Online monitoring and process optimization system and method for electron beam additive manufacturing

Abstract

The invention discloses an online monitoring and process optimization system and method for electron beam additive manufacturing, and aims to solve the problems that in the prior art, the source of monitoring information is single, monitoring and control are disjointed, and the process cannot be adaptively adjusted according to real-time working conditions. Comprising a thermal infrared imager, a machine vision camera, a thermocouple array, a high-temperature strain gauge, a preset process parameter database and a data processing and display control unit. Through cooperative arrangement of multiple sensors and a special mounting structure, synchronous acquisition of surface temperature, molten pool state and stress strain is realized, a three-dimensional thermal sensing model is constructed and operated in real time, data of the multi-source sensors are fused, a temperature field and a strain state in the additive manufacturing process are reconstructed, and the temperature field and the strain state of the additive manufacturing process are obtained. And real-time data is monitored by combining a unified human-computer interaction interface, and the monitored data is directly applied to process control through a real-time sensing-intelligent decision method, so that the process optimization of the manufacturing process is realized, and the stability of the process is remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of additive manufacturing process control technology, specifically an online monitoring and process optimization system and method for electron beam additive manufacturing. Background Technology

[0002] Electron beam powder bed melting is a key technology for manufacturing high-performance metal parts, boasting advantages such as high energy density, fast forming efficiency, and a pure vacuum environment, making it promising for applications in aerospace and biomedical fields. However, in actual manufacturing processes, the electron beam heat source scanning path, power, and other factors can lead to high temperature gradients in the powder bed and formed part areas, as well as residual stress caused by rapid cooling. This can result in metallurgical defects such as warping and cracking, severely impacting the dimensional accuracy and service performance of the parts.

[0003] Currently, monitoring technologies for electron beam powder bed additive manufacturing suffer from limitations. These include limited monitoring methods and difficulty in simultaneously acquiring multi-source information such as temperature and strain. Existing technologies, such as thermocouple measurements, can only reflect the temperature of localized points on the substrate, with a limited number of measurement points, making it difficult to capture the overall temperature field of the powder bed. While high-speed cameras capture the instantaneous morphology and dynamic behavior of the molten pool, their field of view is limited to a small melting area, lacking macroscopic monitoring of the surrounding, broader heat-affected zone. Therefore, a single monitoring method can only reflect the complex manufacturing process from one perspective and cannot construct a complete and mutually corroborating process image.

[0004] For stress monitoring, methods such as X-ray diffraction and drilling are mostly offline and destructive testing methods, which cannot reflect the dynamic accumulation and evolution of stress during the forming process in real time. Although there have been attempts to use sensors such as strain gauges for in-situ measurements, existing technologies generally lack systems that can synchronously and uniformly integrate multi-point, multi-dimensional temperature field monitoring with strain monitoring at key locations, which limits the in-depth understanding of the thermo-mechanical coupling mechanism and the optimization of the process.

[0005] Electron beam additive manufacturing equipment operates in a complex environment characterized by high vacuum, high-temperature radiation, and metal dust, posing significant challenges to the stability and signal quality of monitoring systems. Many existing monitoring solutions lack sufficient consideration for system engineering design. For example, the mounting and fixing methods of sensors (such as thermocouples) may not be robust enough, affecting measurement repeatability; the signal attenuation of optical windows (such as infrared windows) may not be calibrated, introducing systematic errors; and weak analog signals (such as thermocouple and strain gauge signals) are susceptible to electromagnetic interference and distortion during long-distance transmission. These problems all reduce the reliability of monitoring systems and the availability of data in actual production environments.

[0006] Furthermore, the monitoring and process control functions of existing monitoring technologies are typically separated. When a monitoring system identifies process deviations such as uneven temperature distribution, the system itself usually lacks the ability to automatically make decisions and invoke preset strategies to adjust process parameters. Adjustments to process parameters often require manual intervention by operators based on monitoring data, and the timeliness and accuracy of these adjustments are highly dependent on the operator's experience. Therefore, the application of monitoring data is mostly focused on post-event quality analysis, and its real-time optimization potential during the manufacturing process remains largely untapped.

[0007] In summary, there is a current need in the field for a comprehensive monitoring solution that can reliably acquire multi-dimensional information and apply real-time monitoring data to process control, thereby achieving an optimized system for dynamic adjustment of process parameters and further improving the stability and automation level of the electron beam additive manufacturing process. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of existing technologies, such as the reliance on a single monitoring method, the disconnect between monitoring and control, and the inability to adaptively adjust process parameters according to real-time operating conditions. The invention aims to provide a multi-source information monitoring method for electron beam powder bed fusion additive manufacturing processes based on thermocouple arrays, infrared thermal imagers, machine vision cameras, and high-temperature strain gauges. This method enables synchronous and multi-dimensional online monitoring of the temperature of the substrate, the full-width powder bed, and key forming areas, as well as the stress at the substrate-part interface. Furthermore, it integrates online monitoring with process optimization, achieving intelligent and closed-loop control of the electron beam additive manufacturing process.

[0009] This invention provides an online monitoring and process optimization system for electron beam additive manufacturing, the system comprising: An infrared thermal imager is used to acquire the global surface temperature distribution of a powder bed during additive manufacturing. The temperature measurement range of this invention covers at least 200°C to 2000°C. Machine vision cameras are used to monitor the morphology of the molten pool area during additive manufacturing. They can quantify the geometric features of the molten pool and identify spatter, keyhole, and other process defects. A thermocouple array comprising multiple thermocouples for measuring the temperature at multiple preset points on a substrate during additive manufacturing; High-temperature strain gauges are used to measure the strain of the substrate during additive manufacturing, providing measured data for verifying the thermo-mechanical coupling simulation results derived from the thermal simulation process. A process parameter database pre-stores multiple combinations of electron beam currents, scanning speeds, and scanning paths, along with corresponding standard temperature field data. A data processing and display control unit is also included, connected to an infrared thermal imager, machine vision camera, thermocouple array, and high-temperature strain gauges. The data processing and display control unit is configured to: synchronously acquire global surface temperature distribution, molten pool morphology, multi-point temperature, and substrate strain information obtained from various sensors; and construct and run a data-driven three-dimensional thermo-mechanical... The sensing model is configured to: fuse data acquired by an infrared thermal imager module, a thermocouple array, and a machine vision camera module to reconstruct the three-dimensional temperature field of the additive manufacturing process; calculate the theoretical strain at the location of the high-temperature strain gauge based on the three-dimensional temperature field; compare the theoretical strain with the measured strain of the high-temperature strain gauge to verify the model; display multi-source information in an integrated human-machine interface; compare the three-dimensional thermal sensing model with the standard data corresponding to the current process parameters in the process parameter database; and generate electron beam current, scanning speed, or scanning path for correction based on the deviation of the comparison results.

[0010] This invention provides a specific implementation method in which the three-dimensional thermal sensing model is specifically configured as follows when reconstructing a three-dimensional temperature field: The surface temperature field obtained by the infrared thermal imager module is used as the top boundary condition of the model; and the substrate temperature obtained by the thermocouple array is used as the bottom boundary condition of the model.

[0011] This invention provides a specific embodiment in which the data processing and display control unit is further configured as follows: During the preheating stage of additive manufacturing, based on the monitoring data of the thermocouple array, it is determined whether the substrate has reached the preset uniformity and target temperature requirements, and based on the judgment result, it is decided whether to start the melting and solidification stage.

[0012] This invention provides a specific embodiment, which further includes a thermocouple fixing device for mounting a thermocouple array, comprising: A thermocouple socket is provided on the substrate to match the size of the thermocouples in the thermocouple array for precise positioning; and a thermocouple holder is provided to press the sensing end of the thermocouple array against the bottom of the thermocouple socket to ensure good thermal contact.

[0013] This invention provides a specific embodiment, which also includes a signal acquisition device, comprising: The junction box is located outside the vacuum chamber of the additive manufacturing equipment. The thermocouple acquisition module is integrated inside the junction box. The signals from the thermocouple array are connected to the junction box via thermocouple cables. The thermocouple acquisition module acquires and digitizes the signals, which are then led out from the vacuum chamber via an aviation connector.

[0014] This invention provides a specific embodiment in which a high-temperature strain gauge is fixed to a substrate with a high-temperature resistant adhesive and temperature compensation is performed using a full-bridge method.

[0015] This invention provides a specific implementation method, in which the integrated human-computer interaction interface includes: The system provides three-dimensional visualization of the three-dimensional temperature field, including a machine vision camera interface for displaying the morphology of the molten pool region, an infrared imaging interface for displaying the global surface temperature distribution, and a thermocouple array three-dimensional temperature distribution interface for displaying multi-point temperature data in three dimensions.

[0016] The present invention further provides a method for an online monitoring and process optimization system for electron beam additive manufacturing, comprising the following steps: A data-driven 3D thermal sensing model is constructed and run. The model construction steps include: fusing data acquired by an infrared thermal imager module, a thermocouple array, and a machine vision camera module to reconstruct the 3D temperature field of the additive manufacturing process; and executing process optimization decisions. The decision-making steps include: evaluating temperature characteristics based on the high-fidelity data output by the 3D thermal sensing model; and automatically invoking matching optimization strategies when the characteristics deviate from the standards in the preset process parameter database, and generating control signals for dynamically adjusting subsequent electron beam process parameters.

[0017] This invention provides a specific implementation method in which the steps of reconstructing the internal three-dimensional temperature field specifically include: The surface temperature field obtained by the infrared thermal imager module is used as the top boundary condition of the model; and the substrate temperature obtained by the thermocouple array is used as the bottom boundary condition of the model.

[0018] This invention provides a specific embodiment in which, during the preheating stage at the beginning of the method, the method further includes: Based on the monitoring data of the thermocouple array, it is determined whether the substrate has achieved the preset uniformity and target temperature requirements.

[0019] The beneficial effects of this invention are: 1. This invention achieves multi-source data fusion modeling, providing three-dimensional temperature field information. By integrating thermocouple arrays, infrared thermal imagers, machine vision cameras, and high-temperature strain gauges, it simultaneously acquires multi-dimensional information such as the temperature at multiple points inside the substrate, the global temperature field of the powder bed, the morphology and temperature of the melting core area, and the strain on the substrate surface. This combination of multi-source information provides a more comprehensive description of the thermophysical behavior of the additive manufacturing process, which helps to deeply analyze the impact of process parameters on the final forming quality.

[0020] 2. Improve anti-interference capability. Electron beam additive manufacturing is accompanied by interference from strong electromagnetic fields, high temperatures, and metal vapors. The technical solution of this invention includes targeted anti-interference design: by embedding the thermocouple acquisition module inside the vacuum chamber and performing digital processing at the signal source end, the risk of electromagnetic noise introduced during long-distance transmission of analog signals is reduced. At the same time, by pre-calibrating and compensating the transmittance of the infrared observation window, the measurement error caused by window attenuation is reduced.

[0021] 3. Establish a correlation verification mechanism between temperature field and strain measurement. The reconstructed three-dimensional temperature field is used as input to calculate the theoretical strain value at the location of the high-temperature strain gauge and compare it with the measured strain value of the high-temperature strain gauge. This design establishes a mutual verification relationship between temperature measurement and strain measurement.

[0022] 4. To achieve dynamic adjustment of process parameters based on real-time monitoring data, this invention combines a real-time monitoring module with a preset process parameter database to construct a technical path of real-time data evaluation - database strategy invocation - process parameter adjustment; based on the real-time evaluated operating conditions such as temperature uniformity, the corresponding optimization strategy is automatically invoked to adjust subsequent process parameters, thereby directly applying the monitoring data to process control. Attached Figure Description

[0023] The accompanying drawings used in the following description of the embodiments or prior art will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the hardware structure of a multi-source information online monitoring system according to an embodiment of this application.

[0025] 101. Vacuum chamber; 102. Electron gun; 103. Infrared thermal imager; 104. Machine vision camera; 105. Infrared observation window; 106. Molded part; 107. High-temperature strain gauge; 108. Substrate; 109. Thermocouple array; 110. Thermocouple mounting substrate; 111. Strain gauge transmission cable; 112. Thermocouple plug; 113. Thermocouple transmission cable; 114. Thermocouple acquisition module; 115. Thermocouple junction box; 116. Aviation plug; 117. Working platform; 118. Lead screw.

[0026] Figure 2 This is a schematic diagram of a thermocouple array fixing and installation structure according to an embodiment of this application.

[0027] 108. Substrate, 1081. Thermocouple socket, 1082. Fixing screw socket, 1083. Thermocouple holder.

[0028] Figure 3 This is a diagram of the human-machine interface of a data processing and display control unit provided according to an embodiment of this application.

[0029] Human-computer interaction interface, 302. Machine vision camera interface, 303. Infrared imaging interface, 304. Three-dimensional temperature distribution interface of thermocouple array.

[0030] Figure 4 This is an experimental comparison diagram provided according to an embodiment of this application. Detailed Implementation

[0031] The following detailed description of the technical solution of the present invention, in conjunction with preferred embodiments and accompanying drawings, is non-limiting. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0032] An online monitoring and process optimization system for electron beam additive manufacturing, the system comprising: The temperature monitoring module is configured to acquire temperature information from multiple spatial scales during the additive manufacturing process, including: Thermocouple arrays, with multiple temperature probes integrated into the bottom of the additively manufactured substrate, are used to accurately measure the real-time temperature of multiple points on the substrate. The thermocouple arrays employ armored thermocouples to enhance mechanical strength and damage resistance in high-temperature, vacuum, and metal dust environments.

[0033] Furthermore, to address the issues of inaccurate thermocouple positioning and poor contact, a dedicated thermocouple fixing and adjustment device is also included. This device includes a fixing base plate with positioning holes to ensure that each thermocouple sensing end is accurately inserted into the preset temperature measurement position; and an adjustable clamping seat located at the bottom of the base plate to uniformly control the extension length of each thermocouple, ensuring that the array is coplanar and forms good thermal contact with the measurement point.

[0034] Specifically, the signal acquisition module for the thermocouple array includes a multi-channel thermocouple acquisition module integrated and installed inside the vacuum chamber; the acquisition module is encapsulated in a cast aluminum junction box, and the junction box uses compensating wires of the same material as the thermocouples to introduce the analog signals of the external thermocouple array into the acquisition module; the acquisition module directly performs digital processing on the received analog temperature signals in the junction box, converting them into digital signals; the digitally processed digital signals are led out from the vacuum chamber to the external data processing and display control unit through an aviation plug and an external shielded cable.

[0035] An infrared thermal imager, installed at the top of the vacuum chamber, performs non-contact temperature measurement on the entire powder bed surface through an infrared observation window to obtain the global temperature field distribution. The infrared observation window is made of a material with good transmittance in the mid- and far-infrared bands, such as barium fluoride. This ensures that infrared radiation signals can effectively penetrate.

[0036] Furthermore, to compensate for the absorption and attenuation of infrared signals by the infrared observation window, the system also includes infrared window transmittance calibration. Before temperature measurement, the actual transmittance of the infrared observation window is calibrated using a standard heat source (such as a blackbody furnace), and the calibrated transmittance value is preset into the infrared thermal imager to improve the accuracy of temperature measurement.

[0037] The machine vision camera provides high-resolution, high-frame-rate imaging of the high-temperature core area during preheating and forming, capturing images of localized regions. Mounted at the top of the vacuum chamber, adjacent to the electron beam gun, the camera observes the molten area through a visible light viewing window. The machine vision camera has dual functions: first, to capture the molten pool morphology in real time; and second, to perform high-precision temperature measurements of the molten pool based on colorimetric thermometry. For ease of remote operation and precise focusing, the machine vision camera is equipped with a motorized lens, allowing the operator to automatically or manually focus the observation area from outside the vacuum chamber.

[0038] The stress monitoring module includes at least one high-temperature strain gauge for online measurement of the substrate's strain during thermal cycling to reflect the accumulation of residual stress. Specifically, the high-temperature strain gauge is a high-temperature resistance wire strain gauge, which is fixed to the side or a predetermined position of the additive manufacturing substrate with high-temperature adhesive. To eliminate the interference of temperature changes on the measurement results, the high-temperature strain gauge uses a full-bridge method for temperature self-compensation. The signal line of the high-temperature strain gauge is led out from the vacuum chamber via an aviation connector and an external shielded cable, connecting to an external strain acquisition instrument to obtain real-time strain data of the substrate.

[0039] The signal acquisition and data processing unit, consisting of one or more computers and software systems running on them, is configured to: synchronously trigger and acquire multi-channel data, ensuring that data from different sensors are time-aligned; for thermocouple array data: receive and visualize digital temperature signals from the acquisition module inside the vacuum chamber, generating 3D images; for infrared thermal imager data: apply preset window transmittance and material emissivity parameters to process it into a full-frame pseudo-color temperature cloud map; for machine vision camera data: control the camera's acquisition parameters and drive the motorized lens to focus, extract the geometric features of the molten pool from the acquired image, and calculate the molten pool temperature based on colorimetric thermometry; for high-temperature strain gauge data: receive signals from the strain acquisition instrument and display them as strain-time evolution curves.

[0040] A three-dimensional thermal sensing model is constructed and run in real time to achieve the fusion of multi-source information. Specifically, it is configured to: reconstruct the three-dimensional temperature field of the entire forming area using the global surface temperature field as the top heat source input and the substrate temperature distribution as the bottom boundary condition; and calculate the theoretical strain at the high-temperature strain gauge location using the three-dimensional temperature field as input, comparing it with the substrate strain data to verify the accuracy of the model and identify abnormal mechanical events. Process optimization decisions are executed: the homogenization degree is evaluated in real time using the three-dimensional temperature field or the global surface temperature field as input; if the homogenization degree deviates from a preset threshold, an optimization strategy matching the current non-uniformity mode is retrieved from the process parameter database; and process parameters for adjusting subsequent additive manufacturing processes are generated based on the optimization strategy. A human-machine interface is provided, configured to display the calculated three-dimensional temperature image, temperature cloud map, high-definition image of the molten pool, and substrate strain evolution curve, and provide data recording, playback, and export functions.

[0041] The process parameter database is configured as the decision-making core of the system, used to store the correlation between process parameters and monitoring data. Specifically, the process parameter database pre-stores multiple sets of reference process parameters, including electron beam current, scanning speed, and scanning path. The process parameter database also stores standard monitoring data characteristics corresponding to the reference process parameters, as well as optimization strategies for correcting process deviations. The standard monitoring data characteristics include standard temperature distribution uniformity indicators, and the optimization strategies include a set of adjustment instructions for the reference process parameters.

[0042] Furthermore, this invention also provides a method for online monitoring and process optimization based on the above-mentioned system using multi-source information. This method aims to achieve adaptive closed-loop control of the manufacturing process by performing multi-dimensional real-time sensing of the additive manufacturing process and actively adjusting process parameters based on real-time evaluation results. Specifically, it includes the following steps: S1: System Deployment and Pre-calibration Steps. This step is performed before the start of the additive manufacturing task and aims to complete the physical installation and parameter calibration of all sensors.

[0043] S2: Synchronous data acquisition and anti-interference transmission steps, which are carried out throughout the entire additive manufacturing process (including the entire process of substrate preheating, powder spreading, pre-sintering, selective melting and cooling).

[0044] S3: Multi-source data fusion and three-dimensional thermal sensing model construction step. This step is executed by the software system of the data processing and display control unit. Its purpose is to construct and solve a three-dimensional thermal physical model from the multi-source data collected in S2.

[0045] S4: Process optimization decision and execution steps. This step receives the characteristic indicators calculated in the previous step, compares and makes decisions with the preset process database, and actively generates control instructions for adjusting process parameters.

[0046] S5: Multi-dimensional information integration and recording steps, which integrates data from different sensors onto a unified human-machine interface to achieve comprehensive monitoring of the additive manufacturing process and record relevant data.

[0047] Furthermore, S1 includes the following steps: S11: Install the thermocouple array at the preset position on the substrate using a dedicated fixing and adjustment device. Positioning holes ensure accurate temperature measurement point placement, and adjustable clamps standardize the extension height of each thermocouple, ensuring good contact with the substrate. Connect the thermocouple array signal cables to the junction box inside the vacuum chamber, and lead the data from the temperature acquisition module out of the vacuum chamber and connect it to the industrial control computer via cables and aviation connectors.

[0048] S12: The high-temperature strain gauge is attached to the side of the substrate or other designated monitoring points using high-temperature adhesive, and the circuit connection is completed using the full-bridge method to achieve temperature self-compensation. The strain gauge signal cable is connected to an external cable and led out of the vacuum chamber, connecting to the industrial control computer.

[0049] S13: Using a standard heat source such as a blackbody furnace outside the vacuum chamber, measure the apparent radiation temperature before and after passing through the window. Calculate the actual average transmittance of the window based on physical laws. Input the calibrated window transmittance value and the emissivity of the metal powder material to be processed into the infrared thermal imager's software system beforehand. Mount the infrared thermal imager on the bracket at the top of the vacuum chamber, aligning its optical axis perpendicular to the infrared observation window. Adjust the wide-angle lens to completely cover the entire powder bed working area.

[0050] S14: Install a machine vision camera on the top of the vacuum chamber so that it can clearly observe the electron beam melting core area through the visible light observation window.

[0051] Furthermore, S2 includes the following steps: S21: At the same time as the electron beam powder bed additive manufacturing process begins, a synchronous trigger signal is sent to all sensors (thermocouple array acquisition module, infrared thermal imager, machine vision camera, strain acquisition instrument) through the data processing and display control unit to ensure that all data streams are aligned on the time reference.

[0052] S22: The thermocouple acquisition module inside the vacuum chamber acquires the analog voltage signal generated by the thermocouple array in parallel, and performs A / D conversion directly in the junction box to convert the analog signal into a digital signal stream with high anti-interference capability.

[0053] S23: The infrared thermal imager, machine vision camera, and strain acquisition instrument each independently acquire their own raw data.

[0054] Furthermore, S3 includes the following steps: S31: Solve the digital signal from the thermocouple acquisition module to obtain the temperature values ​​of multiple precise temperature measurement points on the substrate.

[0055] S32: Apply the pre-set window transmittance and material emissivity parameters to compensate and correct the raw radiation data collected by the infrared thermal imager, and generate a full-width pseudo-color temperature cloud map that reflects the true temperature distribution.

[0056] S33: Apply a colorimetric temperature measurement algorithm to the image acquired by the machine vision camera to calculate the temperature of the molten pool region.

[0057] S34: Solve the bridge signal measured by the strain acquisition instrument into the micro-strain value of the substrate.

[0058] S35: Using the full-area temperature cloud map generated in S31-33 as the top heat source boundary condition of the model, and the multi-point temperature distribution at the bottom of the substrate as the bottom boundary condition, and combining the geometric features of the molten pool, the internal three-dimensional temperature field distribution of the entire forming area is reconstructed by solving the three-dimensional transient heat conduction equation.

[0059] S36: Using the three-dimensional temperature field output from S35 as input, the theoretical strain evolution curve at the location of the high-temperature strain gauge is calculated through a simplified thermo-mechanical coupling model (e.g., the inherent strain method). This theoretical strain curve is compared with the measured strain-time curve obtained from S34 to verify the accuracy of the entire three-dimensional thermal field model.

[0060] Furthermore, S4 includes the following steps: S41: The system performs quantitative analysis on the three-dimensional temperature field or surface temperature distribution output by S3, and calculates characteristic indicators representing the degree of temperature uniformity through methods such as virtual mesh division.

[0061] S42: When the characteristic index continues to deviate from the preset threshold in the process parameter database, the system determines that optimization is needed and triggers the subsequent process.

[0062] S43: The system first identifies the specific pattern and region of the current non-uniform state. Then, the system identifies the matching abnormal features in the process parameter database and calls the corresponding preset optimization strategy from the optimization strategy library.

[0063] S44: Finally, the system converts the adjustment instruction set in the optimization strategy into control signals that the equipment can directly execute, and sends them to the control system of the additive manufacturing equipment through the standard industrial communication protocol to adjust the subsequent scanning path and process parameters.

[0064] Furthermore, S5 includes the following steps: S51: On the multi-window display of the human-computer interaction interface, based on a unified timeline, the following are displayed synchronously: a 3D temperature view generated from thermocouple array data; a full-area temperature cloud map generated from infrared thermal imager data, with the ability to select any point to view the temperature value; real-time high-definition video of the molten pool generated from a machine vision camera; and a substrate strain-time curve generated from high-temperature strain gauge data. A 3D temperature field view generated from a 3D thermal sensing model is also included.

[0065] S52: Synchronously store all processed multidimensional information and original data to the hard disk for subsequent process quality traceability, data playback, thermal-stress field simulation verification and process optimization.

[0066] Reference Figure 1This embodiment provides the overall structure of an online monitoring and process optimization system and method for electron beam additive manufacturing. The system is deployed in the vacuum chamber 101 of an electron beam additive manufacturing device. Multiple electron guns 102 are installed on the top of the vacuum chamber 101 for melting metal powder.

[0067] The system's hardware includes non-contact sensors, contact sensors, and signal acquisition and transmission devices.

[0068] Non-contact sensors are mounted on top of the vacuum chamber 101, including at least one infrared thermal imager 103 and at least one machine vision camera 104. The infrared thermal imager 103 monitors the global temperature field of the entire powder bed (including the substrate 108 and the molded part 106) through an infrared observation window 105. To ensure temperature measurement accuracy, the infrared observation window 105 is pre-installed using a blackbody furnace, and the radiance before and after the window is measured at 800 °C, 1000 °C, and 1200 °C, respectively. The transmittance is calculated and entered into the thermal imager calibration file. The machine vision camera 104 monitors the high-resolution morphology and dynamic process of the core area of ​​electron beam melting through another observation window.

[0069] The contact sensor includes a high-temperature strain gauge 107 and a thermocouple array 109. The high-temperature strain gauge 107 is bonded to the side of the substrate 108 with high-temperature resistant ceramic adhesive, and is arranged in a four-piece full-bridge configuration: two gauges are symmetrically mounted along the tensile direction and two along the compressive direction to achieve self-compensation for temperature drift. The thermocouple array 109 is arranged below the substrate 108 to accurately measure the multi-point temperature distribution on the bottom surface of the substrate.

[0070] The sensor installation and signal transmission structure are referenced. Figure 1 and Figure 2 The installation and fixation of the thermocouple array 109 are achieved through a dedicated device. For example... Figure 2 As shown, the bottom surface of the substrate 108 is pre-machined with thermocouple insertion holes 1081 for inserting thermocouples and fixing screw holes 1082 for fixing, ensuring the accuracy of the temperature measurement point position. The thermocouple is pressed against the bottom of the thermocouple insertion hole 1081 from below by an adjustable thermocouple holder 1083. The entire thermocouple array 109 is mounted on a thermocouple fixing substrate 110, which, together with the substrate 108, is placed on a work platform 117 that can be lifted and lowered by a lead screw 118.

[0071] The signal transmission path of the thermocouple array 109 is as follows: each thermocouple signal cable 113 is connected to the thermocouple junction box 115 via a thermocouple plug 112. The junction box 115 integrates a multi-channel thermocouple acquisition module 114. This module uses a cast aluminum housing and is equipped with twisted-pair shielded signal cables to enhance its electromagnetic interference resistance. Inside the vacuum chamber, the module converts the received multi-channel analog voltage signals into digital signals, which are then transmitted through an aviation plug 116 out of the vacuum chamber 101 to the data processing and display control unit.

[0072] The data processing and display control unit consists of a high-performance industrial computer and its running software system. This system adopts a modular architecture design, with multiple functional modules working collaboratively to achieve functions such as multi-source information acquisition, fusion analysis, process optimization decision-making, and visualization display. Its functions are implemented as follows: (1) Implementation of data acquisition and synchronization functions: This unit is responsible for the synchronous acquisition of multi-channel data. Specifically, the system communicates with sensors such as infrared thermal imagers and machine vision cameras at the low level through standard device communication protocols (such as GigE Vision and Modbus). A unified time synchronization mechanism (including hardware trigger signals or Network Time Protocol NTP) is used to add timestamps to the data acquired by each sensor.

[0073] (2) Construction and solution of the three-dimensional thermal sensing model: The system feeds the raw acquired data into a 3D thermal sensing model for fusion and modeling. An image extraction unit processes the machine vision camera images. A finite element method based on a general numerical computing platform is used to solve the heat conduction equation in real time to calculate the 3D temperature distribution of the entire region. Combined with a material parameter model, a strain-stress coupling calculation unit is used to calculate the theoretical strain field of the substrate, thereby obtaining the spatial mapping relationship of the thermal distribution.

[0074] (3) Process optimization decision-making: The system uses the temperature parameters output from the 3D thermal sensing model as input to the decision engine. The decision process is as follows: Step 1: Real-time thermal field feature extraction The system performs real-time analysis on the global temperature distribution map acquired by the infrared imaging interface 303, and overlays a virtual grid on the image to divide the powder bed into multiple monitoring areas. For each area, it calculates characteristic quantities such as average temperature, temperature standard deviation, and maximum temperature, generating a feature vector describing the transient thermal state of the powder bed, providing input data for subsequent anomaly pattern recognition.

[0075] Step 2: Rule-based anomaly pattern matching.

[0076] Real-time thermal field feature vectors are input into the decision engine and matched against anomaly feature rules in the process parameter database. These rules are quantifiable conditional-action logic statements, such as: When the average temperature in a certain area is lower than a certain temperature difference threshold of the average temperature of the entire field and the duration exceeds a set time threshold, it is determined to be a "local supercooling" mode. When the average temperature of more than a certain percentage of the monitored areas exceeds the upper limit temperature, it is determined to be in "global overheating" mode. When the average temperature difference between adjacent areas exceeds the gradient threshold, it is determined to be in the "excessive thermal gradient" mode.

[0077] The decision engine identifies the anomaly type based on the matching results and generates a corresponding pattern identifier for subsequent policy invocation.

[0078] Step 3: Optimize strategy invocation and parameter generation.

[0079] When an abnormal pattern is identified and triggered, the system automatically calls the correction function or algorithm model corresponding to that pattern from the optimization strategy library of the process parameter database.

[0080] For example, in the "locally supercooled" mode, the system calculates the electron beam current correction value based on the temperature difference ΔT: .

[0081] Where k is a proportionality coefficient calibrated based on the material's thermophysical parameters. The system uses this coefficient to generate new beam current settings, scanning speed, or path correction parameters to achieve regional process compensation.

[0082] Step 4: Conversion and output of control commands The generated correction parameters are encapsulated into instructions recognizable by the equipment control system via the control signal output module and sent to the electron beam controller via a standard industrial communication protocol (such as Modbus TCP). The control system adjusts the process parameters of the corresponding region based on the received instructions, achieving real-time closed-loop optimization of the powder bed temperature field.

[0083] (4) Implementation of human-computer interaction interface functions: Reference Figure 3 This embodiment provides a human-machine interface 301 for this unit. This interface adopts a multi-window, integrated layout, synchronizing and visually presenting multi-source information from different sensors and process optimization decision results. Specifically, the human-machine interface 301 includes: Machine vision camera interface 302: This window displays high-definition images or videos of the molten pool captured by the machine vision camera 104 in real time, and provides control over the adjustment of camera parameters (such as exposure, gain, etc.).

[0084] Infrared Imaging Interface 303: This window displays a pseudo-color temperature cloud map acquired by the infrared thermal imager 103 in real time. When a process optimization process is triggered, this interface will also highlight the identified abnormal areas and provide text prompts for the currently executing optimization strategy.

[0085] Thermocouple array three-dimensional temperature distribution interface 304: It receives multi-point temperature data from thermocouple acquisition module 114 and visualizes the discrete temperature point data as a three-dimensional temperature field distribution map at the bottom of the substrate through a three-dimensional reconstruction algorithm.

[0086] In this embodiment, the process parameter database pre-stores the correspondence between electron beam current (10–30 mA), scanning speed (700–1500 mm / s), scanning path mode (linear, checkerboard, spiral, etc.) and standard temperature field distribution. Internally, it stores the aforementioned rules and strategy functions, which are calibrated through extensive experimental data and simulation analysis, and support dynamic updates: users can import new experimental data, and the system automatically corrects database entries to improve adaptability and prediction accuracy.

[0087] In this embodiment, the system is used in the additive manufacturing process of GH3536 alloy. Comparative experiments verify that the monitoring and optimization system of this invention can reduce the powder bed temperature uniformity deviation from ±75 ℃ to ±30 ℃. Figure 4 It effectively suppresses hot cracking defects, demonstrating its effectiveness in improving process stability and quality control.

[0088] The description and application in this specification are illustrative and not intended to limit the scope of the invention to the embodiments described above. Modifications and variations to this specification will be apparent to those skilled in the art. Therefore, the claims are intended to cover all such modifications and variations that fall within the true spirit and scope of the invention.

Claims

1. An online monitoring and process optimization system for electron beam additive manufacturing, characterized in that, An infrared thermal imager (103) for acquiring a global surface temperature distribution of a powder bed in an additive manufacturing process; A machine vision camera (104) for monitoring a morphology of a melt pool region in the additive manufacturing process, quantifying geometric features of the melt pool and identifying spatter, keyhole process defects; A thermocouple array (109) comprising a plurality of thermocouples for measuring temperatures of a plurality of preset points of a substrate (108) in the additive manufacturing process; A high-temperature strain gauge (107) for measuring a strain of the substrate (108) in the additive manufacturing process to provide measured data for verifying a thermal-mechanical coupling simulation result derived from a thermal simulation process; A process parameter database pre-storing a plurality of combinations of electron beam current, scanning speed and scanning path and corresponding standard temperature field data; and a data processing and display control unit connected with the infrared thermal imager (103), the machine vision camera (104), the thermocouple array (109) and the high-temperature strain gauge (107), The data processing and display control unit is configured to: synchronously acquire global surface temperature distribution, melt pool region morphology, multi-point temperature and substrate strain information acquired by each sensor; construct and run a data-driven three-dimensional thermal-mechanical perception model configured to: fuse data acquired by the infrared thermal imager module (103), the thermocouple array (109) and the machine vision camera module (104) to reconstruct a three-dimensional temperature field of the additive manufacturing process, and based on the three-dimensional temperature field, calculate a theoretical strain at the position of the high-temperature strain gauge (107), compare it with the measured strain of the high-temperature strain gauge (107) to verify the model; and integrally display the multi-source information on an integrated human-computer interaction interface (301); compare the three-dimensional thermal-mechanical perception model with standard data corresponding to current process parameters in the process parameter database, and according to a deviation of the comparison result, generate an electron beam current, scanning speed or scanning path for correction. The three-dimensional thermal-mechanical perception model, when reconstructing the three-dimensional temperature field, is specifically configured to:

2. The system of claim 1, wherein, use the surface temperature field acquired by the infrared thermal imager module (103) as a top boundary condition of the model; and use the substrate temperature acquired by the thermocouple array (109) as a bottom boundary condition of the model. The data processing and display control unit is further configured to:

3. The system of claim 2, wherein, in a preheating stage of the additive manufacturing, based on monitoring data of the thermocouple array (109), judge whether the substrate (108) reaches preset uniformity and target temperature requirements, and based on a judgment result, decide whether to start a melting and solidification stage. Further comprising a thermocouple fixing device for installing the thermocouple array (109), comprising:

4. The system of claim 1, wherein, ​ A thermocouple socket (1081) is arranged on the substrate (108) to match the size of the thermocouples in the thermocouple array (109) for precise positioning, and a thermocouple clamping seat (1083) is used to tightly top the sensing end of the thermocouple array (109) to the bottom of the thermocouple socket (1081) to ensure good thermal contact.

5. The system of claim 1, wherein, The signal acquisition device comprises: A junction box (115) is arranged outside the vacuum chamber (101) of the additive manufacturing equipment, and a thermocouple acquisition module (114) is integrated inside the junction box (115), wherein the signals of the thermocouple array (109) are connected to the junction box (115) through thermocouple cables (113), and the signal acquisition and digital processing are performed by the thermocouple acquisition module (114) and the signals are led out from the vacuum chamber (101) through an aviation plug (116).

6. The system of claim 2, wherein, The high-temperature strain gauge (107) is fixed on the substrate (108) by high-temperature resistant glue, and full-bridge connection method is used for temperature compensation.

7. The system of claim 1, wherein, The integrated man-machine interface (301) comprises: The three-dimensional temperature field is three-dimensionally visualized, a machine vision camera interface (302) is used to display the topography of the molten pool area, an infrared imaging interface (303) is used to display the global surface temperature distribution, and a thermocouple array three-dimensional temperature distribution interface (304) is used to three-dimensionally visualize the multi-point temperature data.

8. A method for an online monitoring and process optimization system for electron beam additive manufacturing, characterized in that The method is to construct and run a data-driven three-dimensional thermal perception model, and the model construction step comprises: fusing the data acquired by an infrared thermal imager module (103), a thermocouple array (109) and a machine vision camera module (104) to reconstruct a three-dimensional temperature field of the additive manufacturing process, executing process optimization decisions, and the decision step comprises: based on the high-fidelity data output by the three-dimensional thermal perception model, evaluating temperature characteristics; when the characteristics deviate from the standard in the preset process parameter database, automatically calling a matching optimization strategy and generating a control signal for dynamically adjusting subsequent electron beam process parameters.

9. The method of claim 8, wherein, The step of reconstructing the internal three-dimensional temperature field specifically comprises: The surface temperature field acquired by the infrared thermal imager module (103) is used as the top boundary condition of the model, and the substrate temperature acquired by the thermocouple array (109) is used as the bottom boundary condition of the model.

10. The method of claim 8, wherein, The preheating stage at the beginning of the method further comprises: Based on the monitoring data of the thermocouple array (109), it is judged whether the substrate (108) reaches the preset uniformity and target temperature requirement.

Images