Intelligent bolt system based on embedded torque monitoring and artificial intelligence analysis and application method thereof

Abstract

The invention belongs to the technical field of mechanical fasteners and intelligent manufacturing, and discloses an intelligent bolt system based on embedded torque monitoring and artificial intelligence (AI) analysis and an application method of the intelligent bolt system. According to the intelligent bolt, an inner cavity is formed in a screw rod, a strain gauge is fixed in the inner cavity, and a strain signal is led out to a signal output interface in a screw head through a signal transmission line. And an external data terminal obtains real-time strain data from the interface and converts the real-time strain data into tightening torque by using a calibration relation. And further inputting the torque data into an artificial intelligence analysis platform, analyzing and predicting the fatigue life, the loosening risk and the use state of the bolt through a large language model, deep learning and an expert system algorithm, and giving a maintenance strategy suggestion. According to the method, real-time state monitoring and intelligent decision support can be achieved under the normal service state of the bolt without an external additional sensor, the connection quality, the maintenance efficiency and the safety are greatly improved, and the method is suitable for various fields of aerospace, gas turbines, steam turbines, vehicles, bridges, energy equipment, precision machinery and the like.

Description

Technical Field

[0001] The present invention belongs to the field of mechanical fastening and intelligent manufacturing, and specifically relates to an intelligent bolt system and an application method thereof, which can integrate a strain sensing component inside a bolt to realize real-time torque detection, and combine an artificial intelligence model to analyze, predict and support decision-making on data. Background Art

[0002] Bolts are the most basic connectors in mechanical structures, and their tightening quality has a profound impact on the overall safety and performance of the system. Traditional tightening quality control often relies on torque wrenches or fixed torque tools, which have problems such as error deviation, inconvenient measurement, and reliance on manual experience. After the equipment is put into use, the bolts may gradually deteriorate due to fatigue, vibration, corrosion, loosening, etc. Existing technologies are mostly regular shutdown inspections and random inspections, which make it difficult to detect problems in a timely manner, let alone make forward-looking predictions on life and risks.

[0003] With the development of intelligent manufacturing and industrial Internet, users expect to achieve online monitoring and intelligent analysis of key fasteners, thereby improving predictive maintenance and fault prevention capabilities. However, the market lacks a complete solution that can embed the sensing function into the bolt structure itself, does not require external measurement components, and can combine artificial intelligence models for in-depth data analysis. Summary of the invention

[0004] The present invention provides an intelligent bolt system and application method based on embedded torque monitoring and AI analysis. An inner cavity is opened inside the screw rod, and a strain gauge (31) is fixed in the inner cavity. The strain gauge senses the stress state of the bolt in real time, and the signal transmission line (32) outputs the data to the signal output interface (4) in the screw head (2). After obtaining the strain signal, the external data terminal converts it into torque data, and comprehensively processes it through an artificial intelligence analysis platform to provide a reliable basis for maintenance decision-making.

[0005] Beneficial effects of the present invention: Real-time and convenient: Torque data can be obtained when the bolt is in normal use, without disassembly or additional sensing fixtures.

[0006] High-precision detection: Multiple strain gauge distribution and calibration improve data accuracy and stability.

[0007] Intelligent analysis and predictive maintenance: Use AI models to perform trend analysis, fatigue life prediction, loosening risk assessment, and output maintenance recommendations in natural language.

[0008] Wide applicability: It can be used in key connection parts in aerospace, gas turbines, steam turbines, wind power equipment, automobiles, rail transportation, energy equipment, bridge construction and precision machinery. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 : Schematic diagram of the structure of the smart bolt, including a screw rod (1), an inner cavity (11), a strain gauge (31), a signal transmission line (32), a screw head (2), a signal output interface (4), a sealing cover (5) and a mark (6).

[0010] Figure 2 : Schematic diagram of the smart bolt structure (bolt components are represented by simplified nodes and annotations).

[0011] Figure 3 : Smart bolt system structure diagram, showing the data flow between smart bolts, external data terminals, and AI analysis platform.

[0012] Figure 4 : Data processing and AI analysis flow chart, including steps from data collection, preprocessing, model analysis to maintenance suggestion output. DETAILED DESCRIPTION

[0013] (1) Basic structure design and material selection The screw (1) of the present invention can be made of high-strength alloy steel, stainless steel, high-temperature alloy or titanium alloy to ensure that the strength and rigidity requirements are still met after the inner cavity (11) is opened. The size and shape of the inner cavity are determined according to the bolt specifications and the strain gauge size. The strain gauge (31) is attached to the inner cavity wall, requiring the surface to be flat and clean, and using a high-performance adhesive to ensure long-term stability. The signal transmission line (32) needs to have good insulation and fatigue resistance, and be able to withstand the vibration and temperature changes in the service environment of the bolt.

[0014] (2) Strain gauge calibration and torque conversion Before leaving the factory, the bolts with installed strain gauges are calibrated and tested. The relationship between the resistance change of the strain gauge and the actual torque value is recorded under standard torque loading conditions to form a calibration curve or calibration coefficient. After the external data terminal obtains the strain data, it quickly converts the strain amount into a real-time torque value by looking up a table or calculating.

[0015] (3) External data terminal and communication method The external data terminal may be a portable data acquisition instrument, a fixedly installed industrial controller or an embedded electronic module. The terminal obtains data from the signal output interface (4) via a wired plug or a wireless communication module. In the case where a sealing cover (5) is present, the sealing cover (5) needs to be opened for connection during maintenance. When wireless communication is used, a micro wireless transmission module may be integrated into the screw head (2).

[0016] (4) AI analysis platform and data processing flow The AI ​​analysis platform can be deployed on a local server or in the cloud. The data processing steps include: Data preprocessing: filtering, denoising, and outlier detection.

[0017] Historical comparison and feature extraction: Extract bolt fatigue characteristic parameters by referring to previous data, standard specifications and equipment conditions.

[0018] Model analysis and prediction: Use deep learning or big language models to predict bolt fatigue life, loosening trend, and failure risk, and generate natural language maintenance recommendations.

[0019] Decision output: Provide maintenance personnel with clear operation guidance through visual interfaces and reports.

[0020] (5) Marking (6) and ease of maintenance A mark (6) is provided on the screw head (2), which helps maintenance personnel to quickly locate the direction of the signal interface and identify the distribution orientation of the strain gauge on site, thereby ensuring accurate operation during assembly and inspection.

[0021] Example 1 Application scenario: Wind turbine tower flange connection bolts In a wind farm, the tower flange is connected by a large number of bolts. The tightening status of each bolt has a profound impact on the tower safety and power generation efficiency. The traditional maintenance method is to shut down the machine regularly and use a torque wrench to randomly check the tightness of some bolts. This method has blind spots: Difficult to achieve full coverage: The sampling ratio is limited and potential bolts may be missed.

[0022] Cost and time loss: Frequent downtime affects power generation.

[0023] Lack of predictive ability: There is no way to know the cumulative extent of fatigue damage of bolts, and only passive maintenance can be performed afterwards.

[0024] After replacing some or all of the key bolts with the smart bolts of the present invention, the situation is greatly improved. Specific implementation steps: Installation stage: When assembling the tower flange, a certain number of smart bolts are selected to replace ordinary bolts. These bolts have been installed and calibrated with strain gauges (31) in the factory.

[0025] Data collection: When the wind turbine is operating normally, the external data terminal located in the tower is connected to the signal output interface (4) of the smart bolt through a wired connection or wirelessly, and automatically collects strain data every few hours and converts it into real-time torque values. If the wind speed changes drastically or a specific alarm condition occurs (such as a vibration sensor detecting strong vibration), the sampling frequency can be increased.

[0026] AI analysis: The terminal transmits torque data to the cloud AI analysis platform in real time. The platform compares the current data with historical data from the past few months and years, and combines wind farm meteorological records (wind speed, wind direction, temperature changes) and tower design load standards to evaluate the accumulation of bolt fatigue damage. Through the deep learning model, the platform can predict when the bolt may enter the fatigue failure stage. The large language model generates a natural language report based on the analysis results, for example: "The average torque of bolt No. 12 has dropped by about 5% in the past week. The fatigue model shows that its remaining safe service life is about 500 hours. It is recommended to tighten or replace it during the next low wind speed maintenance window." Decision support: After reviewing the report, maintenance engineers can take preventive measures for specific bolts instead of blindly checking all flange bolts. This reduces downtime and reduces the risk of working at height.

[0027] Practical experience shows that through the implementation scheme of the present invention, wind farms can achieve higher operating efficiency and equipment safety. The maintenance cycle is transformed from a fixed schedule to data-driven preventive maintenance, which is conducive to reducing sudden failures and unexpected downtime losses.

[0028] Example 2 Application scenario: Health monitoring of aircraft engine hanger bolts Aircraft engines are connected to the wing or fuselage structure through hanger bolts. The health of these key connectors is directly related to flight safety and reliability. Traditional aircraft maintenance plans adopt a regular overhaul model, and the inspection and maintenance of bolts are also replaced according to the schedule, lacking accurate evaluation based on actual use stress data.

[0029] The smart bolt of the present invention is applied to the connection of the aircraft engine hanger, and the implementation steps are as follows: Manufacturing and assembly: During the final assembly of aircraft engines, the smart bolts of the present invention are used to replace ordinary bolts in key hanger connections. Each smart bolt has been calibrated at the factory to ensure that within the given torque range, the strain gauge output corresponds to the torque value.

[0030] Data collection during flight operations: When the aircraft is put into operation, the onboard data recorder regularly reads the bolt strain signal through the built-in micro data terminal (timely sampling can be performed during level flight, take-off and landing, and turbulent environments), and transmits it to the ground maintenance center through ground inspection or wireless data link.

[0031] AI analysis platform processing: The AI ​​analysis platform of the ground maintenance center integrates the torque data from the bolts and cross-analyzes it with flight parameters (flight hours, number of take-offs and landings, vibration levels, temperature-altitude characteristics) and engine thrust conditions. The deep learning model is used to identify the development trend of fatigue damage. If the torque fluctuation of a hanger bolt increases in the last few flights and the average value deviates from the normal range, the model may determine that its fatigue loss has increased. The large language model gives natural language suggestions based on historical data of similar models, manufacturer standards and expert knowledge: "Taking into account the recent increase in turbulence and frequent thrust adjustments, the fatigue index of hanger bolt No. 5 is close to the limit value. It is recommended to focus on checking and possibly replacing the bolt during the next A inspection." Precision maintenance: Compared with traditional periodic maintenance, this solution allows airlines to replace high-risk bolts in a targeted manner during routine inspections, rather than blindly replacing them in full, saving costs and time. At the same time, it can increase safety margins and avoid uncontrollable failure of bolts during flight.

[0032] Long-term operation data show that through the intelligent bolt solution of the present invention, the operation of the aircraft can be dynamically monitored, and the maintenance strategy is transformed from time-driven to state-driven, reducing unnecessary waste of parts and improving the economy and safety of aircraft operation.

[0033] Expansion and feasibility analysis Based on the above examples, the present invention can also be extended to: Key bolts of automobile chassis and steering system: Monitor the bolt tightening status of vehicles under various road conditions, and assist in unmanned driving and intelligent vehicle maintenance.

[0034] Pressure vessel connecting bolts of petrochemical equipment: long-term online monitoring in high temperature, high pressure and corrosive environments, early warning of possible loosening and leakage risks.

[0035] Key node bolts of bridges or high-rise building structures: Combined with the structural health monitoring system, the load and fatigue process of the bridge can be quantified in real time to reduce the risk of bridge collapse.

[0036] Precision instruments and spacecraft assembly: Continuously monitor the tightness of bolts in extreme environments (such as space and deep sea) to ensure the safety of equipment missions.

[0037] The production process of the present invention can be combined with modern precision processing methods to form a stable inner cavity inside the screw through CNC precision drilling technology, and then install the strain gauge and lay the signal line in an automated manner. In terms of quality control, non-destructive testing methods (such as X-rays or ultrasound) can be used to ensure the quality of inner cavity processing and strain gauge installation.

[0038] In addition, the training of the artificial intelligence analysis platform can be continuously iterated and upgraded based on a large amount of historical data and fault case libraries. With the accumulation of data and model optimization, the accuracy of predictions and the credibility of maintenance recommendations will continue to improve. Through the industrial Internet, data from different equipment and different sites can be collected on a unified platform for big data analysis, which can make more reliable judgments on the common risks of similar models and similar equipment.

[0039] Industrial Applicability The present invention can significantly improve the information and intelligence level of mechanical connectors throughout their life cycle. Through automated measurement, intelligent analysis and natural language interpretation, the present invention helps engineers and maintenance personnel to shift from traditional "periodic maintenance" to "predictive maintenance", thereby improving resource utilization and equipment reliability. It is widely applicable to many industries including but not limited to aerospace, gas turbines, steam turbines, wind power, automobiles, rail transit, petrochemicals, bridge construction and precision manufacturing.

Claims

1. A smart bolt, comprising a screw rod (1) and a screw head (2) connected to one end of the screw rod (1), characterized in that: An inner cavity (11) is provided inside the screw rod (1) along the axial direction, and at least one strain gauge (31) is fixedly installed in the inner cavity (11). The strain gauge (31) is used to detect the strain change generated by the screw rod (1) during the tightening and loading process; the strain gauge (31) is connected to a signal output interface (4) located in the screw head (2) via a signal transmission line (32), and the signal output interface (4) can output the strain signal to an external data terminal.

2. The smart bolt according to claim 1, characterized in that: A plurality of strain gauges (31) are distributed in the inner cavity (11) along the circumference of the screw rod (1), and signals output by the plurality of strain gauges (31) are comprehensively processed to improve the accuracy and stability of torque data.

3. The smart bolt according to claim 1 or 2, characterized in that: The signal output interface (4) is a pluggable electrical connector or a wireless data interface, and a sealing cover (5) is provided outside the screw head (2) for protection.

4. The smart bolt according to any one of claims 1 to 3, characterized in that: A mark (6) is provided on the surface of the screw head (2) for indicating the installation direction of the strain gauge (31) or the bolt alignment position, so as to facilitate quick identification during installation, maintenance and data collection.

5. A smart bolt system, comprising the smart bolt according to any one of claims 1 to 4, an external data terminal and an artificial intelligence analysis platform; wherein: The external data terminal is used to receive the strain signal output by the strain gauge (31) and convert it into bolt torque data; the artificial intelligence analysis platform uses the torque data to analyze and predict the fatigue life, loosening risk and usage status of the bolt, and generates maintenance suggestions.

6. The smart bolt system according to claim 5, characterized in that: The artificial intelligence analysis platform includes a large language model, a deep learning network and an expert knowledge base. It comprehensively analyzes the torque data by integrating historical data, environmental parameters and standard specifications, and outputs analysis results and maintenance recommendations in natural language.

7. A method for applying a smart bolt, comprising the following steps: (1) Installing the smart bolt according to any one of claims 1 to 4 on the components to be connected and performing initial calibration; (2) The external data terminal obtains the strain signal through the signal output interface (4) and converts it into torque data through the calibration relationship; (3) The torque data is input into the artificial intelligence analysis platform for data processing and model analysis to obtain the bolt usage status and preventive maintenance recommendations; (4) According to the maintenance recommendations, corresponding operations are performed on the bolts, such as tightening, replacement or regular re-inspection.

8. The application method according to claim 7, characterized in that: The artificial intelligence analysis platform can update model parameters based on real-time data, and continuously optimize the accuracy of predictions and the practicality of recommendations by comparing historical data and industry standards.

Images