Vacuum degree intelligent compensation method based on temperature change driving

Abstract

The invention relates to the field of vacuum degree compensation, in particular to an intelligent vacuum degree compensation method based on temperature change driving, and the method comprises the steps: arranging an integrated sensor cluster on the inner wall of a falling body cavity, collecting a temperature initial value and a vacuum degree initial value of the falling body cavity when a to-be-monitored falling body is in a static state, and collecting the temperature initial value and the vacuum degree initial value of the falling body cavity when the to-be-monitored falling body is in a falling body state; the method comprises the following steps: collecting real-time temperature data and real-time vacuum degree data of a falling body cavity at preset time nodes at intervals, calculating an operation difference value under the time nodes, carrying out training preprocessing, constructing a deviation monitoring model to learn operation pre-training data, obtaining a deviation fluctuation map, extracting a fluctuation value, comparing the fluctuation value with a fluctuation threshold value, and obtaining a deviation monitoring result. The falling acceleration calculation and / or vacuum degree intelligent compensation are / is performed on the falling body cavity, so that the problem of vacuum degree deviation accumulation caused by incapability of real-time monitoring is avoided, the vacuum degree in the falling body cavity is always kept in a relatively accurate range, intelligent decision of vacuum degree compensation is realized, and the automation degree of vacuum degree compensation is improved.

Description

Technical Field

[0001] The present invention relates to the field of vacuum degree compensation, and particularly to an intelligent vacuum degree compensation method driven by temperature change. Background Art

[0002] Vacuum degree refers to the number of gas molecules per unit volume, and temperature is an important factor affecting the motion state of gas molecules. When the temperature changes, the thermal motion speed of gas molecules also changes accordingly, thereby affecting the vacuum degree in the vacuum chamber.

[0003] The thermal motion of gas molecules intensifies, and the collision frequency between molecules increases. If the number of gas molecules in the vacuum chamber remains unchanged, then an increase in temperature will cause an increase in gas pressure, thereby reducing the vacuum degree. For example, in some high-temperature experimental equipment, when the experimental temperature rises, the chamber that has been pumped to a relatively high vacuum degree may experience a decrease in vacuum degree due to the increase in temperature, affecting the accuracy and reliability of the experiment.

[0004] Chinese Patent Application Publication No.: CN1259282C discloses a ball mill tank with controllable temperature and vacuum degree belonging to the technical field of ceramic colloidal forming process. A vacuum solenoid valve is installed on the right side of the ball mill tank, and a vacuum pressure gauge is installed on the upper part; a temperature controller and a temperature measuring device are installed on the left side; the temperature control medium is installed in the interlayer of the tank wall of the ball mill tank. This device makes the vacuum pressure gauge and the vacuum pumping system plug-and-play, and is convenient to operate; by adjusting the temperature control medium to the required temperature, the temperature control medium can make the temperature of the suspension in the ball mill tank consistent with it. It ensures that the ball mill tank device can not only control the temperature but also achieve a good effect of vacuum defoaming.

[0005] Chinese Patent Authorization Publication No.: CN110887603B discloses a vacuum degree monitoring device, which includes a vacuum gauge (100), a data transmission line (200), a vacuum gauge tube (500) connected to the data transmission line, and a monitoring object connection tube connected to the vacuum gauge tube. The other end of the monitoring object connection tube is connected to a vacuum monitoring object. A first flange (530) is provided at the air inlet end (530a) of the vacuum gauge tube, and a second flange is provided at the air outlet end (410a) of the monitoring object connection tube. The first flange (530) and the second flange (410) are hermetically fitted together, and a clamp (300) is fastened to the outer surface of the connection between the two to fasten the butt joint of the first and second flanges.

[0006] However, the above methods have the following problems: They do not consider the complex relationship between temperature and vacuum degree, have limited compensation accuracy for vacuum degree, and are difficult to meet the requirements of high-precision vacuum application scenarios. Summary of the Invention

[0007] To this end, the present invention provides an intelligent compensation method for vacuum degree driven by temperature change, so as to overcome the problems in the prior art that the complex relationship between temperature and vacuum degree is not considered, the compensation accuracy of vacuum degree is limited, and it is difficult to meet the requirements of high-precision vacuum application scenarios.

[0008] To achieve the above object, the present invention provides an intelligent compensation method for vacuum degree driven by temperature change, including:

[0009] Calculate the central axis of the free-fall chamber, determine the cross-section where the central axis is located, and arrange integrated sensors along the inner wall of the free-fall chamber on the cross-section to form an integrated sensor cluster;

[0010] Place the free-fall object to be monitored at a preset position. When the free-fall object to be monitored is in a stationary state, use the integrated sensors to collect the initial temperature value and the initial vacuum degree value of the free-fall chamber;

[0011] Start the vacuum pump and release the free-fall object to be monitored at the same time. When the free-fall object to be monitored is in a free-fall state, use the integrated sensors to collect the real-time temperature data and the real-time vacuum degree data of the free-fall chamber at preset time nodes. Among them, the vacuum pump is used to pump the free-fall chamber to vacuum at a preset rate uniformly;

[0012] Record the time nodes of single collection by the integrated sensors, calculate the running differences at several time nodes in sequence, and perform training preprocessing on the running differences to obtain corresponding running pre-training data. Among them, the running differences include the temperature difference between the real-time temperature data and the initial temperature value and the vacuum degree difference between the real-time vacuum degree data and the initial vacuum degree value;

[0013] Based on the running pre-training data, determine several deviation features, construct the input layer of the deviation monitoring model, the number of nodes in the input layer is determined according to the number of deviation features, calculate the output nodes of the deviation monitoring model, and set the hidden layer structure of the deviation monitoring model to form the deviation monitoring model. Among them, the deviation features are characteristic parameters reflecting the temperature and vacuum degree changes in the free-fall chamber, including the peak value of the temperature difference, the peak value of the vacuum degree difference, the valley value of the temperature difference, and the valley value of the vacuum degree difference.

[0014] Use the deviation monitoring model to learn the running pre-training data to obtain a deviation fluctuation map, extract the fluctuation values in the deviation fluctuation map, and calculate the falling acceleration and / or perform intelligent compensation for the vacuum degree of the free-fall chamber according to the response result of the fluctuation value and the fluctuation threshold.

[0015] Further, the step of forming the integrated sensor cluster includes:

[0016] Determine the central axis of the falling body cavity according to the geometric shape parameters of the falling body cavity, where

[0017] the geometric shape parameters include the length, width, height and shape profile of the falling body cavity;

[0018] the central axis is the axis of symmetry or the centroid axis of the falling body cavity;

[0019] Select a number of specific positions on the central axis as the reference points of the cross-section. According to the reference points, determine the cross-section perpendicular to the central axis, and arrange the integrated sensors on the inner wall of the falling body cavity at a preset interval along the cross-section to form an integrated sensor cluster.

[0020] Further, the stationary state means that the falling body to be monitored does not displace at the preset position, and the environment in the falling body cavity is in a stable state.

[0021] Further, the preset rate is positively correlated with the length of the central axis, the mass of the falling body to be monitored and / or the response speed of the integrated sensor.

[0022] Further, use a pre-processor to perform training pre-processing on the operation difference, and divide the operation difference according to a preset learning rate to form corresponding operation pre-training data, where

[0023] the preset learning rate is the step size of parameter update during a single learning of the deviation monitoring model.

[0024] Further, set the hidden layer structure of the deviation monitoring model. The hidden layer includes at least one layer of neural network, and the number of nodes in each layer of neural network is related to the complexity and accuracy requirements of the deviation monitoring model.

[0025] Further, compare the fluctuation value with the fluctuation threshold to form a corresponding response result,

[0026] where the fluctuation threshold is set according to the maximum allowable range of the fluctuation value, and is used to judge whether vacuum degree compensation needs to be performed on the falling body cavity.

[0027] Further, when the fluctuation value is lower than the fluctuation threshold, obtain the change amount of the fluctuation value at a continuous number of time nodes. According to the change amount, use the numerical differentiation method to calculate the time derivative of the fluctuation value to obtain the change rate of the fluctuation value. Combine the geometric shape parameters of the falling body cavity, and use the change rate of the fluctuation value to calculate the falling acceleration of the falling body to be monitored at the time node.

[0028] Further, when the fluctuation value is higher than the fluctuation threshold, a compensation instruction is triggered, and the fluctuation value and the time node are input into the deviation monitoring model. Through the calculation of the deviation monitoring model, the preset vacuum degree data is output, and the preset rate is increased until the real-time vacuum degree data of the free-fall chamber is adjusted to the preset vacuum degree data.

[0029] Compared with the prior art, in the present invention, an integrated sensor cluster is arranged on the inner wall of the free-fall chamber. When the free-fall object to be monitored is in a static state, the initial temperature value and the initial vacuum degree value of the free-fall chamber are collected. When the free-fall object to be monitored is in a free-fall state, the real-time temperature data and the real-time vacuum degree data of the free-fall chamber are collected at preset time nodes, the running difference at the time node is calculated and pre-trained, a deviation monitoring model is constructed to learn the pre-trained running data, a deviation fluctuation map is obtained, the fluctuation value is extracted and compared with the fluctuation threshold, and the falling acceleration of the free-fall chamber is calculated and / or the vacuum degree is intelligently compensated, avoiding the problem of vacuum degree deviation accumulation caused by the inability to monitor in real time, keeping the vacuum degree in the free-fall chamber always within a relatively accurate range, realizing the intelligent decision-making of vacuum degree compensation, and improving the automation degree of vacuum degree compensation.

[0030] Further, by arranging a plurality of integrated sensors in the free-fall chamber at preset intervals, multi-point monitoring of different positions inside the free-fall chamber is realized. Compared with a single sensor or simply distributed sensors, this layout can cover the internal space of the free-fall chamber more comprehensively, obtain richer temperature and vacuum degree data, and thus more accurately reflect the overall state inside the free-fall chamber.

[0031] Further, by dividing the running difference according to a preset learning rate, the data can be normalized to make it more suitable for the learning of the model. This helps to reduce the volatility and noise of the data, improve the stability and convergence speed of the model. A reasonable learning rate can accelerate the convergence speed of the model and reduce the training time.

[0032] Further, by triggering a compensation instruction when the fluctuation value is not lower than the fluctuation threshold, the vacuum degree can be adjusted in time, avoiding the influence on the experimental results or the operation of the equipment due to excessive vacuum degree deviation. Using the deviation monitoring model to calculate the preset vacuum degree data of the free-fall chamber at the corresponding time node can more accurately determine the vacuum degree value to be adjusted, improving the accuracy of vacuum degree control. Description of the Drawings

[0033] Figure 1 It is a flowchart of the intelligent vacuum degree compensation method based on temperature change drive of the present invention. Detailed Embodiments

[0034] In order to make the objectives and advantages of the present invention more clear and understandable, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0035] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0036] It should be noted that in the description of the present invention, the terms indicating directions or positional relationships such as "upper", "lower", "left", "right", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings. This is only for convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0037] In addition, it should also be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0038] Please refer to Figure 1 as shown, which is a flowchart of the intelligent vacuum degree compensation method driven by temperature change according to the present invention, including:

[0039] Step S1, calculate the central axis of the falling body cavity, determine the cross-section where the central axis is located, and arrange integrated sensors along the inner wall of the falling body cavity on the cross-section to form an integrated sensor cluster;

[0040] Step S2, place the falling body to be monitored at a preset position. When the falling body to be monitored is in a stationary state, use the integrated sensors to collect the initial temperature value and the initial vacuum degree value of the falling body cavity;

[0041] Step S3, start the vacuum pump and release the falling body to be monitored at the same time. When the falling body to be monitored is in a falling state, use the integrated sensors to collect the real-time temperature data and the real-time vacuum degree data of the falling body cavity at preset time nodes. Among them, the vacuum pump is used to pump the falling body cavity to vacuum at a preset rate uniformly;

[0042] Step S4, record the time nodes of each single acquisition by the integrated sensor, calculate the running differences at several time nodes in sequence, and perform training preprocessing on the running differences to obtain corresponding running pre-training data, where the running differences include the temperature difference between the real-time temperature data and the initial temperature value and the vacuum difference between the real-time vacuum degree data and the initial vacuum degree value;

[0043] Step S5, determine several deviation features based on the running pre-training data, construct the input layer of the deviation monitoring model, the number of nodes in the input layer is determined according to the number of deviation features, calculate the output nodes of the deviation monitoring model, and set the hidden layer structure of the deviation monitoring model to form the deviation monitoring model, where the deviation features are characteristic parameters reflecting the temperature and vacuum degree changes in the falling body cavity, including the peak value of the temperature difference, the peak value of the vacuum difference, the valley value of the temperature difference, and the valley value of the vacuum difference.

[0044] Step S6, use the deviation monitoring model to learn the running pre-training data to obtain a deviation fluctuation map, extract the fluctuation values in the deviation fluctuation map, and perform falling acceleration calculation and / or vacuum degree intelligent compensation on the falling body cavity according to the response result of the fluctuation value and the fluctuation threshold.

[0045] By arranging an integrated sensor cluster on the inner wall of the falling body cavity, when the falling body to be monitored is in a stationary state, the initial temperature value and the initial vacuum degree value of the falling body cavity are collected, and when the falling body to be monitored is in a falling state, the real-time temperature data and the real-time vacuum degree data of the falling body cavity are collected at preset time nodes, calculate the running differences at the time nodes and perform training preprocessing, construct a deviation monitoring model to learn the running pre-training data, obtain a deviation fluctuation map, extract the fluctuation values and compare them with the fluctuation threshold, perform falling acceleration calculation and / or vacuum degree intelligent compensation on the falling body cavity, avoiding the problem of vacuum degree deviation accumulation caused by the inability to monitor in real time, keeping the vacuum degree in the falling body cavity within a relatively accurate range all the time, realizing the intelligent decision-making of vacuum degree compensation, and improving the automation degree of vacuum degree compensation.

[0046] Specifically, the steps of forming the integrated sensor cluster include:

[0047] According to the geometric shape parameters of the falling body cavity, determine the central axis of the falling body cavity through geometric calculation methods, where,

[0048] The geometric shape parameters include the length, width, height and shape contour of the falling body cavity;

[0049] The central axis is the axis of symmetry or the centroid axis of the falling body cavity;

[0050] Select several specific positions on the central axis as the reference points for the cross-section. Based on the reference points, determine the cross-section perpendicular to the central axis, and arrange integrated sensors along the inner wall of the free-fall cavity at a preset interval to form an integrated sensor cluster.

[0051] In a specific implementation, through geometric calculation methods, use the above geometric shape parameters to determine the central axis of the free-fall cavity. The specific calculation methods include:

[0052] Axis of symmetry calculation: If the free-fall cavity has symmetry (such as cylindrical, cuboid, etc.), the central axis can be its axis of symmetry.

[0053] Centroid axis calculation: If the shape of the free-fall cavity is irregular, the central axis can be determined by calculating its centroid. The centroid axis refers to the axis passing through the centroid and being geometrically representative.

[0054] Select several specific positions on the determined central axis. These positions will serve as the reference points for the cross-section. The selection of the reference points can be based on actual needs, for example:

[0055] Uniform distribution: Select several points evenly on the central axis to ensure the uniformity of sensor arrangement.

[0056] Key positions: Select some key positions, such as the top, middle, and bottom of the free-fall cavity, to cover the data acquisition needs at different heights.

[0057] On each cross-section, arrange integrated sensors along the inner wall of the free-fall cavity at a preset interval. The preset interval can be determined according to actual needs and accuracy requirements, for example:

[0058] High-precision requirement: If high-precision data acquisition is required, a smaller interval can be set.

[0059] Cost control: If the cost is limited, the interval can be appropriately increased.

[0060] By setting several integrated sensors at a preset interval in the free-fall cavity, multi-point monitoring of different positions inside the free-fall cavity is achieved. Compared with a single sensor or simply distributed sensors, this layout can cover the internal space of the free-fall cavity more comprehensively, obtain richer temperature and vacuum data, and thus more accurately reflect the overall state inside the free-fall cavity.

[0061] Specifically, the static state means that the free-fall object to be monitored does not move at the preset position, and the environment inside the free-fall cavity is in a stable state.

[0062] Conduct initial data acquisition in the static state to avoid measurement errors caused by the small displacement of the free-fall object or the fluctuation of the environment inside the free-fall cavity.

[0063] A clear definition of the stationary state enables the system to better adapt to different working conditions and application scenarios. Regardless of the changes in the size, shape of the falling body cavity or the type of the falling body to be monitored, as long as the definition of the stationary state is met, the system can collect initial data and monitor according to a unified standard. This adaptability allows the system to be widely applied in various different experimental and industrial environments.

[0064] Specifically, the preset rate is positively correlated with the length of the central axis, the mass of the falling body to be monitored, and / or the response speed of the integrated sensor.

[0065] In a specific implementation, the length of the central axis reflects the size of the falling body cavity. The positive correlation between the preset rate and the length of the central axis means that for a larger falling body cavity, a higher preset rate can be used to reach the required vacuum degree faster. This design enables the system to adapt to falling body cavities of different sizes. Whether it is a small experimental device or a large industrial equipment, the vacuum extraction process can be optimized by adjusting the preset rate.

[0066] The mass of the falling body to be monitored directly affects its motion state in the falling body cavity and its sensitivity to changes in the vacuum degree. The positive correlation between the preset rate and the mass of the falling body means that for a heavier falling body, a higher preset rate can be used for vacuum pumping to ensure that the vacuum degree can quickly reach a stable state during the falling body movement. This design enables the system to flexibly adjust the preset rate according to the mass of the falling body, improving the adaptability of the system.

[0067] The response speed of the integrated sensor determines how quickly the system can detect changes in the vacuum degree and temperature. The positive correlation between the preset rate and the response speed of the sensor means that when the sensor can respond quickly, a higher preset rate can be used for vacuum pumping. This can ensure that when the sensor can provide feedback data in a timely manner, the system can quickly adjust the vacuum degree, reduce waiting time, and improve the overall operating efficiency of the system.

[0068] By reasonably adjusting the preset rate, the system can reduce unnecessary energy consumption while ensuring the control accuracy of the vacuum degree. For example, for a smaller falling body cavity or a lighter falling body, a lower preset rate can be used, thereby reducing the running time and energy consumption of the vacuum pump.

[0069] Specifically, the pre-processor is used to perform training pre-processing on the running difference. The running difference is segmented according to the preset learning rate to form corresponding running pre-training data, where,

[0070] The preset learning rate is the step size of parameter update during a single learning of the deviation monitoring model.

[0071] In a specific implementation, the preset learning rate is the step size for parameter update during a single learning of the deviation monitoring model. The learning rate determines the adjustment amplitude of the model's parameters in each iteration. During the preprocessing process, the running difference is segmented according to the preset learning rate. This means that the running difference data is segmented into multiple small data blocks, and each data block corresponds to a learning step. The running difference is segmented according to the preset learning rate to form multiple running pre-training data blocks. The specific steps are as follows:

[0072] Determine the learning rate: According to the complexity of the model and the training requirements, select a suitable preset learning rate. The learning rate is usually a small positive number, such as 0.01 or 0.001.

[0073] Segment the data: Segment the running difference data according to the step size of the learning rate. For example, if the learning rate is 0.01, then every 0.01 unit of the running difference data will be segmented into an independent data block.

[0074] Form data blocks: Each data block contains the running difference data within a certain range, and these data blocks will be used as the input data for the deviation monitoring model.

[0075] If the learning rate is too large, it may cause the model training to be unstable; if the learning rate is too small, it may cause the training process to be too slow. By adjusting the preset learning rate, it can flexibly adapt to models of different complexities and different training requirements. For example, a larger learning rate can be used at the beginning of training to quickly converge, and then the learning rate can be gradually reduced to improve the accuracy.

[0076] By segmenting the running difference according to the preset learning rate, the data can be normalized, making it more suitable for the model's learning. This helps to reduce the volatility and noise of the data, improve the stability and convergence speed of the model. A reasonable learning rate can accelerate the convergence speed of the model and reduce the training time.

[0077] Specifically, set the hidden layer structure of the deviation monitoring model. The hidden layer includes at least one layer of neural network, and the number of nodes in each layer of the neural network is related to the complexity and accuracy requirements of the deviation monitoring model.

[0078] In a specific implementation, the hidden layer may contain multiple layers of neural networks to enhance the complexity and expressive ability of the model. The number of nodes and the number of layers in each layer of the neural network can be adjusted according to specific needs.

[0079] The number of nodes in each layer of the neural network determines the computing ability and feature extraction ability of that layer. The more nodes there are, the more complex the features the model can learn, but at the same time, it also increases the computing cost and the complexity of the model.

[0080] The complexity of the model is proportional to the number of hidden layers and the number of nodes in each layer. Increasing the number of layers or nodes can improve the model's expressive power, but it will also increase the complexity of the model and the difficulty of training.

[0081] The accuracy requirement of the model determines the design of the hidden layer structure. If high-precision deviation monitoring is required, more hidden layers and nodes may be needed to extract more complex features; if the accuracy requirement is low, the number of layers and nodes can be reduced to simplify the model.

[0082] The number of hidden layers and the number of nodes in each layer should be selected according to specific task requirements and data characteristics. If the task is relatively simple and the data features are obvious, fewer hidden layers and nodes can be used. If the task is complex and the data features are obscure, more hidden layers and nodes are needed to extract complex features. Usually, experiments are required to verify the performance of different hidden layer structures. Different combinations of the number of layers and nodes can be tried, and the optimal structure can be selected through methods such as cross-validation. When designing the hidden layer structure, it is necessary to balance the complexity and accuracy of the model. Too many hidden layers and nodes may lead to overfitting, while too few hidden layers and nodes may lead to underfitting.

[0083] Specifically, the fluctuation value is compared with the fluctuation threshold to form a corresponding response result.

[0084] Among them, the fluctuation threshold is set according to the maximum allowable range of the fluctuation value, and is used to judge whether vacuum degree compensation is required for the falling body cavity.

[0085] In specific implementation, the fluctuation value is extracted from the deviation fluctuation map, which reflects the dynamic characteristics of the temperature and vacuum degree changes in the falling body cavity. These values can be the peak, valley or other statistical characteristics of the temperature difference and vacuum degree difference.

[0086] The fluctuation threshold is set according to the maximum allowable range of the fluctuation value, and is used to judge whether the fluctuation is within the acceptable range. The setting of the fluctuation threshold needs to be determined according to the actual application requirements and safety standards. The maximum allowable range refers to the maximum acceptable value of the fluctuation value under normal operating conditions. This range is usually determined according to experimental data, historical data or safety standards. The setting of the fluctuation threshold needs to consider safety standards to ensure that measures can be taken in time when the fluctuation exceeds the threshold to avoid damage to the equipment or experiment.

[0087] Specifically, when the fluctuation value is lower than the fluctuation threshold, the change amount of the fluctuation value at multiple consecutive time nodes is obtained. According to the change amount, the time derivative of the fluctuation value is calculated by using the numerical differentiation method to obtain the change rate of the fluctuation value. Combining with the geometric shape parameters of the falling body cavity, the falling acceleration of the falling body to be monitored at the time node is calculated by using the change rate of the fluctuation value.

[0088] In specific implementation, when the fluctuation value is lower than the fluctuation threshold, the monitoring platform calculates the falling acceleration of the fluctuation value at the corresponding time node. The formula for calculating the gravitational acceleration is as follows:

[0089] Wherein,

[0090] g t is the falling acceleration of the falling object to be monitored at the corresponding time node, and F t is the falling resistance suffered by the falling object to be monitored at the corresponding time node, and m is the mass of the falling object to be monitored.

[0091] Specifically, the formula for calculating the falling resistance is as follows:

[0092] Wherein,

[0093] A is the total surface area of the falling object to be monitored, ρ is the pressure of the falling object cavity, V is the average velocity of gas molecules in the falling object cavity at the initial temperature value, and ν is the maximum falling velocity of the falling object to be monitored.

[0094] When the fluctuation value is lower than the fluctuation threshold, it indicates that the vacuum degree and temperature change in the falling object cavity are within the normal range. At this time, calculating the falling acceleration and falling resistance can ensure the reliability of the calculation results. Precise control can reduce the wear and impact on the equipment, extend the service life of the equipment, and reduce the maintenance cost and replacement cost of the equipment.

[0095] Specifically, when the fluctuation value is higher than the fluctuation threshold, a compensation instruction is triggered, the fluctuation value and the time node are input into the deviation monitoring model. Through the calculation of the deviation monitoring model, the preset vacuum degree data is output, and the preset rate is increased until the real-time vacuum degree data of the falling object cavity is adjusted to the preset vacuum degree data.

[0096] In specific implementation, by adjusting the preset rate until the real-time vacuum degree data of the falling object cavity is adjusted to the preset vacuum degree data, it can ensure that the vacuum degree in the falling object cavity always remains in a stable state, providing a stable environment for experiments or equipment operation. By precisely controlling the vacuum degree, energy consumption waste caused by inaccurate control can be avoided, and the operation cost of the system is reduced. When necessary, restart the vacuum pump to evacuate the falling object cavity.

[0097] During the intelligent compensation of the vacuum degree, continuously collecting the real-time temperature data and real-time vacuum degree data can provide instant feedback information for the system, enabling the system to timely understand the current temperature and vacuum degree state in the falling object cavity, so as to judge whether the compensation measures are effective and whether further adjustment is needed.

[0098] Compare the real-time vacuum degree data collected with the preset target value of the vacuum degree. If the real-time vacuum degree data can stably approach or reach the target value after compensation and remain within the allowable error range, it indicates that the compensation effect is good. Compare the compensated real-time temperature data and vacuum degree data with the data before compensation. If the fluctuation range of the data after compensation is significantly reduced and closer to the ideal state, it shows that the compensation has played an improvement role. During the compensation process, if the vacuum degree data shows a trend of gradually stabilizing and approaching the target value, and there are no obvious abnormal conditions such as significant rebounds or continuous decreases, it indicates that the compensation measures are working effectively. For example, at the beginning of compensation, the vacuum degree may have short-term fluctuations due to system adjustment, but then gradually stabilizes and approaches the target value, which indicates that the compensation is effective.

[0099] Analyze the change trend of the real-time temperature data. Since temperature changes may affect the vacuum degree, it is necessary to pay attention to whether the temperature remains stable within a reasonable range. If the temperature data remains stable during the compensation process and there is no abnormal upward or downward trend, it indicates that the influence of temperature on the vacuum degree during the compensation process has been effectively controlled and the compensation effect is good.

[0100] Pay attention to the fluctuation frequency of the real-time data, that is, the number of data fluctuations per unit time. If the fluctuation frequency decreases after compensation, it indicates that the compensation measures help reduce the frequent fluctuations of the data and make the system more stable. For example, before compensation, the temperature data fluctuates 5 times per minute and the vacuum degree data fluctuates 3 times per minute. After compensation, the temperature data fluctuates 1 time per minute and the vacuum degree data fluctuates 0.5 times per minute. It can be judged that the compensation has an obvious effect on reducing the fluctuation frequency.

[0101] Use the deviation monitoring model to analyze and predict the real-time collected data, and compare the predicted compensation effect of the model with the actually collected data. If the result predicted by the model is consistent with the actual data, it indicates that the compensation effect meets the expectations. For example, the model predicts that the vacuum degree will stabilize near the target value within 10 minutes after compensation, and the actual data also shows that the vacuum degree tends to be stable after 10 minutes, which indicates that the compensation effect is good.

[0102] Update and optimize the deviation monitoring model according to the continuously collected real-time data. If the model can continuously adjust and optimize according to the new data to make its prediction of the compensation effect more accurate, this also indirectly shows the effectiveness of the compensation measures, because only when the compensation measures are effective and the data is regular can the model better learn and optimize.

[0103] By triggering a compensation instruction when the fluctuation value is not lower than the fluctuation threshold, the vacuum degree can be adjusted in a timely manner, avoiding the influence on the experimental results or the operation of the equipment due to excessive deviation of the vacuum degree. Using the deviation monitoring model to calculate the preset vacuum degree data of the free-fall chamber at the corresponding time nodes can more accurately determine the vacuum degree value to be adjusted, improving the accuracy of vacuum degree control.

[0104] So far, the technical solution of the present invention has been described in combination with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0105] The above are only the preferred embodiments of the present invention and are not used to limit the present invention; for those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for intelligent compensation of vacuum degree driven by temperature change, characterized in that, Including: Calculating the central axis of the free-fall chamber, determining the cross-section where the central axis is located, and arranging integrated sensors along the inner wall of the free-fall chamber on the cross-section to form an integrated sensor cluster; Placing the free-fall to be monitored at a preset position. When the free-fall to be monitored is in a static state, using the integrated sensors to collect the initial temperature value and the initial vacuum degree value of the free-fall chamber; Starting the vacuum pump and simultaneously releasing the free-fall to be monitored. When the free-fall to be monitored is in a free-fall state, using the integrated sensors to collect the real-time temperature data and the real-time vacuum degree data of the free-fall chamber at preset time nodes. Among them, the vacuum pump is used to pump the free-fall chamber to a vacuum at a preset rate uniformly; Recording the time nodes of single collection by the integrated sensors, calculating the running differences at several time nodes in sequence, and performing training preprocessing on the running differences to obtain corresponding running pre-training data. Among them, the running differences include the temperature difference between the real-time temperature data and the initial temperature value and the vacuum degree difference between the real-time vacuum degree data and the initial vacuum degree value; Determining several deviation features based on the running pre-training data, constructing the input layer of the deviation monitoring model. The number of nodes in the input layer is determined according to the number of deviation features, calculating the output nodes of the deviation monitoring model, and setting the hidden layer structure of the deviation monitoring model to form the deviation monitoring model. Among them, the deviation features are characteristic parameters reflecting the temperature and vacuum degree changes in the free-fall chamber, including the peak value of the temperature difference, the peak value of the vacuum degree difference, the valley value of the temperature difference, and the valley value of the vacuum degree difference. Using the deviation monitoring model to learn the running pre-training data to obtain a deviation fluctuation map, extracting the fluctuation values in the deviation fluctuation map, and calculating the falling acceleration and / or performing intelligent compensation for the vacuum degree of the free-fall chamber according to the response result of the fluctuation value and the fluctuation threshold.

2. The intelligent vacuum degree compensation method driven by temperature change according to claim 1, wherein The steps of forming the integrated sensor cluster include: Determining the central axis of the free-fall chamber through geometric calculation methods according to the geometric shape parameters of the free-fall chamber, where The geometric shape parameters include the length, width, height, and shape contour of the free-fall chamber; The central axis is the axis of symmetry or the centroid axis of the free-fall chamber; Selecting several specific positions on the central axis as the reference points of the cross-section, determining the cross-section perpendicular to the central axis according to the reference points, and arranging the integrated sensors along the inner wall of the free-fall chamber at a preset interval on the cross-section to form an integrated sensor cluster.

3. The method for intelligent vacuum degree compensation driven by temperature change according to claim 2, wherein, The static state means that the free-fall to be monitored does not displace at the preset position and the environment in the free-fall chamber is in a stable state.

4. The method for intelligent compensation of vacuum degree driven by temperature change according to claim 3, characterized in that The preset rate is positively correlated with the length of the central axis, the mass of the free-fall to be monitored, and / or the response speed of the integrated sensors.

5. The intelligent vacuum degree compensation method driven by temperature change according to claim 4, wherein Using a preprocessor to perform training preprocessing on the running differences, dividing the running differences according to a preset learning rate to form corresponding running pre-training data, where The preset learning rate is the step size for parameter update during a single learning of the deviation monitoring model.

6. The method for intelligent compensation of vacuum degree driven by temperature change according to claim 5, wherein, Set the hidden layer structure of the deviation monitoring model. The hidden layer includes at least one layer of neural network, and the number of nodes in each layer of neural network is related to the complexity and accuracy requirements of the deviation monitoring model.

7. The intelligent vacuum degree compensation method driven by temperature change according to claim 6, characterized in that Compare the fluctuation value with the fluctuation threshold to form a corresponding response result. Among them, the fluctuation threshold is set according to the maximum allowable range of the fluctuation value, and is used to judge whether it is necessary to perform vacuum degree compensation on the falling body cavity.

8. The method for intelligent compensation of vacuum degree driven by temperature change according to claim 7, wherein When the fluctuation value is lower than the fluctuation threshold, obtain the change amount of the fluctuation value at a plurality of consecutive time nodes. According to the change amount, use the numerical differentiation method to calculate the time derivative of the fluctuation value to obtain the change rate of the fluctuation value. Combine the geometric shape parameters of the falling body cavity, and use the change rate of the fluctuation value to calculate the falling acceleration of the falling body to be monitored at the time node.

9. The method for intelligent compensation of vacuum degree driven by temperature change according to claim 8, characterized in that, When the fluctuation value is higher than the fluctuation threshold, trigger a compensation instruction, input the fluctuation value and the time node into the deviation monitoring model. Through the calculation of the deviation monitoring model, output the preset vacuum degree data, and increase the preset rate until the real-time vacuum degree data of the falling body cavity is adjusted to the preset vacuum degree data.

Images