A multi-scene adaptive working condition retracting and releasing system

Abstract

The application discloses a multi-scene adaptive working condition winding and unwinding system and relates to the technical field of pneumatic winches.The application can accurately identify five working condition categories, namely, empty load winding and unwinding, rated load winding and unwinding, sudden load, strong wind and low temperature, and avoids the problem of high misjudgment rate of the existing fixed threshold judgment mode by extracting multi-dimensional working condition characteristic vectors such as air source pressure fluctuation, tension change gradient, winding and unwinding speed change and environmental parameter fluctuation from a synchronous data sequence and inputting the working condition classification model for judgment.According to the judgment result, corresponding winding and unwinding control parameter sets are called from a working condition-parameter mapping table, adaptive matching of control parameters and actual working conditions is realized, and during the winding and unwinding process, continuous comparison of real-time data and preset ranges triggers dynamic correction of control parameters, and a correction convergence judgment and multi-parameter weighted fusion mechanism are introduced, so that frequent shaking of control instructions is effectively avoided, and the stability and safety of the winding and unwinding process are ensured.

Description

Technical Field

[0001] This invention relates to the field of pneumatic winch technology, specifically to a multi-scenario adaptive winch system. Background Technology

[0002] Pneumatic winches, as lifting and winding devices powered by compressed air, are widely used in marine engineering, mining, field rescue, and military equipment. They use a pneumatic motor to drive a drum to rotate, thus enabling the winding and unwinding of wire ropes. Compared to electric winches, pneumatic winches have advantages such as better explosion-proof performance, simple structure, low maintenance costs, and suitability for humid and flammable / explosive environments, making them irreplaceable in many special operating scenarios.

[0003] In actual operation, pneumatic winches face complex and varied working conditions. Taking marine engineering as an example, pneumatic winches need to withstand alternating switching between various working conditions at different stages of operation, such as no-load deployment and recovery, rated load deployment and recovery, sudden loading, strong wind interference, and low-temperature environments. Under different working conditions, the fluctuation characteristics of air source pressure, the variation law of wire rope tension, the stability requirements of deployment and recovery speed, and the degree of influence of environmental factors on the winch output performance all vary significantly. For example, under sudden load conditions, the wire rope tension will rise sharply in a very short time. If the control system fails to respond in time, it is very easy to cause the wire rope to break due to overload or the winch to stall. Under strong wind conditions, the wind load generates an additional torque on the winch output shaft, causing the actual deployment and recovery speed to deviate from the set value. Under low-temperature conditions, the sealing performance of pneumatic components decreases, and the elastic modulus of the wire rope material changes, both of which affect the accuracy and safety of deployment and recovery control.

[0004] In existing technologies, the deployment and retraction control of pneumatic winches mostly adopts a fixed parameter control strategy. That is, a fixed set of deployment and retraction speeds, tension thresholds, and air source pressure values ​​are preset according to a single operating condition and remain unchanged throughout the entire deployment and retraction process. Although this control method is simple to implement, it cannot adapt to dynamic changes in operating conditions. When the actual operating conditions do not match the preset conditions, there is a large deviation between the control parameters and the actual requirements. This can lead to reduced deployment and retraction efficiency or even safety accidents.

[0005] Some existing technologies attempt to introduce simple threshold judgment mechanisms, reducing the take-up and release speed or shutting down for protection when the tension exceeds a certain fixed threshold. However, this judgment method based on a single parameter threshold has obvious shortcomings: on the one hand, a fixed threshold cannot distinguish between the reasonable range and abnormal range of tension changes under different operating conditions. Normal tension fluctuations under rated load take-up and release conditions may be misjudged as sudden load increases, leading to unnecessary speed reduction or even shutdown; on the other hand, the single-parameter threshold judgment ignores the coupling relationship between multiple sources of information such as air source pressure, take-up and release speed, ambient temperature, and wind speed, and cannot accurately identify the current actual operating condition category, thus failing to retrieve control parameters that truly match the current operating condition.

[0006] Furthermore, existing technologies suffer from different sampling frequencies and inconsistent time bases among sensors, making it impossible to directly align multi-source data in the time dimension and hindering effective joint analysis. Simultaneously, existing technologies lack systematic preprocessing methods for raw sensor data, such as zero drift subtraction, cumulative error correction, temperature hysteresis correction, and gust factor decomposition. Noise and systematic errors embedded in the raw data directly affect the accuracy of operating condition identification and the reliability of control parameter retrieval.

[0007] In terms of dynamic adjustment of control parameters, existing technologies mostly adopt open-loop correction methods, that is, after detecting that the parameter exceeds the limit, they directly switch to another set of preset parameters. They lack the convergence judgment of the correction effect and the multi-parameter weighted fusion mechanism, which can easily cause frequent jitter of control commands and affect the stability of the pneumatic winch's raising and lowering process. Summary of the Invention

[0008] To address the aforementioned technical problems, a multi-scenario adaptive deployment and retraction system is provided. This technical solution resolves the issues raised in the background section.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A multi-scenario adaptive deployment and retraction system, comprising: Data Acquisition Module: Continuously acquires multi-source sensor data during the operation of the pneumatic winch. The multi-source sensor data includes air source pressure data, winch output shaft speed data, wire rope tension data, wire rope winding and unwinding length data, ambient temperature data, and wind speed data. Data processing module: performs timestamp alignment and sampling rate unification processing on the collected multi-source sensor data, mapping sensor data with different sampling frequencies to a unified time base to form a synchronized data sequence; Extraction module: Extracts multi-dimensional operating condition feature vectors from synchronous data sequences, including gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics. Judgment Module: The extracted multi-dimensional working condition feature vector is input into the pre-built working condition classification model to determine the working condition category and obtain the target working condition category of the current pneumatic winch. The target working condition category includes no-load retraction and extension, rated load retraction and extension, sudden load, high wind, and low temperature. Retrieval Module: Based on the determined target working condition category, retrieve the corresponding set of retraction and extension control parameters from the preset working condition-parameter mapping table. The set of retraction and extension control parameters includes the target retraction and extension speed, the target tension threshold, and the air source pressure compensation value. Correction module: Generates pneumatic winch retraction and deployment control commands based on the retrieved retraction and deployment control parameter set, and continuously collects real-time sensor data during the retraction and deployment process. The real-time sensor data is compared with the preset data range under the target working condition category. When the real-time sensor data exceeds the preset data range, dynamic correction of the control parameters is triggered.

[0010] Preferably, the continuous acquisition of multi-source sensor data during the operation of the pneumatic winch specifically includes: Air source pressure data is continuously acquired via a pressure sensor installed at the air inlet of the pneumatic winch at a sampling frequency of no less than 50Hz. The acquired raw pressure data is then processed by a sliding window mean filter, with the sliding window length set to one-fifth of the acquisition period. The filtered data is used as the valid air source pressure data. The calculation method for the sliding window mean filter is as follows: ; in, For a moment The filtered air source pressure value. For a moment The original gas source pressure value collected at the location, The length of the sliding window. Indicates rounding down; The winch output shaft speed data is acquired by an encoder at a sampling frequency of not less than 100Hz. The acquired raw speed data is then subjected to median filtering to eliminate pulse interference. The filter window length is set to three consecutive sampling points. The wire rope tension data is collected by a tension sensor installed at the wire rope guide pulley at a sampling frequency of not less than 20Hz. Zero drift subtraction is performed on the raw tension data. The zero drift value is obtained by averaging the tension data collected continuously for 30 seconds while the pneumatic winch is stationary. The calculation method for zero drift subtraction is as follows: ; in, For a moment The effective tension value after deducting zero drift. The original tension value was collected at time t. The average value of tension data collected continuously for 30 seconds while the object is stationary. The wire rope winding and unwinding length data is collected by a rotary encoder installed on the drum shaft end at a sampling frequency of not less than 10Hz. The original length data is cumulatively corrected by comparing the encoder reading after each winding and unwinding with the physical length of the wire rope at the marked point, calculating the cumulative deviation, and compensating for subsequent length data. The calculation method for the cumulative deviation is as follows: ; in, For the first The cumulative deviation after each placement and retraction. For the first The length value converted from the encoder reading during the next take-up and put-out operation. For the first The actual physical length of the wire rope marking point when it is in place during the second take-up and release; Ambient temperature data is collected by a temperature sensor installed on the surface of the pneumatic winch housing at a sampling frequency of not less than 1Hz. The raw temperature data undergoes hysteresis correction. The hysteresis correction amount is determined by the product of the temperature sensor's response time constant and the current rate of temperature change. The calculation method for hysteresis correction is as follows: ; in, For a moment The temperature value after hysteresis correction. For a moment The original temperature value collected at the location, Let be the response time constant of the temperature sensor. This represents the rate of temperature change of the original temperature data at time t; Wind speed data is collected using an ultrasonic anemometer mounted on top of the pneumatic winch at a sampling frequency of no less than 2Hz. The raw wind speed data is then decomposed into gust factor components, separating the wind speed data into average wind speed components and gust fluctuation components, which are stored separately. The calculation method for gust factor decomposition is as follows: ; in, For a moment The original wind speed value at that location, The average wind speed component, This refers to the pulsating component of the gust.

[0011] Preferably, the step of performing timestamp alignment and sampling rate unification processing on the collected multi-source sensor data, mapping sensor data with different sampling frequencies to a unified time base to form a synchronized data sequence, specifically includes: Using the sampling timestamp of the gas source pressure data as the reference time axis, the timestamps of the data from the other sensors are uniformly converted to this reference time axis; For sensor data with a sampling frequency higher than the reference sampling frequency, a linear interpolation method is used to downsample it to the reference sampling frequency. Specifically, at each time node on the reference time axis, the two closest original sampling points before and after that time node are taken, and the data value at that time node is calculated according to the weighted ratio of time distance. The linear interpolation calculation method is as follows: ; in, The time on the reference time axis The data value obtained by interpolation, and As the reference time node The two most recent original sampling times, and satisfying , and These are the original sampled values ​​at the corresponding times; For sensor data with a sampling frequency lower than the reference sampling frequency, a cubic spline interpolation method is used to upsample it to the reference sampling frequency. Specifically, a cubic spline interpolation function is constructed using all existing sampling points of the sensor as nodes, and the corresponding data value is calculated by substituting the interpolation function at each time node on the reference time axis. After completing the timestamp alignment and sampling rate unification of all sensor data, the air source pressure data, winch output shaft speed data, wire rope tension data, wire rope winding and unwinding length data, ambient temperature data, and wind speed data at the same time node are spliced ​​together in a fixed order to form the synchronous data tuple corresponding to that time node. All synchronous data tuples within a continuous time window are arranged in chronological order to form a synchronous data sequence. The length of the continuous time window is set to the minimum feature extraction time required by the working condition classification model.

[0012] Preferably, the step of extracting a multi-dimensional operating condition feature vector from the synchronous data sequence, including gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics, specifically includes: A data segment of the same length as the continuous time window is extracted from the synchronous data sequence. The pressure fluctuation characteristic value within the sliding window is calculated for the gas source pressure data in this data segment. The sliding window length is set to one-tenth of the continuous time window length, and the sliding step size is set to half the sliding window length. For each sliding window, the absolute value of the difference between the standard deviation of the pressure data and the standard deviation of adjacent sliding windows is calculated. The standard deviations of all sliding windows are arranged in order to form a gas source pressure fluctuation characteristic sub-vector. The absolute values ​​of the differences between the standard deviations of all adjacent sliding windows are arranged in order to form a gas source pressure fluctuation rate characteristic sub-vector. The two sub-vectors are concatenated to obtain the gas source pressure fluctuation characteristic. The calculation method for the standard deviation of the pressure data within the sliding window is as follows: ; in, For the first Standard deviation of gas source pressure within each sliding window This represents the number of sampling points within the sliding window. For the first Within the first sliding window The gas source pressure value at each sampling point For the first The average air pressure data within each sliding window; For the wire rope tension data in the synchronous data segment, calculate the tension difference between two adjacent time nodes in chronological order. Divide the tension difference by the corresponding time interval to obtain the tension change gradient sequence. Calculate the maximum, minimum, mean, and standard deviation of the absolute values ​​of the tension change gradient sequence. Arrange these four statistics in a fixed order to form the tension change gradient feature. The calculation method for the tension change gradient is as follows: ; in, For a moment The tension gradient value at that point, and They are time points and The wire rope tension value at that location; For the wire rope winding and unwinding length data in the synchronous data segment, calculate the length difference between two adjacent time nodes in chronological order. Divide the length difference by the corresponding time interval to obtain the winding and unwinding speed sequence. Calculate the first-order difference sequence of the winding and unwinding speed sequence, and calculate the maximum and mean of the absolute values ​​of the first-order difference sequence. Arrange these two statistics in a fixed order to form the winding and unwinding speed variation characteristics. The calculation method for the winding and unwinding speed is as follows: ; in, For a moment The speed at which the device is extended or retracted. and They are time points and The length of the wire rope taken up and down at the location; The mean and standard deviation of ambient temperature data and wind speed data in the synchronous data segment are calculated within a sliding window. The length of the sliding window is the same as the length of the continuous time window. The mean and standard deviation of the temperature data and the mean and standard deviation of the wind speed data are arranged in a fixed order to form the fluctuation characteristics of the environmental parameters. The gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics are spliced ​​together in a fixed order to form a multi-dimensional operating condition feature vector.

[0013] Preferably, the step of inputting the extracted multi-dimensional working condition feature vector into a pre-constructed working condition classification model for working condition category determination to obtain the target working condition category of the current pneumatic winch specifically includes: The multidimensional working condition feature vector is used as the input of the working condition classification model. The working condition classification model is a multi-classification model trained based on the decision tree ensemble algorithm. The training samples of the model are the multidimensional working condition feature vector corresponding to historical working condition data and the manually labeled working condition category labels. The working condition classification model makes a layer-by-layer judgment on the input multi-dimensional working condition feature vector. First, it determines whether the standard deviation in the gas source pressure fluctuation feature is lower than the first preset threshold and whether the maximum value in the tension change gradient feature is lower than the second preset threshold. If both conditions are met, it is determined to be an unloaded retraction and deployment working condition. If the standard deviation of the air source pressure fluctuation characteristics is between the first preset threshold and the third preset threshold, the mean of the tension change gradient characteristics is between the fourth preset threshold and the fifth preset threshold, and the maximum value of the retraction and extension speed change characteristics is between the sixth preset threshold and the seventh preset threshold, then it is determined to be the rated load retraction and extension condition. If the increment of the maximum value in the tension change gradient feature exceeds the eighth preset threshold within three consecutive sliding windows, and the maximum value in the gas source pressure fluctuation rate feature exceeds the ninth preset threshold, then it is determined to be a sudden load condition. The calculation method for the increment of the maximum value of the tension gradient within three consecutive sliding windows is as follows: ; in, This represents the increment of the maximum tension gradient within three consecutive sliding windows. This represents the maximum tension gradient value of the current sliding window. This represents the maximum tension gradient before the three sliding windows; If the mean wind speed in the environmental parameter fluctuation characteristics exceeds the tenth preset threshold and the standard deviation of the gust pulsation component of the wind speed data exceeds the eleventh preset threshold, it is determined to be a strong wind condition. If the average temperature in the environmental parameter fluctuation characteristics is lower than the twelfth preset threshold and the sliding window standard deviation of the temperature data is lower than the thirteenth preset threshold, it is determined to be a low temperature operating condition. If multiple conditions for determining different operating conditions are met simultaneously, the operating condition category with the highest confidence level among all conditions is selected as the target operating condition category. The confidence level is determined by the probability values ​​of each category output by the operating condition classification model. The calculation method for selecting the category with the highest confidence level is as follows: ; in, To determine the category of the target working condition, c For any category in the set of working condition categories, c 1 to c 5 These correspond to the no-load retraction and extension conditions, the rated load retraction and extension conditions, the sudden load condition, the high wind condition, and the low temperature condition, respectively. This is the probability value of the output category for the working condition classification model given the input multidimensional working condition feature vector F.

[0014] Preferably, the step of retrieving the corresponding set of retraction and extension control parameters from a preset working condition-parameter mapping table based on the determined target working condition category specifically includes: The preset working condition-parameter mapping table stores a set of take-up and release control parameters for each target working condition category. The target take-up and release speed in each set of control parameters is determined based on the quantiles of the historical statistical distribution of the wire rope take-up and release length data under that working condition category. Specifically, the 75th percentile of the historical take-up and release speed data under that working condition category is taken as the target take-up and release speed. The 75th percentile is calculated as follows: ; in, For the target retraction and extension speed, This represents the 75th percentile of the historical distribution of release and take-off speed data. This is the cumulative distribution function of historical launch and recovery speed data under this working condition category; The target tension threshold for each set of control parameters is determined based on the upper limit of the historical statistical distribution of wire rope tension data under that working condition category. Specifically, it is the 90th percentile of the historical tension data under that working condition category multiplied by a safety factor. The safety factor is set according to the hazard level of that working condition category: 1.1 for no-load winding and unwinding, 1.2 for rated load winding and unwinding, 1.3 for sudden load conditions, 1.25 for high wind conditions, and 1.15 for low temperature conditions. The target tension threshold is calculated as follows: ; in, The target tension threshold, This represents the 90th percentile of historical tension data for this operating condition category. The safety factor is the corresponding working condition category; The gas source pressure compensation value for each set of control parameters is determined based on the difference between the mean of the historical statistical distribution of gas source pressure data under that operating condition and the rated gas source pressure. Specifically, the rated gas source pressure is subtracted from the mean of the historical gas source pressure data under that operating condition, and the resulting difference is used as the gas source pressure compensation value. The calculation method for the gas source pressure compensation value is as follows: ; in, This is the gas source pressure compensation value. The rated air supply pressure of the pneumatic winch. This is the average of historical gas source pressure data under this operating condition category; When the target working condition is determined to be a sudden load condition, in addition to retrieving the set of retraction and deployment control parameters corresponding to this working condition, the tension buffer coefficient and speed attenuation coefficient are additionally retrieved from the parameter correction sub-table specifically for sudden load conditions. The retrieved target retraction and deployment speed is multiplied by the speed attenuation coefficient to obtain the corrected target retraction and deployment speed. The calculation method for the corrected target retraction and deployment speed is as follows: ; in, The corrected target release and retraction speed. A velocity attenuation coefficient specifically designed for sudden load conditions; The corrected target tension threshold is obtained by multiplying the retrieved target tension threshold by the tension buffer coefficient. The calculation method for the corrected target tension threshold is as follows: ; in, The corrected target tension threshold, Tension buffer coefficient specifically designed for sudden load conditions.

[0015] Preferably, the pneumatic winch's retraction and deployment control commands are generated based on the retrieved set of retraction and deployment control parameters. Real-time sensor data is continuously collected during the retraction and deployment process, and the real-time sensor data is compared with a preset data range under the target operating condition category. When the real-time sensor data exceeds the preset data range, dynamic correction of the control parameters is triggered, specifically including: The target retraction speed in the set of retraction and extension control parameters is used as the speed command reference value, the target tension threshold is used as the tension command upper limit value, and the air source pressure compensation value is used as the air source adjustment reference value. The three are combined to generate the retraction and extension control command of the pneumatic winch and send it to the actuator. During the execution of the take-up and release control command, real-time sensor data is continuously collected using the same time base as the synchronous data sequence in the steps. The statistical values ​​within the current sliding window are calculated for the real-time collected air source pressure data, wire rope tension data, and winch output shaft speed data. The length of the sliding window is consistent with the length of the sliding window during feature extraction. The current sliding window statistical value of the real-time gas source pressure data is compared with the preset allowable fluctuation range of the gas source pressure under the target working condition category. The allowable fluctuation range is represented by the gas source pressure compensation value under the working condition category as the center and the range of fluctuation above and below the fifteenth preset threshold. The current sliding window statistical value of the real-time wire rope tension data is compared with the target tension threshold under the target working condition category; The real-time winding and unwinding speed, calculated from the real-time winch output shaft speed data, is compared with the target winding and unwinding speed under the target working condition category. When the current sliding window statistical value of the real-time air source pressure data exceeds the allowable fluctuation range of the air source pressure, the ratio of the excess to the boundary value of the allowable fluctuation range is used as the pressure correction coefficient. The air source pressure compensation value is multiplied by the pressure correction coefficient to obtain the corrected air source pressure compensation value. The original air source pressure compensation value is then replaced with the corrected air source pressure compensation value to regenerate the retraction and release control command. The pressure correction coefficient is calculated as follows: ; in, This is the pressure correction factor. This is the statistical value of the current sliding window for real-time gas source pressure data. This represents the upper limit of the allowable fluctuation range of the gas source pressure. This is the lower limit of the allowable fluctuation range of the gas source pressure; When the current sliding window statistical value of the real-time wire rope tension data exceeds the target tension threshold, the ratio of the excess to the target tension threshold is used as the tension correction coefficient. The target take-up and release speed is multiplied by the tension correction coefficient to obtain the corrected target take-up and release speed. The corrected target take-up and release speed replaces the original target take-up and release speed to regenerate the take-up and release control command. The tension correction coefficient is calculated as follows: ; in, This is the tension correction factor. This is the statistical value of the current sliding window for real-time wire rope tension data. The target tension threshold; When the deviation between the real-time deployment and recovery speed and the target deployment and recovery speed exceeds the sixteenth preset threshold, the ratio of the deviation value to the target deployment and recovery speed is used as a speed correction coefficient to superimpose and correct the target deployment and recovery speed. The correction direction is to reduce the deviation. The speed correction coefficient is calculated as follows: ; in, For speed correction factor, For real-time deployment and retraction speed, The target is the speed of expansion and contraction.

[0016] Preferably, the step of triggering dynamic correction of control parameters when real-time sensor data exceeds a preset data range further includes: After triggering dynamic correction of control parameters, the rate of change of each data item is calculated for the real-time sensor data of the last five consecutive time nodes before correction and the real-time sensor data of the first to fifth consecutive time nodes after correction. If the absolute value of the ratio of the rate of change of the corrected air source pressure data to the rate of change of the air source pressure data before correction is less than the seventeenth preset threshold, and the absolute value of the ratio of the rate of change of the corrected wire rope tension data to the rate of change of the wire rope tension data before correction is less than the eighteenth preset threshold, and the absolute value of the ratio of the rate of change of the corrected reeling-out speed to the rate of change of the reeling-out speed before correction is less than the nineteenth preset threshold, then the dynamic correction is deemed effective, and the corrected reeling-out control parameter set continues to be executed. The calculation method for the rate of change ratio is as follows: ; in, The ratio of the rate of change of data item x To correct the rate of change of this data item, The rate of change of this data item before correction; If the absolute value of any of the above three ratios is greater than or equal to the corresponding preset threshold, it is determined that the dynamic correction has not converged. The uncorrected and corrected control parameter sets are then weighted and merged according to preset weights. The weights are determined based on the deviation of the rate of change of each data item. The greater the deviation, the higher the weight of the original parameter set corresponding to the data item. The merged parameter set is used as the new control parameter set to regenerate the control command. The weighted fusion calculation method is as follows: ; in, This is the fused set of control parameters for expansion and contraction. For data items x The corresponding original parameter set or modified parameter set, For data items x The corresponding weights, and satisfying Weight Based on the ratio of the rate of change of each data item Sure, The larger The closer it is to the original parameter set; If the dynamic correction fails to converge after three consecutive attempts, the target working condition category will be forcibly re-entered into the working condition classification model for a second determination. The current control parameter set will be replaced by the set of control parameters corresponding to the second determination result.

[0017] Preferably, the step of comparing real-time sensor data with a preset data range under the target operating condition category further includes: For cases where the target working condition is a sudden load condition, the wire rope tension data range in the preset data range is set with the target tension threshold as the upper limit and 60% of the target tension threshold as the lower limit. The real-time wire rope tension data is compared by comparing the maximum value of the tension data in the current sliding window with the upper limit and the minimum value with the lower limit. When the maximum value exceeds the upper limit or the minimum value is lower than the lower limit, dynamic correction of the control parameters is triggered. For cases where the target operating condition is a strong wind condition, the wind speed data range in the preset data range is capped at the 95th percentile of the historical wind speed data under this operating condition. The real-time wind speed data is compared by comparing the peak value of the gust pulsation component of the wind speed data in the current sliding window with the upper limit. When the peak value exceeds the upper limit, dynamic correction of the control parameters is triggered. At the same time, the air source pressure compensation value is increased by the twenty-first preset threshold to compensate for the influence of wind load on the output torque of the pneumatic winch. For targets operating under low-temperature conditions, the ambient temperature range within the preset data range is limited to the 5th percentile of historical temperature data for that condition. Real-time temperature data is compared by taking the mean of the current sliding window and comparing it to the lower limit. When the mean is lower than the lower limit, the target launch / retract speed is multiplied by a low-temperature speed correction factor. This correction factor is calculated linearly based on the difference between the current temperature and the lower limit temperature; the larger the temperature difference, the smaller the correction factor. The calculation method for the low-temperature speed correction factor is as follows: ; in, This is a low-temperature speed correction factor. This is the speed correction coefficient constant under low-temperature operating conditions. For low-temperature operating conditions, preset the lower limit of the data range. This represents the average value of the current sliding window of real-time ambient temperature data. Simultaneously, the target tension threshold is multiplied by a low-temperature tension correction factor. This low-temperature tension correction factor is determined based on the proportion of change in the elastic modulus of the wire rope material at the current temperature. The calculation method for the low-temperature tension correction factor is as follows: ; in, This is the low-temperature tension correction factor. For the wire rope material at the current real-time temperature The elastic modulus below, For wire rope material at reference temperature The elastic modulus below.

[0018] Compared with the prior art, the present invention provides a multi-scenario adaptive deployment and retraction system, which has the following beneficial effects: This invention solves the problem of direct temporal alignment of multi-sensor data by performing timestamp alignment and unified sampling rate processing on multi-source sensor data such as air source pressure, winch output shaft speed, wire rope tension, retraction / deployment length, ambient temperature, and wind speed, providing a reliable data foundation for subsequent joint analysis. By extracting multi-dimensional operating condition feature vectors such as air source pressure fluctuations, tension change gradients, retraction / deployment speed changes, and environmental parameter fluctuations from synchronous data sequences and inputting them into an operating condition classification model for judgment, it can accurately identify five operating condition categories: no-load retraction / deployment, rated load retraction / deployment, sudden load application, high wind, and low temperature, avoiding the high misjudgment rate problem of existing fixed threshold judgment methods. Based on the judgment results, the corresponding retraction / deployment control parameter set is retrieved from the operating condition-parameter mapping table, achieving adaptive matching between control parameters and actual operating conditions. Furthermore, during the retraction / deployment process, dynamic correction of control parameters is triggered by continuous comparison of real-time data with preset ranges, and a correction convergence judgment and multi-parameter weighted fusion mechanism are introduced to effectively avoid frequent jitter in control commands, ensuring the stability and safety of the retraction / deployment process. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the system module framework of the present invention; Figure 2 This is a schematic diagram of the method flow for S201-S206 in this invention; Figure 3 This is a schematic diagram of the method flow for S301-S305 in this invention; Figure 4 This is a schematic diagram of the method flow for S401-S407 in this invention. Detailed Implementation

[0020] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0021] Example 1 Please refer to Figure 1 As shown, a multi-scenario adaptive deployment and retraction system includes: Data Acquisition Module: Continuously acquires multi-source sensor data during the operation of the pneumatic winch. The multi-source sensor data includes air source pressure data, winch output shaft speed data, wire rope tension data, wire rope winding and unwinding length data, ambient temperature data, and wind speed data. Data processing module: performs timestamp alignment and sampling rate unification processing on the collected multi-source sensor data, mapping sensor data with different sampling frequencies to a unified time base to form a synchronized data sequence; Extraction module: Extracts multi-dimensional operating condition feature vectors from synchronous data sequences, including gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics. Judgment Module: The extracted multi-dimensional working condition feature vector is input into the pre-built working condition classification model to determine the working condition category and obtain the target working condition category of the current pneumatic winch. The target working condition category includes no-load retraction and extension, rated load retraction and extension, sudden load, high wind, and low temperature. Retrieval Module: Based on the determined target working condition category, retrieve the corresponding set of retraction and extension control parameters from the preset working condition-parameter mapping table. The set of retraction and extension control parameters includes the target retraction and extension speed, the target tension threshold, and the air source pressure compensation value. Correction module: Generates pneumatic winch retraction and deployment control commands based on the retrieved retraction and deployment control parameter set, and continuously collects real-time sensor data during the retraction and deployment process. The real-time sensor data is compared with the preset data range under the target working condition category. When the real-time sensor data exceeds the preset data range, dynamic correction of the control parameters is triggered.

[0022] As will be understood by those skilled in the art, this invention solves the problem of the inability to directly align multi-sensor data in the time dimension by performing timestamp alignment and unified sampling rate processing on multi-source sensor data such as air source pressure, winch output shaft speed, wire rope tension, retraction length, ambient temperature, and wind speed, thus providing a reliable data foundation for subsequent joint analysis. By extracting multi-dimensional operating condition feature vectors such as air source pressure fluctuations, tension change gradients, retraction speed changes, and environmental parameter fluctuations from synchronous data sequences and inputting them into the operating condition classification model for judgment, it can accurately identify five operating condition categories: no-load retraction, rated load retraction, sudden load, strong wind, and low temperature, avoiding the high misjudgment rate problem of existing fixed threshold judgment methods. Based on the judgment results, the corresponding retraction control parameter set is retrieved from the operating condition-parameter mapping table, realizing adaptive matching between control parameters and actual operating conditions. Furthermore, during the retraction process, dynamic correction of control parameters is triggered by continuous comparison of real-time data with preset ranges, and a correction convergence judgment and multi-parameter weighted fusion mechanism are introduced to effectively avoid frequent jitter of control commands and ensure the stability and safety of the retraction process.

[0023] Please refer to Figure 2 As shown, multi-source sensor data is continuously collected during the operation of the pneumatic winch, specifically including: S201. Air source pressure data is continuously acquired by a pressure sensor installed at the air inlet of the pneumatic winch at a sampling frequency of not less than 50Hz. The acquired raw pressure data is then processed by a sliding window mean filter. The length of the sliding window is set to one-fifth of the acquisition period. The filtered data is used as the valid air source pressure data. The calculation method for the sliding window mean filter is as follows: ; in, For a moment The filtered air source pressure value. For a moment The original gas source pressure value collected at the location, The length of the sliding window. Indicates rounding down; S202. The winch output shaft speed data is collected by the encoder at a sampling frequency of not less than 100Hz. The collected raw speed data is subjected to median filtering to eliminate pulse interference. The filtering window length is set to three consecutive sampling points. S203. Wire rope tension data is collected by a tension sensor installed at the wire rope guide pulley at a sampling frequency of not less than 20Hz. The raw tension data is processed by zero drift subtraction. The zero drift value is obtained by averaging the tension data collected continuously for 30 seconds while the pneumatic winch is stationary. The calculation method for zero drift subtraction is as follows: ; in, For a moment The effective tension value after deducting zero drift. The original tension value was collected at time t. The average value of tension data collected continuously for 30 seconds while the object is stationary. S204. The wire rope winding and unwinding length data is collected by a rotary encoder installed on the drum shaft end at a sampling frequency of not less than 10Hz. The original length data is cumulatively corrected. The correction method involves comparing the encoder reading after each winding and unwinding with the physical length of the wire rope at the marked point, calculating the cumulative deviation, and compensating for subsequent length data. The calculation method for the cumulative deviation is as follows: ; in, For the first The cumulative deviation after each placement and retraction. For the first The length value converted from the encoder reading during the next take-up and put-out operation. For the first The actual physical length of the wire rope marking point when it is in place during the second take-up and release; S205. Ambient temperature data is acquired by a temperature sensor installed on the surface of the pneumatic winch housing at a sampling frequency of not less than 1Hz. The raw temperature data undergoes hysteresis correction. The hysteresis correction amount is determined by the product of the temperature sensor's response time constant and the current rate of temperature change. The calculation method for hysteresis correction is as follows: ; in, For a moment The temperature value after hysteresis correction. For a moment The original temperature value collected at the location, Let be the response time constant of the temperature sensor. This represents the rate of temperature change of the original temperature data at time t; S206. Wind speed data is collected using an ultrasonic anemometer installed on top of the pneumatic winch at a sampling frequency of not less than 2Hz. The raw wind speed data is decomposed into gust factor components, separating the wind speed data into average wind speed components and gust fluctuation components, which are stored separately. The calculation method for gust factor decomposition is as follows: ; in, For a moment The original wind speed value at that location, The average wind speed component, This refers to the pulsating component of the gust.

[0024] Please refer to Figure 3 As shown, the collected multi-source sensor data undergoes timestamp alignment and sampling rate unification processing, mapping sensor data with different sampling frequencies to a unified time base to form a synchronized data sequence. Specifically, this includes: S301. Using the sampling timestamp of the gas source pressure data as the reference time axis, convert the timestamps of the data from the other sensors to the reference time axis. S302. For sensor data with a sampling frequency higher than the reference sampling frequency, a linear interpolation method is used to downsample it to the reference sampling frequency. Specifically, at each time node on the reference time axis, the two closest original sampling points before and after that time node are taken, and the data value at that time node is calculated according to the weighted ratio of time distance. The linear interpolation calculation method is as follows: ; in, The data values ​​obtained by interpolation at a given time point on the baseline time axis. and These are the two most recent original sampling times before and after the reference time node, and satisfy the following conditions: , and These are the original sampled values ​​at the corresponding times; S303. For sensor data with a sampling frequency lower than the reference sampling frequency, a cubic spline interpolation method is used to upsample it to the reference sampling frequency. Specifically, a cubic spline interpolation function is constructed using all existing sampling points of the sensor as nodes, and the corresponding data value is calculated by substituting the interpolation function at each time node on the reference time axis. S304. After completing the timestamp alignment and sampling rate unification of all sensor data, the air source pressure data, winch output shaft speed data, wire rope tension data, wire rope winding and unwinding length data, ambient temperature data and wind speed data at the same time node are spliced ​​together in a fixed order to form the synchronous data tuple corresponding to that time node. S305. Arrange all synchronous data tuples within the continuous time window in chronological order to form a synchronous data sequence. Set the length of the continuous time window to the shortest feature extraction time required by the working condition classification model.

[0025] A multi-dimensional operating condition feature vector containing gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics is extracted from the synchronous data sequence. Specifically, it includes: A data segment of the same length as the continuous time window is extracted from the synchronous data sequence. The pressure fluctuation characteristic value within the sliding window is calculated for the gas source pressure data in this data segment. The sliding window length is set to one-tenth of the continuous time window length, and the sliding step size is set to half the sliding window length. For each sliding window, the absolute value of the difference between the standard deviation of the pressure data and the standard deviation of adjacent sliding windows is calculated. The standard deviations of all sliding windows are arranged in order to form a gas source pressure fluctuation characteristic sub-vector. The absolute values ​​of the differences between the standard deviations of all adjacent sliding windows are arranged in order to form a gas source pressure fluctuation rate characteristic sub-vector. The two sub-vectors are concatenated to obtain the gas source pressure fluctuation characteristic. The calculation method for the standard deviation of the pressure data within the sliding window is as follows: ; in, For the first Standard deviation of gas source pressure within each sliding window This represents the number of sampling points within the sliding window. For the first Within the first sliding window The gas source pressure value at each sampling point For the first The average air pressure data within each sliding window; For the wire rope tension data in the synchronous data segment, calculate the tension difference between two adjacent time nodes in chronological order. Divide the tension difference by the corresponding time interval to obtain the tension change gradient sequence. Calculate the maximum, minimum, mean, and standard deviation of the absolute values ​​of the tension change gradient sequence. Arrange these four statistics in a fixed order to form the tension change gradient feature. The calculation method for the tension change gradient is as follows: ; in, For a moment The tension gradient value at that point, and They are time points and The wire rope tension value at that location; For the wire rope winding and unwinding length data in the synchronous data segment, calculate the length difference between two adjacent time nodes in chronological order. Divide the length difference by the corresponding time interval to obtain the winding and unwinding speed sequence. Calculate the first-order difference sequence of the winding and unwinding speed sequence, and calculate the maximum and mean of the absolute values ​​of the first-order difference sequence. Arrange these two statistics in a fixed order to form the winding and unwinding speed variation characteristics. The calculation method for the winding and unwinding speed is as follows: ; in, For a moment The speed at which the device is extended or retracted. and They are time points and The length of the wire rope taken up and down at the location; The mean and standard deviation of ambient temperature data and wind speed data in the synchronous data segment are calculated within a sliding window. The length of the sliding window is the same as the length of the continuous time window. The mean and standard deviation of the temperature data and the mean and standard deviation of the wind speed data are arranged in a fixed order to form the fluctuation characteristics of the environmental parameters. The gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics are spliced ​​together in a fixed order to form a multi-dimensional operating condition feature vector.

[0026] Please refer to Figure 4 As shown, the extracted multi-dimensional working condition feature vector is input into a pre-built working condition classification model for working condition category determination, thereby obtaining the target working condition category of the current pneumatic winch, specifically including: S401. Use the multi-dimensional working condition feature vector as the input of the working condition classification model. The working condition classification model is a multi-classification model trained based on the decision tree ensemble algorithm. The training samples of the model are the multi-dimensional working condition feature vector corresponding to the historical working condition data and the manually labeled working condition category labels. S402. The working condition classification model makes a layer-by-layer judgment on the input multi-dimensional working condition feature vector. First, it judges whether the standard deviation in the gas source pressure fluctuation feature is lower than the first preset threshold and whether the maximum value in the tension change gradient feature is lower than the second preset threshold. If both conditions are met at the same time, it is judged as an unloaded retraction and deployment working condition. S403. If the standard deviation of the air source pressure fluctuation characteristics is between the first preset threshold and the third preset threshold, the mean of the tension change gradient characteristics is between the fourth preset threshold and the fifth preset threshold, and the maximum value of the retraction and extension speed change characteristics is between the sixth preset threshold and the seventh preset threshold, then it is determined to be the rated load retraction and extension condition. S404. If the increment of the maximum value in the tension change gradient feature exceeds the eighth preset threshold within three consecutive sliding windows, and the maximum value in the gas source pressure fluctuation rate feature exceeds the ninth preset threshold, then it is determined to be a sudden load condition. The calculation method for the increment of the maximum value of the tension gradient within three consecutive sliding windows is as follows: ; in, This represents the increment of the maximum tension gradient within three consecutive sliding windows. This represents the maximum tension gradient value of the current sliding window. This represents the maximum tension gradient before the three sliding windows; S405. If the average wind speed in the environmental parameter fluctuation characteristics exceeds the tenth preset threshold and the standard deviation of the gust pulsation component of the wind speed data exceeds the eleventh preset threshold, it is determined to be a strong wind condition. S406. If the average temperature in the environmental parameter fluctuation characteristics is lower than the twelfth preset threshold and the sliding window standard deviation of the temperature data is lower than the thirteenth preset threshold, it is determined to be a low temperature working condition. S407. If multiple conditions for determining different operating conditions are met simultaneously, the operating condition category with the highest confidence level among all conditions shall be selected as the target operating condition category. The confidence level is determined by the probability values ​​of each category output by the operating condition classification model. The calculation method for selecting the category with the highest confidence level is as follows: ; in, To determine the category of the target working condition, c For any category in the set of working condition categories, c 1 to c 5 These correspond to the no-load retraction and extension conditions, the rated load retraction and extension conditions, the sudden load condition, the high wind condition, and the low temperature condition, respectively. This is the probability value of the output category for the working condition classification model given the input multidimensional working condition feature vector F.

[0027] Based on the determined target operating condition category, the corresponding set of retraction and extension control parameters is retrieved from the preset operating condition-parameter mapping table, specifically including: The preset working condition-parameter mapping table stores a set of take-up and release control parameters for each target working condition category. The target take-up and release speed in each set of control parameters is determined based on the quantiles of the historical statistical distribution of the wire rope take-up and release length data under that working condition category. Specifically, the 75th percentile of the historical take-up and release speed data under that working condition category is taken as the target take-up and release speed. The 75th percentile is calculated as follows: ; in, For the target retraction and extension speed, This represents the 75th percentile of the historical distribution of release and take-off speed data. This is the cumulative distribution function of historical launch and recovery speed data under this working condition category; The target tension threshold for each set of control parameters is determined based on the upper limit of the historical statistical distribution of wire rope tension data under that working condition category. Specifically, it is the 90th percentile of the historical tension data under that working condition category multiplied by a safety factor. The safety factor is set according to the hazard level of that working condition category: 1.1 for no-load winding and unwinding, 1.2 for rated load winding and unwinding, 1.3 for sudden load conditions, 1.25 for high wind conditions, and 1.15 for low temperature conditions. The target tension threshold is calculated as follows: ; in, The target tension threshold, This represents the 90th percentile of historical tension data for this operating condition category. The safety factor is the corresponding working condition category; The gas source pressure compensation value for each set of control parameters is determined based on the difference between the mean of the historical statistical distribution of gas source pressure data under that operating condition and the rated gas source pressure. Specifically, the rated gas source pressure is subtracted from the mean of the historical gas source pressure data under that operating condition, and the resulting difference is used as the gas source pressure compensation value. The calculation method for the gas source pressure compensation value is as follows: ; in, This is the gas source pressure compensation value. The rated air supply pressure of the pneumatic winch. This is the average of historical gas source pressure data under this operating condition category; When the target working condition is determined to be a sudden load condition, in addition to retrieving the set of retraction and deployment control parameters corresponding to this working condition, the tension buffer coefficient and speed attenuation coefficient are additionally retrieved from the parameter correction sub-table specifically for sudden load conditions. The retrieved target retraction and deployment speed is multiplied by the speed attenuation coefficient to obtain the corrected target retraction and deployment speed. The calculation method for the corrected target retraction and deployment speed is as follows: ; in, The corrected target release and retraction speed. A velocity attenuation coefficient specifically designed for sudden load conditions; The corrected target tension threshold is obtained by multiplying the retrieved target tension threshold by the tension buffer coefficient. The calculation method for the corrected target tension threshold is as follows: ; in, The corrected target tension threshold, Tension buffer coefficient specifically designed for sudden load conditions.

[0028] Based on the retrieved set of retraction and deployment control parameters, the pneumatic winch's retraction and deployment control commands are generated. During the retraction and deployment process, real-time sensor data is continuously collected and compared with a preset data range for the target operating condition. When the real-time sensor data exceeds the preset data range, dynamic correction of the control parameters is triggered, specifically including: The target retraction speed in the set of retraction and extension control parameters is used as the speed command reference value, the target tension threshold is used as the tension command upper limit value, and the air source pressure compensation value is used as the air source adjustment reference value. The three are combined to generate the retraction and extension control command of the pneumatic winch and send it to the actuator. During the execution of the take-up and release control command, real-time sensor data is continuously collected using the same time base as the synchronous data sequence in the steps. The statistical values ​​within the current sliding window are calculated for the real-time collected air source pressure data, wire rope tension data, and winch output shaft speed data. The length of the sliding window is consistent with the length of the sliding window during feature extraction. The current sliding window statistical value of the real-time gas source pressure data is compared with the preset allowable fluctuation range of the gas source pressure under the target working condition category. The allowable fluctuation range is represented by the gas source pressure compensation value under this working condition category, which fluctuates up and down by the fifteenth preset threshold. The current sliding window statistical value of the real-time wire rope tension data is compared with the target tension threshold under the target working condition category; The real-time winding and unwinding speed, calculated from the real-time winch output shaft speed data, is compared with the target winding and unwinding speed under the target working condition category. When the current sliding window statistical value of the real-time air source pressure data exceeds the allowable fluctuation range of the air source pressure, the ratio of the excess to the boundary value of the allowable fluctuation range is used as the pressure correction coefficient. The air source pressure compensation value is multiplied by the pressure correction coefficient to obtain the corrected air source pressure compensation value. The original air source pressure compensation value is then replaced with the corrected air source pressure compensation value to regenerate the retraction and release control command. The pressure correction coefficient is calculated as follows: ; in, This is the pressure correction factor. This is the statistical value of the current sliding window for real-time gas source pressure data. This represents the upper limit of the allowable fluctuation range of the gas source pressure. This is the lower limit of the allowable fluctuation range of the gas source pressure; When the current sliding window statistical value of the real-time wire rope tension data exceeds the target tension threshold, the ratio of the excess to the target tension threshold is used as the tension correction coefficient. The target take-up and release speed is multiplied by the tension correction coefficient to obtain the corrected target take-up and release speed. The corrected target take-up and release speed replaces the original target take-up and release speed to regenerate the take-up and release control command. The tension correction coefficient is calculated as follows: ; in, This is the tension correction factor. This is the statistical value of the current sliding window for real-time wire rope tension data. The target tension threshold; When the deviation between the real-time deployment and recovery speed and the target deployment and recovery speed exceeds the sixteenth preset threshold, the ratio of the deviation value to the target deployment and recovery speed is used as a speed correction coefficient to superimpose and correct the target deployment and recovery speed. The correction direction is to reduce the deviation. The speed correction coefficient is calculated as follows: ; in, For speed correction factor, For real-time deployment and retraction speed, The target is the speed of expansion and contraction.

[0029] When real-time sensor data exceeds the preset data range, dynamic correction of control parameters is triggered, including: After triggering dynamic correction of control parameters, the rate of change of each data item is calculated for the real-time sensor data of the last five consecutive time nodes before correction and the real-time sensor data of the first to fifth consecutive time nodes after correction. If the absolute value of the ratio of the rate of change of the corrected air source pressure data to the rate of change of the air source pressure data before correction is less than the seventeenth preset threshold, and the absolute value of the ratio of the rate of change of the corrected wire rope tension data to the rate of change of the wire rope tension data before correction is less than the eighteenth preset threshold, and the absolute value of the ratio of the rate of change of the corrected reeling-out speed to the rate of change of the reeling-out speed before correction is less than the nineteenth preset threshold, then the dynamic correction is deemed effective, and the corrected reeling-out control parameter set continues to be executed. The calculation method for the rate of change ratio is as follows: ; in, The ratio of the rate of change of data item x To correct the rate of change of this data item, The rate of change of this data item before correction; If the absolute value of any of the above three ratios is greater than or equal to the corresponding preset threshold, it is determined that the dynamic correction has not converged. The uncorrected and corrected control parameter sets are then weighted and merged according to preset weights. The weights are determined based on the deviation of the rate of change of each data item. The greater the deviation, the higher the weight of the original parameter set corresponding to the data item. The merged parameter set is used as the new control parameter set to regenerate the control command. The weighted fusion calculation method is as follows: ; in, This is the fused set of control parameters for expansion and contraction. For data items x The corresponding original parameter set or modified parameter set, For data items x The corresponding weights, and satisfying Weight Based on the ratio of the rate of change of each data item Sure, The larger The closer it is to the original parameter set; If the dynamic correction fails to converge after three consecutive attempts, the target working condition category will be forcibly re-entered into the working condition classification model for a second determination. The current control parameter set will be replaced by the set of control parameters corresponding to the second determination result.

[0030] The comparison of real-time sensor data with a preset data range under the target operating condition category also includes: For cases where the target working condition is a sudden load condition, the wire rope tension data range in the preset data range is set with the target tension threshold as the upper limit and 60% of the target tension threshold as the lower limit. The real-time wire rope tension data is compared by comparing the maximum value of the tension data in the current sliding window with the upper limit and the minimum value with the lower limit. When the maximum value exceeds the upper limit or the minimum value is lower than the lower limit, dynamic correction of the control parameters is triggered. For cases where the target operating condition is a strong wind condition, the wind speed data range in the preset data range is capped at the 95th percentile of the historical wind speed data under this operating condition. The real-time wind speed data is compared by comparing the peak value of the gust pulsation component of the wind speed data in the current sliding window with the upper limit. When the peak value exceeds the upper limit, dynamic correction of the control parameters is triggered. At the same time, the air source pressure compensation value is increased by the twenty-first preset threshold to compensate for the influence of wind load on the output torque of the pneumatic winch. For targets operating under low-temperature conditions, the ambient temperature range within the preset data range is limited to the 5th percentile of historical temperature data for that condition. Real-time temperature data is compared by taking the mean of the current sliding window and comparing it to the lower limit. When the mean is lower than the lower limit, the target launch / retract speed is multiplied by a low-temperature speed correction factor. This correction factor is calculated linearly based on the difference between the current temperature and the lower limit temperature; the larger the temperature difference, the smaller the correction factor. The calculation method for the low-temperature speed correction factor is as follows: ; in, This is a low-temperature speed correction factor. This is the speed correction coefficient constant under low-temperature operating conditions. For low-temperature operating conditions, preset the lower limit of the data range. This represents the average value of the current sliding window of real-time ambient temperature data. Simultaneously, the target tension threshold is multiplied by a low-temperature tension correction factor. This low-temperature tension correction factor is determined based on the proportion of change in the elastic modulus of the wire rope material at the current temperature. The calculation method for the low-temperature tension correction factor is as follows: ; in, This is the low-temperature tension correction factor. For the wire rope material at the current real-time temperature The elastic modulus below, For wire rope material at reference temperature The elastic modulus below.

[0031] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A multi-scenario adaptive deployment and retraction system, characterized in that, include: Data Acquisition Module: Continuously acquires multi-source sensor data during the operation of the pneumatic winch. The multi-source sensor data includes air source pressure data, winch output shaft speed data, wire rope tension data, wire rope winding and unwinding length data, ambient temperature data, and wind speed data. Data processing module: performs timestamp alignment and sampling rate unification processing on the collected multi-source sensor data, mapping sensor data with different sampling frequencies to a unified time base to form a synchronized data sequence; Extraction module: Extracts multi-dimensional operating condition feature vectors from synchronous data sequences, including gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics. Judgment Module: The extracted multi-dimensional working condition feature vector is input into the pre-built working condition classification model to determine the working condition category and obtain the target working condition category of the current pneumatic winch. The target working condition category includes no-load retraction and extension, rated load retraction and extension, sudden load, high wind, and low temperature. Retrieval Module: Based on the determined target working condition category, retrieve the corresponding set of retraction and extension control parameters from the preset working condition-parameter mapping table. The set of retraction and extension control parameters includes the target retraction and extension speed, the target tension threshold, and the air source pressure compensation value. Correction module: Generates pneumatic winch retraction and deployment control commands based on the retrieved retraction and deployment control parameter set, and continuously collects real-time sensor data during the retraction and deployment process. The real-time sensor data is compared with the preset data range under the target working condition category. When the real-time sensor data exceeds the preset data range, dynamic correction of the control parameters is triggered.

2. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 1, characterized in that, The continuous acquisition of multi-source sensor data during the operation of the pneumatic winch specifically includes: Air source pressure data is continuously acquired via a pressure sensor installed at the air inlet of the pneumatic winch at a sampling frequency of no less than 50Hz. The acquired raw pressure data is then processed by a sliding window mean filter, with the sliding window length set to one-fifth of the acquisition period. The filtered data is used as the valid air source pressure data. The calculation method for the sliding window mean filter is as follows: ； in, For a moment The filtered air source pressure value. For a moment The original gas source pressure value collected at the location, The length of the sliding window. Indicates rounding down; The winch output shaft speed data is acquired by an encoder at a sampling frequency of not less than 100Hz. The acquired raw speed data is then subjected to median filtering to eliminate pulse interference. The filter window length is set to three consecutive sampling points. The wire rope tension data is collected by a tension sensor installed at the wire rope guide pulley at a sampling frequency of not less than 20Hz. Zero drift subtraction is performed on the raw tension data. The zero drift value is obtained by averaging the tension data collected continuously for 30 seconds while the pneumatic winch is stationary. The calculation method for zero drift subtraction is as follows: ； in, For a moment The effective tension value after deducting zero drift. For a moment The original tension value collected at the location, The average value of tension data collected continuously for 30 seconds while the object is stationary. The wire rope winding and unwinding length data is collected by a rotary encoder installed on the drum shaft end at a sampling frequency of not less than 10Hz. The original length data is cumulatively corrected by comparing the encoder reading after each winding and unwinding with the physical length of the wire rope at the marked point, calculating the cumulative deviation, and compensating for subsequent length data. The calculation method for the cumulative deviation is as follows: ； in, For the first The cumulative deviation after each placement and retraction. For the first The length value converted from the encoder reading during the next take-up and put-out operation. For the first The actual physical length of the wire rope marking point when it is in place during the second take-up and release; Ambient temperature data is collected by a temperature sensor installed on the surface of the pneumatic winch housing at a sampling frequency of not less than 1Hz. The raw temperature data undergoes hysteresis correction. The hysteresis correction amount is determined by the product of the temperature sensor's response time constant and the current rate of temperature change. The calculation method for hysteresis correction is as follows: ； in, For a moment The temperature value after hysteresis correction. For a moment The original temperature value collected at the location, Let be the response time constant of the temperature sensor. For the raw temperature data at time The rate of temperature change at that location; Wind speed data is collected using an ultrasonic anemometer mounted on top of the pneumatic winch at a sampling frequency of no less than 2Hz. The raw wind speed data is then decomposed into gust factor components, separating the wind speed data into average wind speed components and gust fluctuation components, which are stored separately. The calculation method for gust factor decomposition is as follows: ； in, For a moment The original wind speed value at that location, The average wind speed component, This refers to the pulsating component of the gust.

3. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 2, characterized in that, The process of aligning timestamps and unifying sampling rates on the collected multi-source sensor data maps sensor data with different sampling frequencies to a unified time base, forming a synchronized data sequence. Specifically, this includes: Using the sampling timestamp of the gas source pressure data as the reference time axis, the timestamps of the data from the other sensors are uniformly converted to this reference time axis; For sensor data with a sampling frequency higher than the reference sampling frequency, a linear interpolation method is used to downsample it to the reference sampling frequency. Specifically, at each time node on the reference time axis, the two closest original sampling points before and after that time node are taken, and the data value at that time node is calculated according to the weighted ratio of time distance. The linear interpolation calculation method is as follows: ； in, The time on the reference time axis The data value obtained by interpolation, and As the reference time node The two most recent original sampling times, and satisfying , and These are the original sampled values ​​at the corresponding times; For sensor data with a sampling frequency lower than the reference sampling frequency, a cubic spline interpolation method is used to upsample it to the reference sampling frequency. Specifically, a cubic spline interpolation function is constructed using all existing sampling points of the sensor as nodes, and the corresponding data value is calculated by substituting the interpolation function at each time node on the reference time axis. After completing the timestamp alignment and sampling rate unification of all sensor data, the air source pressure data, winch output shaft speed data, wire rope tension data, wire rope winding and unwinding length data, ambient temperature data, and wind speed data at the same time node are spliced ​​together in a fixed order to form the synchronous data tuple corresponding to that time node. All synchronous data tuples within a continuous time window are arranged in chronological order to form a synchronous data sequence. The length of the continuous time window is set to the minimum feature extraction time required by the working condition classification model.

4. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 3, characterized in that, The extraction of a multi-dimensional operating condition feature vector from the synchronous data sequence, including gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics, specifically includes: A data segment of the same length as the continuous time window is extracted from the synchronous data sequence. The pressure fluctuation characteristic value within the sliding window is calculated for the gas source pressure data in this data segment. The sliding window length is set to one-tenth of the continuous time window length, and the sliding step size is set to half the sliding window length. For each sliding window, the absolute value of the difference between the standard deviation of the pressure data and the standard deviation of adjacent sliding windows is calculated. The standard deviations of all sliding windows are arranged in order to form a gas source pressure fluctuation characteristic sub-vector. The absolute values ​​of the differences between the standard deviations of all adjacent sliding windows are arranged in order to form a gas source pressure fluctuation rate characteristic sub-vector. The two sub-vectors are concatenated to obtain the gas source pressure fluctuation characteristic. The calculation method for the standard deviation of the pressure data within the sliding window is as follows: ； in, For the first Standard deviation of gas source pressure within each sliding window This represents the number of sampling points within the sliding window. For the first Within the first sliding window The gas source pressure value at each sampling point For the first The average air pressure data within each sliding window; For the wire rope tension data in the synchronous data segment, calculate the tension difference between two adjacent time nodes in chronological order. Divide the tension difference by the corresponding time interval to obtain the tension change gradient sequence. Calculate the maximum, minimum, mean, and standard deviation of the absolute values ​​of the tension change gradient sequence. Arrange these four statistics in a fixed order to form the tension change gradient feature. The calculation method for the tension change gradient is as follows: ； in, For a moment The tension gradient value at that point, and They are time points and The wire rope tension value at that location; For the wire rope winding and unwinding length data in the synchronous data segment, calculate the length difference between two adjacent time nodes in chronological order. Divide the length difference by the corresponding time interval to obtain the winding and unwinding speed sequence. Calculate the first-order difference sequence of the winding and unwinding speed sequence, and calculate the maximum and mean of the absolute values ​​of the first-order difference sequence. Arrange these two statistics in a fixed order to form the winding and unwinding speed variation characteristics. The calculation method for the winding and unwinding speed is as follows: ； in, For a moment The speed at which the device is extended or retracted. and They are time points and The length of the wire rope taken up and down at the location; The mean and standard deviation of ambient temperature data and wind speed data in the synchronous data segment are calculated within a sliding window. The length of the sliding window is the same as the length of the continuous time window. The mean and standard deviation of the temperature data and the mean and standard deviation of the wind speed data are arranged in a fixed order to form the fluctuation characteristics of the environmental parameters. The gas source pressure fluctuation characteristics, tension change gradient characteristics, retraction and expansion speed change characteristics, and environmental parameter fluctuation characteristics are spliced ​​together in a fixed order to form a multi-dimensional operating condition feature vector.

5. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 4, characterized in that, The step of inputting the extracted multi-dimensional working condition feature vector into a pre-constructed working condition classification model for working condition category determination, to obtain the target working condition category of the current pneumatic winch, specifically includes: The multidimensional working condition feature vector is used as the input of the working condition classification model. The working condition classification model is a multi-classification model trained based on the decision tree ensemble algorithm. The training samples of the model are the multidimensional working condition feature vector corresponding to historical working condition data and the manually labeled working condition category labels. The working condition classification model makes a layer-by-layer judgment on the input multi-dimensional working condition feature vector. First, it determines whether the standard deviation in the gas source pressure fluctuation feature is lower than the first preset threshold and whether the maximum value in the tension change gradient feature is lower than the second preset threshold. If both conditions are met, it is determined to be an unloaded retraction and deployment working condition. If the standard deviation of the air source pressure fluctuation characteristics is between the first preset threshold and the third preset threshold, the mean of the tension change gradient characteristics is between the fourth preset threshold and the fifth preset threshold, and the maximum value of the retraction and extension speed change characteristics is between the sixth preset threshold and the seventh preset threshold, then it is determined to be the rated load retraction and extension condition. If the increment of the maximum value in the tension change gradient feature exceeds the eighth preset threshold within three consecutive sliding windows, and the maximum value in the gas source pressure fluctuation rate feature exceeds the ninth preset threshold, then it is determined to be a sudden load condition. The calculation method for the increment of the maximum value of the tension gradient within three consecutive sliding windows is as follows: ； in, This represents the increment of the maximum tension gradient within three consecutive sliding windows. This represents the maximum tension gradient value of the current sliding window. This represents the maximum tension gradient before the three sliding windows; If the mean wind speed in the environmental parameter fluctuation characteristics exceeds the tenth preset threshold and the standard deviation of the gust pulsation component of the wind speed data exceeds the eleventh preset threshold, it is determined to be a strong wind condition. If the average temperature in the environmental parameter fluctuation characteristics is lower than the twelfth preset threshold and the sliding window standard deviation of the temperature data is lower than the thirteenth preset threshold, it is determined to be a low temperature operating condition. If multiple conditions for determining different operating conditions are met simultaneously, the operating condition category with the highest confidence level among all conditions is selected as the target operating condition category. The confidence level is determined by the probability values ​​of each category output by the operating condition classification model. The calculation method for selecting the category with the highest confidence level is as follows: ； in, To determine the category of the target working condition, For any category in the set of working condition categories, to These correspond to the no-load retraction and extension conditions, the rated load retraction and extension conditions, the sudden load condition, the high wind condition, and the low temperature condition, respectively. For the working condition classification model, input multi-dimensional working condition feature vector Output category under conditions The probability value.

6. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 5, characterized in that, The step of retrieving the corresponding set of retraction and extension control parameters from a preset working condition-parameter mapping table based on the determined target working condition category specifically includes: The preset working condition-parameter mapping table stores a set of take-up and release control parameters for each target working condition category. The target take-up and release speed in each set of control parameters is determined based on the quantiles of the historical statistical distribution of the wire rope take-up and release length data under that working condition category. Specifically, the 75th percentile of the historical take-up and release speed data under that working condition category is taken as the target take-up and release speed. The 75th percentile is calculated as follows: ； in, For the target retraction and extension speed, This represents the 75th percentile of the historical distribution of release and take-off speed data. This is the cumulative distribution function of historical launch and recovery speed data under this working condition category; The target tension threshold for each set of control parameters is determined based on the upper limit of the historical statistical distribution of wire rope tension data under that working condition category. Specifically, it is the 90th percentile of the historical tension data under that working condition category multiplied by a safety factor. The safety factor is set according to the hazard level of that working condition category: 1.1 for no-load winding and unwinding, 1.2 for rated load winding and unwinding, 1.3 for sudden load conditions, 1.25 for high wind conditions, and 1.15 for low temperature conditions. The target tension threshold is calculated as follows: ； in, The target tension threshold, This represents the 90th percentile of historical tension data for this operating condition category. The safety factor is the corresponding working condition category; The gas source pressure compensation value for each set of control parameters is determined based on the difference between the mean of the historical statistical distribution of gas source pressure data under that operating condition and the rated gas source pressure. Specifically, the rated gas source pressure is subtracted from the mean of the historical gas source pressure data under that operating condition, and the resulting difference is used as the gas source pressure compensation value. The calculation method for the gas source pressure compensation value is as follows: ； in, This is the gas source pressure compensation value. The rated air supply pressure of the pneumatic winch. This is the average of historical gas source pressure data under this operating condition category; When the target working condition is determined to be a sudden load condition, in addition to retrieving the set of retraction and deployment control parameters corresponding to this working condition, the tension buffer coefficient and speed attenuation coefficient are additionally retrieved from the parameter correction sub-table specifically for sudden load conditions. The retrieved target retraction and deployment speed is multiplied by the speed attenuation coefficient to obtain the corrected target retraction and deployment speed. The calculation method for the corrected target retraction and deployment speed is as follows: ； in, The corrected target release and retraction speed. A velocity attenuation coefficient specifically designed for sudden load conditions; The corrected target tension threshold is obtained by multiplying the retrieved target tension threshold by the tension buffer coefficient. The calculation method for the corrected target tension threshold is as follows: ； in, The corrected target tension threshold, Tension buffer coefficient specifically designed for sudden load conditions.

7. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 6, characterized in that, The pneumatic winch's deployment and retrieval control commands are generated based on the retrieved deployment and retrieval control parameter set. During the deployment and retrieval process, real-time sensor data is continuously collected and compared with a preset data range under the target operating condition category. When the real-time sensor data exceeds the preset data range, dynamic correction of the control parameters is triggered, specifically including: The target retraction speed in the set of retraction and extension control parameters is used as the speed command reference value, the target tension threshold is used as the tension command upper limit value, and the air source pressure compensation value is used as the air source adjustment reference value. The three are combined to generate the retraction and extension control command of the pneumatic winch and send it to the actuator. During the execution of the take-up and release control command, real-time sensor data is continuously collected using the same time base as the synchronous data sequence in the steps. The statistical values ​​within the current sliding window are calculated for the real-time collected air source pressure data, wire rope tension data, and winch output shaft speed data. The length of the sliding window is consistent with the length of the sliding window during feature extraction. The current sliding window statistical value of the real-time gas source pressure data is compared with the preset allowable fluctuation range of the gas source pressure under the target working condition category. The allowable fluctuation range is represented by the gas source pressure compensation value under the working condition category as the center and the range of fluctuation above and below the fifteenth preset threshold. The current sliding window statistical value of the real-time wire rope tension data is compared with the target tension threshold under the target working condition category; The real-time winding and unwinding speed, calculated from the real-time winch output shaft speed data, is compared with the target winding and unwinding speed under the target working condition category. When the current sliding window statistical value of the real-time air source pressure data exceeds the allowable fluctuation range of the air source pressure, the ratio of the excess to the boundary value of the allowable fluctuation range is used as the pressure correction coefficient. The air source pressure compensation value is multiplied by the pressure correction coefficient to obtain the corrected air source pressure compensation value. The original air source pressure compensation value is then replaced with the corrected air source pressure compensation value to regenerate the retraction and release control command. The pressure correction coefficient is calculated as follows: ； in, This is the pressure correction factor. This is the statistical value of the current sliding window for real-time gas source pressure data. This represents the upper limit of the allowable fluctuation range of the gas source pressure. This is the lower limit of the allowable fluctuation range of the gas source pressure; When the current sliding window statistical value of the real-time wire rope tension data exceeds the target tension threshold, the ratio of the excess to the target tension threshold is used as the tension correction coefficient. The target take-up and release speed is multiplied by the tension correction coefficient to obtain the corrected target take-up and release speed. The corrected target take-up and release speed replaces the original target take-up and release speed to regenerate the take-up and release control command. The tension correction coefficient is calculated as follows: ； in, This is the tension correction factor. This is the statistical value of the current sliding window for real-time wire rope tension data. The target tension threshold; When the deviation between the real-time deployment and recovery speed and the target deployment and recovery speed exceeds the sixteenth preset threshold, the ratio of the deviation value to the target deployment and recovery speed is used as a speed correction coefficient to superimpose and correct the target deployment and recovery speed. The correction direction is to reduce the deviation. The speed correction coefficient is calculated as follows: ； in, For speed correction factor, For real-time deployment and retraction speed, The target is the speed of expansion and contraction.

8. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 7, characterized in that, The step of triggering dynamic correction of control parameters when real-time sensor data exceeds a preset data range also includes: After triggering dynamic correction of control parameters, the rate of change of each data item is calculated for the real-time sensor data of the last five consecutive time nodes before correction and the real-time sensor data of the first to fifth consecutive time nodes after correction. If the absolute value of the ratio of the rate of change of the corrected air source pressure data to the rate of change of the air source pressure data before correction is less than the seventeenth preset threshold, and the absolute value of the ratio of the rate of change of the corrected wire rope tension data to the rate of change of the wire rope tension data before correction is less than the eighteenth preset threshold, and the absolute value of the ratio of the rate of change of the corrected reeling-out speed to the rate of change of the reeling-out speed before correction is less than the nineteenth preset threshold, then the dynamic correction is deemed effective, and the corrected reeling-out control parameter set continues to be executed. The calculation method for the rate of change ratio is as follows: ； in, The ratio of the rate of change of data item x To correct the rate of change of this data item, The rate of change of this data item before correction; If the absolute value of any of the above three ratios is greater than or equal to the corresponding preset threshold, it is determined that the dynamic correction has not converged. The uncorrected and corrected control parameter sets are then weighted and merged according to preset weights. The weights are determined based on the deviation of the rate of change of each data item. The greater the deviation, the higher the weight of the original parameter set corresponding to the data item. The merged parameter set is used as the new control parameter set to regenerate the control command. The weighted fusion calculation method is as follows: ； in, This is the fused set of control parameters for expansion and contraction. For data items The corresponding original parameter set or modified parameter set, For data items The corresponding weights, and satisfying Weight Based on the ratio of the rate of change of each data item Sure, The larger The closer it is to the original parameter set; If the dynamic correction fails to converge after three consecutive attempts, the target working condition category will be forcibly re-entered into the working condition classification model for a second determination. The current control parameter set will be replaced by the set of control parameters corresponding to the second determination result.

9. The pneumatic winch deployment and recovery system with multi-scenario adaptive operating conditions according to claim 8, characterized in that, The comparison of real-time sensor data with a preset data range under the target operating condition category also includes: For cases where the target working condition is a sudden load condition, the wire rope tension data range in the preset data range is set with the target tension threshold as the upper limit and 60% of the target tension threshold as the lower limit. The real-time wire rope tension data is compared by comparing the maximum value of the tension data in the current sliding window with the upper limit and the minimum value with the lower limit. When the maximum value exceeds the upper limit or the minimum value is lower than the lower limit, dynamic correction of the control parameters is triggered. For cases where the target operating condition is a strong wind condition, the wind speed data range in the preset data range is capped at the 95th percentile of the historical wind speed data under this operating condition. The real-time wind speed data is compared by comparing the peak value of the gust pulsation component of the wind speed data in the current sliding window with the upper limit. When the peak value exceeds the upper limit, dynamic correction of the control parameters is triggered. At the same time, the air source pressure compensation value is increased by the twenty-first preset threshold to compensate for the influence of wind load on the output torque of the pneumatic winch. For targets operating under low-temperature conditions, the ambient temperature range within the preset data range is limited to the 5th percentile of historical temperature data for that condition. Real-time temperature data is compared by taking the mean of the current sliding window and comparing it to the lower limit. When the mean is lower than the lower limit, the target launch / retract speed is multiplied by a low-temperature speed correction factor. This correction factor is calculated linearly based on the difference between the current temperature and the lower limit temperature; the larger the temperature difference, the smaller the correction factor. The calculation method for the low-temperature speed correction factor is as follows: ； in, This is a low-temperature speed correction factor. This is the speed correction coefficient constant under low-temperature operating conditions. For low-temperature operating conditions, preset the lower limit of the data range. This represents the average value of the current sliding window of real-time ambient temperature data. Simultaneously, the target tension threshold is multiplied by a low-temperature tension correction factor. This low-temperature tension correction factor is determined based on the proportion of change in the elastic modulus of the wire rope material at the current temperature. The calculation method for the low-temperature tension correction factor is as follows: ； in, This is the low-temperature tension correction factor. For the wire rope material at the current real-time temperature The elastic modulus below, For wire rope material at reference temperature The elastic modulus below.

Images