Steel rail flash welding process detection and evaluation device

Abstract

The invention provides a steel rail flash welding process detection and evaluation device, the device comprises a data detection module, a process analysis module and a quality inspection expert module, the data detection module is installed on a steel rail flash welding machine, and the data detection module transmits data with the process analysis module and the quality inspection expert module through wireless transmission. According to the invention, through the sensor, the communication module and the storage module which are installed on the rail welding equipment, field welding data are obtained in real time and a related database is established. Meanwhile, a welding model analysis module based on mechanical learning, a neural network and a fuzzy control algorithm is established, control over heat input in the welding process and the steel rail burning amount is achieved, the actual heat requirement is more effectively matched, and therefore the quality of a steel rail flash welding head is improved; and by comparing the actual deformation amount, the deformation resistance, the deformation speed and the predicted amount at the end of welding, the quality of the welding joint is evaluated by applying a machine learning or neural network algorithm, so that quantitative analysis of the quality of the joint is realized.

Description

Technical Field

[0001] The invention belongs to the field of railway rail welding process detection and evaluation, and in particular relates to a rail flash welding process detection and evaluation device. Background Art

[0002] Rail flash welding equipment includes mobile rail flash welders, fixed rail flash welders, mobile rail flash welding and heat treatment all-in-one machines and other equipment. Among the above equipment, mobile rail welding equipment is distributed in railway construction sites around the world, and fixed rail welding equipment is distributed in rail welding bases in different regions. The on-site construction environment is poor, resulting in the quality of welded joints cannot be guaranteed due to complex reasons during on-site welding. In the flash welding process, the main principle is to use the current to generate heating through the internal resistance of the rail, and to complete the welding by pressurized upsetting. The quality of flash welding depends on many factors, including the heat input and temperature distribution during the welding process, whether the liquid metal layer formed on the end face of the rail at the end of the flash welding is sufficient, the flatness of the end face of the rail during upsetting, and the appropriate upsetting force. These factors are the key to the success of flash welding.

[0003] The current rail flash welding process detection and on-site welding joint quality assessment technology is not perfect enough to improve the quality of rail flash welding joints. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention proposes a rail flash welding process detection and evaluation device.

[0005] A rail flash welding process detection and evaluation device, the device comprises a data detection module, a process analysis module and a quality inspection expert module, the data detection module is installed on a rail flash welding machine, and the data detection module transmits data with the process analysis module and the quality inspection expert module through wireless or wired transmission.

[0006] Among them, the data detection module is used to collect key data of existing rail flash welding equipment and establish a database for data retrieval and storage.

[0007] The process analysis module is used to establish a welding process model, analyze key parameters of the rail flash welding process and predict key results.

[0008] The quality inspection expert module is used to establish a welding process model, analyze key parameters of the rail flash welding process and predict key results.

[0009] Wherein, the method comprises:

[0010] The data detection module acquires key data of the welding process in real time through acquisition devices such as displacement sensors, voltage sensors, current sensors, pressure sensors, vibration sensors, temperature sensors, and flow sensors, and establishes a welding key database through communication technology and storage technology for the storage and call of key data;

[0011] The process analysis module uses this key data to predict the real-time temperature field of the welding process, and compares and corrects it with the target temperature field. On the basis of real-time monitoring and control, it realizes intelligent adjustment of welding process parameters, thereby improving the matching of effective welding heat input and temperature field.

[0012] The quality inspection expert module evaluates the quality of welded joints by comparing the actual deformation, deformation resistance, deformation speed and predicted value at the end of welding, and applies machine learning or neural network algorithms to achieve quantitative analysis of joint quality.

[0013] Among them, the key data information of the flash welding process obtained by the data detection module includes welding voltage, welding current, welding time, movable frame speed, rail consumption, upsetting amount, upsetting speed, upsetting force, etc.

[0014] Among them, the process analysis module is equipped with an intelligent heat and deformation demand adaptation system and a welding parameter model analysis system based on a combination of machine learning, neural networks and fuzzy control algorithms.

[0015] Therein, a quality assessment score is formed for each weld joint.

[0016] The present invention obtains on-site welding data in real time and establishes a relevant database through sensors, communication modules and storage modules installed on rail welding equipment. At the same time, a welding model analysis module based on machine learning, neural network and fuzzy control algorithm is established to retrieve and calculate key rail welding data in real time, predict the real-time temperature field during rail welding and predict the deformation, deformation resistance and deformation speed at the end of welding. By comparing the real-time temperature field prediction value with the process target temperature field at that moment, the fuzzy control algorithm is used to adjust the welding process parameters, realize the control of the heat input of the welding process and the burning amount of the rail, and more effectively match the actual heat demand, so as to improve the quality of the flash welding joint of the rail; according to the flash welding results, through type inspection, production inspection data sampling and targeted process tests, combined with expert experience to study the weight distribution plan of quality-related parameters, use machine learning or neural network technology to establish a quality prediction model, and form a quality assessment score for each welding joint. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is an overall schematic diagram of a rail flash welding process detection and evaluation device.

[0018] Figure 2 This is a structural diagram of the data detection module of the rail flash welding process detection and evaluation device.

[0019] Figure 3 This is a schematic diagram of the structure of the model analysis and quality inspection expert module of the rail flash welding process detection and evaluation device. DETAILED DESCRIPTION

[0020] A rail flash welding process detection and evaluation device provided by the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0021] Attached Figure 1-3 As shown, a rail flash welding process detection and evaluation device includes a data detection module, a process analysis module and a quality inspection expert module. The data detection module is installed on a rail flash welding machine, and the data detection module transmits data to the process analysis module and the quality inspection expert module via wireless transmission.

[0022] The data detection module is used to collect key data of existing rail flash welding equipment and establish a database for data retrieval and storage.

[0023] The process analysis module is used to establish a welding process model, analyze key parameters of the rail flash welding process and predict key results.

[0024] The quality inspection expert module is used to establish a welding process model, analyze key parameters of the rail flash welding process and predict key results.

[0025] The method comprises,

[0026] The data detection module acquires key data of the welding process in real time through acquisition devices such as displacement sensors, voltage sensors, current sensors, pressure sensors, vibration sensors, temperature sensors, and flow sensors, and establishes a welding key database through communication technology and storage technology for the storage and call of key data;

[0027] The process analysis module uses this key data to predict the real-time temperature field of the welding process, and compares and corrects it with the target temperature field. On the basis of real-time monitoring and control, it realizes intelligent adjustment of welding process parameters, thereby improving the matching of effective welding heat input and temperature field.

[0028] The quality inspection expert module evaluates the quality of welded joints by comparing the actual deformation, deformation resistance, deformation speed and predicted value at the end of welding, and applies machine learning or neural network algorithms to achieve quantitative analysis of joint quality.

[0029] The key data information of the flash welding process acquired by the data detection module includes welding voltage, welding current, welding time, moving frame speed, rail consumption, upsetting amount, upsetting speed, upsetting force, etc.

[0030] The process analysis module is equipped with an intelligent heat and deformation demand adaptation system and a welding parameter model analysis system based on a combination of machine learning, neural networks and fuzzy control algorithms.

[0031] A quality assessment score is generated for each weld joint.

[0032] In actual operation, the data monitoring module may include measuring devices such as current transformers, voltage transformers, displacement sensors, and database modules equipped in the welding machine control system. The measuring equipment acquires key data in the rail flash welding process in real time, and the MCU built into the welding control system processes and analyzes the data through specific algorithms, and the database built into the welding machine control system stores and calls the data. Data communication technology and storage technology are used to ensure the real-time acquisition and data accuracy of key data in the welding process.

[0033] The process analysis module calls the data monitoring module to obtain the key data of the rail flash welding process in real time - the weight parameters of the voltage, current, time, consumption and other parameters at different stages as the input of the machine learning and neural network prediction model, which is used to predict the flash welding temperature field and use the predicted value as the output of the model. By comparing the predicted temperature field with the target temperature field, the fuzzy control algorithm is used to adjust and correct the welding process parameters. The fuzzy control algorithm is used to adjust the welding process parameters to achieve the control of the heat input of the welding process and the amount of rail burnout, so as to more effectively match the actual heat demand, thereby improving the quality of the rail flash welding joint.

[0034] The quality inspection expert module is based on the welding quality control process of the process analysis module. It uses the measured generalized correlation parameters and welding process curves as inputs, and through type inspection, production inspection data sampling and targeted process tests, combined with expert experience to study the quality correlation parameter weight distribution scheme, and uses machine learning and neural network technology to establish a quality prediction model. By comparing and analyzing the difference and model calculation between the actual generalized correlation parameters in the flash welding process and the generalized correlation parameters during type inspection, a quality assessment score is formed for each welding joint.

[0035] The present invention can be applied to welding control software of equipment such as mobile rail flash welding machines, fixed rail flash welding machines and mobile rail flash welding and heat treatment all-in-one machines, to assist on-site operators in making process adjustments and improve the quality of welded joints.

[0036] The rail flash welding process detection and evaluation device can collect, store and call the key data of the welding process in real time, thereby providing a large amount of reliable key data for welding model establishment, welding result prediction and joint quality evaluation.

Claims

1. A rail flash welding process detection and evaluation device, characterized in that: The device comprises a data detection module, a process analysis module and a quality inspection expert module. The data detection module is installed on the rail flash welding machine. The data detection module transmits data with the process analysis module and the quality inspection expert module through wireless or wired transmission.

2. A rail flash welding process detection and evaluation device according to claim 1, characterized in that: The data detection module is used to collect key data of existing rail flash welding equipment and establish a database for data retrieval and storage.

3. A rail flash welding process detection and evaluation device according to claim 1, characterized in that: The process analysis module is used to establish a welding process model, analyze key parameters of the rail flash welding process and predict key results.

4. A rail flash welding process detection and evaluation device according to claim 1, characterized in that: The quality inspection expert module is used to establish a welding process model, analyze key parameters of the rail flash welding process and predict key results.

5. A rail flash welding process detection and evaluation device according to claim 1, characterized in that: The method comprises, The data detection module acquires key data of the welding process in real time through acquisition devices such as displacement sensors, voltage sensors, current sensors, pressure sensors, vibration sensors, temperature sensors, and flow sensors, and establishes a welding key database through communication technology and storage technology for the storage and call of key data; The process analysis module uses this key data to predict the real-time temperature field of the welding process, and compares and corrects it with the target temperature field. On the basis of real-time monitoring and control, it realizes intelligent adjustment of welding process parameters, thereby improving the matching of effective welding heat input and temperature field. The quality inspection expert module evaluates the quality of welded joints by comparing the actual deformation, deformation resistance, deformation speed and predicted value at the end of welding, and applies machine learning or neural network algorithms to achieve quantitative analysis of joint quality.

6. A rail flash welding process detection and evaluation device according to claim 5, characterized in that: The key data information of the flash welding process acquired by the data detection module includes welding voltage, welding current, welding time, moving frame speed, rail consumption, upsetting amount, upsetting speed, upsetting force, etc.

7. A rail flash welding process detection and evaluation device according to claim 5, characterized in that: The process analysis module is equipped with an intelligent welding process parameter demand adaptation system and a welding parameter model analysis system based on a combination of machine learning, neural networks and fuzzy control algorithms.

8. A rail flash welding process detection and evaluation device according to claim 5, characterized in that: The quality inspection expert module uses machine learning neural network technology to establish a quality prediction model. By comparing and analyzing the differences and model calculations between the actual generalized correlation parameters during flash welding and the generalized correlation parameters (target parameters) during type inspection, a quality assessment score is formed for each weld joint.

Images