Intelligent control method and system for mixed-tower prestress tensioning

Abstract

This invention belongs to the field of control, specifically an intelligent control method and system for prestressed tower tensioning. The method involves: real-time acquisition of tension force, elongation, anchorage zone strain, and tower vibration signals; fuzzy inference based on the deviation between the input values ​​and the target; and weighted aggregation of the current tensioning stage and strain state to generate a weighted fuzzy control increment set. Subsequently, in the defuzzification stage, multiple constraints are introduced to calculate the target quantity for the next cycle: a safety constraint function is constructed, and the weight of the dangerous interval is attenuated during centroid calculation to obtain the basic control increment; simultaneously, a damping factor is generated based on the historical increment dispersion, and a frequency safety factor is generated based on vibration spectrum analysis of resonance risk; finally, the basic increment is multiplied and corrected using the damping and frequency dual factors, and then the current actual control quantity is superimposed to output a safe and smooth target control quantity for the next cycle.

Description

Technical Field

[0001] This application belongs to the field of control, and in particular relates to an intelligent control method and system for prestressed tensioning of mixed towers. Background Technology

[0002] Prestressed tensioning control is based on the "dual control" principle, using tension force as the primary control index and strand elongation as the verification index. On-site operators rely on experience for control, resulting in a relatively crude and low-level intelligent control process. PID control algorithms often lead to overshoot or response lag when dealing with time-varying and material parameter uncertainties during tensioning, affecting the uniformity of the tensioning effect. While using fuzzy control algorithms to improve adaptability, the algorithm's input parameters are singular and cannot reflect the true state of the structure, preventing the control system from timely predicting and responding to risks of localized stress concentration or overall dynamic instability. The centroid method in the fuzzy control stage fails to account for the potential risk differences corresponding to different control decisions, posing structural safety hazards. Furthermore, existing control strategies overlook the correlation effect between the tensioning equipment and the flexible tower structure, potentially causing resonance during tensioning due to the excitation frequency being close to the structure's natural frequency, threatening structural safety. Therefore, how to integrate multi-source real-time information and develop an advanced intelligent control method that can take into account control stability, structural safety and resonance avoidance is a technical problem that urgently needs to be solved in the field of mixed tower prestressed tensioning. Summary of the Invention

[0003] To address the challenge of existing technologies in simultaneously achieving control stability, structural safety, and resonance avoidance.

[0004] In the first aspect, the present invention proposes an intelligent control method for prestressed tensioning of mixed towers, comprising the following steps: The current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower are obtained as real-time input parameters. Based on the preset membership function and fuzzy rule base, the deviation between the real-time input parameters and the target value is fuzzified and inferred. According to the current stage of the tensioning process and the strain of the anchoring zone, the output of each activated fuzzy rule is weighted and aggregated to generate a weighted fuzzy control incremental output set. Based on the weighted fuzzy control increment output set, the following method of fusing multiple constraints is used for defuzzification to calculate the target control quantity for the next control cycle: The dispersion of historical control increments within a preset time window is calculated to generate a damping factor; a safety constraint function is constructed to map the predicted control output quantity to the structural safety level; based on real-time spectrum analysis of the tower vibration signal and comparison with the preset structural natural frequency, the structural resonance risk under the current state is assessed, and a frequency safety factor for smooth adjustment of the tensioning rate is generated; when calculating the centroid of the weighted fuzzy control increment output set, the safety constraint function is used as a weight to attenuate and suppress the membership degree of the dangerous interval, resulting in a basic control increment; the actual control increment is obtained based on the basic control increment, the damping factor, and the frequency safety factor, and the actual control increment is added to the current actual control quantity to obtain the target control quantity for the next cycle with a smooth transition.

[0005] On the other hand, the present invention proposes an intelligent control system for prestressed tower tensioning, comprising the following modules: The acquisition module is used to acquire the current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower as real-time input parameters. The aggregation module is used to perform fuzzification processing and reasoning on the deviation between the real-time input parameters and the target value based on the preset membership function and fuzzy rule library, and to weight the output of each activated fuzzy rule according to the current stage of the tensioning process and the strain of the anchoring zone, and aggregate to generate a weighted fuzzy control incremental output set. The calculation module is used to defuzzify the weighted fuzzy control increment output set using the following method of fusing multiple constraints, and calculate the target control quantity for the next control cycle: calculate the dispersion of historical control increments within a preset time window to generate a damping factor; construct a safety constraint function that maps the predicted control output quantity to the structural safety level; assess the structural resonance risk under the current state based on real-time spectrum analysis of the tower vibration signal and comparison with the preset structural natural frequency, and generate a frequency safety factor for smooth adjustment of the tensioning rate; when calculating the centroid of the weighted fuzzy control increment output set, use the safety constraint function as a weight to attenuate and suppress the membership degree of the dangerous interval to obtain a basic control increment; obtain the actual control increment based on the basic control increment, the damping factor, and the frequency safety factor, and add the actual control increment to the current actual control quantity to obtain the target control quantity for the next cycle with a smooth transition.

[0006] This invention utilizes a weight adjustment mechanism for different tensioning stages and anchorage zone states in fuzzy inference, and incorporates multiple constraints in the defuzzification process. By using a damping factor correlated with the stability of historical control quantities, the control output is smoothed, enhancing operational stability. A risk penalty function is constructed to proactively avoid high-risk over-tensioning operations when calculating control quantities, improving structural safety. Simultaneously, a frequency safety factor is generated through real-time spectrum analysis to avoid structural resonance risks during tensioning. The generated control quantities are not only highly accurate but also consider the stability of the control process, the ultimate bearing capacity safety of the structure, and the avoidance of resonance, achieving intelligent control of the prestressed tower tensioning process. Attached Figure Description

[0007] Figure 1 This is a flowchart of an intelligent control method for prestressed tower tensioning according to an embodiment of the present invention. Detailed Implementation

[0008] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0009] An intelligent control method for prestressed tensioning of a hybrid tower is proposed in this embodiment of the invention, such as... Figure 1 As shown, it includes the following steps: S1 acquires the current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower as real-time input parameters.

[0010] Tension force signals are acquired using pressure sensors or hydraulic sensors mounted on intelligent jacks; steel strand elongation signals are acquired using wire displacement sensors or grating rulers; strain signals in the anchorage zone are acquired using resistance strain gauges pre-attached to the concrete surface of the anchorage area; and tower vibration signals are acquired using triaxial accelerometers deployed at key locations on the tower body. Synchronous acquisition and analog-to-digital conversion of multi-channel analog signals are achieved through the NI-DAQmx data acquisition card interface function library. A Kalman filter algorithm is then used to perform real-time noise reduction and state estimation on the raw data from each sensor to obtain real-time input parameters.

[0011] S2, based on the preset membership function and fuzzy rule base, the deviation between the real-time input parameters and the target value is fuzzified and inferred, and according to the current stage of the tensioning process and the strain of the anchoring zone, the output of each activated fuzzy rule is weighted and aggregated to generate a weighted fuzzy control incremental output set.

[0012] The real-time input parameters of tension force and elongation are subtracted from the target values ​​of the tensioning process control to obtain the deviation value E and the deviation change rate EC. Triangular or Gaussian membership functions are constructed using the trimf or gaussmf functions from the scikit-fuzzy library to fuzzify the deviation E and deviation change rate EC into negative large, negative medium, negative small, zero, positive small, positive medium, and positive large fuzzy variables. Based on the Mamdani inference engine, the following operation is performed: IF E is A AND EC is B THEN Fuzzy rules of the form C, where A, B, and C are fuzzy sets. Representing the control increment, such as the pressure change step size, the activation intensity and fuzzy output of each rule are obtained. Then, based on the current tensioning process in the initial tensioning, main tensioning, or load-bearing anchoring stage, and considering whether the strain value in the anchoring zone exceeds the safety threshold, a weighting coefficient is multiplied by the fuzzy output of each activation rule using a lookup table or a preset weighting function. The max-min synthesis method is then used to aggregate all weighted fuzzy outputs to obtain a weighted fuzzy control increment output set.

[0013] As an optional implementation, the step of weighting the outputs of each activated fuzzy rule based on the current stage of the tensioning process and the strain in the anchorage zone includes: The tensioning process is divided into the initial pre-tensioning stage, the main tensioning stage, and the load observation and formal anchoring stage. In the initial pretensioning stage, the weighting coefficient of the fuzzy rule related to the tension force and the elongation of the steel strand is set to 1.3, and the weighting coefficient of the fuzzy rule related to the strain in the anchorage zone is set to 0.8. During the main tensioning phase, the weight coefficients of all activation rules are set to the baseline value of 1.0; During the load observation and formal anchoring phase, if the instantaneous increase in strain in the anchoring zone exceeds 200με, the weight coefficient of all fuzzy rules with the output "reduce oil pump pressure" will be forcibly set to 2.5, and the weight coefficient of the remaining rules will be set to 1.0.

[0014] A finite state machine for the tensioning process is pre-established, and the state machine monitors the tension force in real time. With target tension The ratio switching state. When At this point, the state is defined as the initial pre-tensioning stage, eliminating strand relaxation. The weighting coefficients of the rules related to the two control indicators, tension force and elongation, are set to a higher value of 1.3. However, since the anchorage zone is not yet fully stressed at this stage, the strain reference value is low, so the weighting coefficients of the relevant rules are set to a lower value of 0.8. At this point, the main tension control stage begins. In this stage, all input parameters—tension force, elongation, and anchorage zone strain—are equally important and require coordinated control. Therefore, the weighting coefficients of all activation rules are set to the baseline value of 1.0. Then, the load observation and formal anchoring phase begins.

[0015] Once the load-bearing observation and formal anchoring phase begins, the focus shifts to ensuring anchoring safety and preventing slippage. High-frequency acquisition of strain data from the primary anchoring zone is then performed. and the strain value of the previous sampling point Compare and calculate the instantaneous increment. .like Exceeding a preset threshold, preferably 200 με, indicates potential microcracks in the anchorage zone concrete or signs of steel strand slippage, representing a high-risk signal. In this case, immediate emergency intervention is implemented. From the current fuzzy inference results, all fuzzy rules whose output conclusion is to reduce the oil pump pressure (i.e., a negative control increment or a significant reduction in oil pump pressure) are identified, and their weighting coefficients are forcibly increased to 2.5. The weighting coefficients for all other rules, such as maintaining or increasing pressure, are set to 1.0. Through this weighting, the aggregated fuzzy control increment output set is biased towards outputting negative increments to reduce oil pump pressure, thereby unloading part of the load and preventing safety accidents.

[0016] S3, based on the weighted fuzzy control increment output set, the fuzziness is de-defined by the following method of fusing multiple constraints, and the target control quantity for the next control cycle is calculated: the dispersion of historical control increments within a preset time window is calculated, and a damping factor is generated.

[0017] Create a first-in, first-out queue (collections.deque) of length N to store the control increments for the most recent N cycles. For each control cycle, use the numpy.std algorithm to calculate the standard deviation of the data in this queue as the dispersion index. Through function ,in As an adjustable parameter, the dispersion is mapped to a damping factor D between 0 and 1.

[0018] In an optional embodiment, the step of calculating the dispersion of historical control increments within a preset time window to generate a damping factor includes: Set the time window to the most recent 10 control cycles, and obtain the sequence of control increment output values ​​within the 10 cycles; Calculate the standard deviation σ of the sequence; Through formula The damping factor D is calculated, where the value of D is limited to the interval [0.4, 1.0]. This is the sensitivity coefficient.

[0019] At the end of each control cycle, the latest control increment will be... Values ​​are pushed into the queue, and the oldest value is removed. Calculate the standard deviation of the 10 incremental values ​​in the queue. Optionally, the preset maximum allowable increment range of the system can be used. The standard deviation is normalized using a dimensionless method, and the exponential function is: In this embodiment, the sensitivity coefficient k_D is set to 2.5, so that when the incremental output is stable, When the value is small, the D value is very close to 1, and the damping effect is weak; however, when the control increment oscillation intensifies, As the value increases, the D value decreases, resulting in an inhibitory effect.

[0020] For example, suppose the system is set to a maximum permissible increment per step. MPa / period. If the incremental sequence is kept stable, the standard deviation... It can be 0.08 MPa, and its normalized relative volatility is... It is 0.016. Substituting this into the formula, the damping factor is... A factor close to 1 indicates that the system is stable and no additional damping is required.

[0021] Conversely, if the control increment sequence exhibits violent alternating positive and negative oscillations, for example, {+2.0, -2.5, +2.2, -2.8, +1.9} MPa / period, the calculated standard deviation... It will be very large, for example MPa. After normalization, its relative volatility is 0.48. Substituting into the formula, we get... The above normalization process ensures that the relative volatility remains unchanged even if the system switches the unit to Pa or kPa. Since the value of D is limited to [0.4, 1.0], the value of D is forcibly set to the lower limit of 0.4. A smaller damping factor will proportionally reduce the control increment amplitude in the next cycle, which is equivalent to applying strong damping to the control system, thereby smoothing the decompression process and guiding the system to return to stability.

[0022] S4. Construct a safety constraint function that maps the predictive control output to the structural safety level.

[0023] Define a piecewise function as a safety constraint function. When the predicted tension obtained by superimposing the current tension force with the control increment is within the safe range, the function value is 1; when the predicted tension force is lower than the allowable range, the value decreases linearly with the increase of the deviation; when the predicted tension force enters the over-tension danger zone, the function value switches to a minimum value based on exponential decay, rapidly approaching 0, so as to cut off the control weight towards the danger zone.

[0024] In one embodiment, constructing a safety constraint function that maps the predictive control output to the structural safety level includes: Set target tension ; When the current tension force plus the estimated control increment corresponds to the predicted tension force F located at... When the safety interval is within the safe range, the safety constraint function value S(F) is 1; When F is lower At that time, the value of the safety constraint function ; When F is higher At that time, the value of the safety constraint function .

[0025] The implementation of the function requires an increment from the control output, such as the change in oil pump pressure. The mapping model to the predicted tension F can be simplified to: ,in This represents the current actual tension force, and K is the jack calibration gain coefficient. Assume the target tension force... If the value is 1500kN, then the safe range can be set to [1470kN, 1545kN]. When a fuzzy control increment outputs the estimated tension... Within this interval, the safety constraint value S(F) = 1. Using a constant of 1 indicates that this interval conforms to the allowable error range in the prestressed construction specification, i.e., -2% to +3%, representing no additional safety risk. Therefore, the output of the fuzzy inference is fully trusted, and its full weight is retained during defuzzification.

[0026] When F is below 1470kN, the safety constraint function switches to a linear attenuation formula with a lower limit truncation. Undertension mainly affects the long-term load-bearing performance and service life of the structure, but it usually does not cause catastrophic engineering damage at the moment of construction. Therefore, as the deviation increases, it is only necessary to gradually decrease the weight corresponding to the dangerous increment. For example, if the estimated tension F is 1460kN, substituting it into the formula yields... This indicates that the system's trust weight for the control increment has decreased to 33%.

[0027] When F exceeds 1545 kN, the safety constraint function switches to a steep attenuation formula based on the square of the exponent. Over-tensioning is an extremely high-risk condition in prestressed tensioning operations, which can easily lead to serious instantaneous safety accidents such as jack overload bursting, concrete crushing under the anchor plate, and even steel strand breakage and slippage. The direct threat of over-tensioning to structural safety is far greater than that of under-tensioning; therefore, a steep attenuation formula based on the square of the exponent is adopted. For example, if the estimated tension force... The value is 1560kN. Substitute this value into the formula to calculate the constraint value. If the resistance reaches 1575kN, the constraint value will drop exponentially. When the system aggregates and defuzzifies, it can effectively veto any dangerous control increments that attempt to significantly increase the system pressure, thus ensuring absolute safety in hydraulic operation.

[0028] S5. Based on real-time spectrum analysis of the tower vibration signal and comparison with the preset structural natural frequency, assess the structural resonance risk under the current state and generate a frequency safety factor for smooth adjustment of the tensioning rate.

[0029] The tower vibration acceleration signal for the most recent time window is captured, and a Fast Fourier Transform (FFT) is performed using the FFT function from the scipy.fft library to obtain the signal's spectrum. The vibration energy within the frequency band of one or more preset structural natural frequencies is extracted and compared with preset warning and alarm thresholds. Based on the comparison result, a frequency safety factor is output. It is used to smoothly intervene in the tensioning rate, i.e., the magnitude of the control increment, rather than directly causing sudden changes in system pressure.

[0030] In an optional embodiment, the step of assessing the structural resonance risk under the current state based on real-time spectral analysis of the tower vibration signal and comparison with a preset structural natural frequency, and generating a frequency safety factor for smoothly adjusting the tensioning rate, includes: A fast Fourier transform is performed on the real-time acquired tower vibration signal to obtain the power spectral density map of the vibration energy. Extract the vibration energy within the frequency band of one or more preset structural natural frequencies, and compare it with preset warning thresholds and alarm thresholds; If the vibration energy in any frequency band does not exceed the corresponding warning threshold, it is determined to be a safe state, and the frequency safety factor is set to 1.0; If the warning threshold is exceeded but the alarm threshold is not exceeded, it is determined to be a warning state, and the frequency safety factor is set to 0.5; If the alarm threshold is exceeded, the system is in an alarm state. If the current basic control increment is positive, the frequency safety factor is set to 0.0. If the current basic control increment is negative, the frequency safety factor is set to 1.0.

[0031] For example, vibration signals are continuously acquired at a sampling rate of 200Hz. A Fast Fourier Transform (FFT) is performed on the latest data segment every 5 seconds, and the power spectral density (PSD) is calculated. Assuming the first natural frequency of the mixing tower is known to be 0.5Hz through finite element analysis, a monitoring frequency band [0.4Hz, 0.6Hz] is preset, and a warning threshold is set for this frequency band. For example, 0.08 (m / s) 2 ) 2 / Hz and an alarm threshold For example, 0.12 (m / s) 2 ) 2 / Hz.

[0032] In each analysis cycle, the total vibration energy within the monitoring frequency band is calculated. .like The system is deemed to be in a safe state; frequency safety factor. Set to 1.0 to allow normal tensioning rate. If... The system is flagged as a warning state, indicating a potential risk of resonance. Setting it to 0.5 halves the tension pressure increment in the current control cycle, achieving smooth deceleration. If The system is flagged as an alarm state. At this point, further loading is strictly prohibited. If the basic control increment solved by the current fuzzy system is positive, requiring a voltage boost, then a forced setting will be implemented. Setting it to 0.0 resets the increment to zero, effectively pausing the pressure holding process; if the current increment is negative, requiring pressure release, then set it to 0.0. Setting it to 1.0 allows the system to unload normally. If the current basic control increment is exactly zero, in order to prevent small positive fluctuations caused by high-frequency noise or numerical truncation errors from the underlying sensors from triggering malfunctions, the critical case of a zero basic control increment is treated the same as when the basic control increment is positive, and the frequency safety factor is also set to 0.0.

[0033] S6. When calculating the centroid of the weighted fuzzy control increment output set, the safety constraint function is used as a weight to attenuate and suppress the membership degree of the danger interval, resulting in a basic control increment. The actual control increment is obtained based on the basic control increment, the damping factor, and the frequency safety factor. The actual control increment is then added to the current actual control quantity to obtain the target control quantity for the next cycle with a smooth transition.

[0034] The weighted centroid method is used for defuzzification calculation, and the basic control increment is calculated. The summation is performed on discrete increment points over the entire output universe. , For the corresponding safety constraint function value, The corresponding membership degree; through multiplication operation Complete the correction of the basic control increment, through The absolute control target signal is obtained and output to the hydraulic pump frequency converter.

[0035] In an optional embodiment, obtaining the actual control increment based on the basic control increment, the damping factor, and the frequency safety factor includes: The membership function formed by the weighted fuzzy control increment output set The safety constraint value mapped to the predicted physical quantity corresponding to the control output universe of discourse variable Multiply the values, truncate the membership of high-risk areas, and calculate the basic control increment using the centroid method.

[0036] Assume control of the incremental universe of discourse For example, the allowable change in oil pump pressure within a single cycle is [-5MPa, +5MPa], and it is discretized into multiple points. After fuzzy reasoning and weighting, the membership value at each discrete point is obtained. At the same time, for each Based on the superimposed estimated tension, the safety constraint value is calculated. If a certain positive increment MPa may cause the estimated tension to enter the over-tension danger zone. The value of will become extremely small according to the exponential decay formula, for example, 0.001. In the calculation At that time, the contribution item of the danger point This will be greatly weakened, equivalent to the membership degree of the danger zone being truncated. Weighted centroid towards... The safety control zone is offset by approximately 1. Therefore, a base increment that might originally be +3.0 MPa will be automatically corrected and reduced to, for example, +1.2 MPa.

[0037] Based on the incremental control measures that take into account structural safety risks Then, it enters the correction phase. For example, the calculated... It is +1.2 MPa. At this point, the damping factor D and frequency safety factor for this cycle are obtained from the aforementioned steps. Scenario 1: Recent control is stable with D=0.95 and vibration is normal. Then the actual control increment MPa. Scenario 2: Stable control D=0.95, but vibration energy enters the warning zone. ,but MPa, tensioning speed is gently halved. Scenario 3: Control oscillation exists D=0.5, and the vibration enters the alarm zone, then... The system pauses pressurization. The actual control increment will be... The actual pressure superimposed on the current system Up, output the instruction for the next cycle.

[0038] This invention also proposes an intelligent control system for prestressed tower tensioning, comprising the following modules: The acquisition module is used to acquire the current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower as real-time input parameters. The acquisition module is used to acquire the current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower as real-time input parameters. The aggregation module is used to perform fuzzification processing and reasoning on the deviation between the real-time input parameters and the target value based on the preset membership function and fuzzy rule library, and to weight the output of each activated fuzzy rule according to the current stage of the tensioning process and the strain of the anchoring zone, and aggregate to generate a weighted fuzzy control incremental output set. The calculation module is used to defuzzify the weighted fuzzy control increment output set using the following method of fusing multiple constraints, and calculate the target control quantity for the next control cycle: calculate the dispersion of historical control increments within a preset time window to generate a damping factor; construct a safety constraint function that maps the predicted control output quantity to the structural safety level; assess the structural resonance risk under the current state based on real-time spectrum analysis of the tower vibration signal and comparison with the preset structural natural frequency, and generate a frequency safety factor for smooth adjustment of the tensioning rate; when calculating the centroid of the weighted fuzzy control increment output set, use the safety constraint function as a weight to attenuate and suppress the membership degree of the dangerous interval to obtain a basic control increment; obtain the actual control increment based on the basic control increment, the damping factor, and the frequency safety factor, and add the actual control increment to the current actual control quantity to obtain the target control quantity for the next cycle with a smooth transition.

[0039] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0040] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An intelligent control method for prestressed tensioning of a mixed tower, characterized in that, Includes the following steps: The current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower are obtained as real-time input parameters. Based on the preset membership function and fuzzy rule base, the deviation between the real-time input parameters and the target value is fuzzified and inferred. According to the current stage of the tensioning process and the strain of the anchoring zone, the output of each activated fuzzy rule is weighted and aggregated to generate a weighted fuzzy control incremental output set. Based on the weighted fuzzy control increment output set, the following method of fusing multiple constraints is used to defuzzify the data and calculate the target control quantity for the next control cycle: calculate the dispersion of historical control increments within a preset time window and generate a damping factor; construct a safety constraint function that maps the predicted control output quantity to the structural safety level; based on real-time spectrum analysis of the tower vibration signal and comparison with the preset structural natural frequency, assess the structural resonance risk under the current state and generate a frequency safety factor for smoothly adjusting the tensioning rate. In calculating the centroid of the weighted fuzzy control increment output set, the safety constraint function is used as a weight to attenuate the membership of the dangerous interval, obtaining a basic control increment, wherein the membership function formed by the weighted fuzzy control increment output set The safety constraint value mapped by the predicted physical quantity corresponding to the control output universe variable is multiplied, the membership of the high-risk area is truncated, and the basic control increment is calculated by the centroid method. The actual control increment is obtained based on the basic control increment, the damping factor, and the frequency safety factor. This actual control increment is then added to the current actual control quantity to obtain the target control quantity for the next cycle with a smooth transition. The weighted centroid method is used for defuzzification calculation, and the basic control increment... The summation is performed on discrete increment points over the entire output universe. , For the corresponding safety constraint function value, The corresponding membership degree; through multiplication operation Complete the correction of the basic control increment, through The absolute control target signal is obtained and output to the hydraulic pump frequency converter.

2. The method according to claim 1, characterized in that, The step of weighting the outputs of each activated fuzzy rule based on the current stage of the tensioning process and the strain in the anchorage zone includes: The tensioning process is divided into the initial pre-tensioning stage, the main tensioning stage, and the load observation and formal anchoring stage. In the initial pretensioning stage, the weighting coefficient of the fuzzy rule related to the tension force and the elongation of the steel strand is set to 1.3, and the weighting coefficient of the fuzzy rule related to the strain in the anchorage zone is set to 0.

8. During the main tensioning phase, the weight coefficients of all activation rules are set to the baseline value of 1.0; During the load observation and formal anchoring phase, if the instantaneous increase in strain in the anchoring zone exceeds 200με, the weight coefficient of all fuzzy rules with the output "reduce oil pump pressure" will be forcibly set to 2.5, and the weight coefficient of the remaining rules will be set to 1.

0.

3. The method according to claim 1, characterized in that, The calculation of the dispersion of historical control increments within a preset time window generates a damping factor, including: Set the time window to the most recent 10 control cycles, and obtain the sequence of control increment output values ​​within the 10 cycles; Calculate the standard deviation σ of the sequence; Through formula The damping factor D is calculated, where the value of D is limited to the interval [0.4, 1.0]. This is the sensitivity coefficient.

4. The method according to claim 1, characterized in that, The process involves real-time spectral analysis of the tower vibration signal, comparison with a preset structural natural frequency, assessment of the structural resonance risk under the current state, and generation of a frequency safety factor for smooth adjustment of the tensioning rate, including: A fast Fourier transform is performed on the real-time acquired tower vibration signal to obtain the power spectral density map of the vibration energy. Extract the vibration energy within the frequency band of one or more preset structural natural frequencies, and compare it with preset warning thresholds and alarm thresholds; If the vibration energy in any frequency band does not exceed the corresponding warning threshold, it is determined to be a safe state, and the frequency safety factor is set to 1.0; If the warning threshold is exceeded but the alarm threshold is not exceeded, it is determined to be a warning state, and the frequency safety factor is set to 0.5; If the alarm threshold is exceeded, the system is in an alarm state. If the current basic control increment is positive, the frequency safety factor is set to 0.

0. If the current basic control increment is negative, the frequency safety factor is set to 1.

0.

5. An intelligent control system for prestressed tensioning of mixed-stove towers, characterized in that, Includes the following modules: The acquisition module is used to acquire the current tension, steel strand elongation, anchorage zone strain, and tower vibration signal of the hybrid tower as real-time input parameters. The aggregation module is used to perform fuzzification processing and reasoning on the deviation between the real-time input parameters and the target value based on the preset membership function and fuzzy rule library, and to weight the output of each activated fuzzy rule according to the current stage of the tensioning process and the strain of the anchoring zone, and aggregate to generate a weighted fuzzy control incremental output set. The calculation module is used to defuzzify the weighted fuzzy control increment output set using the following method of fusing multiple constraints, and calculate the target control quantity for the next control cycle: calculate the dispersion of historical control increments within a preset time window to generate a damping factor; construct a safety constraint function that maps the predicted control output quantity to the structural safety level; assess the structural resonance risk under the current state based on real-time spectrum analysis of the tower vibration signal and comparison with the preset structural natural frequency, and generate a frequency safety factor for smoothing the tensioning rate; when calculating the centroid of the weighted fuzzy control increment output set, use the safety constraint function as a weight to attenuate and suppress the membership degree of the dangerous interval, and obtain a basic control increment; The actual control increment is obtained based on the basic control increment, the damping factor, and the frequency safety factor. The actual control increment is then added to the current actual control quantity to obtain the target control quantity for the next cycle with a smooth transition.

6. The system according to claim 5, characterized in that, The step of weighting the outputs of each activated fuzzy rule based on the current stage of the tensioning process and the strain in the anchorage zone includes: The tensioning process is divided into the initial pre-tensioning stage, the main tensioning stage, and the load observation and formal anchoring stage. In the initial pretensioning stage, the weighting coefficient of the fuzzy rule related to the tension force and the elongation of the steel strand is set to 1.3, and the weighting coefficient of the fuzzy rule related to the strain in the anchorage zone is set to 0.

8. During the main tensioning phase, the weight coefficients of all activation rules are set to the baseline value of 1.0; During the load observation and formal anchoring phase, if the instantaneous increase in strain in the anchoring zone exceeds 200με, the weight coefficient of all fuzzy rules with the output "reduce oil pump pressure" will be forcibly set to 2.5, and the weight coefficient of the remaining rules will be set to 1.

0.

7. The system according to claim 5, characterized in that, The calculation of the dispersion of historical control increments within a preset time window generates a damping factor, including: Set the time window to the most recent 10 control cycles, and obtain the sequence of control increment output values ​​within the 10 cycles; Calculate the standard deviation σ of the sequence; Through formula The damping factor D is calculated, where the value of D is limited to the interval [0.4, 1.0]. This is the sensitivity coefficient.

8. The system according to claim 5, characterized in that, The process involves real-time spectral analysis of the tower vibration signal, comparison with a preset structural natural frequency, assessment of the structural resonance risk under the current state, and generation of a frequency safety factor for smooth adjustment of the tensioning rate, including: A fast Fourier transform is performed on the real-time acquired tower vibration signal to obtain the power spectral density map of the vibration energy. Extract the vibration energy within the frequency band of one or more preset structural natural frequencies, and compare it with preset warning thresholds and alarm thresholds; If the vibration energy in any frequency band does not exceed the corresponding warning threshold, it is determined to be a safe state, and the frequency safety factor is set to 1.0; If the warning threshold is exceeded but the alarm threshold is not exceeded, it is determined to be a warning state, and the frequency safety factor is set to 0.5; If the alarm threshold is exceeded, the system is in an alarm state. If the current basic control increment is positive, the frequency safety factor is set to 0.

0. If the current basic control increment is negative, the frequency safety factor is set to 1.

0.

9. The system according to claim 5, characterized in that, The process of obtaining the actual control increment based on the basic control increment, the damping factor, and the frequency safety factor includes: The membership function formed by the weighted fuzzy control increment output set The safety constraint value mapped to the predicted physical quantity corresponding to the control output universe of discourse variable Multiply the values, truncate the membership of high-risk areas, and calculate the basic control increment using the centroid method.

Images