A method and system for laser additive temperature control of rail defects

Abstract

This invention provides a method and system for controlling the temperature of laser additive manufacturing in rail defects. Relating to the field of laser additive manufacturing technology, the system acquires real-time acoustic and temperature signals from the molten region during the laser additive process. Based on the acoustic signals, it analyzes for any interference events and then assesses the reliability of the current temperature signal. According to the reliability assessment results, the system can intelligently adjust the laser output power, thereby precisely controlling the laser additive manufacturing temperature.

Description

Technical Field

[0001] This application relates to the field of laser additive manufacturing technology, and more specifically, to a method and system for controlling the temperature of laser additive manufacturing of rail defects. Background Technology

[0002] In railway infrastructure maintenance, laser additive manufacturing technology is widely used for repairing rail defects due to its unique advantages. This technology uses a high-energy laser beam to melt metal powder, forming a new material layer at the rail defect site to restore its original properties.

[0003] When a high-energy laser beam begins to scan and heat the surface of the rail, it first interacts with these residual surface impurities. These impurities, such as some organic matter or oxides, have different absorption rates of laser energy compared to the rail's bulk material. Some impurities may have higher absorption efficiency for specific wavelengths of laser light, leading to localized instantaneous overheating, or even rapid decomposition or vaporization. This rapid vaporization process produces tiny gas plumes or wisps of smoke that are rapidly ejected from the surface of the molten area or its vicinity.

[0004] These suddenly generated gas plumes or fumes not only disturb the surface morphology of the molten area, making it unstable, but may also form tiny bubbles within the molten area. If these bubbles fail to escape before the metal solidifies, they will create pores in the repair layer, known as "porosity defects." These fumes carrying high-temperature particles or gases can directly obstruct or interfere with the line of sight of infrared thermometers. Infrared sensors are designed to accurately measure the radiant temperature of the molten area's surface, but now they may receive radiant signals from the fumes, or the signals they receive from the molten area may be scattered or attenuated by the fumes. This results in the sensor's output temperature reading no longer being the true temperature of the molten area, but rather a "mixed" temperature that includes fumes, or an underestimated temperature.

[0005] When the control system receives such distorted and inaccurate temperature feedback signals, it adjusts the power based on this erroneous information. More challenging is the fact that the distribution of these surface impurities is often uneven. The control system must not only handle normal temperature fluctuations in the molten zone but also identify and process nonlinear and random temperature feedback anomalies caused by external disturbances in real time. Traditional PID control or preset model control, whose design logic is usually based on stable and predictable feedback signals, struggles to effectively distinguish between temperature changes within the molten zone itself and sensor signal anomalies caused by impurity interference. This confusion prevents the control system from making accurate judgments and timely compensations, ultimately leading to uneven distribution of problems within the repair layer, such as excessively high porosity in localized areas or coarse microstructure in localized areas. This creates new safety hazards on the repaired rail, reducing the reliability and durability of the repair.

[0006] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0007] This application discloses a method and system for temperature control in laser additive manufacturing of rail defects, aiming to solve the problem that in the existing laser additive repair process of rail defects, the infrared temperature measurement device feedback signal is inaccurate due to interference from impurities on the rail surface, which in turn affects the adjustment of laser output power and ultimately leads to a decline in the quality of the repair layer.

[0008] The technical solution of this application is as follows:

[0009] In a first aspect, this application discloses a method for controlling the temperature of laser additive manufacturing for rail defects, the method comprising:

[0010] Real-time acquisition of acoustic and temperature signals from the molten region during laser additive manufacturing;

[0011] The sound signal is analyzed to determine whether there are any interference events, and the reliability of the temperature signal at the current moment is assessed based on the analysis results.

[0012] Based on the reliability assessment results, the laser output power is adjusted to control the laser additive temperature.

[0013] Furthermore, based on the above, the sound signal is analyzed to determine whether there are any interference events, and based on the analysis results, the reliability of the temperature signal at the current moment is assessed, including: if there are interference events, the reliability assessment is determined to be a low reliability mode; if there are no interference events, the reliability assessment is determined to be a high reliability mode.

[0014] Based on this, and according to the reliability assessment results, the laser output power is adjusted to control the laser additive temperature, including:

[0015] In response to the high reliability mode result of the reliability assessment, the laser output power is adjusted based on the deviation between the current temperature signal and the target temperature and the calculation based on PID control.

[0016] If the reliability assessment result is a low reliability mode, the temperature signal corresponding to the high reliability mode before the current temperature signal is obtained and used as the current temperature reference value to calculate and adjust the laser output power.

[0017] In some preferred embodiments, analysis is performed on the sound signal to determine whether any interference events are present, including:

[0018] Acquire multi-channel audio signals and perform preprocessing; perform time-frequency analysis on the preprocessed acoustic signals to obtain time-frequency spectra;

[0019] Extract multidimensional feature vectors from the time-frequency spectrum and calculate the similarity between the multidimensional feature vectors and the preset acoustic features of impurities;

[0020] When the similarity between the multidimensional feature vector and the preset impurity acoustic features is lower than the preset similarity threshold, it is determined to be an undetermined interference event;

[0021] When an event is identified as an undetermined interference event, the temperature signal is dynamically compensated based on the degree of deviation between the multidimensional feature vector and the normal sound baseline.

[0022] Furthermore, when an event is determined to be an undetermined interference event, dynamic compensation is performed on the temperature signal based on the degree of deviation between the multidimensional feature vector and the normal sound baseline, including:

[0023] Obtain the degree of deviation between the multidimensional feature vector and the normal sound baseline;

[0024] Based on the degree of deviation and the thermal response characteristics of the molten region, the compensation amount of the temperature signal is dynamically adjusted.

[0025] The compensation amount for the temperature signal is adjusted based on the current rate of temperature change in the melting zone and the expected target temperature.

[0026] Based on the above, and according to the current rate of temperature change in the melting zone and the expected target temperature, the compensation amount is adjusted, including:

[0027] Real-time calculation of the instantaneous rate of temperature change in the molten zone, as well as acquisition of ambient wind speed and temperature distribution in the rail preheating zone;

[0028] The reference range of the melting zone temperature is dynamically adjusted based on the ambient wind speed and the temperature distribution in the rail preheating area.

[0029] The instantaneous rate of change is compared with the reference range. Based on the degree and direction of the instantaneous rate of change exceeding the reference range, the compensation amount of the temperature signal is adjusted in the opposite direction to guide the current rate of temperature change in the molten region back to the reference range.

[0030] As a technological improvement, the compensation amount is adjusted based on the current rate of temperature change in the melting zone and the expected target temperature, and also includes:

[0031] Calculate the deviation between the current temperature of the molten region and the expected target temperature;

[0032] The compensation amount of the temperature signal is corrected according to the magnitude and direction of the deviation, so that the temperature of the melting zone approaches the expected target temperature.

[0033] As a further improvement, the reference range of the molten zone temperature is dynamically adjusted based on the ambient wind speed and the temperature distribution in the rail preheating area, including:

[0034] Based on the instantaneous rate of temperature change in the melting zone and historical data, the ambient wind speed and temperature distribution in the rail preheating zone are corrected.

[0035] The reference range of the melting zone temperature is dynamically adjusted based on the corrected ambient wind speed and the temperature distribution in the rail preheating zone.

[0036] As a system extension, based on the instantaneous rate of temperature change in the molten zone and historical data, corrections are made to the ambient wind speed and the temperature distribution in the rail preheating zone, including:

[0037] Real-time monitoring of the instantaneous change rate of temperature in the molten zone, laser output power, and powder feeding rate; the powder feeding rate is the rate at which alloy powder is injected into the molten zone during laser additive manufacturing.

[0038] In response to abnormal fluctuations in the instantaneous rate of change, such as abnormal fluctuations in laser output power or powder feeding rate, the correction range for ambient wind speed and temperature distribution in the rail preheating zone is adjusted according to the degree of influence of the fluctuations in laser output power or powder feeding rate on the instantaneous rate of change of temperature in the molten zone.

[0039] Secondly, this application also discloses a temperature control system for laser additive manufacturing of rail defects. The system includes: an acquisition module for acquiring sound signals and temperature signals of the molten area during the laser additive manufacturing process in real time; an evaluation module for analyzing the sound signals to determine whether there are any impurity interference events, and for evaluating the reliability of the temperature signal at the current moment based on the analysis results; and a control module for adjusting the laser output power based on the reliability evaluation results, thereby controlling the laser additive manufacturing temperature.

[0040] In summary, the laser additive manufacturing temperature control method and system for rail defects provided in this application acquires sound and temperature signals from the molten region during the laser additive manufacturing process in real time. Based on the sound signal analysis, it checks for any interference events and then assesses the reliability of the current temperature signal. According to the reliability assessment results, the system can intelligently adjust the laser output power, thereby precisely controlling the laser additive manufacturing temperature.

[0041] By introducing sound signals as an auxiliary judgment criterion, this application can identify and distinguish between temperature measurement anomalies caused by impurity interference and actual temperature fluctuations. When impurity interference is detected, the system can adopt corresponding strategies, such as reducing the reliability weight of the temperature signal or using historical data for reference, to avoid power adjustment based on erroneous feedback. Compared with the traditional PID control method that relies solely on infrared temperature measurement, the technical solution of this application has significant advantages. It can effectively cope with complex and changing field environments, reduce the impact of randomness and nonlinear interference on temperature control, and improve the accuracy and robustness of temperature control. Accordingly, the porosity and uniformity of the repair layer are significantly improved, thereby enhancing the reliability and durability of laser additive repair of rail defects and providing more reliable technical support for the maintenance of railway infrastructure. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart illustrating the steps of the laser additive manufacturing temperature control method for rail defects disclosed in an embodiment of the present invention.

[0044] Figure 2 This is a schematic diagram of the temperature control system for laser additive manufacturing of rail defects disclosed in an embodiment of the present invention. Detailed Implementation

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which these embodiments belong; the terminology used herein and in the specification of the application is for the purpose of describing particular embodiments only and is not intended to limit these embodiments; the terms "comprising" and "having," and any variations thereof, in the specification of these embodiments and the foregoing drawings, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification of these embodiments and the foregoing drawings are used to distinguish different objects, not to describe a particular order.

[0046] The implementation details of the technical solution in this embodiment are described in detail below:

[0047] In railway infrastructure maintenance, laser additive manufacturing technology is widely used for repairing rail defects. However, in actual railway field environments, rail surfaces may be covered with trace amounts of oil, fine dust, oxide layers, and other impurities that are difficult to completely remove. When a high-energy laser beam scans and heats the rail surface, these impurities interact with the laser, potentially causing localized areas to overheat, rapidly decompose, or vaporize, producing tiny gas plumes or fumes. These gas plumes or fumes not only disturb the surface morphology of the molten area but may also obstruct or interfere with the infrared temperature measurement device, leading to inaccurate temperature readings from the sensor. Traditional temperature control systems adjust power based on these inaccurate feedback signals, causing the temperature of the molten area to deviate from the ideal process window, affecting the material structure and mechanical properties of the repair layer, and even creating new safety hazards.

[0048] In response, this application proposes a method for controlling the temperature of laser additive manufacturing for rail defects, the method comprising:

[0049] S101, real-time acquisition of sound and temperature signals in the molten region during laser additive manufacturing;

[0050] S102, Analyze the sound signal to determine whether there is any interference event, and evaluate the reliability of the temperature signal at the current moment based on the analysis results;

[0051] S103 adjusts the laser output power based on the reliability assessment results, thereby controlling the laser additive temperature.

[0052] The temperature control method for laser additive manufacturing of rail defects proposed in this application aims to solve the problems of inaccurate temperature measurement and control failure caused by surface impurities during laser additive manufacturing. The "molten zone" refers to the area where the laser beam acts on the rail surface, melting the metal material to form a liquid pool; this is the key area for material deposition and metallurgical bonding during laser additive manufacturing. The "acoustic signal" refers to the sound wave signals generated in the molten zone and its surrounding environment during laser additive manufacturing; these signals may include acoustic characteristics related to impurity vaporization, splashing, etc. The "temperature signal" refers to the temperature data of the molten zone acquired in real time by sensors such as infrared thermometers.

[0053] Specifically, the implementation of this application can be as follows: First, it is necessary to acquire the sound and temperature signals of the molten region during the laser additive manufacturing process in real time. Sound signals can be acquired by deploying a high-sensitivity microphone array near the laser additive manufacturing equipment; these microphones can capture the faint sounds generated by the molten region. Temperature signals can be acquired using an infrared thermometer, which is aimed at the molten region to measure its surface radiation temperature in real time. After these signals are acquired, they are transmitted to the control system via a data acquisition card for further processing.

[0054] Secondly, the sound signal is analyzed to determine if any interference events occur, and the reliability of the current temperature signal is assessed based on the analysis results. For example, spectral analysis of the acquired sound signal can identify abnormal noise within a specific frequency range, which may be related to the rapid vaporization or decomposition of impurities. When such abnormal noise is detected, an interference event can be preliminarily identified. Based on this assessment, the reliability of the current temperature signal is evaluated. If an interference event is present, the reliability of the current temperature signal is considered low; if no interference event is present, the reliability of the current temperature signal is considered high.

[0055] Finally, based on the reliability assessment results, the laser output power is adjusted to control the laser additive manufacturing temperature. When the temperature signal is assessed as having high reliability, the control system can calculate and adjust the laser output power using a standard PID control algorithm based on the deviation between the current temperature signal and the target temperature to precisely control the temperature of the molten region. However, when the temperature signal is assessed as having low reliability, it is not advisable to directly use this signal for control because the current temperature signal may be inaccurate due to interference from impurities. In this case, the control system can acquire historical temperature signals corresponding to the high reliability mode prior to the current temperature signal, use them as the current temperature reference value for calculation, and adjust the laser output power accordingly. This method avoids using distorted temperature signals for control, thereby improving the robustness of temperature control.

[0056] The laser additive manufacturing temperature control method for rail defects disclosed in this application effectively identifies potential interference events caused by impurities during the laser additive manufacturing process by introducing real-time analysis of the acoustic signals from the molten region. Traditional methods primarily rely on temperature signals obtained from infrared thermometers for feedback control. However, when impurities such as oil or dust are present on the rail surface, these impurities generate smoke or gas under laser irradiation, directly interfering with the infrared thermometer's measurements and causing distortion in the temperature feedback signal. This distortion prevents traditional PID control methods from accurately determining the true temperature of the molten region, potentially leading to incorrect laser power adjustments and affecting repair quality.

[0057] The core innovation of this application lies in its use of sound signals as an auxiliary basis for judgment, rather than relying solely on temperature signals. By analyzing sound signals, abnormal acoustic characteristics caused by impurities can be detected in real time and non-contactly, such as unique sound waves generated by vaporization or explosion. Once these impurity interference events are identified, the system can perform a reliability assessment of the current temperature signal. When the temperature signal is judged to be of low reliability, the system no longer blindly adjusts the power based on the potentially distorted temperature signal, but instead uses historical temperature data from previous high-reliability modes as a reference, or adopts other more robust control strategies. This reliability assessment mechanism based on multi-source information fusion enables the laser additive manufacturing temperature control system to intelligently cope with complex and changing field environments, avoiding control failures caused by sensor signal interference.

[0058] Compared with existing technologies, the advantages of this application lie in its significantly improved environmental adaptability and robustness of control. Traditional methods may suffer from large temperature control deviations and unstable repair quality under the interference of impurities, and may even lead to defects such as porosity and uneven structure in the repair layer. However, this application, by introducing acoustic signals to assist in judgment, can promptly detect and avoid the unreliability of temperature signals, thus maintaining precise temperature control of the molten zone even in the presence of impurities. This not only improves the quality and reliability of laser additive repair of rail defects but also reduces the complexity of on-site operations and the stringent requirements for rail surface pretreatment, possessing significant practical application value.

[0059] In some embodiments described above in this application, a method is proposed to analyze the sound signal to determine whether an interference event exists and to perform a reliability assessment on the temperature signal. Specifically, the step of analyzing the sound signal to determine whether an interference event exists and, based on the analysis results, performing a reliability assessment on the temperature signal at the current moment includes: if an interference event exists, determining the reliability assessment as a low reliability mode; if no interference event exists, determining the reliability assessment as a high reliability mode.

[0060] Specifically, in the laser additive manufacturing process, the acoustic signals from the molten region can reflect various information about the internal physical processes. When acoustic signal analysis indicates the presence of interference events, such as foreign objects entering the molten pool or abnormal phenomena like splashing, the temperature signal acquired by the temperature sensor may be affected, leading to reduced accuracy or representativeness. In this case, classifying the reliability assessment of the current temperature signal as low reliability means that this temperature signal should not be directly used as the basis for precise control. Conversely, if acoustic signal analysis indicates the absence of interference events, the current molten region is considered to be in a relatively stable state, and the acquired temperature signal has high reliability. Therefore, classifying the reliability assessment as high reliability indicates that the temperature signal can be trusted and used for subsequent precise control.

[0061] This application's solution achieves dynamic judgment of temperature signal quality by directly linking the analysis results of the sound signal with the reliability assessment of the temperature signal. When impurity interference events are detected, the temperature signal is marked as a low-reliability mode because such interference may cause temperature measurement distortion or inaccuracy, thereby avoiding control decisions based on inaccurate data. Conversely, when there is no impurity interference, the temperature signal is considered high-reliability and can be used as a reliable control input. This mechanism ensures that subsequent laser output power adjustments are based on an accurate judgment of the temperature signal's reliability, thereby improving the robustness and accuracy of temperature control.

[0062] The specific control logic in this embodiment is as follows:

[0063] 1. Acoustic signal acquisition and analysis unit:

[0064] Hardware Configuration: A high-sensitivity microelectromechanical systems (MEMS) microphone array is installed near the molten area of ​​the laser repair head. This microphone array is capable of capturing sound waves in the 20 Hz to 20 kHz range and has good immunity to ambient noise. The microphone output signal is preprocessed by a low-noise amplifier and then fed into a high-speed analog-to-digital converter (ADC) to convert the analog acoustic signal into a digital signal. For example, an ADC with a sampling rate of at least 48 kHz and a resolution of at least 16 bits can be selected to ensure accurate capture of transient acoustic events.

[0065] Signal Processing: The digitized acoustic signal is fed into an embedded digital signal processor (DSP), such as Texas Instruments' (TI) C2000 series or NXP's i.MX RT series microcontrollers, which possess powerful floating-point arithmetic capabilities. This processor performs the following processing on the acoustic signal in real time:

[0066] Temporal feature extraction: The instantaneous energy intensity of the acoustic signal is monitored in real time by calculating the root mean square (RMS) value or peak value of the signal. At the same time, short-time energy and zero-crossing rate analysis are used to identify abrupt changes in the signal.

[0067] Frequency domain feature extraction: A Fast Fourier Transform (FFT) is performed on the acquired acoustic signal to obtain its spectral information. By analyzing the energy distribution of different frequency components, it is possible to identify any anomalous energy concentrations within specific frequency ranges. For example, when a laser interacts with impurities to produce a burst, a significant energy boost typically occurs in the high-frequency band between 5 kHz and 15 kHz.

[0068] Acoustic Feature Recognition: This unit has a pre-set set of acoustic feature recognition rules. When the instantaneous energy intensity of a real-time acoustic signal increases rapidly within a very short time (e.g., less than 1 millisecond), and its spectral analysis shows a significant energy peak in a pre-set high-frequency band (e.g., 5 kHz to 15 kHz), accompanied by a very short signal duration (e.g., less than 10 milliseconds), the unit determines that a "pop" event has occurred. These thresholds and frequency ranges can be calibrated through prior experiments using acoustic data generated under different types of impurities (such as oil, dust, and oxide layers) and laser irradiation.

[0069] 2. Temperature feedback reliability assessment unit:

[0070] This unit operates in parallel with the acoustic signal acquisition and analysis unit, and receives real-time temperature readings from the infrared thermometer. The infrared thermometer typically employs a non-contact infrared sensor, such as a photodiode or thermopile sensor, with a response time in the millisecond range to match the dynamic changes in the laser additive manufacturing process.

[0071] Reliability determination logic: Once the acoustic signal acquisition and analysis unit identifies an event matching the characteristics of a "pop" sound, the temperature feedback reliability assessment unit immediately determines that the temperature readings transmitted by the infrared thermometer are "low" in terms of reliability at the current moment and within a very short time window immediately following (e.g., 50 milliseconds from the start of the pop). This 50 millisecond time window is set based on the typical timescale of smoke generation, obstruction, and dissipation to ensure coverage of the entire interference process.

[0072] For periods when no "pop" event was detected, the reliability of the temperature readings remained "high".

[0073] 3. Adaptive temperature control decision unit:

[0074] This unit is the core of the entire control system. It receives temperature readings from the infrared thermometer and reliability information provided by the temperature feedback reliability assessment unit. This unit can be implemented by a high-performance industrial-grade programmable logic controller (PLC) or a dedicated motion controller to ensure the real-time performance and accuracy of control commands.

[0075] Power regulation in high reliability mode: When the reliability assessment unit indicates that the reliability of the current temperature reading is "high", the adaptive temperature control decision unit will accurately calculate and adjust the output power of the laser based on the deviation between the current temperature reading and the target temperature, according to the preset proportional-integral-derivative (PID) control logic or multi-step model control strategy.

[0076] PID control: The laser power output P is calculated using the following formula:

[0077]

[0078] Where P(t) is the laser power output at the current moment; e(t) is the temperature deviation at the current moment (target temperature - actual measured temperature). , , These are the proportional, integral, and differential coefficients, which are determined through system identification and debugging.

[0079] Multi-step model control: Based on a pre-set process parameter table, the model adjusts the laser power in steps according to the magnitude of the deviation between the current temperature and the target temperature, as well as the duration of the deviation. For example, when the temperature deviates from the target value by more than 5 degrees Celsius, the power adjustment step is 20 watts; when the deviation is within 2 degrees Celsius, the power adjustment step is 5 watts.

[0080] Power regulation in low reliability mode: When the reliability assessment unit indicates that the reliability of the current temperature reading is "low", the adaptive temperature control decision unit will take the following strategies to avoid drastic power fluctuations caused by erroneous feedback:

[0081] Pause or reduce weight: Temporarily ignore the role of this "low reliability" temperature reading in PID or multi-step model calculations, or set its weight close to zero.

[0082] Alternative values ​​can be used: Instead, a valid temperature reading deemed "highly reliable" immediately before the "pop" event can be used as the current temperature reference for power adjustment. Alternatively, a simple moving average calculated based on several (e.g., the most recent 5) highly reliable temperature readings prior to the event can be used as a reference.

[0083] Maintenance or Mild Adjustment: During this brief period of "low reliability," the control system maintains the laser power at the level before the event, or makes mild, fine adjustments based on a previously stable trend. For example, if the power was stable before the event, it remains unchanged; if the power is slowly rising or falling, it continues that trend in small steps until reliability is restored.

[0084] Auxiliary judgment of the state of the molten region:

[0085] By further analyzing the intensity and frequency of the "popping sound," this method can also indirectly assess the degree of disturbance to the molten zone caused by impurities. For example, if the acoustic signal acquisition and analysis unit detects continuous and high-intensity popping sounds, it may indicate a high impurity content on the rail surface or a violent gasification reaction, thus suggesting a potentially high risk of porosity formation in the molten zone. This auxiliary information can be recorded as additional physical evidence for assessing the quality of the repair layer. For instance, after repair, this acoustic data can be used to conduct targeted non-destructive testing on specific areas to verify the presence of defects such as porosity.

[0086] Through the above technical solution, this application can intelligently classify the reliability of the real-time acquired temperature signal based on the actual working conditions of the molten region during laser additive manufacturing, especially the presence of impurities. This classification mechanism provides a clear decision-making basis for subsequent laser output power adjustment, avoiding blind control when the temperature signal is interfered with, thereby effectively improving the accuracy, stability, and adaptability of laser additive manufacturing temperature control and reducing additive manufacturing quality problems caused by measurement errors.

[0087] In some embodiments described above, the sound signal is analyzed to determine the presence of interference events, and the reliability of the current temperature signal is assessed based on the analysis results, thereby determining whether the reliability assessment is in a high-reliability mode or a low-reliability mode. However, in actual laser additive manufacturing processes, simply assessing the reliability of the temperature signal is insufficient. Different laser output power adjustment strategies are needed based on different reliability modes to ensure accurate and stable temperature control under various operating conditions. If the laser output power is not adjusted specifically, directly adjusting the power based on the low-reliability temperature signal may lead to control instability, thereby affecting the additive manufacturing quality.

[0088] In response, this application further proposes a method for adjusting the laser output power based on the reliability assessment result to control the laser additive temperature, comprising: in response to the reliability assessment result being a high reliability mode, adjusting the laser output power based on the deviation between the current temperature signal and the target temperature and using PID control; in response to the reliability assessment result being a low reliability mode, obtaining the temperature signal corresponding to the high reliability mode before the current temperature signal and using it as the current temperature reference value to calculate and adjust the laser output power.

[0089] Specifically, when the reliability assessment result is in high reliability mode, it indicates that the currently acquired temperature signal is accurate and reliable. In this case, a proportional-integral-derivative (PID) control algorithm can be used to precisely adjust the laser output power. PID control is a feedback control method widely used in industrial control. It calculates the deviation between the current temperature signal and the preset target temperature, and combines the proportional, integral, and derivative terms of the deviation to generate a control quantity, thereby achieving real-time, continuous, and precise adjustment of the laser output power. The proportional term is used to quickly respond to the current deviation, the integral term is used to eliminate steady-state error, and the derivative term is used to predict the trend of deviation changes and suppress overshoot. Through PID control, the temperature of the molten region can quickly approach and stabilize at the target temperature, ensuring the stability and consistency of the additive manufacturing process.

[0090] When the reliability assessment result is in a low reliability mode, it usually means that the current temperature signal may be affected by interference events, leading to reduced accuracy or abnormal fluctuations. In this case, if PID control still relies directly on the current temperature signal, it may introduce erroneous control commands, causing drastic fluctuations in laser output power and even damaging additive manufacturing quality. Therefore, this application proposes a more robust strategy: acquiring the temperature signal corresponding to a high reliability mode prior to the current temperature signal and using it as the current temperature reference value. This "temperature signal in a high reliability mode" can be understood as the temperature value recorded during a recent period of interference-free and reliable temperature signals, such as the average temperature during that period, the last stable temperature value, or a temperature value obtained through trend prediction. Based on this reliable historical temperature reference value, the system can calculate and adjust the laser output power, for example, by using a preset power adjustment curve, power adjustment based on model prediction, or simple open-loop control, to maintain the basic stability of the additive manufacturing process and avoid malfunctions caused by unreliable signals.

[0091] This application's solution effectively addresses the problem of maintaining stable temperature control during laser additive manufacturing when the temperature signal becomes unreliable due to impurities. Specifically, when the temperature signal is in a high-reliability mode, precise PID control enables rapid response and accurate stabilization of the molten region temperature, ensuring additive manufacturing quality. Conversely, when the temperature signal is in a low-reliability mode, switching to a historically high-reliability temperature signal as a reference avoids the risks associated with directly relying on potentially distorted current temperature signals, effectively preventing control system instability and additive manufacturing defects. This adaptive control strategy significantly improves the robustness and reliability of the laser additive manufacturing process.

[0092] In some preferred embodiments, a specific example is given below. Assume that during the laser additive manufacturing process for rail defects, the target melting temperature is set to 1500°C. During the normal additive manufacturing phase, acoustic signal analysis of the molten region shows no interference events, therefore the temperature signal is assessed as being in a high-reliability mode. At this time, the system monitors the current temperature signal in real time, for example, detecting a temperature of 1495°C, which deviates from the target temperature by -5°C. Based on this deviation, the PID controller calculates and outputs a command to increase the laser power to raise the temperature back to 1500°C. If the temperature rises to 1505°C, the PID controller calculates and outputs a command to decrease the laser power.

[0093] However, at a certain moment, the acoustic signal analysis detects a noticeable splashing sound from impurities. The system determines that an impurity interference event has occurred and assesses the current temperature signal as a low-reliability mode. At this point, even if the temperature sensor might report an abnormal instantaneous temperature value (e.g., a sudden jump to 1600°C or 1400°C due to splash obstruction or reflection), the system will not directly use this abnormal value for PID control. Instead, the system acquires the most recent stable temperature value (e.g., 1500°C) before the impurity interference occurred, i.e., during the most recent period of high-reliability mode, and uses this as the current temperature reference value. Based on this historical reference value, the system executes a preset or model-based power adjustment strategy, such as maintaining the current laser power or making a small preset adjustment to ensure the stability of the additive process during the interference and avoid over- or under-power adjustments due to erroneous temperature readings. Once the impurity interference event ends, the acoustic signal returns to normal, and the temperature signal is once again assessed as high-reliability mode, the system switches back to PID control based on the current temperature signal to continue precisely adjusting the laser output power.

[0094] In some embodiments described above, acoustic signal analysis is used to determine the presence of impurity interference events, and the reliability of the temperature signal is assessed accordingly. However, in actual laser additive manufacturing processes, in addition to obvious impurity interference events, there may be atypical acoustic anomalies that are difficult to classify directly as impurities. These anomalies can also affect the accuracy of the temperature signal. If a binary reliability assessment is performed solely based on the presence of explicit impurity interference events, it may not be sufficient to address all potential signal interference situations, thereby affecting the precision of temperature control. Therefore, this application further proposes a more refined acoustic signal analysis method, aiming to more comprehensively identify and process various acoustic anomalies to improve the accuracy of temperature signal reliability assessment and to dynamically compensate for the temperature signal.

[0095] The above-mentioned analysis of the sound signal to determine whether there are any interference events includes:

[0096] Acquire multi-channel audio signals and perform preprocessing; perform time-frequency analysis on the preprocessed acoustic signals to obtain time-frequency spectra;

[0097] Extract multidimensional feature vectors from the time-frequency spectrum and calculate the similarity between the multidimensional feature vectors and preset impurity acoustic features;

[0098] When the similarity between the multidimensional feature vector and the preset impurity acoustic features is lower than a preset similarity threshold, it is determined to be an undetermined interference event;

[0099] When a pending interference event is identified, the temperature signal is dynamically compensated based on the degree of deviation between the multidimensional feature vector and the normal sound baseline.

[0100] Specifically, acquiring multi-channel acoustic signals refers to simultaneously acquiring acoustic data at different locations around the molten region using multiple microphones or acoustic sensors. Multi-channel acquisition helps capture spatial information of the sound field, improving signal integrity and interference resistance. Preprocessing typically includes noise suppression, signal filtering (e.g., removing ambient background noise or mechanical noise at specific frequencies), and signal normalization to ensure the accuracy of subsequent analysis.

[0101] The time-frequency analysis of the preprocessed acoustic signal to obtain a time-frequency spectrum can be understood as converting the time-domain signal into a two-dimensional representation that simultaneously contains time and frequency information. Commonly used time-frequency analysis methods include Short-Time Fourier Transform (STFT) or Wavelet Transform, which aim to reveal the frequency component changes of the acoustic signal at different time points, thereby more comprehensively reflecting the acoustic characteristics of the molten region. Time-frequency spectra, such as spectrograms, can visually display the distribution of energy in time and frequency.

[0102] In practical applications, extracting multidimensional feature vectors from time-frequency spectra refers to quantifying and extracting a set of numerical values ​​that characterize the acoustic signal properties from the time-frequency spectrum, such as Mel-frequency cepstral coefficients (MFCC), spectral centroid, spectral bandwidth, spectral roll-off point, and zero-crossing rate. These feature vectors can effectively capture information such as the timbre, rhythm, and energy distribution of sound. The similarity between these multidimensional feature vectors and preset impurity acoustic features is typically calculated using methods such as cosine similarity, Euclidean distance, or correlation coefficient. The purpose is to quantify the degree of matching between the current acoustic signal and the features of known impurities (such as sounds produced by defects like pores, inclusions, and cracks). The preset impurity acoustic features are pre-established based on a large amount of experimental data and experience, representing the typical acoustic performance of different types of impurities.

[0103] When the similarity between the multidimensional feature vector and the preset impurity acoustic features is lower than a preset similarity threshold, it is determined to be a pending interference event. This means that when the current acoustic signal has low similarity with all known typical impurity acoustic features, but deviates from the sound baseline of a normal additive manufacturing process, the system identifies it as a "pending" anomaly that requires further attention. This "pending interference event" may not be a typical impurity, but it still indicates that there is some abnormal state in the molten region.

[0104] When an event is identified as an pending interference event, the temperature signal is dynamically compensated based on the degree of deviation between the multidimensional feature vector and the normal sound baseline. The normal sound baseline refers to the typical set of acoustic signal characteristics generated in the molten region during an ideal and stable laser additive manufacturing process. The degree of deviation can be quantified by calculating the distance or difference between the current multidimensional feature vector and the normal sound baseline. Dynamic compensation refers to correcting the real-time acquired temperature signal according to the degree and direction of this deviation. For example, if the deviation is large and the indicated temperature may be underestimated, the temperature signal is appropriately corrected upwards, and vice versa. The purpose is to obtain temperature data closer to the true value even under atypical interference, avoiding temperature measurement errors caused by acoustic anomalies.

[0105] This application's solution, through the introduction of multi-channel acoustic signal acquisition, refined time-frequency analysis, and multi-dimensional feature vector extraction, enables the system to capture acoustic information of the molten region more comprehensively and meticulously. It is precisely because of the in-depth feature extraction of the acoustic signal and the similarity calculation with preset impurity acoustic features that the system can initially determine whether typical impurity interference exists. Furthermore, when the similarity between the acoustic signal and known impurity features is not high, but it deviates significantly from the normal sound baseline, the system can identify it as a "pending interference event," which compensates for the shortcomings of relying solely on explicit impurity features for binary judgment. By dynamically compensating the temperature signal based on the degree of deviation of the acoustic signal from the normal sound baseline, this application can refine the processing of atypical acoustic anomalies that may still affect the accuracy of temperature measurement, thereby avoiding the problem of simply marking the temperature signal as low reliability in these cases, leading to a decrease in control accuracy. This mechanism allows the temperature signal to maintain high accuracy under a wider range of operating conditions, providing a more reliable basis for subsequent laser output power adjustment.

[0106] In some preferred embodiments, a specific example is given below. Suppose that during laser additive manufacturing, slight splashing or bubble bursting suddenly occurs in the molten region. The acoustic signals generated by these phenomena may not be sufficiently identified as typical "impurity interference events," but their acoustic characteristics are subtly different from the normal, stable additive manufacturing sound.

[0107] First, a multi-channel acoustic sensor synchronously acquires these acoustic signals and preprocesses them to remove environmental noise. Then, a short-time Fourier transform is performed on the preprocessed signals to generate a time-frequency spectrum. From this spectrum, feature vectors such as MFCC and spectral centroid are extracted. Next, the system calculates the similarity between these extracted feature vectors and preset typical impurity acoustic features (e.g., features of large-sized pore rupture, frictional sound features generated by inclusions). If the calculation shows that the similarity between the current acoustic signal and all preset impurity features is below a preset threshold (e.g., below 0.6), the system will not classify it as a clear impurity interference event. However, the system further compares the currently extracted feature vectors with a pre-established normal additive manufacturing process acoustic baseline (e.g., the average acoustic feature value of a stable molten pool) and calculates the degree of deviation (e.g., Euclidean distance). If the deviation exceeds a small threshold (e.g., 0.1), the system classifies the current event as a "pending interference event".

[0108] Once a potential interference event is identified, the system dynamically compensates the real-time temperature signal based on the magnitude of the deviation. For example, if the deviation indicates a slight "muffled" characteristic in the acoustic signal, this might mean that slight fluctuations on the molten pool surface are causing a slight deviation in the infrared thermometry. The system might then make a small upward correction to the current temperature signal (e.g., an increase of 0.5°C). This dynamic compensation mechanism allows the system to finely adjust the temperature measurement without completely switching to a low-reliability mode, thus maintaining high-precision temperature control and avoiding over- or under-power adjustments due to misjudgments.

[0109] In response, this application further proposes a step of dynamically compensating the temperature signal based on the degree of deviation between the multidimensional feature vector and the normal sound baseline when the event is determined to be an undetermined interference event. This step includes:

[0110] Obtain the degree of deviation between the multidimensional feature vector and the normal sound baseline;

[0111] Based on the degree of deviation and the thermal response characteristics of the molten region, the compensation amount of the temperature signal is dynamically adjusted.

[0112] The compensation amount of the temperature signal is adjusted based on the current rate of temperature change in the melting zone and the expected target temperature.

[0113] Specifically, obtaining the deviation of the multidimensional feature vector from the normal sound baseline refers to quantifying the degree of abnormality of the current sound signal by calculating the distance or similarity index between the multidimensional feature vector extracted at the current moment and the pre-established normal sound baseline. For example, methods such as Euclidean distance, Mahalanobis distance, or cosine similarity can be used to calculate this, reflecting the difference between the current acoustic state and the normal stable state.

[0114] The dynamic adjustment of the temperature signal compensation amount based on the degree of deviation and the thermal response characteristics of the molten region can be understood as follows: after determining the degree of abnormality in the sound signal, it is not simply a matter of linear compensation, but rather further consideration of the thermophysical parameters of the molten region material, such as thermal conductivity, specific heat capacity, and latent heat of phase change, as well as the influence of process parameters such as laser power and scanning speed on the temperature response. For example, when the thermal inertia of the molten region material is large, even if the degree of deviation in the sound signal is large, the temperature compensation amount should be adjusted slowly to avoid overshoot; conversely, for materials with rapid thermal response, the compensation amount can be adjusted more aggressively. This can be achieved through a pre-established thermal model, lookup table, or machine learning-based prediction model, enabling the compensation amount to more accurately adapt to the actual thermal behavior of the molten region.

[0115] In practical applications, adjusting the compensation amount of the temperature signal based on the current rate of temperature change in the molten zone and the expected target temperature means incorporating consideration of the dynamic temperature trend after initially determining the compensation amount. For example, if the current molten zone temperature is rising rapidly and has approached or exceeded the expected target temperature, even if the audible signal indicates the need for compensation, the compensation amount should be appropriately reduced or reversed to prevent temperature overshoot; conversely, if the temperature is dropping too quickly and is far from the expected target temperature, the compensation amount should be increased. This correction mechanism makes temperature control not only responsive but also forward-looking and target-oriented, effectively suppressing temperature fluctuations and ensuring that the molten zone temperature remains stable within the target range.

[0116] This application's solution incorporates consideration of the thermal response characteristics of the molten region, ensuring that temperature signal compensation is no longer solely based on the degree of anomalousness in the sound signal. Instead, it combines the material's inherent physical properties and heat transfer laws, making the compensation amount more consistent with the actual physical process. Furthermore, by monitoring the rate of temperature change in the molten region in real time and comparing it with the expected target temperature, the compensation amount can be dynamically corrected. This gives the temperature control system the ability to predict and adjust for future trends. It is precisely this multi-dimensional and dynamic compensation mechanism that enables more accurate, timely, and stable temperature signal compensation when impurity interference occurs, avoiding the overcompensation or undercompensation problems that may occur with traditional methods.

[0117] In some preferred embodiments, a specific example is given below. Suppose that during laser additive manufacturing, a sound sensor detects an abnormal acoustic signal. After time-frequency analysis and feature extraction, this signal is identified as an undetermined interference event, and the deviation of the multidimensional feature vector from the normal sound baseline is calculated to a specific value (e.g., deviation X). At this point, the system first calculates a preliminary temperature compensation amount (e.g., the temperature signal needs to be compensated upwards by Y degrees) based on this deviation X and a pre-established thermal response model of the molten region (which considers the thermal conductivity, specific heat capacity, etc., of the current rail material).

[0118] Furthermore, the system monitors the instantaneous rate of temperature change in the current molten region in real time as Z℃ / second, while the expected target temperature is T_target. If the Z value indicates that the temperature is rising rapidly and the current temperature is close to T_target, the system will correct the initially calculated compensation amount Y based on this trend. For example, to avoid temperature overshoot, the compensation amount Y may be appropriately reduced, or if the temperature rises too quickly, it may even be corrected to a negative value to encourage the temperature to fall back. Conversely, if the Z value indicates that the temperature is falling rapidly and is far from T_target, the compensation amount Y may be appropriately increased to accelerate the temperature recovery. Through this dynamic correction, even under the interference of impurities, the temperature of the molten region can be accurately guided and stabilized near the target temperature, thereby ensuring the quality of additive manufacturing.

[0119] This application further proposes the compensation amount for the aforementioned modified temperature signal, including:

[0120] Real-time calculation of the instantaneous rate of temperature change in the molten zone, as well as acquisition of ambient wind speed and temperature distribution in the rail preheating zone;

[0121] The reference range of the temperature in the melting zone is dynamically adjusted based on the ambient wind speed and the temperature distribution in the rail preheating zone.

[0122] The instantaneous rate of change is compared with the reference range, and the compensation amount of the temperature signal is adjusted in reverse according to the degree and direction of the instantaneous rate of change exceeding the reference range, so as to guide the current rate of change of temperature in the melting region back to the reference range.

[0123] Specifically, in the laser additive manufacturing process, the instantaneous rate of temperature change in the molten region is a key indicator reflecting its dynamic thermal characteristics. Real-time calculation of this rate of change provides immediate dynamic information about the current thermal process. Simultaneously, obtaining ambient wind speed and the temperature distribution in the rail preheating area is crucial. Ambient wind speed directly affects the convective heat dissipation efficiency of the molten region; higher wind speeds result in faster heat dissipation and a more pronounced temperature decrease in the molten region. The temperature distribution in the rail preheating area reflects the heat accumulation of the substrate; higher preheating temperatures lead to less heat loss in the molten region, and vice versa. These external factors have a significant and dynamic impact on the instantaneous rate of temperature change in the molten region.

[0124] Furthermore, based on the ambient wind speed and the temperature distribution in the rail preheating zone, the reference range of the molten zone temperature can be dynamically adjusted. This reference range is not a fixed threshold, but rather adaptively adjusted according to actual environmental conditions and the thermal state of the substrate. For example, when the wind speed is high or the rail preheating temperature is low, the normal temperature change rate in the molten zone may be too high or too low. In this case, the reference range needs to be widened or narrowed accordingly to more accurately reflect the ideal temperature change trend under the current operating conditions.

[0125] Therefore, the instantaneous rate of change calculated in real time is compared with the dynamically adjusted reference range. When the instantaneous rate of change exceeds the reference range, it indicates that the temperature change trend in the molten region has deviated from the expected trend. At this time, the compensation amount of the temperature signal is adjusted in reverse according to the degree and direction of exceeding the reference range. For example, if the instantaneous rate of change is too high (temperature rises too quickly or falls too slowly), negative compensation is performed to reduce laser output power or increase heat dissipation; if the instantaneous rate of change is too low (temperature rises too slowly or falls too quickly), positive compensation is performed to increase laser output power or reduce heat dissipation. The purpose of this reverse adjustment is to effectively guide the current rate of temperature change in the molten region back within the reference range, thereby maintaining the stability and controllability of the temperature in the molten region.

[0126] This application's solution addresses the problem of insufficient temperature control accuracy under complex and variable conditions by introducing ambient wind speed and temperature distribution in the rail preheating zone as the basis for dynamically adjusting the reference range of the molten zone temperature. When external environmental factors change, the ideal temperature change rate in the molten zone also changes. By acquiring and utilizing this external information in real time, the system can establish a dynamic reference benchmark that better reflects the actual thermodynamic process. When the instantaneous rate of temperature change in the molten zone deviates from this dynamic reference range, the system can promptly and accurately identify the anomaly and adjust the compensation amount of the temperature signal in the opposite direction based on the degree and direction of the deviation. This adaptive feedback mechanism ensures that even under fluctuating environmental conditions, the rate of temperature change in the molten zone can be effectively constrained within the desired range, thereby avoiding temperature overshoot or undershoot caused by external disturbances and significantly improving the robustness and accuracy of temperature control.

[0127] In some preferred embodiments, a specific example is given below. Assume that during the initial stage of the laser additive manufacturing process, the ambient wind speed is low and the rail is preheated uniformly. The instantaneous rate of change of temperature in the molten region is set to a reference range of [R1, R2]. As the additive manufacturing process progresses, a sudden strong wind blows, causing a sharp increase in the heat dissipation efficiency of the molten region. Without intervention, the temperature may drop rapidly.

[0128] At this point, the system detects a significant increase in ambient wind speed in real time and, combined with the temperature distribution in the rail preheating area, dynamically adjusts the reference range of the molten zone temperature to [R3, R4], where R3 and R4 may be lower than R1 and R2 to accommodate a faster heat dissipation rate. Simultaneously, the system calculates the instantaneous rate of change of the molten zone temperature in real time and finds that it has fallen below the adjusted reference range R3, indicating that the temperature is dropping too rapidly.

[0129] Based on the degree and direction of the instantaneous temperature change rate falling below the reference range R3, the system immediately performs a reverse adjustment to the compensation amount of the temperature signal, for example, increasing the compensation amount of the laser output power. The purpose of this reverse adjustment is to counteract the additional heat dissipation caused by strong winds, guiding the instantaneous temperature change rate of the molten region back within the reference range [R3, R4]. Through this dynamic adjustment and reverse compensation mechanism, even under sudden environmental changes, the temperature of the molten region can be effectively controlled, avoiding additive manufacturing defects caused by sudden temperature drops and ensuring the stability of the additive manufacturing process and product quality.

[0130] In some embodiments described above, the instantaneous rate of change of the molten zone temperature is calculated in real time, and the reference range of the molten zone temperature is dynamically adjusted based on the ambient wind speed and the temperature distribution in the rail preheating area. The instantaneous rate of change is then compared with the reference range, and the compensation amount of the temperature signal is adjusted in reverse according to the degree and direction of the instantaneous rate of change exceeding the reference range, thereby effectively guiding the current rate of change of the molten zone temperature back to the reference range. However, in actual laser additive manufacturing, simply controlling the rate of temperature change within a certain range may not be sufficient to ensure that the actual temperature of the molten zone accurately approaches and stabilizes at the expected target temperature. If there is a continuous small deviation between the actual temperature and the target temperature, even if the rate of change is controlled, it may lead to poor additive layer quality, uneven microstructure, or other defects. Therefore, this application further proposes a method for correcting the temperature signal compensation amount to more accurately control the molten zone temperature to approach the expected target temperature.

[0131] According to the above-mentioned method for controlling the temperature of laser additive manufacturing for rail defects, the correction of the compensation amount also includes:

[0132] Calculate the deviation between the current melting zone temperature and the expected target temperature;

[0133] Based on the magnitude and direction of the deviation, the compensation amount of the temperature signal is corrected so that the temperature of the melting region approaches the expected target temperature.

[0134] Specifically, "calculating the deviation between the current molten region temperature and the expected target temperature" refers to the system continuously monitoring the real-time temperature of the molten region and comparing it with a pre-set ideal target temperature for the laser additive manufacturing process to obtain the numerical difference between the two. This deviation can be positive (actual temperature higher than the target temperature) or negative (actual temperature lower than the target temperature), and its magnitude reflects the degree of deviation. The "expected target temperature" is an ideal temperature value determined comprehensively based on factors such as the characteristics of the rail material, additive manufacturing process requirements, and the required additive layer performance, aiming to ensure that the metallurgical reaction and solidification behavior during the additive manufacturing process are in optimal condition.

[0135] Furthermore, the phrase "correcting the compensation amount of the temperature signal according to the magnitude and direction of the deviation" means that after obtaining the temperature deviation, the system adjusts the previously calculated temperature signal compensation amount based on the absolute value and sign (positive or negative) of the deviation. For example, if the actual temperature is lower than the target temperature, the compensation amount needs to be increased to improve the laser output power, thereby raising the temperature; conversely, if the actual temperature is higher than the target temperature, the compensation amount needs to be decreased to reduce the laser output power, thereby lowering the temperature. The purpose of this "correction" is to enable the actual temperature of the molten region to dynamically and accurately converge towards the expected target temperature through this deviation-based feedback mechanism.

[0136] This application's solution effectively compensates for the shortcomings of relying solely on temperature change rate control by introducing a direct feedback mechanism for the deviation between the actual temperature of the molten region and the expected target temperature. Specifically, in the above embodiments, although controlling the instantaneous change rate can maintain the dynamic stability of the temperature, it cannot guarantee that the temperature will eventually stabilize precisely at a specific value. By calculating the deviation between the current molten region temperature and the expected target temperature in real time, and correcting the compensation amount of the temperature signal according to the magnitude and direction of this deviation, the system can form a closed-loop control. This allows the adjustment of the laser output power to not only consider the trend of temperature change but also directly respond to the degree of deviation between the temperature and the target value. Thus, when the molten region temperature deviates from the expected target temperature, this deviation information is immediately used to adjust the compensation amount, thereby driving the laser output power to change in the correct direction and causing the molten region temperature to converge towards the expected target temperature. This dual control strategy, combining control of the temperature change rate and precise correction of the absolute temperature value, ensures more refined and accurate temperature control in the additive manufacturing process.

[0137] In some preferred embodiments, a specific example is given below. Assume that during the laser additive manufacturing process for rail defects, the preset target temperature is 1550°C. At a certain moment, the system monitors the current temperature of the molten region as 1530°C. At this time, the system calculates the deviation between the current molten region temperature and the target temperature as -20°C (1530°C - 1550°C). Based on this negative deviation, the system determines that the current temperature is too low and requires increasing the laser output power to raise the temperature. Therefore, the system corrects the temperature signal compensation amount previously determined by the temperature change rate control based on the magnitude and direction of this -20°C deviation; for example, by adding a positive compensation value. This corrected compensation amount will be used to adjust the laser output power so that it can raise the molten region temperature to 1550°C more quickly while maintaining the temperature change rate within a reasonable range. Conversely, if the system monitors the current molten region temperature as 1570°C, the deviation is +20°C. At this point, the system will correct the temperature signal compensation based on the positive deviation, for example, by decreasing a positive compensation value or increasing a negative compensation value, to reduce the laser output power and thus lower the temperature of the molten region back to 1550°C. Through this real-time, deviation-based correction mechanism, even with minor fluctuations in the external environment or internal process parameters, the temperature of the molten region can be precisely guided and stabilized near the expected target temperature, thereby ensuring the stability of the additive manufacturing process and product quality.

[0138] This application further proposes a method for dynamically adjusting the reference range of the molten zone temperature based on ambient wind speed and the temperature distribution in the rail preheating zone, which includes:

[0139] Based on the instantaneous rate of change of temperature in the melting zone and historical data, the ambient wind speed and the temperature distribution in the rail preheating zone are corrected.

[0140] The reference range of the temperature in the melting zone is dynamically adjusted based on the corrected ambient wind speed and the temperature distribution in the rail preheating zone.

[0141] Specifically, the aforementioned correction of the ambient wind speed and the temperature distribution in the rail preheating zone based on the instantaneous rate of change of the molten zone temperature and historical data refers to calibrating or optimizing the input parameters of the ambient wind speed and the temperature distribution in the rail preheating zone by analyzing the real-time trend of the molten zone temperature and past temperature control data. The instantaneous rate of change of the molten zone temperature can be collected and calculated in real time by temperature sensors, while historical data can include temperature change curves, laser output power, powder feeding rate, and corresponding environmental parameter records during the laser additive manufacturing process over a past period. Through comprehensive analysis of this data, a model or algorithm can be established to assess the accuracy of the current ambient wind speed and the temperature distribution in the rail preheating zone, and adaptively correct it. For example, if historical data shows that the instantaneous rate of change of the molten zone temperature consistently deviates from expectations at a specific ambient wind speed and preheating temperature, the effective values ​​of the ambient wind speed or preheating temperature can be adjusted to better reflect the actual thermal response characteristics.

[0142] This application's solution corrects the ambient wind speed and rail preheating zone temperature distribution by incorporating the instantaneous change rate of the molten zone temperature and historical data. This addresses the problem in traditional methods where environmental parameters may be inaccurate or fail to fully reflect the actual thermal environment. Specifically, the instantaneous change rate of the molten zone temperature directly reflects the actual heat loss or absorption in the molten zone, while historical data provides long-term insights into the system's thermal behavior. By combining these real, dynamic feedback information with environmental parameters, the ambient wind speed and rail preheating zone temperature distribution can be more accurately estimated and corrected. Consequently, the corrected environmental parameters more accurately characterize the true thermal environment of the molten zone, allowing for more reliable input when dynamically adjusting the reference range of the molten zone temperature. This ensures a more reasonable and precise reference range setting, laying the foundation for accurate temperature control in laser additive manufacturing.

[0143] In some preferred embodiments, a specific example is given below. Assume that during laser additive manufacturing, the initial ambient wind speed acquired by the system is 2 m / s, and the temperature distribution in the rail preheating area is a uniform 200 degrees Celsius. However, in actual operation, the instantaneous rate of temperature change in the molten area is consistently higher than expected; for example, the temperature decreases faster than predicted by the model. In this case, the system will determine, based on the instantaneous rate of temperature change in the molten area (e.g., a sustained rapid decrease) and historical data (e.g., in similar past situations, even with an ambient wind speed of 2 m / s, the actual cooling effect was stronger), that the current effective ambient wind speed may be higher than 2 m / s, or that the heat dissipation effect in the rail preheating area is more significant than expected. Based on this determination, the system will correct the ambient wind speed, for example, to 2.5 m / s, or correct the temperature distribution in the rail preheating area to reflect its stronger heat dissipation capacity. Subsequently, the system will dynamically adjust the reference range of the molten area temperature based on this corrected ambient wind speed (2.5 m / s) and the corrected temperature distribution in the rail preheating area. For example, the reference range for temperature drop can be appropriately widened to accommodate faster cooling rates, thereby enabling subsequent temperature compensation and laser power adjustment to more accurately guide the temperature of the molten region to the target range, avoiding excessively low or high temperatures due to inaccurate estimation of environmental parameters.

[0144] This application further proposes steps for correcting the ambient wind speed and temperature distribution in the rail preheating zone based on the instantaneous rate of change of the molten zone temperature and historical data, including:

[0145] The instantaneous rate of temperature change in the molten region, the laser output power, and the powder feeding rate are monitored in real time; the powder feeding rate is the rate at which alloy powder is injected into the molten region during laser additive manufacturing.

[0146] In response to abnormal fluctuations in the instantaneous rate of change, the laser output power or powder feeding rate will fluctuate abnormally. The correction range for the ambient wind speed and the temperature distribution of the rail preheating zone will be adjusted according to the degree of influence of the fluctuations in the laser output power or the powder feeding rate on the instantaneous rate of change of the temperature in the melting zone.

[0147] Specifically, real-time monitoring of the instantaneous rate of change of the molten zone temperature, laser output power, and powder feeding rate refers to continuously acquiring real-time data of these key process parameters through corresponding sensors or control system interfaces. The instantaneous rate of change of the molten zone temperature can be obtained through differential calculation of temperature data acquired by devices such as infrared thermometers; the laser output power is provided by the laser control system; and the powder feeding rate is determined by the flow meter or control parameters of the powder feeding device. The powder feeding rate can be understood as the mass or volume of alloy powder injected into the molten zone per unit time during laser additive manufacturing, which directly affects the size, shape, and heat absorption of the molten pool.

[0148] When abnormal fluctuations in the instantaneous rate of change of temperature in the molten zone are detected, the system further checks whether there are synchronous abnormal fluctuations in the laser output power or powder feeding rate. Such synchronous fluctuations indicate that the temperature anomaly may be related to the stability of internal process parameters. Based on this, the correction magnitudes for ambient wind speed and the temperature distribution in the rail preheating area are adjusted according to the degree of influence of fluctuations in laser output power or powder feeding rate on the instantaneous rate of change of temperature in the molten zone. For example, if it is found that the temperature fluctuation is mainly caused by drastic changes in laser power, the correction magnitudes for ambient wind speed and the temperature distribution in the rail preheating area may be reduced, or adjusted to focus more on the control of internal parameters to avoid excessive or inaccurate corrections to external environmental factors. The degree of influence can be assessed using pre-established physical models, empirical formulas, or machine learning models, which can quantify the contribution of fluctuations in different parameters to the rate of change of temperature in the molten zone.

[0149] This application's solution introduces real-time monitoring of internal process parameters (such as laser output power and powder feeding rate) and correlates them with fluctuations in the instantaneous rate of temperature change in the molten zone, enabling more accurate identification of the root cause of temperature fluctuations. When temperature fluctuations are primarily caused by internal parameters, the system can correspondingly adjust the correction magnitude for external environmental factors (ambient wind speed, temperature distribution in the rail preheating zone), avoiding unnecessary or inaccurate external corrections due to misjudgment. This achieves refined management of the temperature control strategy, ensuring robustness and accuracy of temperature control under complex and variable operating conditions.

[0150] In some preferred embodiments, a specific example is given below. Suppose that during laser additive manufacturing, the system detects a sudden and significant drop in the instantaneous rate of temperature change in the molten region. At this point, the system further examines the laser output power and powder feeding rate. If it finds that the laser output power also shows a significant drop at the same time, while the ambient wind speed and the temperature distribution in the rail preheating area remain relatively stable, the system determines that the main cause of this temperature drop is the fluctuation in laser output power. In this case, the system will, based on the degree of impact of the laser output power drop on the rate of temperature change, correspondingly reduce the correction magnitude for the ambient wind speed and the temperature distribution in the rail preheating area, or shift the focus of correction to stabilizing the laser output power. For example, if the drop in laser power causes a decrease in the rate of temperature change of X degrees / second, and according to the model, corrections to the ambient wind speed and preheating temperature typically result in a compensation of Y degrees / second, then the correction magnitude might be adjusted to YX degrees / second, or the temperature might be corrected directly by adjusting the laser power, rather than by significantly modifying environmental parameters. This refined correction strategy ensures the accuracy of temperature control and avoids introducing new instabilities due to excessive correction of external environmental factors.

[0151] This application also discloses a temperature control system for laser additive manufacturing of rail defects, such as... Figure 2 As shown, the system includes:

[0152] The acquisition module 201 is used to acquire the sound signal and temperature signal of the molten area in real time during the laser additive manufacturing process;

[0153] The evaluation module 202 is used to analyze the sound signal, determine whether there are any impurity interference events, and evaluate the reliability of the temperature signal at the current moment based on the analysis results.

[0154] The control module 203 is used to adjust the laser output power based on the reliability assessment results, thereby controlling the laser additive temperature.

[0155] The rail defect laser additive manufacturing temperature control system proposed in this application aims to solve the problems of inaccurate temperature measurement and control failure caused by surface impurities during laser additive manufacturing by integrating multi-source information sensing and intelligent evaluation mechanisms. The system acquires sound and temperature signals from the molten region simultaneously through an acquisition module, intelligently judges the reliability of the temperature signals through an evaluation module, and finally, the control module adaptively adjusts the laser output power based on the evaluation results, thereby ensuring precise temperature control of the molten region under complex working conditions.

[0156] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for controlling the temperature of laser additive manufacturing for rail defects, characterized in that, The method includes: Real-time acquisition of acoustic and temperature signals from the molten region during laser additive manufacturing; The sound signal is analyzed to determine whether there are any interference events, and the reliability of the temperature signal at the current moment is assessed based on the analysis results. Based on the reliability assessment results, the laser output power is adjusted to control the laser additive temperature. The step of analyzing the sound signal to determine whether there are any interference events, and based on the analysis results, assessing the reliability of the temperature signal at the current moment, includes: If an impurity interference event exists, the reliability assessment is determined to be in low reliability mode; if no impurity interference event exists, the reliability assessment is determined to be in high reliability mode. The step of adjusting the laser output power based on the reliability assessment results to control the laser additive temperature includes: In response to the reliability assessment result indicating a high reliability mode, the laser output power is adjusted based on the deviation between the current temperature signal and the target temperature, and calculated using PID control. In response to the reliability assessment result being a low reliability mode, the temperature signal corresponding to the high reliability mode prior to the current temperature signal is obtained and used as the current temperature reference value to calculate and adjust the laser output power.

2. The method for controlling the temperature of laser additive manufacturing for rail defects according to any one of claims 1, characterized in that, The step of analyzing the sound signal to determine whether there are any interference events includes: Acquire multi-channel audio signals and perform preprocessing; perform time-frequency analysis on the preprocessed acoustic signals to obtain time-frequency spectra; Extract multidimensional feature vectors from the time-frequency spectrum and calculate the similarity between the multidimensional feature vectors and preset impurity acoustic features; When the similarity between the multidimensional feature vector and the preset impurity acoustic features is lower than a preset similarity threshold and deviates from the normal sound baseline, it is determined as an undetermined interference event. When a pending interference event is identified, the temperature signal is dynamically compensated based on the degree of deviation between the multidimensional feature vector and the normal sound baseline. When a potential interference event is identified, dynamic compensation is performed on the temperature signal based on the degree of deviation between the multidimensional feature vector and the normal sound baseline, including: Obtain the degree of deviation between the multidimensional feature vector and the normal sound baseline; Based on the degree of deviation and the thermal response characteristics of the molten region, the compensation amount of the temperature signal is dynamically adjusted. The compensation amount of the temperature signal is adjusted based on the current rate of temperature change in the melting zone and the expected target temperature.

3. The method for controlling the temperature of laser additive manufacturing for rail defects according to claim 2, characterized in that, The compensation amount for correcting the temperature signal based on the current rate of temperature change in the melting zone and the expected target temperature includes: Real-time calculation of the instantaneous rate of temperature change in the molten zone, as well as acquisition of ambient wind speed and temperature distribution in the rail preheating zone; The reference range of the temperature in the melting zone is dynamically adjusted based on the ambient wind speed and the temperature distribution in the rail preheating zone. The instantaneous rate of change is compared with the reference range, and the compensation amount of the temperature signal is adjusted in reverse according to the degree and direction of the instantaneous rate of change exceeding the reference range, so as to guide the current rate of change of temperature in the melting region back to the reference range; The reference range for dynamically adjusting the temperature of the melting zone based on the ambient wind speed and the temperature distribution in the rail preheating zone includes: Based on the instantaneous rate of change of the temperature in the melting zone and historical data, the ambient wind speed and the temperature distribution in the rail preheating zone are corrected; based on the corrected ambient wind speed and the temperature distribution in the rail preheating zone, the reference range of the temperature in the melting zone is dynamically adjusted. The step of correcting the ambient wind speed and the temperature distribution in the rail preheating zone based on the instantaneous rate of change of the temperature in the melting zone and historical data includes: The instantaneous rate of temperature change in the molten region, the laser output power, and the powder feeding rate are monitored in real time; the powder feeding rate is the rate at which alloy powder is injected into the molten region during laser additive manufacturing. In response to abnormal fluctuations in the instantaneous rate of change, the laser output power or powder feeding rate will also experience abnormal fluctuations. Therefore, the correction range for the ambient wind speed and the temperature distribution in the rail preheating zone will be adjusted according to the degree of influence of the fluctuations in the laser output power or the powder feeding rate on the instantaneous rate of change of the temperature in the melting zone.

4. The method for controlling the temperature of laser additive manufacturing for rail defects according to claim 3, characterized in that, The correction of the compensation amount based on the current rate of temperature change in the molten region and the expected target temperature further includes: Calculate the deviation between the current melting zone temperature and the expected target temperature; Based on the magnitude and direction of the deviation, the compensation amount of the temperature signal is corrected so that the temperature of the melting region approaches the expected target temperature.

5. A temperature control system for laser additive manufacturing of rail defects, used to execute the temperature control method for laser additive manufacturing of rail defects as described in any one of claims 1-4, characterized in that, The system includes: The acquisition module is used to acquire sound signals and temperature signals of the molten area in real time during the laser additive manufacturing process; The evaluation module is used to analyze the sound signal, determine whether there are any interference events, and evaluate the reliability of the temperature signal at the current moment based on the analysis results. The control module is used to adjust the laser output power based on the reliability assessment results, thereby controlling the laser additive temperature.

Images