequently, the original rope length sequence is projected to reconstruct the harmonic components, obtaining a sequence of reconstructed rope length measurements for each frequency, achieving fine-grained frequency domain decomposition of the signal. By performing power spectrum analysis based on projection coefficients on the acceleration sequence, the dominant vibration frequency is identified, and the upper limit of the high-frequency disturbance frequency is determined accordingly. This avoids the risk of insufficient bandwidth caused by traditional fixed filter band design and can adaptively filter out marine vibration interference. Harmonic components are accumulated within the frequency range not exceeding this upper limit to obtain the low-frequency main trend rope length sequence, accurately extracting the slow-changing trend. Finally, by subtracting the low-frequency main trend value and the high-frequency error component from the original rope length value, a high-frequency dynamic error vector and a compensated rope length sequence are generated, achieving dual compensation for winch meter counting errors. Compared with the fixed low-pass or high-pass filters commonly used in existing technologies, this method combines projection reconstruction and spectrum analysis within the same framework to adaptively separate components of different frequency bands, preserving the slowly changing true rope length trend while efficiently eliminating vibration noise.
[0066] The process involves synchronously sampling via an angular displacement acquisition mechanism and an acceleration acquisition mechanism along the rope direction to obtain the drum angular displacement time series and acceleration time series, and then converting the drum angular displacement time series into the original rope time series. Specifically, this includes:
[0067] An angular displacement acquisition mechanism is set up along the winch rope exit path. The angular displacement acquisition mechanism is implemented by an angular displacement encoder. An equal interval sampling period is selected, the total number of sampling points in a single data window is determined, and multiple sampling times are obtained sequentially according to the sampling period at the start of sampling. The drum angular displacement measurement value is recorded at each sampling time.
[0068] Obtain the radius parameter of the winch drum, multiply the drum angular displacement measurement value at each sampling time with the drum radius parameter, and generate the original rope length measurement value corresponding to each sampling time in chronological order;
[0069] An acceleration acquisition mechanism along the rope direction is set up on the winch rope exit path. The acceleration acquisition mechanism uses an acceleration sensor to keep the acceleration sampling and rope length sampling synchronized in time. At each sampling moment, the acceleration sampling value along the rope direction is recorded and formed into an acceleration time series according to the sampling order.
[0070] An encoder is installed along the winch rope exit path to measure the drum angular displacement. Sampling is performed at equal intervals, with the time interval between two adjacent sampling points set to... The total number of sampling points within a single data window is set to The sampling time is obtained: , ;in, In the first The roller angular displacement corresponding to each sampling point; For discrete-time sampling index; In order to be with the first The physical moments on the continuous time axis corresponding to each sampling index; This is the start time of sampling;
[0071] Obtain the radius of the winch drum, denoted as . ;
[0072] The angular displacement data is converted into the original rope length signal, specifically as follows:
[0073] ;in, For at any time The corresponding original rope length measurement value;
[0074] An acceleration sensor is installed along the winch's rope exit path to record the acceleration signal along the rope direction at the rope exit point. The sampling was synchronized with the rope length signal, resulting in the following discrete acceleration sequence:
[0075] ;in, For at any time Discrete values of acceleration along the rope direction, measured by an accelerometer.
[0076] Within a single data window, constructing a one-dimensional rope time series vector and a one-dimensional acceleration time series vector according to the sampling order specifically includes:
[0077] Arrange all the original rope length measurements obtained in a single data window in the order of sampling to form a one-dimensional rope length time sequence vector containing only the original rope length measurements.
[0078] All acceleration sample values obtained within a single data window are arranged sequentially according to the sampling order to form a one-dimensional acceleration time series vector containing only acceleration sample values.
[0079] The long-time sequence vector of the rope is constructed as follows: ;in, It is a one-dimensional vector composed of all the original rope length measurements;
[0080] The acceleration response time series vector is constructed as follows: ;in, It is a one-dimensional vector composed of all discrete acceleration values.
[0081] The discrete angular frequency is set based on the total number of sampling points and the sampling period, generating sine basis function vectors and cosine basis function vectors to form a harmonic decomposition orthogonal basis function set, specifically including:
[0082] Based on the total number of sampling points within a single data window, the number of discrete angular frequencies used in harmonic decomposition is set to one-quarter of the total number of sampling points;
[0083] Based on the total number of sampling points and the sampling period, perform an angular frequency calculation operation for each discrete angular frequency index. Multiply the angular constant representing a complete period with the discrete angular frequency index, and divide the result with the product of the total number of sampling points and the sampling period to obtain the discrete angular frequency corresponding to the discrete angular frequency index.
[0084] For each discrete angular frequency, the corresponding sinusoidal basis function values are calculated at all sampling times and arranged into a sinusoidal basis function vector according to the sampling order;
[0085] For each discrete angular frequency, the corresponding cosine basis function values are calculated at all sampling times and arranged into a cosine basis function vector according to the sampling order, so that the sine basis function vector and the cosine basis function vector constitute a harmonic decomposition orthogonal basis function set within a finite time window.
[0086] Set the number of discrete frequencies used in harmonic decomposition to ;
[0087] Construct the first The discrete angular frequencies are , ;in, Frequency index;
[0088] Perform steps S301 and S302 for each discrete angular frequency. At all times Construct the following orthogonal basis function vectors:
[0089] S301. Construct the sinusoidal basis function vector as follows:
[0090] ;in, For the first The sinusoidal basis functions at all sampling times The vector of values on;
[0091] S302. Construct the cosine basis function vector as follows:
[0092] ;in, For the first The cosine basis functions at all sampling times The vector that takes values on.
[0093] The process of performing projection operations on the original long-time rope sequence at various discrete angular frequencies to obtain sine and cosine component coefficients, reconstructing and accumulating these coefficients to obtain a harmonic reconstructed rope length measurement sequence specifically includes:
[0094] For each discrete angular frequency, at all sampling times, the original rope length measurement value is multiplied by the corresponding sine basis function value at each sampling time. The product result of all sampling points within a single data window is summed. The sum is multiplied by a constant two and then divided by the total number of sampling points to obtain the sine component coefficients at the current discrete angular frequency.
[0095] For each discrete angular frequency, at all sampling times, the original rope length measurement value is multiplied by the cosine basis function value at the corresponding sampling time. The product result of all sampling points within a single data window is summed. The sum is multiplied by a constant two and then divided by the total number of sampling points to obtain the cosine component coefficients at the current discrete angular frequency.
[0096] At each sampling time, the sine basis function value and the sine component coefficients at the corresponding sampling time are multiplied, and the cosine basis function value and the cosine component coefficients at the corresponding sampling time are multiplied. The two product results are added together to obtain the harmonic reconstruction component at the current discrete angular frequency.
[0097] At each sampling time, the harmonic reconstruction components obtained at all discrete angular frequencies are accumulated to obtain a sequence of harmonic reconstruction rope length measurements.
[0098] Perform steps S401 and S402 for each discrete angular frequency. The projection coefficients of the original rope length signal under its orthogonal basis are constructed as follows:
[0099] S401, The sinusoidal component coefficients are constructed as follows:
[0100] ;in, The original rope length signal is in the first... Projection coefficients on sinusoidal basis functions;
[0101] S402, the cosine component coefficients are:
[0102] ;in, The original rope length signal is in the first... Projection coefficients on each cosine basis function;
[0103] Constructing discrete angular frequencies At any moment The specific refactoring items are:
[0104] ;in, For at any time Frequency index is Harmonic reconstruction components at time;
[0105] All frequency components at time The sum of the above constitutes the overall harmonic reconstruction value:
[0106] ;in, For at any time Index all frequencies The total harmonic reconstruction rope length after the harmonic components are accumulated.
[0107] The process involves performing spectral analysis on the acceleration time series, calculating the acceleration power spectrum intensity and the average power spectrum intensity, identifying the dominant discrete angular frequency index set, and selecting the discrete angular frequency with the largest angular frequency value as the upper limit of the high-frequency disturbance frequency. Specifically, this includes:
[0108] For each discrete angular frequency, at all sampling times, the acceleration sample value is multiplied by the corresponding sine basis function value at each sampling time. The product results at all sampling points are summed to obtain the cumulative value of the sine component, and the cumulative value of the sine component is squared. At the same time, the acceleration sample value is multiplied by the corresponding cosine basis function value at each sampling time. The product results at all sampling points are summed to obtain the cumulative value of the cosine component, and the cumulative value of the cosine component is squared. The two squared results are then added together to obtain the acceleration power spectrum intensity at the current discrete angular frequency.
[0109] The acceleration power spectrum intensity obtained at all discrete angular frequencies is averaged. The sum of all acceleration power spectrum intensities is then divided by the number of discrete angular frequencies to obtain the average power spectrum intensity.
[0110] Among all discrete angular frequencies, the discrete angular frequency indices with acceleration power spectrum intensity greater than average power spectrum intensity are selected to form a set of vibration-dominant discrete angular frequency indices.
[0111] When the set of vibration-dominant discrete angular frequencies is not empty, the discrete angular frequency with the largest angular frequency value is selected from the set of vibration-dominant discrete angular frequencies as the upper limit of the vibration-dominant frequency; when the set of vibration-dominant discrete angular frequencies is empty, the discrete angular frequency with the largest angular frequency value is selected from all discrete angular frequencies as the upper limit of the vibration-dominant frequency, and the upper limit of the vibration-dominant frequency is used as the upper limit of the high-frequency disturbance frequency.
[0112] For each discrete angular frequency Calculate the power spectral intensity of the acceleration sequence at this frequency, specifically:
[0113] ;in, For the acceleration signal at the frequency index Discrete angular frequency The power spectral intensity at the following values;
[0114] Calculate the average power spectral intensity as ;
[0115] Construct a set of dominant vibration frequencies Specifically:
[0116] ;
[0117] Construct the upper limit of the dominant vibration frequency Specifically:
[0118] ;in, For frequency index The angular frequency is the maximum angular frequency in the frequency set.
[0119] The step of accumulating harmonic components within discrete angular frequencies not exceeding the upper limit of the high-frequency disturbance frequency to obtain a sequence of low-frequency main trend rope length measurements specifically includes:
[0120] In all discrete angular frequency indices, select those whose corresponding angular frequencies are not higher than the upper limit of the high-frequency disturbance frequency to form a low-frequency index set.
[0121] At each sampling time, for each discrete angular frequency in the low-frequency index set, the sine component coefficient corresponding to each discrete angular frequency is multiplied with the sine basis function value at each sampling time, and the cosine component coefficient corresponding to each discrete angular frequency is multiplied with the cosine basis function value at each sampling time. The two product results are added together to obtain the low-frequency harmonic components at each sampling time and each discrete angular frequency.
[0122] At each sampling moment, the low-frequency harmonic components obtained at all discrete angular frequencies in the low-frequency index set are accumulated to form a sequence of low-frequency main trend rope length measurements representing the low-frequency main trend of winch rope length change over time.
[0123] Construct a low-frequency index set as ;
[0124] At each sampling time, according to The low-frequency trend signal is calculated as follows:
[0125] ;in, For at any time The low-frequency main trend rope length signal.
[0126] The step of subtracting the corresponding low-frequency main trend rope length measurement value from the original rope length measurement value to obtain the high-frequency dynamic error component, and forming an error vector according to the sampling order, specifically includes:
[0127] At each sampling moment, the low-frequency main trend rope length measurement value corresponding to each sampling moment is subtracted from the original rope length measurement value to obtain the high-frequency dynamic error component corresponding to each sampling moment.
[0128] According to the sampling time sequence, the high-frequency dynamic error components of all sampling times are arranged sequentially to form a high-frequency dynamic error vector.
[0129] The high-frequency error at each moment is set as the difference between the original signal and the low-frequency signal, specifically: ;in, For at any time High-frequency vibration error components;
[0130] The constructed error vector is: ;in, It is a one-dimensional vector composed of the high-frequency vibration error components at all times.
[0131] The process of subtracting the corresponding high-frequency dynamic error component from the original rope length measurement value to obtain the compensated rope length measurement value, and forming a sequence of compensated rope length measurement values according to the sampling order, serves as the output result of the dynamic error compensation for the marine winch meter counting, specifically including:
[0132] At each sampling moment, the original rope length measurement value is subtracted from the high-frequency dynamic error component corresponding to each sampling moment to obtain the compensated rope length measurement value corresponding to each sampling moment.
[0133] According to the sampling time sequence, the compensated rope length measurement values of all sampling times are arranged sequentially to form a sequence of compensated rope length measurement values, which serves as the output result of dynamic error compensation for ocean winch meter counting.
[0134] Every moment The length of the rope after compensation is: ;in, For at any time The measured value of the rope length after compensation;
[0135] The compensated output sequence is constructed as follows: ;in, It is a one-dimensional vector composed of the rope length measurements after compensation at all times.
[0136] 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.
[0137] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A method for compensating dynamic errors in meter counting of marine winches, characterized in that, include: By synchronously sampling the angular displacement acquisition mechanism and the acceleration acquisition mechanism along the rope direction, the time series of the drum angular displacement and the time series of acceleration are obtained, and the time series of the drum angular displacement is converted into the original rope time series. Within a single data window, a one-dimensional rope time series vector and a one-dimensional acceleration time series vector are constructed according to the sampling order. Based on the total number of sampling points and the sampling period, a discrete angular frequency is set to generate sinusoidal basis function vectors and cosine basis function vectors, which constitute the orthogonal basis function set of harmonic decomposition. Projection operations are performed on the original long-time rope sequence at each discrete angular frequency to obtain the sine component coefficients and cosine component coefficients. The reconstructed and accumulated rope length measurement values are obtained by the harmonic reconstruction. Perform spectral analysis on the acceleration time series, calculate the acceleration power spectrum intensity and the average power spectrum intensity, identify the set of vibration-dominant discrete angular frequency indices, and select the discrete angular frequency with the largest angular frequency value as the upper limit of the high-frequency disturbance frequency. By accumulating harmonic components within a discrete angular frequency range not exceeding the upper limit of the high-frequency disturbance frequency, a sequence of low-frequency main trend rope length measurements is obtained. Subtract the corresponding low-frequency main trend rope length measurement value from the original rope length measurement value to obtain the high-frequency dynamic error component, and form an error vector according to the sampling order; The original rope length measurement value is subtracted from the corresponding high-frequency dynamic error component to obtain the compensated rope length measurement value. The compensated rope length measurement value sequence is formed according to the sampling order and used as the output result of dynamic error compensation for ocean winch meter counting.
2. The method for compensating dynamic errors in meter counting for marine winches according to claim 1, characterized in that, The process involves synchronously sampling via an angular displacement acquisition mechanism and an acceleration acquisition mechanism along the rope direction to obtain the drum angular displacement time series and acceleration time series, and then converting the drum angular displacement time series into the original rope time series. Specifically, this includes: An angular displacement acquisition mechanism is set up along the winch rope exit path. The angular displacement acquisition mechanism is implemented by an angular displacement encoder. An equal interval sampling period is selected, the total number of sampling points in a single data window is determined, and multiple sampling times are obtained sequentially according to the sampling period at the start of sampling. The drum angular displacement measurement value is recorded at each sampling time. Obtain the radius parameter of the winch drum, multiply the drum angular displacement measurement value at each sampling time with the drum radius parameter, and generate the original rope length measurement value corresponding to each sampling time in chronological order; An acceleration acquisition mechanism along the rope direction is set up on the winch rope exit path. The acceleration acquisition mechanism uses an acceleration sensor to keep the acceleration sampling and rope length sampling synchronized in time. At each sampling moment, the acceleration sampling value along the rope direction is recorded and formed into an acceleration time series according to the sampling order.
3. The method for compensating dynamic errors in meter counting for marine winches according to claim 2, characterized in that, Within a single data window, constructing a one-dimensional rope time series vector and a one-dimensional acceleration time series vector according to the sampling order specifically includes: Arrange all the original rope length measurements obtained in a single data window in the order of sampling to form a one-dimensional rope length time sequence vector containing only the original rope length measurements. All acceleration sample values obtained within a single data window are arranged sequentially according to the sampling order to form a one-dimensional acceleration time series vector containing only acceleration sample values.
4. The method for compensating dynamic errors in meter counting of marine winches according to claim 3, characterized in that, The discrete angular frequency is set based on the total number of sampling points and the sampling period, generating sine basis function vectors and cosine basis function vectors to form a harmonic decomposition orthogonal basis function set, specifically including: Based on the total number of sampling points within a single data window, the number of discrete angular frequencies used in harmonic decomposition is set to one-quarter of the total number of sampling points; Based on the total number of sampling points and the sampling period, perform an angular frequency calculation operation for each discrete angular frequency index. Multiply the angular constant representing a complete period with the discrete angular frequency index, and divide the result with the product of the total number of sampling points and the sampling period to obtain the discrete angular frequency corresponding to the discrete angular frequency index. For each discrete angular frequency, the corresponding sinusoidal basis function values are calculated at all sampling times and arranged into a sinusoidal basis function vector according to the sampling order; For each discrete angular frequency, the corresponding cosine basis function values are calculated at all sampling times and arranged into a cosine basis function vector according to the sampling order, so that the sine basis function vector and the cosine basis function vector constitute a harmonic decomposition orthogonal basis function set within a finite time window.
5. A method for compensating dynamic errors in meter counting for marine winches according to claim 4, characterized in that, The process of performing projection operations on the original long-time rope sequence at various discrete angular frequencies to obtain sine and cosine component coefficients, reconstructing and accumulating these coefficients to obtain a harmonic reconstructed rope length measurement sequence specifically includes: For each discrete angular frequency, at all sampling times, the original rope length measurement value is multiplied by the corresponding sine basis function value at each sampling time. The product result of all sampling points within a single data window is summed. The sum is multiplied by a constant two and then divided by the total number of sampling points to obtain the sine component coefficients at the current discrete angular frequency. For each discrete angular frequency, at all sampling times, the original rope length measurement value is multiplied by the cosine basis function value at the corresponding sampling time. The product result of all sampling points within a single data window is summed. The sum is multiplied by a constant two and then divided by the total number of sampling points to obtain the cosine component coefficients at the current discrete angular frequency. At each sampling time, the sine basis function value and the sine component coefficients at the corresponding sampling time are multiplied, and the cosine basis function value and the cosine component coefficients at the corresponding sampling time are multiplied. The two product results are added together to obtain the harmonic reconstruction component at the current discrete angular frequency. At each sampling time, the harmonic reconstruction components obtained at all discrete angular frequencies are accumulated to obtain a sequence of harmonic reconstruction rope length measurements.
6. A method for compensating dynamic errors in meter counting for marine winches according to claim 5, characterized in that, The process involves performing spectral analysis on the acceleration time series, calculating the acceleration power spectrum intensity and the average power spectrum intensity, identifying the dominant discrete angular frequency index set, and selecting the discrete angular frequency with the largest angular frequency value as the upper limit of the high-frequency disturbance frequency. Specifically, this includes: For each discrete angular frequency, at all sampling times, the acceleration sample value is multiplied by the corresponding sine basis function value at each sampling time. The product results at all sampling points are summed to obtain the cumulative value of the sine component, and the cumulative value of the sine component is squared. At the same time, the acceleration sample value is multiplied by the corresponding cosine basis function value at each sampling time. The product results at all sampling points are summed to obtain the cumulative value of the cosine component, and the cumulative value of the cosine component is squared. The two squared results are then added together to obtain the acceleration power spectrum intensity at the current discrete angular frequency. The acceleration power spectrum intensity obtained at all discrete angular frequencies is averaged. The sum of all acceleration power spectrum intensities is then divided by the number of discrete angular frequencies to obtain the average power spectrum intensity. Among all discrete angular frequencies, the discrete angular frequency indices with acceleration power spectrum intensity greater than average power spectrum intensity are selected to form a set of vibration-dominant discrete angular frequency indices. When the set of vibration-dominant discrete angular frequencies is not empty, the discrete angular frequency with the largest angular frequency value is selected from the set of vibration-dominant discrete angular frequencies as the upper limit of the vibration-dominant frequency; when the set of vibration-dominant discrete angular frequencies is empty, the discrete angular frequency with the largest angular frequency value is selected from all discrete angular frequencies as the upper limit of the vibration-dominant frequency, and the upper limit of the vibration-dominant frequency is used as the upper limit of the high-frequency disturbance frequency.
7. A method for compensating dynamic errors in meter counting for marine winches according to claim 6, characterized in that, The step of accumulating harmonic components within discrete angular frequencies not exceeding the upper limit of the high-frequency disturbance frequency to obtain a sequence of low-frequency main trend rope length measurements specifically includes: In all discrete angular frequency indices, select those whose corresponding angular frequencies are not higher than the upper limit of the high-frequency disturbance frequency to form a low-frequency index set. At each sampling time, for each discrete angular frequency in the low-frequency index set, the sine component coefficient corresponding to each discrete angular frequency is multiplied with the sine basis function value at each sampling time, and the cosine component coefficient corresponding to each discrete angular frequency is multiplied with the cosine basis function value at each sampling time. The two product results are added together to obtain the low-frequency harmonic components at each sampling time and each discrete angular frequency. At each sampling moment, the low-frequency harmonic components obtained at all discrete angular frequencies in the low-frequency index set are accumulated to form a sequence of low-frequency main trend rope length measurements representing the low-frequency main trend of winch rope length change over time.
8. A method for compensating dynamic errors in meter counting for marine winches according to claim 7, characterized in that, The step of subtracting the corresponding low-frequency main trend rope length measurement value from the original rope length measurement value to obtain the high-frequency dynamic error component, and forming an error vector according to the sampling order, specifically includes: At each sampling moment, the low-frequency main trend rope length measurement value corresponding to each sampling moment is subtracted from the original rope length measurement value to obtain the high-frequency dynamic error component corresponding to each sampling moment. According to the sampling time sequence, the high-frequency dynamic error components of all sampling times are arranged sequentially to form a high-frequency dynamic error vector.
9. A method for compensating dynamic errors in meter counting for marine winches according to claim 8, characterized in that, The process of subtracting the corresponding high-frequency dynamic error component from the original rope length measurement value to obtain the compensated rope length measurement value, and forming a sequence of compensated rope length measurement values according to the sampling order, serves as the output result of the dynamic error compensation for the marine winch meter counting, specifically including: At each sampling moment, the original rope length measurement value is subtracted from the high-frequency dynamic error component corresponding to each sampling moment to obtain the compensated rope length measurement value corresponding to each sampling moment. According to the sampling time sequence, the compensated rope length measurement values of all sampling times are arranged sequentially to form a sequence of compensated rope length measurement values, which serves as the output result of dynamic error compensation for ocean winch meter counting.