Wind power mixed tower concrete pouring perpendicularity deviation control method and system

Abstract

The invention relates to the technical field of wind power mixed tower construction, in particular to a wind power mixed tower concrete pouring perpendicularity deviation control method and system. According to the method, multi-source monitoring data of a mixed tower template system is obtained, an extended Kalman filtering algorithm is adopted to fuse the data to obtain a six-degree-of-freedom pose state, real perpendicularity is calculated in combination with environment compensation, and a long-short-term memory network algorithm is adopted to predict a perpendicularity deviation trend; executing first-stage regulation and control according to the real perpendicularity to adjust pouring parameters, executing second-stage regulation and control according to the inclination angle deviation to adjust a supporting system, and executing third-stage regulation and control to re-plan a process when a predicted value exceeds a limit; and closed-loop control is formed according to a control period, regulation and control data are recorded, and regulation and control parameters are optimized by adopting a reinforcement learning algorithm. According to the invention, real-time monitoring and prediction and accurate regulation and control of the perpendicularity are realized, and the construction quality and efficiency of the mixed tower are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind power mixed tower construction, in particular to a wind power mixed tower concrete pouring verticality deviation control method and system. BACKGROUND

[0002] As a key supporting structure of a wind turbine generator, the verticality precision of a wind power mixed tower directly affects the operation safety and service life of the whole machine, and high-precision verticality control is a core technical problem of mixed tower construction.

[0003] The traditional mixed tower construction adopts a discrete control mode of "pouring-maintenance-measurement-adjustment", which has the following technical defects: the verticality measurement is performed after the initial setting of the concrete, and the deviation cannot be effectively adjusted when it is found, and the feedback is seriously lagged; the formwork support is adjusted depending on the experience of workers, and there is a lack of quantitative basis and precise control means; it is difficult to establish the correlation between the pouring parameters and the verticality because the influence of environmental factors such as temperature and wind load cannot be compensated in real time; and the post-correction often requires the removal of concrete, which is costly and affects the structural integrity. The existing technology cannot meet the requirements of wind power engineering for construction precision. SUMMARY

[0004] The present application provides a wind power mixed tower concrete pouring verticality deviation control method and system, which solves the problems of feedback lag and insufficient control precision of the traditional method, and realizes real-time and accurate control of the verticality during construction.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions: The present application provides a wind power mixed tower concrete pouring verticality deviation control method, which comprises: S100: obtaining formwork posture data, support force data, formwork displacement data and concrete side pressure data of a mixed tower formwork system; S200: using an extended Kalman filter algorithm to fuse the formwork posture data, support force data, formwork displacement data and concrete side pressure data to obtain a six-degree-of-freedom pose state of the mixed tower formwork system, and combining with environmental influence factor compensation calculation to obtain a real verticality; using a long short-term memory network algorithm to analyze historical verticality data to obtain a verticality deviation prediction value in a future period; S300: performing primary regulation and control according to the real verticality to adjust the pouring speed, pouring position and vibration intensity; performing secondary regulation and control according to the inclination deviation in the six-degree-of-freedom pose state to adjust the support force of the hydraulic support point and the cable tension; and performing tertiary regulation and control when the verticality deviation prediction value exceeds an alarm value to re-plan the pouring sequence and construction joint position; S400: repeat S100-S300 according to the preset control cycle to form a closed-loop control; record the control data to build a case library, and automatically optimize the control parameters based on the historical control effect by using a reinforcement learning algorithm.

[0006] As a preferred technical solution of the present application, S100 comprises: In the four corners and the middle of the tower template system, an inclination sensor is arranged to obtain the template posture data; In the support system and the cable system of the tower template system, a strain sensor is arranged to obtain the support force data; A laser emitter is arranged at the reference point, and a target is arranged in the tower template system, and the template displacement data is obtained by laser ranging; A pressure sensor is arranged inside the tower template system to obtain the concrete side pressure data.

[0007] As a preferred technical solution of the present application, the calculation step of the true verticality comprises: Obtain the measured verticality based on the six-degree-of-freedom pose state; Obtain the current temperature data, wind load data and construction load data; Calculate the temperature compensation, wind load compensation and construction load compensation respectively; Add the measured verticality to the temperature compensation, wind load compensation and construction load compensation to obtain the true verticality.

[0008] As a preferred technical solution of the present application, the primary control comprises: According to the difference between the true verticality and the verticality deviation threshold, the adjusted pouring speed is calculated according to the following formula: ; Wherein, is the planned pouring speed, is the speed adjustment coefficient, is the current true verticality, is the verticality deviation threshold; Based on the force balance analysis of the tower template system, the optimal pouring point is calculated, and the pouring position is adjusted according to the optimal pouring point; According to the difference between the true verticality and the verticality deviation threshold, the vibration area and the vibration intensity are determined.

[0009] As a preferred technical solution of the present application, the secondary control comprises: Extract the inclination deviation of each support point in the six-degree-of-freedom pose state, and according to the inclination deviation of each support point and the corresponding influence area, the adjusted support force of each hydraulic support point is calculated according to the following formula: ; in, For the first The adjusted support strength of each support point Basic support, This is the support force adjustment coefficient. For the first Inclination deviation of each support point For the first The area affected by each support point; Adjust the support force of each hydraulic support point according to the adjusted support force, and establish the transfer function between cable tension and verticality; The tension adjustment amount of each cable is calculated based on the transfer function and the actual verticality. Adjust the tension of each cable according to the aforementioned tension adjustment amount.

[0010] As a preferred embodiment of the present invention, the three-level regulation includes: Determine whether the predicted verticality deviation value exceeds the warning value; When the predicted verticality deviation exceeds the warning value, concrete pouring shall be suspended. The pouring sequence of the remaining concrete will be replanned based on the current distribution of poured concrete and the actual verticality. Adjust the construction joint position based on the predicted verticality deviation value; Adjust the maintenance start time according to the actual verticality.

[0011] As a preferred embodiment of the present invention, the optimized control parameters include: Record the actual verticality, adjustment action, and verticality change after each adjustment as adjustment data; The control data is stored in the case library; Extract historical control data from the aforementioned case library; The reward value is calculated based on the historical regulation data using a reinforcement learning algorithm; The speed adjustment coefficient and the support force adjustment coefficient are updated based on the reward value.

[0012] This invention also proposes a verticality deviation control system for concrete pouring in wind power hybrid towers, comprising: The data acquisition module is used to acquire template posture data, support force data, template displacement data, and concrete lateral pressure data of the mixed tower formwork system. The state calculation and prediction module is used to fuse the template attitude data, support force data, template displacement data and concrete lateral pressure data using the extended Kalman filter algorithm to obtain the six-degree-of-freedom pose state of the mixed tower template system, and to calculate the true verticality by combining environmental influence factor compensation; and to analyze the historical verticality data using the long short-term memory network algorithm to obtain the predicted value of verticality deviation for future periods. The graded control module is used to perform first-level control based on the actual verticality, adjusting the pouring speed, pouring position, and vibration intensity; perform second-level control based on the tilt angle deviation in the six-degree-of-freedom pose state, adjusting the support force of the hydraulic support point and the tension of the cable; and perform third-level control when the predicted verticality deviation exceeds the warning value, replanning the pouring sequence and construction joint position. The closed-loop optimization module is used to repeatedly execute the data acquisition module to the hierarchical control module to form a closed-loop control according to a preset control cycle; it records the control data to build a case library, and uses a reinforcement learning algorithm to automatically optimize the control parameters based on historical control effects.

[0013] The beneficial effects of this invention are: 1. This invention eliminates the influence of temperature, wind load, and other interferences on measurements by fusing multi-source sensor data through extended Kalman filtering and combining it with dynamic compensation for environmental factors, thereby obtaining accurate true verticality. By combining long short-term memory networks with time-series analysis of historical data to predict deviation trends, quality control is transformed from post-measurement to real-time monitoring and predictive early warning, achieving proactive preventative control.

[0014] 2. This invention constructs a three-tiered progressive control strategy encompassing pouring process adjustment, support system compensation, and construction plan replanning, employing differentiated control methods for varying degrees of verticality deviation. Through a closed-loop control mechanism, the three levels of control work synergistically, achieving precise response throughout the entire process, from fine-tuning to emergency responses. Compared to traditional single-control methods, this approach demonstrates greater adaptability and effectiveness.

[0015] 3. This invention constructs a case library by recording control process data and uses a reinforcement learning algorithm to automatically optimize key parameters such as speed adjustment coefficient and support force adjustment coefficient based on historical control effects. The system has self-learning capabilities, and the control strategy is continuously improved with the accumulation of construction experience, realizing a technological upgrade from fixed parameter control to intelligent adaptive control. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of a method for controlling the verticality deviation of concrete pouring in a wind power hybrid tower according to the present invention. Figure 2 This is a structural schematic diagram of a method for controlling the verticality deviation of concrete pouring in a wind power hybrid tower according to the present invention. Detailed Implementation

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0018] Example 1: As Figure 1 As shown, the present invention provides a method for controlling the verticality deviation of concrete pouring in wind power hybrid towers, comprising: S100: Acquire template posture data, support force data, template displacement data, and concrete lateral pressure data of the mixed tower formwork system; Further, S100 includes: Tilt sensors are arranged at the four corners and the middle of the mixed tower template system to acquire the template attitude data; Strain sensors are arranged in the support system and cable system of the mixed tower formwork system to obtain the support force data; A laser emitter is set at a reference point, and a target is set in the mixed tower template system. The template displacement data is obtained by laser ranging. Pressure sensors are arranged inside the concrete formwork system to obtain the concrete side pressure data.

[0019] Specifically, in this embodiment, a multi-source sensor array is deployed in the mixed tower formwork system to collect monitoring data in real time during the construction process.

[0020] Acquiring template attitude data: Inclination sensors are arranged at the four corners and the center of the mixed tower template system to acquire template attitude data. The inclination sensors monitor the tilt angle of the template in two orthogonal directions and output the tilt angle value at each measuring point.

[0021] Acquiring Support Force Data: Strain sensors are deployed in the support and cable systems of the hybrid tower formwork system to acquire support force data. At the hydraulic support points, strain sensors monitor the strain values ​​of the support rods, which are then converted into support force based on the strain-stress relationship. In the cable system, tension sensors directly monitor the tension values ​​of each cable.

[0022] Acquiring template displacement data: A laser emitter is set at a reference point, and a target is set in the mixed tower template system. Template displacement data is acquired through laser ranging. The laser ranging system measures the distance change of the target relative to the reference point to obtain the displacement of each measuring point in the horizontal and vertical directions.

[0023] Acquiring concrete lateral pressure data: Pressure sensors are arranged inside the concrete formwork system to acquire concrete lateral pressure data. The pressure sensors are arranged in layers along the height and circumference of the formwork to monitor the distribution of lateral pressure applied to the formwork in real time during concrete pouring.

[0024] All sensor data is collected and transmitted uniformly through a data acquisition system, providing input for subsequent data fusion and verticality calculation.

[0025] S200: The extended Kalman filter algorithm is used to fuse the template attitude data, support force data, template displacement data and concrete lateral pressure data to obtain the six-degree-of-freedom pose state of the mixed tower template system, and the true verticality is obtained by combining environmental influence factor compensation calculation; the long short-term memory network algorithm is used to analyze the historical verticality data to obtain the predicted value of verticality deviation for future periods. Furthermore, the calculation steps for the true verticality include: The verticality is measured based on the six-degree-of-freedom pose state. Obtain current temperature data, wind load data, and construction load data; Calculate the temperature compensation, wind load compensation, and construction load compensation separately; The measured verticality is added to the temperature compensation, wind load compensation, and construction load compensation to obtain the true verticality.

[0026] Specifically, the extended Kalman filter algorithm is used to fuse the data acquired in S100. The extended Kalman filter algorithm establishes a state-space model of the hybrid tower formwork system, takes the formwork attitude data, support force data, formwork displacement data, and concrete lateral pressure data as observation inputs, and estimates the six-degree-of-freedom pose state of the hybrid tower formwork system through iterative calculations in prediction and update steps.

[0027] The six-degree-of-freedom pose state includes three translational degrees of freedom and three rotational degrees of freedom, specifically: X-direction displacement, Y-direction displacement, Z-direction displacement, rotation angle around the X-axis (pitch angle), rotation angle around the Y-axis (roll angle), and rotation angle around the Z-axis (yaw angle). Among these, the rotation angles around the X-axis and Y-axis reflect the tilt state of the template and are key parameters for calculating verticality.

[0028] The extended Kalman filter algorithm can effectively fuse data from different types of sensors, reduce the measurement error of a single sensor, and improve the accuracy and robustness of pose estimation.

[0029] The measured perpendicularity is obtained based on the six-degree-of-freedom pose state. The measured perpendicularity is calculated based on the rotation angles around the X-axis and Y-axis, representing the deviation angle of the template relative to the vertical direction.

[0030] Environmental factors such as temperature changes, wind loads, and construction loads can affect sensor measurements and formwork deformation, necessitating compensation and correction for the measured verticality. Specific steps include: Acquire current temperature data, wind load data, and construction load data. Temperature data is acquired through a temperature sensor, wind load data is acquired through an anemometer, and construction load data is calculated based on the current pouring progress and the self-weight of the concrete.

[0031] Calculate the temperature compensation, wind load compensation, and construction load compensation separately. The temperature compensation calculation considers the impact of temperature changes on the thermal expansion and contraction of the formwork and on the zero-point drift of the sensors. Based on the difference between the current temperature and the calibrated temperature, combined with the linear expansion coefficient of the formwork material and the temperature coefficient of the sensors, calculate the displacement and tilt angle measurement deviation caused by temperature, and convert them into verticality compensation. The wind load compensation calculation considers the horizontal force and overturning moment exerted on the formwork by wind pressure. Based on wind speed and direction data, combined with the windward area and centroid height of the formwork, calculate the formwork deformation and tilt angle caused by wind load, as the wind load compensation. The construction load compensation calculation considers the formwork deformation caused by the self-weight of concrete and the loads from construction equipment. Based on the current pouring height and concrete density, calculate the lateral pressure distribution, combined with the formwork stiffness and the stiffness of the support system, calculate the formwork deformation and tilt angle changes caused by the load, as the construction load compensation.

[0032] The measured verticality is added to the temperature compensation, wind load compensation, and construction load compensation to obtain the true verticality. The calculation formula is: ; in, For true verticality, For measuring perpendicularity based on six-degree-of-freedom pose state, This is the temperature compensation amount. For wind load compensation, This is the amount of compensation for construction load. It is a time variable.

[0033] The true verticality eliminates the interference of environmental factors and truly reflects the actual verticality status of the mixed tower template system, providing an accurate basis for subsequent control decisions.

[0034] A Long Short-Term Memory (LSTM) network algorithm is used to perform time-series analysis on historical verticality data to predict future verticality trends. Historical verticality data is arranged in time series to construct training samples for the LTM network. The training samples include verticality data within a past time window as input and corresponding future verticality changes as output labels. The LTM network model is trained using these training samples. The LTM network, through its unique gating mechanism, can capture long-term dependencies in time-series data, effectively learning the dynamic patterns of verticality changes. The trained LTM network model is then used to process current verticality data, outputting predicted verticality deviation values ​​for future periods. These predicted deviation values ​​represent the expected deviation of the template verticality from the standard value at a future point in time.

[0035] By predicting vertical deviation, the risk of exceeding vertical limits can be identified in advance, providing an early warning basis for emergency response in the three-level control system and realizing the transformation from passive correction to proactive prevention.

[0036] S300: Perform Level 1 control based on the actual verticality to adjust the pouring speed, pouring position, and vibration intensity; perform Level 2 control based on the tilt angle deviation in the six-degree-of-freedom pose state to adjust the support force of the hydraulic support point and the tension of the cable; when the predicted verticality deviation exceeds the warning value, perform Level 3 control to re-plan the pouring sequence and construction joint position. Specifically, the present invention adopts a three-level adaptive control strategy, which progressively controls the verticality of the tower construction process from three levels: casting process parameters, support system compensation, and process scheme replanning, based on the degree and predicted trend of verticality deviation, to ensure that the verticality is always under control during the construction process of the mixed tower.

[0037] Furthermore, the primary regulation includes: Based on the difference between the actual verticality and the verticality deviation threshold, the adjusted pouring speed is calculated according to the following formula: ; in, To plan the pouring speed, This is the speed adjustment coefficient. This represents the current actual verticality. This is the verticality deviation threshold. The optimal pouring point is calculated based on the force balance analysis of the mixed tower formwork system, and the pouring position is adjusted according to the optimal pouring point. The vibration area and vibration intensity are determined based on the difference between the actual verticality and the verticality deviation threshold.

[0038] Specifically, when (Current actual verticality) is close to or exceeds When the verticality deviation threshold is reached, the pouring speed is reduced accordingly to slow down the accumulation rate of verticality deviation; when much smaller At the same time, the pouring speed is maintained near the planned value to ensure construction efficiency. The speed adjustment coefficient... Initial values ​​are set based on engineering experience and optimized using reinforcement learning algorithms in S400.

[0039] The optimal pouring point is calculated based on the stress balance analysis of the hybrid tower formwork system, and the pouring position is adjusted according to the optimal pouring point. Specifically, by analyzing the current stress state and verticality deviation direction of the formwork, the concrete pouring position that makes the formwork stress tend to be balanced is calculated. When the verticality deviation on one side of the formwork is large, concrete is poured preferentially on the opposite side or the side with smaller deviation, and the self-balancing of the formwork posture is achieved by adjusting the load distribution.

[0040] The vibration area and vibration intensity are determined based on the difference between the actual verticality and the verticality deviation threshold. When the verticality deviation is large, the vibration intensity on the side of the deviation direction is reduced or vibration is avoided in that area to prevent the formwork displacement caused by vibration from further aggravating the deviation. When the verticality is good, construction is carried out according to the standard vibration process to ensure the compactness of the concrete.

[0041] Furthermore, the secondary regulation includes: Extract the tilt angle deviation of each support point in the six-degree-of-freedom pose state. Based on the tilt angle deviation of each support point and the corresponding influence area, calculate the adjusted support force of each hydraulic support point according to the following formula: ; in, For the first The adjusted support strength of each support point Basic support, This is the support force adjustment coefficient. For the first Inclination deviation of each support point For the first The area affected by each support point; Adjust the support force of each hydraulic support point according to the adjusted support force, and establish the transfer function between cable tension and verticality; The tension adjustment amount of each cable is calculated based on the transfer function and the actual verticality. Adjust the tension of each cable according to the aforementioned tension adjustment amount.

[0042] Specifically, the basic support force Calculated based on the self-weight of the concrete and the self-weight of the formwork, it is used to bear the basic load. When the inclination angle at a certain support point deviates... When the area is large, the supporting force at the support point is increased, applying an upward adjusting torque to the template, causing the template to tend towards a vertical orientation. The affected area... This indicates the effective range of influence of the support point on the template posture. The larger the area of ​​influence, the more significant the adjustment effect of the support point.

[0043] Finally, the support force of each hydraulic support point is adjusted according to the aforementioned support force adjustment amount. Through pressure control of the hydraulic system, the support force of each support point is adjusted to the target value in real time, achieving active compensation through multi-point coordination.

[0044] A transfer function is established between cable tension and verticality. This transfer function describes the influence of cable tension variation on template verticality and is obtained through structural mechanics analysis or fitting of field measurement data.

[0045] The tension adjustment amount of each cable is calculated based on the transfer function and the actual verticality. When there is a deviation in the actual verticality, the required change in cable tension is derived from the transfer function, so that the template returns to a vertical state under the action of the cable force.

[0046] The tension of each cable is adjusted according to the stated tension adjustment amount. The tension value of each cable is adjusted in real time via the cable tightening or loosening device, achieving coordinated control between the cable system and the hydraulic support system.

[0047] Furthermore, the three-level regulation includes: Determine whether the predicted verticality deviation value exceeds the warning value; When the predicted verticality deviation exceeds the warning value, concrete pouring shall be suspended. The pouring sequence of the remaining concrete will be replanned based on the current distribution of poured concrete and the actual verticality. Adjust the construction joint position based on the predicted verticality deviation value; Adjust the maintenance start time according to the actual verticality.

[0048] Specifically, the three-level control is triggered when the verticality is predicted to exceed the limit. It is an emergency control mechanism that avoids verticality loss of control by suspending construction and replanning the process.

[0049] First, determine whether the predicted verticality deviation exceeds the warning value. The warning value is set based on the engineering allowable deviation and safety margin, and is usually set to 80% to 90% of the allowable deviation.

[0050] When the predicted verticality deviation exceeds the warning value, the system automatically triggers a three-level control mechanism to suspend concrete pouring. The purpose of suspending pouring is to prevent construction from continuing as the verticality deviation continues to increase, thus avoiding the accumulation of deviation that could become uncorrectable.

[0051] During the construction suspension period, the pouring sequence of the remaining concrete is replanned based on the current distribution of poured concrete and the actual verticality. By analyzing the distribution of poured concrete and the direction of the current verticality deviation, the areas and order in which subsequent pouring should be prioritized are determined, so that the load distribution generated by the subsequent pouring process is conducive to correcting the current deviation.

[0052] The location of the construction joint is adjusted based on the predicted verticality deviation. The location of the construction joint affects the continuity of concrete pouring and the overall structural integrity. By adjusting the location of the construction joint, necessary correction operations can be performed after the current segment is completed, creating better initial conditions for the construction of the next segment.

[0053] Adjust the curing start time according to the actual verticality. When the verticality deviation is large, appropriately delay or advance the curing start time to control the concrete strength development process, reserve a time window for adjusting the support system, or fix the corrected formwork posture through early strength development.

[0054] After completing the process adjustment plan, concrete pouring was resumed, and primary and secondary control measures were continued to ensure that verticality remained under control during subsequent construction.

[0055] S400: Repeat S100 to S300 according to the preset control cycle to form a closed-loop control; record the control data to build a case library, and use reinforcement learning algorithm to automatically optimize the control parameters based on the historical control effect.

[0056] Furthermore, the optimized control parameters include: Record the actual verticality, adjustment action, and verticality change after each adjustment as adjustment data; The control data is stored in the case library; Extract historical control data from the aforementioned case library; The reward value is calculated based on the historical regulation data using a reinforcement learning algorithm; The speed adjustment coefficient and the support force adjustment coefficient are updated based on the reward value.

[0057] Specifically, steps S100 to S300 are repeated according to a preset control cycle, forming a closed-loop control process of "monitoring-calculation-regulation-verification". The preset control cycle is set according to the pouring speed and the formwork response characteristics, and is preferably 5 to 10 minutes in this embodiment.

[0058] Within each control cycle, the system first executes S100 to acquire the current template posture data, support force data, template displacement data, and concrete lateral pressure data; then executes S200 to fuse and calculate the multi-source data to obtain the six-degree-of-freedom pose state and the true verticality, and predicts the verticality deviation value for future periods; next, it executes S300 to perform corresponding first-level, second-level, or third-level control based on the true verticality and the predicted verticality deviation value; finally, it verifies the control effect and feeds back the verticality change after control to the next control cycle.

[0059] Through periodic closed-loop control, the system can respond to changes in verticality in real time, continuously correct deviations, and ensure that the verticality remains within the allowable range throughout the construction of the mixed tower. Compared with the traditional discrete measurement and adjustment mode, closed-loop control shifts the quality control node from "after the fact" to "during the process," significantly improving the control effect.

[0060] During the closed-loop control process, relevant data from each adjustment is recorded to construct a case library. Specifically, the actual verticality, adjustment action, and change in verticality after adjustment are recorded as adjustment data for each adjustment.

[0061] The actual verticality record reflects the verticality state before adjustment, indicating the degree and direction of deviation of the current template. The adjustment action record records the specific adjustment measures implemented, including adjustments to pouring speed, pouring position, and vibration intensity in the first-level adjustment; adjustments to the support force of each support point and the tension of each cable in the second-level adjustment; and adjustments to the process plan in the third-level adjustment. The verticality change after adjustment records the actual change in verticality after the adjustment is implemented, used to evaluate the adjustment effect.

[0062] The controlled data is stored in a case library. The case library uses a structured data storage method, with each case record containing information such as the state before control, control action parameters, post-control effects, and environmental parameters. As construction progresses, the data accumulated in the case library becomes increasingly rich, providing a data foundation for subsequent intelligent optimization.

[0063] The system employs reinforcement learning algorithms to automatically optimize control parameters based on historical control effects. Reinforcement learning learns the optimal control strategy through the interaction between the agent and the environment, enabling the system to continuously improve its control performance based on historical experience.

[0064] Historical control data is extracted from the aforementioned case library. This historical control data contains the actual effects of different control actions taken under different vertical conditions, and this data constitutes the training samples for reinforcement learning.

[0065] A reinforcement learning algorithm is used to calculate reward values ​​based on the historical control data. These reward values ​​are used to evaluate the effectiveness of the control actions; a positive reward value is assigned when the vertical deviation decreases after control, and a negative reward value is assigned when the vertical deviation increases. The magnitude of the reward value is proportional to the change in vertical deviation; the more significant the improvement in deviation, the higher the reward value.

[0066] The control parameters are updated based on the reward value. The control parameters include the speed adjustment coefficient of the first-level control in S300. and the support adjustment coefficient of secondary regulation Reinforcement learning algorithms iteratively update the algorithm by maximizing the cumulative reward value. and The value of this value allows the adjustment action to produce a better correction effect under the same verticality deviation conditions.

[0067] In this embodiment, and The initial value is set based on engineering experience. As the case study database accumulates and the reinforcement learning algorithm is continuously optimized, and The parameters gradually converge to the optimal value. The optimized control parameters can more accurately reflect the relationship between pouring speed, support force adjustment and verticality changes under the current engineering conditions, achieving adaptive control.

[0068] By combining closed-loop control with learning optimization, this invention not only achieves real-time verticality monitoring and control, but also possesses the ability to learn and continuously improve itself, ensuring that the control accuracy continuously improves with the accumulation of construction experience, and ultimately realizing high-precision, intelligent verticality control for wind power hybrid tower construction.

[0069] Example 2: This example uses a 100-meter-high wind turbine hybrid tower construction project in a coastal area as an example to illustrate the application effect of the method of the present invention in actual engineering. The wind turbine hybrid tower adopts a segmented casting method, with each segment being 3-4 meters high, and the design verticality requirement is H / 1000. The construction site experiences frequent wind speed changes and significant day-night temperature differences.

[0070] The project initially employed traditional construction methods, and measurements taken after the third section was poured revealed a cumulative verticality deviation of 68mm. Key issues included: verticality deviation could only be measured after the concrete had initially set, making adjustment impossible by then; failure to compensate for nighttime temperature variations and wind load effects; and low precision and time-consuming manual adjustments to the support system.

[0071] Starting from paragraph 4, the project adopts a verticality deviation control system for concrete pouring of wind power hybrid towers according to the present invention, including: The data acquisition module is used to acquire template posture data, support force data, template displacement data, and concrete lateral pressure data of the mixed tower formwork system. The state calculation and prediction module is used to fuse the template attitude data, support force data, template displacement data and concrete lateral pressure data using the extended Kalman filter algorithm to obtain the six-degree-of-freedom pose state of the mixed tower template system, and to calculate the true verticality by combining environmental influence factor compensation; and to analyze the historical verticality data using the long short-term memory network algorithm to obtain the predicted value of verticality deviation for future periods. The graded control module is used to perform first-level control based on the actual verticality, adjusting the pouring speed, pouring position, and vibration intensity; perform second-level control based on the tilt angle deviation in the six-degree-of-freedom pose state, adjusting the support force of the hydraulic support point and the tension of the cable; and perform third-level control when the predicted verticality deviation exceeds the warning value, replanning the pouring sequence and construction joint position. The closed-loop optimization module is used to repeatedly execute the data acquisition module to the hierarchical control module to form a closed-loop control according to a preset control cycle; it records the control data to build a case library, and uses a reinforcement learning algorithm to automatically optimize the control parameters based on historical control effects.

[0072] Specifically, during the fifth stage of pouring, the system detected a measured verticality of 31 mm, while simultaneously monitoring a 2°C temperature drop and a wind speed increase to level 4. Through environmental compensation calculations, the corrected actual verticality was 34 mm. The Long Short-Term Memory (LSTM) network predicted that the verticality deviation would approach the warning value within the next hour.

[0073] The system implemented Level 1 control, reducing the pouring speed from 3.0 m³ / h to 2.5 m³ / h and adjusting the pouring position to the side with the smaller deviation. Simultaneously, Level 2 control was implemented, adjusting the hydraulic support force and cable tension according to the inclination deviation of each support point. Ten minutes after the adjustment, the verticality deviation was reduced to 26 mm.

[0074] During the sixth stage of construction, the system predicted that the verticality would exceed the warning value, triggering a level-three control measure, suspending pouring and replanning the remaining pouring sequence, and adjusting the location of the construction joint.

[0075] The system continuously executes closed-loop control according to an 8-minute control cycle. After 8 phases of construction, approximately 400 control records have been accumulated in the case library, and the reinforcement learning algorithm has optimized the speed adjustment coefficient. (Pouring speed adjustment parameters in primary control) and support force adjustment coefficient (Support force adjustment parameters in secondary regulation) enable continuous improvement in regulation accuracy.

[0076] After using the system of this invention, a 100-meter-high construction was completed, and the final cumulative verticality deviation was 92mm, achieving an accuracy of H / 1087, which is better than the design requirements.

[0077] Compared to traditional methods, verticality control accuracy is improved by approximately 70%, and the average deviation increment per segment is reduced from 26mm to approximately 8mm. Real-time control avoids large-scale post-construction corrections, significantly improving construction continuity. The prediction module's over-limit warning accuracy reaches over 90%, realizing a shift from passive correction to proactive prevention.

[0078] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wind power hybrid tower concrete pouring verticality deviation control method, characterized in that, Comprise: S100: Obtain the template posture data, support force data, template displacement data and concrete lateral pressure data of the formwork system; S200: The extended Kalman filter algorithm is adopted to fuse the template posture data, support force data, template displacement data and concrete lateral pressure data to obtain the six-degree-of-freedom pose state of the formwork system, and the real verticality is calculated by combining the environmental influence factor compensation; The long short-term memory network algorithm is used to analyze the historical verticality data to obtain the verticality deviation prediction value of the future period; S300: According to the real verticality, perform primary regulation and control, adjust the pouring speed, pouring position and vibration intensity; According to the inclination deviation in the six-degree-of-freedom pose state, perform secondary regulation and control, adjust the support force of the hydraulic support point and the tension of the cable; When the verticality deviation prediction value exceeds the warning value, perform tertiary regulation and control, and re-plan the pouring sequence and construction joint position; S400: Repeat S100 to S300 according to the preset control period to form a closed-loop control; Record the regulation and control data to construct a case library, and automatically optimize the regulation and control parameters based on the historical regulation and control effect by using the reinforcement learning algorithm.

2. The wind power hybrid tower concrete pouring verticality deviation control method according to claim 1, characterized in that, S100 comprises: Inclination sensors are arranged at the four corners and the middle of the formwork system to obtain the template posture data; Strain sensors are arranged on the support system and the cable system of the formwork system to obtain the support force data; A laser emitter is set at the reference point, and a target is set on the formwork system, and the template displacement data is obtained by laser ranging; Pressure sensors are arranged on the inside of the formwork system to obtain the concrete lateral pressure data.

3. The wind power hybrid tower concrete pouring verticality deviation control method according to claim 1, characterized in that, The calculation steps of the real verticality include: Obtain the measured verticality based on the six-degree-of-freedom pose state; Obtain the current temperature data, wind load data and construction load data; Calculate the temperature compensation amount, wind load compensation amount and construction load compensation amount respectively; Add the measured verticality to the temperature compensation amount, wind load compensation amount and construction load compensation amount to obtain the real verticality.

4. The wind power hybrid tower concrete pouring verticality deviation control method according to claim 1, characterized in that, The primary regulation and control includes: According to the difference between the real verticality and the verticality deviation threshold, calculate the adjusted pouring speed according to the following formula: ； wherein, is a planned pouring speed, is a speed adjustment coefficient, is a current real verticality, is a verticality deviation threshold value; Calculate the optimal pouring point based on the force balance analysis of the formwork system, and adjust the pouring position according to the optimal pouring point; Determine the vibration area and vibration intensity according to the difference between the real verticality and the verticality deviation threshold.

5. The wind power hybrid tower concrete pouring verticality deviation control method according to claim 1, characterized in that, The secondary regulation and control includes: Extract the inclination deviation of each support point in the six-degree-of-freedom pose state, and calculate the adjusted support force of each hydraulic support point according to the inclination deviation of each support point and the corresponding influence area according to the following formula: ； wherein, is an adjusted support force of the th support point, is a base support force, is a support force adjustment coefficient, is an inclination angle deviation of the th support point, is an influence area of the th support point; Adjust the support force of each hydraulic support point according to the adjusted support force, and establish the transfer function between the cable tension and the verticality; Calculate the tension adjustment amount of each cable according to the transfer function and the real verticality; Adjust the tension of each cable according to the tension adjustment amount.

6. The wind power hybrid tower concrete pouring verticality deviation control method according to claim 1, characterized in that, The tertiary regulation and control includes: Determine whether the verticality deviation prediction value exceeds the warning value; When the verticality deviation prediction value exceeds the warning value, pause the concrete pouring; replanning the pouring sequence of the remaining concrete according to the current poured concrete distribution and the real verticality; adjusting the construction joint position according to the verticality deviation prediction value; adjusting the maintenance starting time according to the real verticality.

7. The method according to claim 4 or 5, characterized in that, The optimization control parameters include: recording the real verticality, control action and verticality change after control of each control as control data; storing the control data to a case library; extracting historical control data from the case library; calculating a reward value based on the historical control data by using a reinforcement learning algorithm; updating the speed adjustment coefficient and the support force adjustment coefficient according to the reward value.

8. A wind power hybrid tower concrete pouring verticality deviation control system, characterized in that, comprise: a data acquisition module, configured to acquire formwork posture data, support force data, formwork displacement data and concrete lateral pressure data of a formwork-tower system; a state calculation and prediction module, configured to fuse the formwork posture data, support force data, formwork displacement data and concrete lateral pressure data by using an extended Kalman filtering algorithm to obtain a six-degree-of-freedom pose state of the formwork-tower system, and to calculate a real verticality by combining environmental influence factor compensation; and configured to analyze historical verticality data by using a long short-term memory network algorithm to obtain a verticality deviation prediction value in a future period; a hierarchical control module, configured to perform primary control according to the real verticality, and to adjust pouring speed, pouring position and vibration intensity; configured to perform secondary control according to an inclination deviation in the six-degree-of-freedom pose state, and to adjust support force of a hydraulic support point and cable tension; and configured to perform tertiary control when the verticality deviation prediction value exceeds an alarm value, and to replan a pouring sequence and a construction joint position; a closed-loop optimization module, configured to repeatedly execute the data acquisition module to the hierarchical control module according to a preset control period to form a closed-loop control; configured to record control data to construct a case library; and configured to automatically optimize control parameters based on historical control effects by using a reinforcement learning algorithm.

Images