A segment vibrating parameter self-adaptive optimization method based on vibrating process data

Abstract

The present application belongs to the technical field of concrete prefabricated component production control, and particularly relates to a pipe segment vibrating parameter self-adaptive optimization method based on vibrating process data, comprising: synchronously acquiring a molding image sequence and a transient active power sequence; performing feature extraction on the molding image sequence to obtain a bubble quantity change rate, and obtaining an active power change rate according to the transient active power sequence; performing discrete time axis translation transformation and correlation calculation on the active power change rate and the bubble quantity change rate in a historical time window to determine a bubble floating delay time; performing time dimension extrapolation on the current active power change rate by using the bubble floating delay time, predicting potential resistance loss, and combining the initial maximum power with the current transient active power to calculate a over-vibration risk index to perform vibrating shutdown control. The present application can effectively break the visual delay illusion, realize self-adaptive shutdown control, and prevent pipe segment bottom layer over-vibration segregation.

Description

Technical Field

[0001] This invention relates to the field of precast concrete component production control technology. More specifically, this invention relates to an adaptive optimization method for segment vibration parameters based on vibration process data. Background Technology

[0002] In the tunnel construction industry, tunnel segments are the core load-bearing components, and their casting quality directly determines the structural safety and waterproof lifespan of the tunnel. In the precast production of tunnel segments, the vibration process after concrete pouring is a key step to remove internal air and increase density. In order to ensure that the concrete is uniformly dense before demolding to prevent water leakage later, the industry usually uses fixed vibration parameters or surface vision-based monitoring technology for automated shutdown control.

[0003] Chinese patent application CN120182912A discloses a method and system for monitoring concrete vibration based on image recognition and 3D point clouds. It acquires images and 3D point cloud data of the concrete surface, uses a target detection algorithm to identify and count the number of coarse aggregates and air bubbles on the surface, extracts the elevation difference of the vibrated surface from the point cloud data, and establishes a correlation model between visual monitoring results and the performance of hardened concrete to evaluate the vibration effect. However, this patent application mainly relies on surface visual features for post-event correlation evaluation. Because the escape of surface air bubbles and elevation changes during vibration inherently have a lag, the system is prone to triggering feedback only after internal segregation has already occurred, making real-time interruption of the vibration process impossible.

[0004] Chinese patent document CN110608769B discloses a real-time monitoring system and method for concrete vibration. It involves laying optical fibers within the pouring area and using an optical fiber sensing unit to receive and analyze vibration location, amplitude, and duration data. The system determines under-vibration or over-vibration areas based on a comparison of the accumulated amplitude values ​​at the vibration location with a preset threshold. However, this patent requires pre-laying a large number of optical fibers on the reinforcing steel frame, making construction and deployment cumbersome. Furthermore, the fibers are easily damaged by the vibration equipment in complex field environments. Its monitoring logic is highly dependent on the spatial distribution statistics of external vibration energy. In the enclosed environment of the segment fixing mold, the collected vibration amplitude is easily interfered with by the mold's own vibration frequency, making it difficult for the system to accurately distinguish between the decrease in internal concrete resistance and signal fluctuations caused by mold structural resonance, leading to misjudgments of the vibration status.

[0005] In existing technologies, although some solutions attempt to control the vibration quality of tunnel segments through surface visual statistics or distributed fiber optic amplitude monitoring, these solutions all rely on static threshold comparisons of apparent phenomena or external vibration signals, lacking dynamic correlation calculations of the decay rate of internal rheological state and surface visual lag time. In the precast production of tunnel segments, the disintegration of internal concrete resistance inevitably precedes the rupture of surface bubbles. Existing methods cannot overcome the illusion of visual delay, often passively waiting for apparent compaction, resulting in redundant vibration time and causing the coarse aggregate at the bottom of the segment to settle and segregate. At the same time, fixed parameter control cannot adapt to the rheological differences of concrete with different mix proportions, resulting in large fluctuations in vibration quality and making it difficult to achieve adaptive shutdown control that balances compaction and segregation prevention. Summary of the Invention

[0006] To address the technical problem of existing technologies being unable to dynamically calculate the hysteresis time between internal rheological decay and surface visual appearance, leading to passive delays in vibration shutdown and potentially causing over-vibration segregation of the tunnel lining segments, this invention provides an adaptive optimization method for tunnel lining vibration parameters based on vibration process data, comprising: The system synchronously acquires and records the forming image sequence of the segment surface dynamic process, as well as the transient active power sequence driving the segment forming mold vibration; it extracts features from the forming image sequence to obtain the bubble number change rate, which characterizes the surface visual phenomenon; it extracts features from the transient active power sequence to obtain the active power change rate, which characterizes the internal rheological state; it performs discrete time axis translation transformation and correlation calculation on the active power change rate and bubble number change rate within the historical time window to determine the bubble rising delay time, which reflects the change in internal rheological state preceding the surface visual phenomenon; it extrapolates the current active power change rate in the time dimension using the bubble rising delay time to predict the potential resistance loss in the future delay blind zone, and calculates the over-vibration risk index by combining the initial maximum power and the current transient active power; when the over-vibration risk index is greater than the shutdown control threshold, it outputs a speed reduction shutdown command to stop the vibration operation of the excitation motor.

[0007] This invention simultaneously acquires images of the forming surface of the tunnel segment and the transient active power of the excitation motor, extracting their respective rates of change. By using discrete time axis translation calculation, it accurately obtains the delay time in which internal rheological decay precedes the escape of surface bubbles. Then, it extrapolates the current resistance reduction magnitude using this delay time to construct an over-vibration risk index. When the over-vibration risk index exceeds the standard, it triggers a speed reduction and shutdown. This transforms the vibration control logic from passively waiting for surface phenomena to actively predicting the trend of internal resistance loss, effectively eliminating the ineffective vibration of the bottom layer caused by visual delay blind spots, preventing the sedimentation and segregation of coarse aggregate at the bottom of the tunnel segment, and realizing adaptive shutdown control based on the actual liquefaction process of concrete, thereby improving the precast compactness and structural safety of the tunnel segment.

[0008] Preferably, the step of synchronously acquiring the forming image sequence recording the dynamic process of the tube segment surface and the transient active power sequence driving the tube segment forming mold vibration includes: using the system clock to stamp a first physical timestamp on each frame of the continuously acquired forming image sequence; reading the instantaneous voltage sequence and instantaneous current sequence of the three-phase power supply driving the tube segment forming mold vibration in real time, and stamping them with a second physical timestamp that is on the same reference as the first physical timestamp; and performing element-wise multiplication on the instantaneous voltage sequence and instantaneous current sequence aligned based on the second physical timestamp to obtain the transient active power sequence.

[0009] Preferably, the step of extracting features from the molding image sequence to obtain the bubble number change rate characterizing the surface visual phenomenon includes: removing the reflective area of ​​the mold metal edge in the molding image sequence and retaining the internal slurry area as the effective observation area; performing binarization processing and morphological opening operation on the effective observation area to extract the contour of the connected components, and taking the total number of connected components that meet the preset circularity constraint and area empirical interval constraint as the total number of surface bubbles in the current frame image; calculating the difference between the total number of surface bubbles in the current frame image and the total number of surface bubbles in the previous frame image, and dividing the absolute value of the difference by the physical time interval between two adjacent image samplings to obtain the bubble number change rate.

[0010] This invention effectively eliminates visual interference caused by light fluctuations and horizontal bubble slippage by removing reflective areas at the edge of the mold and using morphological opening operations combined with geometric contour constraints to filter connected components. This ensures that the statistical target corresponds only to the physical target of the real ruptured and escaped bubble, thereby obtaining a pure and reliable dynamic change rate of surface bubbles. This provides an accurate apparent data benchmark for subsequent calculation of the time series differences between internal and external states.

[0011] Preferably, the step of extracting features from the transient active power sequence to obtain the active power change rate characterizing the internal rheological state includes: smoothing the transient active power sequence using a moving average filtering algorithm, calculating the difference between the smoothed transient active power at the current moment and the smoothed transient active power at the previous sampling moment, and dividing the difference by the physical time interval between two adjacent samplings to obtain the active power change rate.

[0012] This invention utilizes a moving average filtering algorithm to smooth transient active power and calculate the difference between adjacent sampling times. It can filter out high-frequency electromagnetic interference and noise generated by grid voltage fluctuations, highlight the true work envelope of the excitation motor overcoming internal frictional resistance, thereby stably capturing the resistance loss rate caused by the collapse of the slurry flocculation structure and providing a continuous and interference-resistant internal rheological attenuation index.

[0013] Preferably, determining the bubble rising delay time, which reflects the change in internal rheological state preceding the surface visual phenomenon, includes: extracting the rate of change of active power at each discrete moment within a sliding time window length before the current moment to form a first data sequence; extracting the rate of change of the number of bubbles after shifting backward by a test offset step on the time axis to form a second data sequence; calculating the correlation coefficient between the first data sequence and the second data sequence, and using the correlation coefficient as the time offset correlation; finding the optimal test offset step number corresponding to the time offset correlation reaching its maximum value through traversal calculation; and multiplying the optimal test offset step number by the physical time interval between two adjacent image samples to obtain the bubble rising delay time.

[0014] This invention dynamically aligns the actual time difference between resistance breakdown and surface phenomena under specific working conditions by performing time-displacement traversal and correlation extreme value search on the internal power decay sequence and the surface bubble escape sequence within a historical sliding window. It automatically adapts to the differences in slurry viscosity characteristics caused by different water-cement ratios and ambient temperatures, thereby obtaining the hysteresis time that evolves in real time with the site conditions, providing a time-series matching basis for breaking the limitations of fixed empirical parameters.

[0015] Preferably, the step of finding the optimal number of test offset steps that maximizes the time offset correlation through traversal calculation includes: limiting the lower limit of the traversal range of the test offset steps to 0; dividing the depth of the tube mold by the minimum empirical velocity of the bubble to obtain the theoretical maximum delay time, and limiting the upper limit of the traversal range of the test offset steps to the theoretical maximum delay time; and finding the optimal number of test offset steps within the set traversal range.

[0016] Preferably, the step of extrapolating the current rate of change of active power using the bubble rising delay time to predict the potential drag loss in the future delay blind zone includes: multiplying the absolute value of the rate of change of active power at the current moment by the bubble rising delay time, constructing a first-order linear extrapolation model based on the current rheological gradient, obtaining the potential predicted attenuation of power in the future delay blind zone, and using the potential predicted attenuation as the potential drag loss.

[0017] Preferably, the over-vibration risk index satisfies the expression: In the formula, This represents the over-vibration risk index at the current moment. This represents the current rate of change of active power. This is the time delay for the bubbles to rise; This is the initial maximum power; This represents the current transient active power. It is the absolute value symbol; It is a function for maximizing the value; This is the preset dissipation dead zone threshold.

[0018] This invention constructs a ratio relationship between the extrapolated predicted future resistance attenuation and the total historical dissipated resistance, and uses the maximum value function to forcibly extract the dissipation dead zone threshold as the lower limit of the denominator. This prevents the risk of exponential amplification and false shutdown caused by the extremely small attenuation of transient active power in the early stage of vibration, which leads to the denominator approaching 0. It can automatically amplify the structural instability signal when the slurry is close to complete liquefaction as the vibration process progresses, thereby providing a unified and sensitive shutdown judgment benchmark at the critical point of the bottom coarse aggregate settlement.

[0019] Preferably, the method for obtaining the dissipation dead zone threshold is as follows: extract the smoothed transient active power sequence after the excitation motor enters the stable vibration start stage; calculate the maximum absolute value of the maximum deviation of the transient active power at each sampling point in the smoothed transient active power sequence relative to the initial highest power; multiply the maximum absolute value of the deviation by a preset safety margin coefficient to obtain the preset dissipation dead zone threshold.

[0020] Preferably, the method for obtaining the minimum empirical velocity of the bubble is as follows: by observing the limit of the upward movement distance and the physical time of the microbubbles in a slurry with a specific ratio through a transparent acrylic scale model test, and calibrating the ratio of the limit of the upward movement distance to the physical time as the minimum empirical velocity of the bubble.

[0021] The beneficial effects of this invention are as follows: By synchronously acquiring images of the forming surface of the pipe segment and the transient active power of the excitation motor, this invention dynamically calculates the hysteresis time of the internal rheological decay leading the surface bubble rupture, and extrapolates the future resistance loss based on this time dimension difference to construct an over-vibration risk index. When the over-vibration risk index reaches the critical state, it actively triggers a speed reduction and shutdown. This transforms the vibration control logic from passively waiting based on apparent phenomena to actively predicting the collapse of internal resistance, effectively blocking the ineffective vibration of the bottom layer caused by the visual delay blind zone, eliminating the hidden danger of coarse aggregate settling and segregation at the bottom of the pipe segment, and realizing adaptive shutdown control that adjusts according to the actual liquefaction state of concrete, thereby improving the forming density and structural safety performance of the precast pipe segment. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating an adaptive optimization method for segment vibration parameters based on vibration process data according to the present invention. Figure 2 This is a schematic diagram illustrating the change in the active power of segment vibration in this invention; Figure 3 This is a schematic diagram illustrating the change of the over-vibration risk index in this invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0025] This invention discloses an adaptive optimization method for segment vibration parameters based on vibration process data, referring to... Figure 1 This includes steps S1 to S4: S1: Obtain the formed image sequence and transient active power sequence.

[0026] It should be noted that, due to the high yield stress exhibited in the initial vibration stage after the concrete segment is poured, the exciter motor needs to output a large power to break the cement flocculation structure. As vibration proceeds, the concrete gradually liquefies, causing its resistance to the exciter motor to decrease rapidly. At the same time, the air bubbles accumulated inside overcome the viscosity and begin to slowly float to the surface and burst. This results in a serious time delay misalignment between the visual observation of the surface and the actual internal liquefaction state. Therefore, this invention synchronously acquires the active power sequence characterizing the rheological state of the internal slurry and the forming image sequence characterizing the bursting phenomenon of surface air bubbles under the same time reference, laying a physical data foundation for subsequent exploration of the spatiotemporal differences between the inside and outside and preventing over-vibration of the bottom layer.

[0027] Specifically, an industrial camera is deployed on a high-frequency vibration isolation bracket directly above the segment forming mold. RGB images of the concrete surface are continuously acquired at a fixed sampling frequency and converted into grayscale images to form a forming image sequence. The forming image sequence refers to a continuous set of grayscale images that record the dynamic process of surface slurry rolling and bubble bursting. The system clock of the industrial control motherboard is used to stamp the first physical timestamp on each frame of the forming image sequence.

[0028] Meanwhile, through the data communication interface of the frequency converter directly connected to the excitation motor, the instantaneous voltage sequence and instantaneous current sequence of the three-phase power supply driving the vibration of the tube forming mold are read in real time at the same sampling frequency, and a second physical timestamp with the same reference as the first physical timestamp is added to the instantaneous voltage sequence and instantaneous current sequence in the industrial control motherboard.

[0029] The transient active power sequence is obtained by multiplying the instantaneous voltage sequence and the instantaneous current sequence element by element after alignment based on the second physical timestamp.

[0030] S2: Obtain the rate of change of the number of bubbles and the rate of change of active power.

[0031] It should be noted that, because concrete initially exhibits a high yield stress as a plastic solid, it contains a flocculated structure formed by the mutual adsorption of cement particles and physical interlocking forces between coarse aggregates. When driving the mold to perform periodic vibration, the exciter must overcome extremely high internal frictional resistance to dissipate energy, resulting in an extremely high transient active power sequence at the initial moment. As the high-frequency excitation energy continues to disrupt the flocculated structure of the cement paste, the concrete liquefies and transforms from a plastic solid to a fluid. The frictional resistance between internal particles rapidly dissipates, and the external damping load that the exciter needs to overcome to maintain the same vibration amplitude decreases significantly, leading to a corresponding decrease in the transient active power sequence. Synchronous fallback, therefore, this invention uses the decay rate of transient active power sequence, i.e., the rate of change of active power, to characterize the rate of frictional loss caused by the collapse of the flocculated structure inside the concrete. At the same time, considering that surface bubbles may only undergo positional translation without actual rupture during the segment vibration process, and that there is light fluctuation interference in the industrial site, directly using pixel grayscale difference can easily misjudge background noise as bubble rupture signal, this invention introduces morphological contour detection method to anchor the physical target of real bubbles, and extracts the rate of change of bubble number and the rate of change of active power to characterize the pure dynamic trend, placing the macroscopic real visual phenomenon and the microscopic mechanical decay process on an equivalent time domain benchmark for comparison.

[0032] Specifically, for the molding image sequence, the reflective areas of the mold's metal edges are first cropped and removed, leaving only the internal slurry area as the effective observation area. The Otsu thresholding algorithm is used to binarize the effective observation area, and morphological opening operations are used to remove small, isolated light spot noise. Then, the contours of connected components in the binarized image are extracted, and the total number of connected components that meet the preset circularity constraint and area empirical interval distribution constraint is counted, which is taken as the total number of surface bubbles in the current frame image. The difference between the total number of surface bubbles in the current frame image and the total number of surface bubbles in the previous frame image is calculated, and the absolute value of this difference is divided by the physical time interval between two adjacent image samplings to obtain the bubble number change rate. The bubble number change rate reflects the dynamic change speed of the actual overflow or bursting of surface bubbles at the current moment, eliminating static visual deception interference from bubble horizontal slippage and light flicker.

[0033] For the preset roundness and area experience range, implementers can extract standard bubble still image samples from multiple sets of historical vibration videos, and use statistical normal distribution fitting to calculate the upper and lower limit thresholds of their geometric characteristics for offline calibration. The reason for choosing this statistical method is that the surface tension of bubbles varies due to different water-cement ratios. Based on real statistical samples, it can approximate the geometric and physical contour boundary of actual bubbles to the greatest extent, thereby improving the accuracy of connected domain screening. In other embodiments, implementers can dynamically scale and correct the preset roundness and area experience range according to the changes in slurry tension caused by the real-time dosage of water-reducing agent.

[0034] Furthermore, for the transient active power sequence, to eliminate the noise caused by high-frequency electromagnetic interference and grid voltage fluctuations, a moving average filtering algorithm is used to smooth the transient active power and extract the power envelope reflecting the true work trend. The difference between the smoothed transient active power at the current moment and the smoothed transient active power at the previous sampling moment is calculated, and this difference is divided by the physical time interval between two adjacent samplings to obtain the rate of change of active power. As the flocculated structure inside the concrete collapses rapidly, the coarse aggregate gradually loses the viscous support of the cement paste and tends to sink under gravity. This microstructural instability process directly leads to a sharp decrease in the internal frictional damping that the excitation motor needs to overcome, resulting in an accelerated decline in transient active power. Therefore, the absolute value of the rate of change of active power measures the rate of loss of the aforementioned frictional damping, directly mapping the rate of disintegration of the rheological state inside the concrete from the perspective of energy dissipation, providing a mechanical evolution basis for predicting the trend of coarse aggregate sinking and segregation in advance.

[0035] Because the electrical signal of the excitation motor is accompanied by high-frequency switching glitches and noise, it can easily mask the true low-frequency trend reflecting the change in slurry resistance. In order to effectively filter out high-frequency glitches while preserving the effective low-frequency trend to the greatest extent and to prevent the true signal from being excessively smoothed out, the filtering window of the moving average filtering algorithm must stably cover the motor power supply cycle in the time dimension. Based on the above reasons, this embodiment associates the filtering window length of the moving average filtering algorithm with a preset time span. The implementer calculates the filtering window length by multiplying the sampling frequency by the preset time span. Specifically, to meet the condition of stably covering the motor power supply cycle, the empirical value of the preset time span is 0.1 seconds to 0.2 seconds. In other embodiments, the implementer can adaptively scale the preset time span according to the real-time signal-to-noise ratio of the power envelope, and then update the filtering window length.

[0036] S3: Calculate the correlation of the rate of change to determine the delay time for the bubble to rise.

[0037] It should be noted that since the complete collapse of internal frictional resistance will inevitably occur much earlier than the large-scale rupture and escape of surface bubbles on the time axis, the surface visual phenomenon has an irresistible physical delay deception. Therefore, this invention extracts historical data sequences and applies a discrete time axis translation transformation to them. By calculating the highest correlation coefficient between the rate of change of active power and the rate of change of the number of bubbles through misalignment, the specific bubble rising delay time under the current working condition is derived in reverse. This is used as the core boundary condition for subsequent anticipatory control to break the illusion of visual delay.

[0038] Specifically, a sliding time window length prior to the current moment is extracted. Within the range covered by this sliding time window, the rate of change of active power at each discrete moment in history is extracted to form a first data sequence. The rate of change of the number of bubbles after shifting backward by a specific number of test offset steps on the time axis is extracted to form a second data sequence. The correlation coefficient between the first and second data sequences is calculated, and this correlation coefficient is used as the time offset correlation at the current number of test offset steps. By continuously changing the number of test offset steps and performing a comprehensive traversal calculation, the number of test offset steps corresponding to the maximum extreme value of the time offset correlation is found, which is taken as the optimal number of test offset steps. During the traversal search for the maximum extreme value, based on the physical causal relationship that changes in internal rheological state necessarily precede the occurrence of surface visual phenomena, the lower limit of the traversal range of the number of test offset steps is limited to 0, and the upper limit of the traversal range of the number of test offset steps is limited to the theoretical maximum delay time obtained by dividing the segment mold depth by the minimum empirical velocity of the bubble. This forces the elimination of spurious correlation peak interference caused by occasional system noise.

[0039] Multiplying the optimal test offset step by the physical time interval between two adjacent image samples yields the bubble rise delay time, which describes the actual hysteresis time experienced by the bubble at the bottom of the segment as it overcomes the frictional resistance of the slurry to reach the surface and burst.

[0040] Regarding the sliding time window length, implementers can set the sliding time window length by statistically analyzing the average time span consumed from the sudden change in resistance to the concentrated escape of surface bubbles during multiple vibration processes. The empirical range is 3 to 8 seconds. Regarding the minimum empirical bubble velocity, implementers can calibrate the minimum empirical bubble velocity by observing the ultimate upward distance and physical time consumed by microbubbles in a specific slurry mix through transparent acrylic scale model tests. The empirical range is 5 to 10 millimeters per second.

[0041] S4: Calculate the vibration risk index for vibration control.

[0042] It should be noted that traditional visual shutdown is based solely on passively waiting for the apparent number of bubbles to stop changing significantly, which completely ignores the bubble rising delay time. As a result, the bottom layer of the tube segment has already experienced ineffective over-vibration damage equivalent to the delay time. Therefore, this invention breaks the conventional passive feedback mechanism. Based on the calculated bubble rising delay time, it performs a forward evolution calculation on the current active power change rate in the time dimension to obtain the over-vibration risk index. At the moment when the internal structure is about to undergo irreversible segregation, a shutdown command is issued in advance, forming an adaptive protection closed loop that blocks the over-vibration blind zone.

[0043] Specifically, after the exciter motor starts and passes the initial current surge stage, the initial maximum power of the stable vibration start stage is recorded. The initial maximum power represents the actual power peak recorded after the exciter motor enters the stable vibration start stage, reflecting the energy dissipation level of the segment concrete under high yield stress in the early stage of vibration, and serving as a physical benchmark for subsequent calculation of the internal resistance loss progress of the segment concrete.

[0044] Based on the initial maximum power, the current transient active power, the bubble rise delay time, and the current rate of change of active power, calculate the over-vibration risk index: ; In the formula, The current vibration risk index reflects the degree of damage that the internal structure will suffer within the visual blind spot where the bubble will overflow in the future. The current rate of change of active power represents the change in current transient active power over physical time. The bubble rise delay time represents the actual physical time it takes for the bubble to overcome resistance and reach the surface, as measured by the test. The initial maximum power represents the actual peak power recorded after the excitation motor enters the stable vibration initial stage; The current transient active power represents the time node. The instantaneous value of active power after time-smoothing; It is the absolute value symbol; It is a function for maximizing the value; The preset dissipation dead zone threshold represents the inherent background power fluctuation limit before the initial flocculation structure inside the segment concrete is substantially broken. It is used to prevent the denominator from approaching 0 due to the extremely small attenuation of transient active power in the early stage of vibration, which could lead to the mathematical singularity amplification of the over-vibration risk index and the risk of accidental shutdown.

[0045] Among them, the dissipation dead zone threshold Calibration is performed using extreme value statistics based on initial steady-state data under historical standard operating conditions. Specifically, the smoothed active power sequence of the exciter motor during the first 3 to 5 seconds of the initial stable phase after passing the initial current surge peak is extracted in real time, and the relative peak power of each sampling point in this sequence is calculated. The maximum absolute value of the deviation is calculated, and this maximum absolute value is multiplied by a preset safety margin coefficient. The result is used as the dissipation dead zone threshold. In this embodiment, the safety margin factor is set to 1.2 to fully cover the background noise and prevent non-physical divergence of the initial over-vibration risk index and false alarms. In other embodiments, the implementer can flexibly adjust this factor within the range of 1.2 to 1.8 according to the actual electromagnetic interference level on site. At this point, it indicates that the high yield stress skeleton inside the tunnel segment has not yet undergone substantial collapse. The system then uses a maximum value function to forcibly extract... As a denominator, it physically absorbs early electromagnetic noise and minor rheological interference, ensuring the stability of the algorithm in the low-dissipation stage.

[0046] Among them, the absolute value of the current rate of change of active power. Delay time for bubble to rise Multiplication was used to construct a first-order linear extrapolation model based on the current rheological gradient, which calculated the potential predicted power attenuation during the delay blind period. This value is based on the most severe transient attenuation trend and indirectly maps the structural failure trend before the internal resistance is fully released; in the denominator The total dissipated power characterizing the current progress of resistance loss is the current transient active power as the concrete gradually transforms from a plastic solid to a fluid during the continuous vibration process of the segment mold. The subsequent continuous decline leads to a decrease in the total dissipated power in the denominator, which represents the current progress of resistance loss. As the accumulation continues to increase, during this fluidization transition, if the numerator term calculates the potential predicted power attenuation during the delay dead zone... The relatively large value indicates that the coarse aggregate at the bottom of the segment is losing its viscous support from the slurry and is rapidly settling, leading to a relatively larger ratio of the numerator to the denominator, and consequently increasing the over-vibration risk index. The value increases accordingly; through the division relationship between the above physical parameters, the ratio of potential resistance loss in the future delay blind zone to the total historical dissipated resistance is actually calculated. Using this ratio, the destructive trend of internal coarse aggregate sinking can be objectively measured in advance before the visual phenomenon of the complete disappearance of surface bubbles occurs, thereby blocking the danger of bottom layer over-vibration induced by visual appearance delay.

[0047] Furthermore, the over-vibration risk index Real-time transmission to the programmable logic controller, when the vibration risk index When the speed exceeds the preset shutdown control threshold, the programmable logic controller sends a speed reduction and shutdown command to the frequency converter to stop the vibration operation of the excitation motor.

[0048] Regarding the shutdown control threshold, implementers can extract and calibrate it through on-site extreme failure tests. The specific calibration method is as follows: under a specific segment mold and the same batch of concrete mix ratio, multiple stepped test vibrations are performed and the data of the entire process is recorded. The machine is forcibly stopped at distinctly different points where the over-vibration risk index increases. After the segment is formed and demolded, the density and compressive strength of the bottom of the segment are tested in detail using the ultrasonic rebound comprehensive method. Finally, the maximum over-vibration risk index when the bottom strength fully meets the design requirements and there are no signs of coarse aggregate settling and segregation inside is selected as the conventional shutdown control threshold under this specific working condition. In other embodiments, implementers can dynamically adjust and set the shutdown control threshold based on the real-time concrete slump test results.

[0049] For example, Figure 2 This is a schematic diagram illustrating the change in the active power of segment vibration in this invention. Figure 2 It can be seen that the active power decreases steadily with the liquefaction process in the early stage of vibration. When the active power decreases rapidly due to the instability and sinking of coarse aggregate in the later stage of vibration, the stop command point can be locked at the moment of the sudden decrease. This shows that the present invention can keenly capture the instability characteristics of the internal rheological structure and achieve a high-precision early warning response that surpasses traditional visual feedback.

[0050] Figure 3 This is a schematic diagram illustrating the change in the over-vibration risk index in this invention. Figure 3 It can be seen that when the over-vibration risk index exceeds the shutdown control threshold, the system immediately outputs a shutdown command, which can block the over-vibration blind zone at the bottom of the segment.

Claims

1. An adaptive optimization method for segment vibration parameters based on vibration process data, characterized in that, include: Simultaneously acquire and record the forming image sequence of the dynamic process on the surface of the tube segment, as well as the transient active power sequence that drives the vibration of the tube segment forming mold; Feature extraction is performed on the formed image sequence to obtain the rate of change of the number of bubbles, which characterizes the surface visual phenomenon; feature extraction is performed on the transient active power sequence to obtain the rate of change of active power, which characterizes the internal rheological state. Discrete time axis translation transformation and correlation calculation were performed on the rate of change of active power and the rate of change of bubble number within the historical time window to determine the bubble rising delay time, which reflects the change of internal rheological state before the surface visual phenomenon. By extrapolating the current rate of change of active power in the time dimension using the bubble rising delay time, the potential drag loss in the future delay blind zone is predicted, and the over-vibration risk index is calculated by combining the initial maximum power and the current transient active power. When the over-vibration risk index exceeds the shutdown control threshold, a speed reduction and shutdown command is output to stop the vibration operation of the excitation motor.

2. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 1, characterized in that, The synchronous acquisition of the forming image sequence recording the dynamic process of the tube segment surface, and the transient active power sequence driving the vibration of the tube segment forming mold, includes: The system clock is used to assign a first physical timestamp to each frame of the continuously acquired molding image sequence; the instantaneous voltage sequence and instantaneous current sequence of the three-phase power supply driving the vibration of the tube forming mold are read in real time, and a second physical timestamp with the same reference as the first physical timestamp is assigned to them; the instantaneous voltage sequence and instantaneous current sequence aligned based on the second physical timestamp are multiplied element-wise to obtain the transient active power sequence.

3. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 1, characterized in that, The step of extracting features from the formed image sequence to obtain the rate of change in the number of bubbles, which characterizes the surface visual phenomenon, includes: The reflective areas of the mold metal edge in the forming image sequence are removed, and the internal slurry area is retained as the effective observation area. The effective observation area is binarized and morphological opening is performed to extract the contour of the connected components. The total number of connected components that meet the preset circularity constraint and area empirical interval constraint is taken as the total number of surface bubbles in the current frame image. The difference between the total number of surface bubbles in the current frame image and the total number of surface bubbles in the previous frame image is calculated, and the absolute value of the difference is divided by the physical time interval between two adjacent image samplings to obtain the bubble number change rate.

4. The adaptive optimization method for segment vibration parameters based on vibration process data according to claim 1, characterized in that, The process of extracting features from the transient active power sequence to obtain the rate of change of active power characterizing the internal rheological state includes: The transient active power sequence is smoothed using a moving average filtering algorithm. The difference between the smoothed transient active power at the current time and the smoothed transient active power at the previous sampling time is calculated, and the difference is divided by the physical time interval between two adjacent sampling times to obtain the active power change rate.

5. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 1, characterized in that, The determination of the bubble rise delay time, which reflects the change in internal rheological state preceding the surface visual phenomenon, includes: Within the range covered by a sliding time window before the current moment, the rate of change of active power at each discrete moment is extracted to form a first data sequence; the rate of change of the number of bubbles after shifting backward by a test offset step on the time axis is extracted to form a second data sequence; the correlation coefficient between the first data sequence and the second data sequence is calculated, and this correlation coefficient is used as the time offset correlation; the optimal test offset step number is found by iterative calculation to maximize the time offset correlation; the optimal test offset step number is multiplied by the physical time interval between two adjacent image samples to obtain the bubble rising delay time.

6. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 5, characterized in that, The step of finding the optimal number of test offset steps that maximizes the time offset correlation through traversal calculation includes: The lower limit of the traversal range of the test offset steps is set to 0; the theoretical maximum delay time is obtained by dividing the depth of the tube mold by the minimum empirical velocity of the bubble, and the upper limit of the traversal range of the test offset steps is set to the theoretical maximum delay time; the optimal number of test offset steps is found within the set traversal range.

7. The adaptive optimization method for segment vibration parameters based on vibration process data according to claim 1, characterized in that, The method of extrapolating the current rate of change of active power in the time dimension using the bubble rising delay time to predict potential drag loss in the future delay blind zone includes: The absolute value of the rate of change of active power at the current moment is multiplied by the bubble rising delay time to construct a first-order linear extrapolation model based on the current rheological gradient, and the potential predicted attenuation of power in the future delay blind zone is obtained. The potential predicted attenuation is then used as the potential drag loss.

8. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 1, characterized in that, The over-vibration risk index satisfies the expression: ； In the formula, This represents the over-vibration risk index at the current moment. This represents the current rate of change of active power. This is to delay the time for the bubbles to rise. This is the initial maximum power; This represents the current transient active power. It is the absolute value symbol; It is a function for maximizing the value; This is the preset dissipation dead zone threshold.

9. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 8, characterized in that, The method for obtaining the dissipation dead zone threshold is as follows: Extract the smoothed transient active power sequence after the excitation motor enters the stable vibration start stage; calculate the maximum absolute value of the transient active power at each sampling point in the smoothed transient active power sequence relative to the initial highest power; multiply the maximum absolute value of the deviation by the preset safety margin coefficient to obtain the preset dissipation dead zone threshold.

10. The method for adaptive optimization of segment vibration parameters based on vibration process data according to claim 6, characterized in that, The method for obtaining the minimum empirical velocity of the bubble is as follows: The ultimate upward distance and physical time of microbubbles in a specific slurry were observed through a transparent acrylic scale model test, and the ratio of the ultimate upward distance to the physical time was calibrated as the minimum empirical velocity of the bubble.

Images