Cable-stayed bridge self-resetting intelligent control method and system based on SMA and AI prediction

Abstract

The invention relates to the technical field of bridges, in particular to a cable-stayed bridge self-resetting intelligent control method and system based on SMA and AI prediction. Comprising the steps that cantilever beams are poured along the two sides of a zero block of a box girder, pre-tightened first-class SMA inhaul cables are installed between the cantilever beams and a bridge tower and between the cantilever beams and a bridge pier, and pre-tightened second-class SMA inhaul cables are installed on adjacent prefabricated box girder sections; multi-source monitoring data are obtained, the multi-source monitoring data comprise stress data and temperature data of an SMA inhaul cable, lateral deviation displacement data and strain data of a prefabricated box girder section are obtained, and axial force data and bridge deck load data of a bridge tower are obtained; constructing an SMA stress attenuation prediction model based on a time sequence neural network, calculating a self-resetting critical intervention threshold value of the SMA cable, comparing the stress attenuation curve with the critical intervention threshold value, and judging a stress node which is about to be lower than the threshold value; and the SMA stress attenuation prediction model sends an activation instruction to the PLC control module. The problem of dynamic lateral resistance of the prefabricated box and the bridge tower is solved, and self-resetting intelligent control over the cable-stayed bridge is achieved.

Description

Technical Field

[0001] This invention relates to the field of bridge technology, and in particular to a self-resetting intelligent control method and system for cable-stayed bridges based on SMA and AI prediction. Background Technology

[0002] Cable-stayed bridges, with their advantages of large span adaptability and efficient force transfer, have become core transportation infrastructure in scenarios such as crossing rivers and seas and complex mountainous terrain. However, during long-term service, the issue of eccentric loading in two-way traffic scenarios has become a core safety hazard that could lead to the collapse of cable-stayed bridges. Specifically, the load distribution on the bridge deck of a two-way cable-stayed bridge is easily affected by the type of vehicles traveling on it. When one lane is consistently used by heavy vehicles while the other lane is mainly used by light vehicles or is empty, a continuous eccentric loading effect will occur. Although prestressing is applied to the precast box girder during the construction stage to improve its compressive strength, this prestressing is mainly designed for vertical loads and is difficult to balance the horizontal tilting force generated by long-term eccentric loading.

[0003] On the one hand, the main beam offset causes uneven stress on the stay cables, with some cables exceeding their tension limits while others are in a relaxed state, reducing the overall lateral stiffness of the structure. On the other hand, long-term offset can lead to stress concentration in local areas, exacerbating damage such as concrete cracking and steel corrosion, ultimately causing a continuous decline in the overall stability of the cable-stayed bridge.

[0004] Shape memory alloys (SMAs) possess unique superelasticity, shape memory, high damping, and flag-shaped hysteresis properties. SMAs can serve as limiters and self-resetting agents.

[0005] Therefore, it is necessary to design a self-resetting intelligent control method and system for cable-stayed bridges based on SMA and AI prediction for the scenario of heavy vehicle eccentric loading in two-way lanes, so as to solve the dynamic lateral resistance problem of precast box girder and bridge tower. Summary of the Invention

[0006] To address the aforementioned problems, this invention provides a self-resetting intelligent control method and system for cable-stayed bridges based on SMA and AI prediction.

[0007] In a first aspect, the present invention provides a self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction, comprising: S1. Cast the No. 0 box girder on the tower plate of the bridge tower, cast the cantilever beam along both sides of the No. 0 box girder, and gradually splice the precast box girder segments along both sides of the No. 0 box girder. S2. Cast cantilever beams on both sides of the zero block of the box girder, install pre-tightened first-type SMA cables between the cantilever beams and the bridge towers and piers, and install pre-tightened second-type SMA cables in adjacent precast box girder sections; S3. Acquire multi-source monitoring data, including stress data and temperature data of SMA cables, lateral displacement data and strain data of precast box girder segments, and axial force data and bridge deck load data of bridge towers. S4. Construct an SMA stress attenuation prediction model based on a time-series neural network. Input multi-source monitoring data into the SMA stress attenuation prediction model. After training, output the stress attenuation curves of the first type of SMA cable and the second type of SMA cable within a future preset period. Calculate the self-resetting critical intervention threshold of the SMA cable. Compare the stress attenuation curve with the critical intervention threshold to determine the stress nodes that are about to fall below the threshold. The S5.SMA stress attenuation prediction model sends an activation command to the PLC control module. Based on the stress difference corresponding to the stress node, the PLC control module activates the heating element built into the SMA cable to heat the first and second type of SMA cables to restore their pre-tension state, thereby driving the zero block of the box girder and the precast box girder segment to correct lateral deviation and achieve self-resetting of the cable-stayed bridge.

[0008] Furthermore, the cantilever beams cast along both sides of the zero block of the box girder, and the pre-tensioned first-type SMA cables installed between the cantilever beams and the bridge towers and piers, include: Cantilever beam steel reinforcement cages are arranged along the transverse sides of the zero block of the box girder and along the longitudinal direction close to the bridge tower and pier. The steel reinforcement cages are welded and fixed to the pre-embedded steel reinforcements of the zero block of the box girder, the bridge tower and the pier to form an integrated load-bearing structure. SMA cable anchoring holes are pre-set at the ends of the cantilever beam for cantilever beam casting; One end of the first type of SMA cable is passed through the anchorage hole at the end of the cantilever beam and fixed by the anchorage. The other end extends and is anchored to the preset anchorage point of the bridge tower and pier. The first type of SMA cable is pre-tightened using tensioning equipment.

[0009] Furthermore, the installation of pre-tensioned second-type SMA cables in adjacent precast box girder segments includes: Holes are machined on the side of the precast box girder segment. After the holes of two adjacent precast box girder segments are connected, they form a U-shape that is symmetrical towards the center of the interior. Multiple U-shaped holes are distributed longitudinally in a stepped manner. The second type of SMA cable is inserted into the U-shaped duct entrance on one side of the beam segment and exited from the duct exit on the other side of the beam segment. Temporary anchors are fitted at both ends of the cable for fixation. After each preset number of precast box girder segments are installed, steel wire locks are suspended between the precast box girder segments and the bridge towers. One end of the steel wire lock is anchored to a preset lifting point on the bridge tower, and the other end is anchored to a preset lifting lug on the precast box girder segment. This process is repeated until the precast box girder segment extends to the preset closure position in the middle of the river and the closure is completed. The prestressing force is adjusted according to the real-time strain of the precast box girder segment. At that time, according to the formula Calculate the preload adjustment amount, where, To preset the strain triggering coefficient, To preset the pre-tightening adjustment coefficient, The elastic modulus of SMA Let SMA cable cross-sectional area be 1. Design initial stress for type II SMA cables. The cross-sectional area of ​​the beam segment is used to apply an adjustment amount through a tensioning device until the strain of the beam segment is less than [a certain value]. ,in, This is the preset strain control coefficient.

[0010] Furthermore, the first type of SMA cable and the second type of SMA cable include an anchor body, SMA wires connected between the anchor bodies, a heating element and a copper foil heat-conducting layer disposed in the anchor. The anchor body has a through hole for passing through the SMA wire. An annular groove is formed on the inner wall of the through hole in the circumferential direction. The heating element is embedded in the annular groove. The copper foil heat-conducting layer wraps around the heating element and the outside of the SMA wire. The outside of the copper foil heat-conducting layer is in close contact with the inner wall of the through hole.

[0011] Furthermore, the acquisition of multi-source monitoring data includes: Foil strain sensors were attached to the middle section of the first type of SMA cable, and thermocouple sensors were wrapped around the first type of SMA cable near the anchor body. Stress and temperature data of the first type of SMA cable were collected respectively. Laser displacement sensors were installed on both sides of the zero block of the box girder to collect lateral displacement data, and strain gauges with full-bridge connection were attached to the bottom of the precast box girder segment to collect strain data. An axial force sensor is installed on the underside of the bridge tower to collect axial force data of the bridge tower; Weighing sensors are installed on the bridge deck to collect bridge deck load data; All sensor data is transmitted to the data acquisition terminal via an industrial bus.

[0012] Furthermore, the construction of the SMA stress attenuation prediction model based on a time-series neural network includes: An LSTM time series prediction model was constructed and trained using historical multi-source monitoring data. After the model training is completed, the stress decay curves of the two types of SMA cables are output, and the curves are fitted according to the exponential decay formula: , among which, among which for SMA cable stress at any time For the initial preload stress, The attenuation coefficient is... For service time, For residual stress, the first type of SMA cable Positively correlated with the axial force of the bridge tower, the second type of SMA cable Positively correlated with the frequency of bridge deck load data; The critical intervention threshold for the first type of SMA cable, combined with span correction, is calculated using the following formula: ,in , To allow for stress loss rate, This is the span correction factor; The critical intervention threshold for type II SMA cables, combined with temperature difference correction, is calculated using the following formula: ,in , This is the temperature difference correction factor; Comparing the stress decay curves of the two types of SMA cables with their respective critical intervention thresholds, if a time point exists... ,and Then determine For stress nodes, where ; The node determination for the first type of SMA cable must satisfy the lateral displacement of the box girder. The node determination for the second type of SMA cable must meet the strain requirements of the precast box girder segment. ,in, To preset the lateral deviation ratio coefficient, This is the preset strain ratio coefficient.

[0013] Furthermore, the construction of the LSTM time series prediction model, and the training of the LSTM time series prediction model using multi-source monitoring data, includes: Historical multi-source monitoring data is obtained as training data and divided into training set, validation set and test set; The LSTM time series prediction model is trained using the Adam optimization algorithm, with the root mean square error as the loss function, as shown in the formula below. Where N is the number of samples, To predict stress, For the first The actual stress of each sample; Input the test set data into the trained model and calculate the prediction error for the first type of SMA cable. Prediction error of type II SMA cable If If the model satisfies the generalization requirement, This is the preset error threshold.

[0014] Furthermore, the process for determining the allowable stress loss rate includes: Fatigue load data for two types of SMA cables were obtained. The fatigue load for the first type of SMA cable was the axial force fluctuation value of the bridge tower. The fatigue load of the second type of SMA cable is the average number of heavy vehicles passing over the bridge deck per day. ; The fatigue influence coefficient is calculated using the following formula: The fatigue influence coefficient of the first type of SMA cable is calculated using the following formula: in, Based on the stress loss rate, The influence coefficient of axial force on bridge towers. max is the maximum fluctuation value of the axial force of the bridge tower; The fatigue influence coefficient of the second type of SMA cable is calculated using the following formula: in, This is the bridge deck load influence coefficient. This represents the maximum daily number of heavy vehicles that can pass over the bridge. Determine the allowable stress loss rate for Class I SMA cables. Type II SMA cable .

[0015] Furthermore, the PLC control module activates the heating element built into the SMA cable based on the stress difference corresponding to the stress node, including: The PLC control module receives the activation command sent by the AI ​​model, extracts the stress difference value corresponding to the stress node, and the stress difference value of the first type of SMA cable. Stress difference of type II SMA cable ; The heating temperature increment is calculated based on the stress difference using the following formula: ,in The stress temperature coefficient of SMA; The heating power of the first type of SMA cable is adjusted according to the axial force of the bridge tower. When the axial force is not less than the preset axial force value, the preset rated power is used, and when the axial force is less than the preset axial force value, the full rated power is used. The heating power of the second type of SMA cable is adjusted according to the bridge deck load. When the load is not less than the preset load, the preset rated power is used, and when the load is less than the preset load, the full rated power is used. During the heating process, the temperature is fed back in real time through thermocouples. When the target temperature is reached, the mode is switched to heat preservation mode and the heat preservation time is continued for a preset time. Monitoring the stress recovery of the first type of SMA cable to And the lateral displacement of the box girder is less than The stress in the second type of SMA cable rebounded to And the strain of the precast box girder segment is less than If the self-reset is completed, the PLC control module will cut off the heating circuit.

[0016] Secondly, a self-resetting intelligent control system for cable-stayed bridges based on SMA and AI prediction includes: The construction module is configured to cast the No. 0 block of the box girder on the tower plate of the bridge tower, cast the cantilever beam along both sides of the No. 0 block of the box girder, and gradually splice the precast box girder segments along both sides of the No. 0 block of the box girder. The SMA cable installation pretensioning module is configured to cast cantilever beams along both sides of the zero block of the box girder, install pretensioned first-type SMA cables between the cantilever beams and the bridge towers and piers, and install pretensioned second-type SMA cables in adjacent precast box girder segments. The data acquisition module is configured to acquire multi-source monitoring data, including stress data and temperature data of SMA cables, lateral displacement data and strain data of precast box girder segments, and axial force data and bridge deck load data of bridge towers. The stress node judgment module is configured to construct an SMA stress attenuation prediction model based on a time-series neural network, input multi-source monitoring data into the SMA stress attenuation prediction model, output the stress attenuation curves of the first type of SMA cable and the second type of SMA cable within a preset period after training, calculate the self-resetting critical intervention threshold of the SMA cable, compare the stress attenuation curve with the critical intervention threshold, and judge the stress node that is about to fall below the threshold. The PLC control module is configured such that the SMA stress attenuation prediction model sends an activation command to the PLC control module. Based on the stress difference corresponding to the stress node, the PLC control module activates the heating element built into the SMA cable to heat the first and second type of SMA cables to restore their pretension state, thereby driving the zero block of the box girder and the precast box girder segment to correct lateral deviation and achieve self-resetting of the cable-stayed bridge.

[0017] In summary, the present invention has the following beneficial technical effects: 1. The cantilever beam, the zero block of the box girder, the bridge tower, and the pier are welded and fixed with pre-embedded steel bars to form an integrated force-bearing system. Simultaneously, the first type of SMA cables are arranged in four directions, with one end anchored to a pre-set anchoring hole at the end of the cantilever beam, and the other end connected to pre-set anchoring points on the bridge tower and pier respectively. After pre-tensioning by a tensioning device, a horizontal constraint force field surrounding the zero block can be formed. The four-way cables can synchronously balance the horizontal tilting force generated by eccentric loading, preventing the bridge tower and pier from accumulating lateral deflection towards the heavy vehicle side. The second type of SMA cables, with a stepped distribution, can directly constrain the relative offset between beam segments. Furthermore, multiple U-shaped channels are distributed longitudinally in a stepped pattern, which can match the cable pre-tensioning force according to the strain differences at different locations of the beam segment. When the actual strain of the beam segment exceeds the preset trigger threshold, the pre-tensioning force adjustment can be quantitatively calculated based on the difference between the actual strain and the allowable strain of the beam segment and precisely applied through a tensioning device to ensure that the strain of the beam segment is controlled within a safe range.

[0018] 2. This invention triggers the shape memory effect of SMA filaments by heating with heating elements, quickly restoring preload without requiring traffic interruption for manual replacement or tensioning, unlike traditional steel cables. During heating, thermocouples provide real-time feedback on the SMA cable temperature, automatically switching to insulation mode once the target temperature is reached. During self-reset, it monitors in real-time whether the stress in the first type of SMA cable rises to the critical intervention threshold and whether the lateral displacement of the zero block of the box girder is less than the allowable value; and whether the stress in the second type of SMA cable rises to the critical intervention threshold and whether the strain of the precast box girder segment is less than the allowable value. Upon completion of self-reset, the heating circuit is cut off. This reduces maintenance costs by 60% without interrupting traffic, significantly improving the operational reliability and economy of cable-stayed bridges.

[0019] 3. This invention integrates an LSTM time-series prediction model with multi-source data to proactively mitigate stress attenuation risks. The constructed SMA stress attenuation prediction model uses multi-source data, including stress and temperature of the first and second types of SMA cables, lateral displacement of the zero block of the box girder, strain of the precast box girder segment, axial force of the bridge tower, and bridge deck load, as input. The Adam optimization algorithm is used for training, controlling the model's prediction error within a preset error threshold, and outputting future SMA cable stress attenuation curves. Simultaneously, a span correction coefficient is determined based on the span characteristics of the cable-stayed bridge, and a temperature difference correction coefficient is determined based on the annual average temperature difference of the service environment. This allows for the calculation of the self-resetting critical intervention thresholds for the two types of SMA cables. By comparing the stress attenuation curves with the critical intervention thresholds, stress nodes about to fall below the thresholds are accurately identified, improving prediction accuracy. Attached Figure Description

[0020] Figure 1 This is a flowchart of a self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction, according to Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the overall structure of the cable-stayed bridge according to Embodiment 1 of the present invention; Figure 3 This is a front view of the structure of a single bridge tower of the cable-stayed bridge according to Embodiment 1 of the present invention; Figure 4 This is a side view of the cable-stayed bridge according to Embodiment 1 of the present invention; Figure 5 This is a front view of the connection relationship of the precast box girder segments in Embodiment 1 of the present invention; Figure 6 This is a top view of the connection relationship of the precast box girder segments in Embodiment 1 of the present invention; Figure 7 This is a schematic diagram of the heating element structure of the first type SMA cable and the second type SMA cable in Embodiment 1 of the present invention.

[0021] Among them, 1. Bridge tower; 2. Bridge pier; 3. Box girder zero block; 4. Precast box girder segment; 5. Cantilever beam; 6. Type I SMA cable; 7. Type II SMA cable; 701. Anchor body; 702. SMA wire; 703. Heating element; 704. Copper foil heat-conducting layer. Detailed Implementation

[0022] The present invention will be further described in detail below with reference to the accompanying drawings.

[0023] Example 1 Reference Figure 1 This embodiment of a self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction includes: Reference Figure 2 and Figure 3 S1. Cast box girder No. 0 block 3 on the tower plate of bridge tower 1, cast cantilever beam 5 along both sides of box girder No. 0 block 3, and gradually splice precast box girder segments 4 along both sides of box girder No. 0 block 3. Step S1 is the main construction process for cable-stayed bridges in existing technology. It involves binding the reinforcing steel skeleton of box girder block 3, with the skeleton fixed to the pre-embedded reinforcing steel of bridge tower 1 and pier 2 using double-sided lap welding. Box girder block 3 is poured using C60 high-performance concrete, employing a layered vibration compaction process. After pouring, it is covered with geotextile and cured with water. When pouring cantilever beams 5 along both sides of box girder block 3, the length of cantilever beams 5 is determined to be 1.1 times the width of the bridge deck, with a rectangular cross-section. Their reinforcing steel skeleton is mechanically connected to the reinforcing steel skeleton of box girder block 3, which can be achieved using sleeve extrusion connections. C60 concrete is also used for pouring, and curing is performed according to the aforementioned requirements. When gradually splicing precast box girder segments 4 along both sides of box girder block 3, the precast box girder segments 4 are first transported to the splicing position using a bridge erecting machine, and epoxy adhesive is applied to the splicing surface. During the splicing process, a total station is used to monitor the axial deviation of the beam segments in real time to ensure that the main beam linearity meets the design requirements.

[0024] S2. Cast cantilever beams 5 along both sides of box girder zero block 3, install pre-tightened first type SMA cables 6 between cantilever beams 5 and bridge towers 1 and piers 2, and install pre-tightened second type SMA cables 7 in adjacent precast box girder sections 4. When installing the pre-tensioned first type SMA cable 6 between the cantilever beam 5 and the bridge tower 1 and pier 2, the first type SMA cable 6 uses NiTi shape memory alloy wire bundle, which consists of 30 NiTi alloy wires with a diameter of 5mm. The phase transformation temperature range is: martensitic phase transformation temperature Ms=-18℃~-15℃, austenitic phase transformation temperature As=58℃~62℃, and hyperelastic strain range 8%-10%. When installing the pre-tensioned second type SMA cable 7 in the adjacent precast box girder segment 4, the second type SMA cable 7 uses NiTi alloy wire bundle of the same material, which consists of 5 NiTi alloy wires with a diameter of 5mm. The cable surface is covered with a polytetrafluoroethylene protective layer to prevent corrosion during service.

[0025] The cantilever beams 5 are cast along both sides of the zero block 3 of the box girder, and pre-tensioned first-type SMA cables are installed between the cantilever beams 5 and the bridge towers 1 and piers 2, including: S21. Along the transverse sides of the zero block 3 of the box girder and in the longitudinal direction close to the bridge tower 1 and the bridge pier 2, the steel reinforcement cage of the cantilever beam 5 is respectively arranged. The steel reinforcement cage is welded and fixed to the pre-embedded steel reinforcement of the zero block 3 of the box girder, the bridge tower 1 and the bridge pier 2 to form an integrated load-bearing structure. Reference Figure 4 S22. Pre-set SMA cable anchor holes at the ends of cantilever beam 5 and then pour cantilever beam 5. One end of the first type SMA cable 6 is passed through the anchor hole at the end of the cantilever beam 5 and fixed by the anchor. The other end extends and is anchored to the preset anchor points of the bridge tower 1 and the bridge pier 2. The first type SMA cable 6 is pre-tightened using a tensioning device.

[0026] When pre-setting SMA cable anchorage holes at the ends of cantilever beam 5, Q355 steel sleeves are pre-embedded. During installation, positioning steel bars are used to fix the welded steel reinforcement skeleton of cantilever beam 5. After passing one end of the first type SMA cable 6 through the anchorage hole at the end of cantilever beam 5, it is fixed by cold-cast anchorage; the other end extends and is anchored to the pre-set anchorage points of bridge tower 1 and bridge pier 2. A 2000kN electric jack is used to pre-tighten the first type SMA cable 6. The pre-tightening is carried out in 3 stages: the first stage tensioning is to 30% of the design pre-tightening force and holding for 5 minutes; the second stage tensioning is to 70% and holding for 5 minutes; the third stage tensioning is to 100% and holding for 10 minutes. During the pre-tightening process, a built-in pressure sensor is used to monitor the pre-tightening force in real time.

[0027] Reference Figure 5 and Figure 6 The second type of SMA cable 7 pre-tensioned and installed in adjacent precast box girder segments 4 includes: S23. Drains are machined on the side of the precast box girder segment 4. After the drainage channels of two adjacent precast box girder segments 4 are connected, they form a U-shape that is symmetrical towards the center of the interior. Multiple U-shaped drainage channels are distributed longitudinally in a stepped manner. When processing the ducts on the side of the precast box girder segment 4, the ducts of two adjacent precast box girder segments 4 are joined to form a U-shape that is symmetrical towards the inward center. Multiple U-shaped ducts are distributed in a stepped manner along the longitudinal direction. Care should be taken to avoid connecting with the ducts through which steel reinforcement chains are installed. The steel reinforcement chains are used to connect adjacent precast box girder segments 4 and also need to be pre-tightened. The steel reinforcement chains are existing technology and will not be discussed here. The U-shaped ducts within the same cross section are arranged symmetrically.

[0028] S24. Insert the second type SMA cable 7 into the U-shaped duct entrance of one side beam segment and out of the duct exit of the other side beam segment, and fix the cable with temporary anchors at both ends; Before inserting the second type SMA cable 7 into the U-shaped duct entrance of one side of the beam segment, apply silicone grease lubricant to the inner wall of the duct. Use an electric cable inserter to insert the cable into the duct, and monitor the cable tension in real time during the insertion process. After the cable exits from the duct outlet of the other side of the beam segment, install wedge-shaped temporary anchors at both ends of the cable for fixation. Place neoprene rubber pads between the temporary anchors and the side of the beam segment to avoid direct contact between the anchors and the beam segment, which could cause concrete damage. Apply an initial preload of 5kN during temporary fixation to prevent the cable from slipping during subsequent construction.

[0029] Reference Figure 1 S25. After each preset number of precast box girder segments 4 are installed, a steel wire lock is suspended between the precast box girder segment 4 and the bridge tower 1. One end of the steel wire lock is anchored to the preset lifting point of the bridge tower 1, and the other end is anchored to the preset lifting lug of the precast box girder segment 4. This is repeated until the precast box girder segment 4 extends to the preset closing position in the middle of the river and the closing is completed. After every three precast box girder segments 4 are installed, steel wire locks are suspended between the precast box girder segment 4 and the bridge tower 1. After suspension, a tension meter is used to monitor the tension of the steel wire locks to ensure that the tension value is controlled between 20-30kN. The above installation and suspension steps are repeated until the precast box girder segment 4 extends to the preset closing position in the middle of the river. When closing, the elevation of the two ends of the beam segment is adjusted first, then the temporary steel support is welded, and finally the concrete of the closing section is poured to complete the closing.

[0030] S26. Adjust the prestressing force according to the real-time strain of the precast box girder segment. When the strain of the girder segment... At that time, according to the formula Calculate the preload adjustment amount, where, To preset the strain triggering coefficient, To preset the pre-tightening adjustment coefficient, The elastic modulus of SMA Let SMA cable cross-sectional area be 1. Design initial stress for type II SMA cables. The cross-sectional area of ​​the beam segment is used to apply an adjustment amount through a tensioning device until the strain of the beam segment is less than [a certain value]. ,in, This is the preset strain control coefficient.

[0031] When adjusting the preload based on the real-time strain of the precast box girder segment, firstly, a full-bridge strain gauge (model BF120-3AA) is attached to the neutral axis at the bottom of the precast box girder segment. The strain of the girder segment is then collected in real-time using a data acquisition device. Based on the real-time strain of the girder segment, the preload adjustment amount is calculated according to the formula. Take 60 GPa (the room temperature elastic modulus of NiTi shape memory alloy). Calculated based on the number of filament bundles and the cross-sectional area of ​​a single filament. (Take 300MPa (determined based on the lateral resistance requirements of the beam segment). Calculated based on the actual cross-sectional dimensions of the box girder segment. A preload of 0.8 is used. The calculated preload adjustment is applied using a 1000kN electric jack. During tensioning, the beam segment strain is monitored at a frequency of 1Hz until the beam segment strain is less than 0.8. , The preset strain control coefficient is set to 0.8, i.e. strain ≤ 1200με. After adjustment, the cable anchor is locked with a torque wrench, and the adjusted preload and strain data are recorded for subsequent parameter optimization of the AI ​​model.

[0032] Reference Figure 7 The first type of SMA cable 6 and the second type of SMA cable 7 include an anchor body 701, SMA wires 702 connected between the anchor bodies 701, a heating element 703 disposed in the anchor, and a copper foil heat-conducting layer 704. The anchor body 701 has a through hole for passing through the SMA wire 702. The inner wall of the through hole has an annular groove along the circumference. The heating element 703 is embedded in the annular groove. The copper foil heat-conducting layer 704 wraps around the heating element 703 and the SMA wire 702. The outer side of the copper foil heat-conducting layer 704 is in close contact with the inner wall of the through hole.

[0033] S3. Acquire multi-source monitoring data, including stress data and temperature data of SMA cables, lateral displacement data and strain data of precast box girder segments, and axial force data and bridge deck load data of bridge towers. The acquisition of multi-source monitoring data includes: S31. A foil strain sensor is attached to the middle section of the first type of SMA cable, and a thermocouple sensor is wrapped around the first type of SMA cable near the anchor body. Stress data and temperature data of the first type of SMA cable are collected respectively. S32. Install laser displacement sensors on both sides of the transverse side of the zero block of the box girder to collect lateral displacement data, and attach full-bridge strain gauges to the bottom of the precast box girder segment to collect strain data. S33. Install an axial force sensor on the underside of the bridge tower to collect axial force data of the bridge tower; S34. Install weighing sensors on the bridge deck to collect bridge deck load data; S35. Transmit all sensor data to the data acquisition terminal via industrial bus.

[0034] A foil strain sensor is attached to the middle section of the first type SMA cable. The sensor is a BF350-3AA type foil strain gauge. A thermocouple sensor is wrapped around the first type SMA cable near the anchor body. A K-type thermocouple is selected. The thermocouple probe is bonded to the cable surface with thermally conductive silicone grease. When collecting stress data of the first type SMA cable, the strain signal is converted into stress value through a strain gauge. Temperature data is directly transmitted through a thermocouple transmitter.

[0035] Laser displacement sensors were installed on both sides of the zero block of the box girder; laser triangular displacement sensors were selected. Full-bridge strain gauges were attached to the bottom of the precast box girder segment; BX120-5AA type full-bridge strain gauges were selected.

[0036] An axial force sensor, specifically a vibrating wire axial force gauge, is installed on the underside of the bridge tower. The sensor is connected to the pre-embedded steel plate in the bridge tower foundation via a flange. During installation, ensure that the sensor axis coincides with the central axis of the bridge tower to prevent collision damage. When collecting axial force data from the bridge tower, the frequency signal is read using a vibrating wire reader, and then converted into an axial force value according to the calibration formula.

[0037] Weighing sensors, employing quartz crystals, are installed on the bridge deck. Two sensors are deployed along the transverse direction of the bridge deck for each lane, embedded within the pavement layer and sealed with epoxy resin. When collecting bridge load data, the weighing controller calculates vehicle axle load, total weight, and travel speed in real time, and records the travel time and lane position information of heavy vehicles.

[0038] S4. Construct an SMA stress attenuation prediction model based on a time-series neural network. Input multi-source monitoring data into the SMA stress attenuation prediction model. After training, output the stress attenuation curves of the first type of SMA cable and the second type of SMA cable within a future preset period. Calculate the self-resetting critical intervention threshold of the SMA cable. Compare the stress attenuation curve with the critical intervention threshold to determine the stress nodes that are about to fall below the threshold. The construction of the SMA stress attenuation prediction model based on a time-series neural network includes: S41. Construct an LSTM time series prediction model and train the LSTM time series prediction model using historical multi-source monitoring data; When constructing the LSTM time-series prediction model, the model structure includes: an input layer (number of neurons = feature dimension, i.e., 6 features correspond to 6 neurons), two hidden layers, with the first LSTM layer having 64 neurons and the second LSTM layer having 32 neurons, both using the ReLU activation function, a dropout layer (dropout rate 0.2 to prevent overfitting), and an output layer (1 neuron, outputting the predicted stress value). When training the model using historical multi-source monitoring data, the historical data duration is ≥1 year to ensure coverage of different seasons and traffic load conditions. Before training, the data is divided into time series segments (using a 24-hour sliding window, i.e., the data from the previous 24 hours predicts the stress for the following hour).

[0039] S42. After the model training is complete, the stress decay curves of two types of SMA cables are output, and the curves are fitted according to the exponential decay formula: , among which, among which for SMA cable stress at any time For the initial preload stress, The attenuation coefficient is... For service time, For residual stress, the first type of SMA cable Positively correlated with the axial force of the bridge tower, the second type of SMA cable Positively correlated with the frequency of bridge deck load data; The stress decay curve output after model training, when fitted according to the exponential decay formula, represents the initial preload stress of the first type of SMA cable. Take 300MPa (consistent with the construction preload), residual stress The residual stress after complete relaxation of the SMA material is taken as 50 MPa, and the attenuation coefficient k1 ranges from 0.001 to 0.005 per day (k1 increases by 0.0005 per day for every 1000 kN increase in the axial force of the bridge tower); the initial preload stress of the second type of SMA cable is... Take 250 MPa, residual stress With a pressure of 40 MPa, the attenuation coefficient k2 ranges from 0.002 to 0.008 per day. For every 50 additional heavy vehicle passages per day on the bridge surface, k2 increases by 0.001 per day. The least squares method is used for curve fitting, and a goodness of fit R² ≥ 0.95 is considered a valid fit.

[0040] S43. Calculate the critical intervention threshold for the first type of SMA cable, combined with span correction, using the following formula: ,in , To allow for stress loss rate, This is the span correction factor; When calculating the critical intervention threshold for the first type of SMA cable, (Design initial stress) is taken as 300 MPa, and the allowable stress loss rate is determined according to S431 - S434 (the value range is 5% - 20%); the span correction coefficient is determined according to the main span L of the cable-stayed bridge. When L ≤ 200 m, = 1.0; when 200 m < L ≤ 500 m, = 1.1; when L > 500 m, = 1.2. The larger the span, the higher the requirement for lateral deviation control, and the corresponding threshold increases. For example: for a cable-stayed bridge with a main span of 300 m, when ω = 10%, = 300×(1 - 10%) = 270 MPa, = 270×1.1 = 297 MPa.

[0041] S44. Calculate the critical intervention threshold of the second type of SMA cable, combined with temperature difference correction. The formula is: , where , is the temperature difference correction coefficient; When calculating the critical intervention threshold of the second type of SMA cable, (Design initial stress) is taken as 250 MPa, and the allowable stress loss rate is the same as S431 - S434 (the value range is 8% - 25%); the temperature difference correction coefficient is determined according to the annual maximum temperature difference ΔT at the bridge site: when ΔT ≤ 30 °C, =​​​​​​​​​​​​​​​​​​​​​​​​​​​​The stress over -24 hours remains above the threshold. A buffer mechanism is implemented during the judgment process: if three consecutive data points (calculated at 1 hour / point) meet the requirements... Critical, and corresponding Stress at time > Only then was the final judgment made. This is to identify stress nodes and avoid misjudgments caused by errors in a single measurement.

[0043] S46. The node determination for the first type of SMA cable must satisfy the lateral displacement of the box girder. The node determination for the second type of SMA cable must meet the strain requirements of the precast box girder segment. ,in, To preset the lateral deviation ratio coefficient, This is the preset strain ratio coefficient.

[0044] In the node determination of the first type of SMA cable, (Maximum allowable lateral displacement of box girder zero block) is set to 50mm, with a preset lateral displacement ratio coefficient. =0.8, meaning when the lateral displacement is ≥40mm, the node is determined based on stress conditions; in the node determination of the second type of SMA cable, (Maximum allowable strain of precast box girder segment) is taken as 1500με, with a preset strain ratio coefficient. =0.7, meaning when the strain ≥1050με, the node is determined in conjunction with the stress condition. Both types of node determination require simultaneous satisfaction of stress, displacement, and strain conditions to ensure the accuracy of intervention.

[0045] The construction of the LSTM time series prediction model, and the training of the LSTM time series prediction model using multi-source monitoring data, includes: S411. Obtain historical multi-source monitoring data as training data, and divide it into training set, validation set and test set; S412. The LSTM time series prediction model is trained using the Adam optimization algorithm, with the root mean square error as the loss function, as shown in the formula. Where N is the number of samples, To predict stress, For the first The actual stress of each sample; S413. Input the test set data into the trained model and calculate the prediction error of the first type of SMA cable. Prediction error of type II SMA cable If If the model satisfies the generalization requirement, This is the preset error threshold.

[0046] The process for determining the allowable stress loss rate includes: S431. Obtain fatigue load data for two types of SMA cables. The fatigue load for the first type of SMA cable is the axial force fluctuation value of the bridge tower. The fatigue load of the second type of SMA cable is the average number of heavy vehicles passing over the bridge deck per day. ; When obtaining fatigue load data for two types of SMA cables, the axial force fluctuation value of the bridge tower is... The daily difference between the maximum and minimum axial force of the bridge towers is used, with a data collection period of 30 consecutive days. The average daily number of heavy vehicles passing over the bridge deck, N, represents the number of vehicles with a total weight ≥ 30t per day, also collected over 30 consecutive days, and the average value is used. This represents the maximum fluctuation value of the axial force of the bridge towers for the first type of SMA cables. Based on a design value of 5000kN, the maximum daily heavy vehicle traffic on the bridge deck for Class II SMA cables is... The location is determined based on the traffic planning for the bridge site.

[0047] S432. Calculate the fatigue influence coefficient using the following formula: The fatigue influence coefficient of the first type of SMA cable is calculated using the following formula: in, Based on the stress loss rate, The influence coefficient of axial force on bridge towers. `max` represents the maximum fluctuation value of the axial force in the bridge tower. When calculating the fatigue influence coefficient of the first type of SMA cable, the foundation stress loss rate is also considered. =5%, natural stress loss under no fatigue load, influence coefficient of axial force in bridge towers. =0.15, reflecting the weight of the influence of axial force fluctuation on cable fatigue.

[0048] S433. Calculate the fatigue influence coefficient of the second type of SMA cable using the following formula: in, This is the bridge deck load influence coefficient. The maximum daily number of heavy vehicles passing over the bridge deck; when calculating the fatigue influence coefficient of the second type of SMA cable, the foundation stress loss rate. =8%, Bridge deck load influence coefficient =0.17.

[0049] S434. Determine the allowable stress loss rate for Class I SMA cables. Type II SMA cable When determining the allowable stress loss rate, the first type of SMA cable... The value ranges from 5% to 20%, for the second type of SMA cable. The value ranges from 8% to 25%. The allowable stress loss rate must meet the following requirement: the residual preload of the cable at this loss rate can still balance 70% of the design off-center load, ensuring that the structure is in a safe state before intervention.

[0050] The S5.SMA stress attenuation prediction model sends an activation command to the PLC control module. Based on the stress difference corresponding to the stress node, the PLC control module activates the heating element built into the SMA cable to heat the first and second type of SMA cables to restore their pre-tension state, thereby driving the zero block of the box girder and the precast box girder segment to correct lateral deviation and achieve self-resetting of the cable-stayed bridge.

[0051] The PLC control module activates the heating element built into the SMA cable based on the stress difference corresponding to the stress node and the cable number, including: The S51.PLC control module receives the activation command sent by the AI ​​model, extracts the stress difference value corresponding to the stress node, and the stress difference value of the first type of SMA cable. Stress difference of type II SMA cable ; S52. Calculate the heating temperature increment based on the stress difference, using the following formula: ,in The stress temperature coefficient of SMA; When calculating the heating temperature increment based on the stress difference, the stress temperature coefficient of SMA The value is 2.0 MPa / ℃, based on measured data of NiTi shape memory alloy, with an error range of ±0.2 MPa / ℃. The calculated result is rounded up.

[0052] Simultaneously, the heating temperature increment must satisfy: target temperature (current temperature + ΔT) ≤ austenite termination temperature Af of the SMA cable, where Af = 65℃ (to avoid excessive temperature leading to material property degradation). If the calculated value exceeds this, it will be forcibly... Let Af be the difference between the current temperature and the current temperature. For example, when the current temperature is 60℃, the maximum ΔT is 5℃.

[0053] S53. Adjust the heating power of the first type of SMA cable according to the axial force of the bridge tower. When the axial force is not less than the preset axial force value, use the preset rated power; when the axial force is less than the preset axial force value, use the full rated power. S54. Adjust the heating power of the second type of SMA cable according to the bridge deck load. When the load is not less than the preset load, the preset rated power is used. When the load is less than the preset load, the full rated power is used. During the heating process, the temperature is fed back in real time through thermocouples. When the target temperature is reached, switch to heat preservation mode and continue to preserve for the preset time. When adjusting the heating power of the second type of SMA cable, a preset load value of 30t is used, and the current maximum single-lane load is monitored in real time by bridge deck weighing sensors. The heating element parameters of the second type of SMA cable are the same as those of the first type, with a preset rated power of 500W and a full rated power of 800W. When the bridge deck load is ≥30t, the preset rated power is used for heating; when the bridge deck load is <30t, full rated power is used for heating. During the heating process, thermocouple sensors provide feedback on the cable temperature at a frequency of 1Hz, and the PLC adjusts the heating power through a PID algorithm. After reaching the target temperature, the system switches to heat preservation mode and maintains the heat preservation for a preset time of 30 minutes to ensure that the SMA material fully completes the austenitic phase transformation and stably restores the preload.

[0054] S55. Monitor the stress recovery of the first type of SMA cable to... And the lateral displacement of the box girder is less than The stress in the second type of SMA cable rebounded to And the strain of the precast box girder segment is less than If the self-reset is completed, the PLC control module will cut off the heating circuit.

[0055] During self-resetting monitoring, the stress of the first type of SMA cable was acquired in real time using a foil strain sensor, and the lateral displacement of the box girder was acquired using a laser displacement sensor. =50mm, the maximum allowable lateral displacement of the zero block of the box girder; the stress of the second type of SMA cable is also collected by strain sensors, and the strain of the precast box girder segment is collected by full-bridge strain gauges, among which =1500με, the maximum allowable strain of the beam segment. The criterion is: within 30 consecutive seconds, the stress of the first type of cable ≥ Furthermore, the lateral displacement is <50mm, and the stress of the second type of cable is ≥ Furthermore, if the strain of the beam segment is less than 1500με, then the self-resetting process is considered complete if the above conditions are met.

[0056] The PLC control module then cuts off the heating circuit, the output module disconnects the relay, and records the self-reset duration, energy consumption, final stress, displacement, and strain values, storing these data in the data acquisition terminal as feedback data for AI model optimization. If the heating time exceeds 2 hours and the judgment condition is still not met, the PLC automatically triggers a fault alarm and stops heating, awaiting manual intervention.

[0057] Example 2 A self-resetting intelligent control system for cable-stayed bridges based on SMA and AI prediction includes: The construction module is configured to cast the No. 0 block of the box girder on the tower plate of the bridge tower, cast the cantilever beam along both sides of the No. 0 block of the box girder, and gradually splice the precast box girder segments along both sides of the No. 0 block of the box girder. The SMA cable installation pretensioning module is configured to cast cantilever beams along both sides of the zero block of the box girder, install pretensioned first-type SMA cables between the cantilever beams and the bridge towers and piers, and install pretensioned second-type SMA cables in adjacent precast box girder segments. The data acquisition module is configured to acquire multi-source monitoring data, including stress data and temperature data of SMA cables, lateral displacement data and strain data of precast box girder segments, and axial force data and bridge deck load data of bridge towers. The stress node judgment module is configured to construct an SMA stress attenuation prediction model based on a time-series neural network, input multi-source monitoring data into the SMA stress attenuation prediction model, output the stress attenuation curves of the first type of SMA cable and the second type of SMA cable within a preset period after training, calculate the self-resetting critical intervention threshold of the SMA cable, compare the stress attenuation curve with the critical intervention threshold, and judge the stress node that is about to fall below the threshold. The PLC control module is configured such that the SMA stress attenuation prediction model sends an activation command to the PLC control module. Based on the stress difference corresponding to the stress node, the PLC control module activates the heating element built into the SMA cable to heat the first and second type of SMA cables to restore their pretension state, thereby driving the zero block of the box girder and the precast box girder segment to correct lateral deviation and achieve self-resetting of the cable-stayed bridge.

[0058] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction, characterized in that, include: S1. Cast the No. 0 box girder on the tower plate of the bridge tower, cast the cantilever beam along both sides of the No. 0 box girder, and gradually splice the precast box girder segments along both sides of the No. 0 box girder. S2. Cast cantilever beams along both sides of the zero block of the box girder, install pre-tightened first-type SMA cables between the cantilever beams and the bridge towers and piers, and install pre-tightened second-type SMA cables in adjacent precast box girder sections. S3. Acquire multi-source monitoring data, including stress data and temperature data of SMA cables, lateral displacement data and strain data of precast box girder segments, and axial force data and bridge deck load data of bridge towers. S4. Construct an SMA stress attenuation prediction model based on a time-series neural network. Input multi-source monitoring data into the SMA stress attenuation prediction model. After training, output the stress attenuation curves of the first type of SMA cable and the second type of SMA cable within a future preset period. Calculate the self-resetting critical intervention threshold of the SMA cable. Compare the stress attenuation curve with the critical intervention threshold to determine the stress nodes that are about to fall below the threshold. S5. The SMA stress attenuation prediction model sends an activation command to the PLC control module. The PLC control module activates the heating element built into the SMA cable according to the stress difference corresponding to the stress node, heats the first type of SMA cable and the second type of SMA cable to restore their pre-tension state, drives the zero block of the box girder and the precast box girder segment to correct the lateral deviation, so as to realize the self-resetting of the cable-stayed bridge.

2. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 1, characterized in that, The cantilever beams are cast along both sides of the zero block of the box girder, and pre-tensioned first-type SMA cables are installed between the cantilever beams and the bridge towers and piers, including: Cantilever beam steel reinforcement cages are arranged along the transverse sides of the zero block of the box girder and along the longitudinal direction close to the bridge tower and pier. The steel reinforcement cages are welded and fixed to the pre-embedded steel reinforcements of the zero block of the box girder, the bridge tower and the pier to form an integrated load-bearing structure. SMA cable anchoring holes are pre-set at the ends of the cantilever beam for cantilever beam casting; One end of the first type of SMA cable is passed through the anchorage hole at the end of the cantilever beam and fixed by the anchorage. The other end extends and is anchored to the preset anchorage point of the bridge tower and pier. The first type of SMA cable is pre-tightened using tensioning equipment.

3. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 1, characterized in that, The second type of SMA cable, which is pre-tensioned and installed in adjacent precast box girder segments, includes: Holes are machined on the side of the precast box girder segment. After the holes of two adjacent precast box girder segments are connected, they form a U-shape that is symmetrical towards the center of the interior. Multiple U-shaped holes are distributed longitudinally in a stepped manner. The second type of SMA cable is inserted into the U-shaped duct entrance on one side of the beam segment and exited from the duct exit on the other side of the beam segment. Temporary anchors are fitted at both ends of the cable for fixation. After each preset number of precast box girder segments are installed, steel wire locks are suspended between the precast box girder segments and the bridge towers. One end of the steel wire lock is anchored to a preset lifting point on the bridge tower, and the other end is anchored to a preset lifting lug on the precast box girder segment. This process is repeated until the precast box girder segment extends to the preset closure position in the middle of the river and the closure is completed.

4. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 1, characterized in that, The first type of SMA cable and the second type of SMA cable include an anchor body, SMA wires connected between the anchor bodies, a heating element and a copper foil heat-conducting layer disposed in the anchor. The anchor body has a through hole for passing through the SMA wire. An annular groove is formed on the inner wall of the through hole in the circumferential direction. The heating element is embedded in the annular groove. The copper foil heat-conducting layer wraps around the heating element and the outside of the SMA wire. The outside of the copper foil heat-conducting layer is in close contact with the inner wall of the through hole.

5. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 1, characterized in that, The acquisition of multi-source monitoring data includes: Foil strain sensors were attached to the middle section of the first type of SMA cable, and thermocouple sensors were wrapped around the first type of SMA cable near the anchor body. Stress and temperature data of the first type of SMA cable were collected respectively. Laser displacement sensors were installed on both sides of the zero block of the box girder to collect lateral displacement data, and strain gauges with full bridging method were attached to the bottom of the precast box girder segment to collect strain data. An axial force sensor is installed on the underside of the bridge tower to collect axial force data of the bridge tower; Weighing sensors are installed on the bridge deck to collect bridge deck load data; All sensor data is transmitted to the data acquisition terminal via an industrial bus.

6. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 1, characterized in that, The construction of the SMA stress attenuation prediction model based on a time-series neural network includes: An LSTM time series prediction model was constructed and trained using historical multi-source monitoring data. After the model training is completed, the stress decay curves of the two types of SMA cables are output, and the curves are fitted according to the exponential decay formula: , among which, among which for SMA cable stress at any time For the initial preload stress, The attenuation coefficient is... For service time, For residual stress, the first type of SMA cable Positively correlated with the axial force of the bridge tower, the second type of SMA cable Positively correlated with the frequency of bridge deck load data; The critical intervention threshold for the first type of SMA cable, combined with span correction, is calculated using the following formula: ,in , To allow for stress loss rate, This is the span correction factor; The critical intervention threshold for type II SMA cables, combined with temperature difference correction, is calculated using the following formula: ,in , This is the temperature difference correction factor; Comparing the stress decay curves of the two types of SMA cables with their respective critical intervention thresholds, if a time point exists... ,and Then determine For stress nodes, where ; The node determination for the first type of SMA cable must satisfy the lateral displacement of the box girder. The node determination for the second type of SMA cable must meet the strain requirements of the precast box girder segment. ,in, To preset the lateral deviation ratio coefficient, This is the preset strain ratio coefficient.

7. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 6, characterized in that, The construction of the LSTM time series prediction model, and the training of the LSTM time series prediction model using multi-source monitoring data, includes: Historical multi-source monitoring data is obtained as training data and divided into training set, validation set and test set; The LSTM time series prediction model is trained using the Adam optimization algorithm, with the root mean square error as the loss function, as shown in the formula below. Where N is the number of samples, To predict stress, For the first The actual stress of each sample; Input the test set data into the trained model and calculate the prediction error of the first type of SMA cable. Prediction error of type II SMA cable If If the model satisfies the generalization requirement, This is a preset error threshold.

8. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 7, characterized in that, The process for determining the allowable stress loss rate includes: Fatigue load data for two types of SMA cables were obtained. The fatigue load for the first type of SMA cable was the axial force fluctuation value of the bridge tower. The fatigue load of the second type of SMA cable is the average number of heavy vehicles passing over the bridge deck per day. ; The fatigue influence coefficient is calculated using the following formula: The fatigue influence coefficient of the first type of SMA cable is calculated using the following formula: in, Based on the stress loss rate, The influence coefficient of axial force on bridge towers. max is the maximum fluctuation value of the axial force of the bridge tower; The fatigue influence coefficient of the second type of SMA cable is calculated using the following formula: in, This is the bridge deck load influence coefficient. This represents the maximum daily number of heavy vehicles that can pass over the bridge. Determine the allowable stress loss rate for Class I SMA cables. Type II SMA cable .

9. The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction according to claim 8, characterized in that, The PLC control module activates the heating element built into the SMA cable based on the stress difference corresponding to the stress node, including: The PLC control module receives the activation command sent by the AI ​​model, extracts the stress difference value corresponding to the stress node, and the stress difference value of the first type of SMA cable. Stress difference of type II SMA cable ; The heating temperature increment is calculated based on the stress difference using the following formula: ,in The stress temperature coefficient of SMA; The heating power of the first type of SMA cable is adjusted according to the axial force of the bridge tower. When the axial force is not less than the preset axial force value, the preset rated power is used, and when the axial force is less than the preset axial force value, the full rated power is used. The heating power of the second type of SMA cable is adjusted according to the bridge deck load. When the load is not less than the preset load, the preset rated power is used, and when the load is less than the preset load, the full rated power is used. During the heating process, the temperature is fed back in real time through thermocouples. When the target temperature is reached, the mode is switched to heat preservation mode and the heat preservation time is continued for a preset time.

10. A self-resetting intelligent control system for cable-stayed bridges based on SMA and AI prediction, characterized in that, The self-resetting intelligent control method for cable-stayed bridges based on SMA and AI prediction, as described in any one of claims 1-9, includes: The construction module is configured to cast the No. 0 block of the box girder on the tower plate of the bridge tower, cast the cantilever beam along both sides of the No. 0 block of the box girder, and gradually splice the precast box girder segments along both sides of the No. 0 block of the box girder. The SMA cable installation pretensioning module is configured to cast cantilever beams along both sides of the zero block of the box girder, install pretensioned first-type SMA cables between the cantilever beams and the bridge towers and piers, and install pretensioned second-type SMA cables in adjacent precast box girder segments. The data acquisition module is configured to acquire multi-source monitoring data, including stress data and temperature data of SMA cables, lateral displacement data and strain data of precast box girder segments, and axial force data and bridge deck load data of bridge towers. The stress node judgment module is configured to construct an SMA stress attenuation prediction model based on a time-series neural network, input multi-source monitoring data into the SMA stress attenuation prediction model, output the stress attenuation curves of the first type of SMA cable and the second type of SMA cable within a preset period after training, calculate the self-resetting critical intervention threshold of the SMA cable, compare the stress attenuation curve with the critical intervention threshold, and judge the stress node that is about to fall below the threshold. The PLC control module is configured such that the SMA stress attenuation prediction model sends an activation command to the PLC control module. Based on the stress difference corresponding to the stress node, the PLC control module activates the heating element built into the SMA cable to heat the first and second type of SMA cables to restore their pretension state, thereby driving the zero block of the box girder and the precast box girder segment to correct lateral deviation and achieve self-resetting of the cable-stayed bridge.

Images