Measurement of radiant heat energy density in grid areas on the build surface
Optical sensing techniques for monitoring thermal energy density in additive manufacturing processes address the lack of non-destructive quality verification, enhancing part quality by detecting defects and optimizing parameters.
Patent Information
- Application Number
- JP2023208352
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2018-03-15
- Filing Date
- 2023-12-11
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2038-08-01
AI Technical Summary
Current additive manufacturing processes lack non-destructive methods to accurately verify the mechanical, geometric, and metallurgical properties of production parts, leading to potential part failures due to undetected defects.
Implement optical sensing techniques to monitor thermal energy density during the additive manufacturing process using optical sensors, such as pyrometers and photodiodes, to track the evolution of in-process physical phenomena, allowing for real-time detection of manufacturing defects and adjustments to process parameters.
Enables non-destructive quality assurance by identifying potential defects and optimizing process parameters, thereby improving the consistency and quality of manufactured parts.
Smart Images

Figure 0007576873000016 
Figure 0007576873000017 
Figure 0007576873000018
Abstract
Description
[Technical field]
[0001]
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application Nos. 62 / 540,016, filed August 1, 2017, 62 / 633,487, filed February 21, 2018, and 62 / 643,457, filed March 15, 2018, each entitled “Systems And Methods For Measuring Energy Input During An Additive Manufacturing Operation,” the disclosures of which are incorporated herein by reference in their entireties for all purposes. [Background technology]
[0002] Additive manufacturing, or the continuous assembly or construction of parts by a combination of material addition and applied energy, takes many forms and currently exists in many specific implementations and embodiments. Additive manufacturing can be performed using any of a number of different processes that involve the formation of three-dimensional parts of virtually any shape. The various processes generally involve the sintering, hardening, or melting of liquid, powder, or granular raw materials layer by layer using ultraviolet light, high power lasers, or electron beams, respectively. Unfortunately, there are limited established processes for determining the quality of parts produced in this manner. Traditional quality assurance testing generally involves post-process measurements of the mechanical, geometric, or metallurgical properties of the parts, which often result in the destruction of the parts. Destructive testing is an accepted method for verifying part quality because it allows for close examination of various internal features of the parts, but such testing is not applicable to production parts for obvious reasons. Therefore, a non-destructive and accurate method for verifying the mechanical, geometric, and metallurgical properties of production parts produced by additive manufacturing is desired. Summary of the Invention
[0003]
[0003] The described embodiments relate to additive manufacturing, which involves the use of an energy source in the form of a moving field of intense thermal energy. When this thermal energy causes physical melting of the added material, these processes are widely known as welding processes. In welding processes, material is added incrementally and continuously and melted by an energy source in a manner similar to fusion welding. Exemplary welding processes suitable for use with the described embodiments include processes using a scanning energy source with a powder bed, and wire-fed processes using an arc, laser or electron beam as the energy source.
[0004]
[0004] When the added material takes the form of layers of powder, after each incremental layer of powder material is added in succession to the part being built, the scanning energy source melts the incrementally added powder by welding areas of the powder layers together to create moving melt areas (hereinafter referred to as melt pools) so that, upon solidification, they become part of the previously added, melted and solidified layers in succession below the new layers to form the part being built. Because additive machining processes are time-consuming and can involve multiple passes of the melt pool, it can be difficult to avoid at least slight variations in the size and temperature of the melt pool when using the melt pool to solidify the part. The embodiments described herein reduce or minimize discontinuities caused by variations in the size and temperature of the melt pool. It is noted that due to the high speed movement of the heating elements and the complex patterns required to form the three-dimensional structures, the additive manufacturing process can be driven by one or more processors associated with a computer numerical control (CNC).
[0005]
[0005] The overall objective of the described embodiments is to apply optical sensing techniques, for example for quality estimation, process control, or both, to additive manufacturing processes. Optical sensors can be used to track the evolution of in-process physical phenomena by tracking the evolution of relevant in-process physical variables. As used herein, optical can include portions of the electromagnetic spectrum including near infrared (IR), visible, and near ultraviolet (UV). Generally, the optical spectrum is considered to be wavelengths from 380 nm to 780 nm. However, near UV and near IR can extend to wavelengths as short as 1 nm and as long as 3000 nm, respectively. Sensor readings collected from optical sensors can be used to determine process quality metrics (IPQMs). One such IPQM is thermal energy density (TED), which helps characterize the amount of energy applied to different regions of a part.
[0006]
[0006] TED is a metric that is sensitive to user-defined laser powder bed fusion process parameters, such as laser power, laser speed, and hatch spacing. This metric can be used for analysis using IPQM comparisons to a baseline data set. The resulting IPQM can be calculated for each scan and displayed graphically or in three dimensions using point clouds. Also, IPQM comparisons to a baseline data set that indicate manufacturing defects can be used to generate control signals for process parameters. In some embodiments, if detailed thermal analysis is desired, thermal energy density can be determined for individual portions of each scan. In some embodiments, thermal energy data from multiple scans can be divided into separate grid regions of a grid, with each grid region reflecting the total amount of energy received at each grid region for one layer or a predefined number of layers.
[0007]
[0007] An additive manufacturing method is disclosed that includes: generating multiple scans of an energy source across a build surface; measuring an amount of energy radiated from the build surface during each of the multiple scans using an optical sensor monitoring the build surface; determining an area of the build surface traversed during the multiple scans; determining a thermal energy density of the area of the build surface traversed by the multiple scans based on the amount of energy radiated and the area of the build surface traversed by the multiple scans; mapping the thermal energy density to one or more locations on the build surface; determining that the thermal energy density is characterized by a density outside a range of density values; and then adjusting a subsequent scan of the energy source across or proximate to the one or more locations on the build surface.
[0008]
[0008] An additive manufacturing method is disclosed that includes the steps of: generating a scan of an energy source across a build surface; measuring the amount of energy emitted from the powder bed during the scan using an optical sensor monitoring the powder bed; determining an area associated with the scan; determining a thermal energy density in the area of the scan based on the amount of energy emitted and the area of the scan; determining that the thermal energy density is characterized by a density outside a range of density values; and then adjusting a subsequent scan of the energy source across the build surface.
[0009]
[0009] An additive manufacturing method is disclosed that includes: performing an additive manufacturing operation using an energy source; receiving sensor data associated with a photodiode during a scan of the energy source across a powder bed; receiving drive signal data indicating that the energy source has been powered on; using the energy source drive signal data to identify sensor data collected when the energy source was powered on; dividing the sensor data into a plurality of sample sections, each sample section corresponding to a portion of the scan; determining a thermal energy density of each of the plurality of sample sections; and identifying one or more portions of a part that are most likely to contain a manufacturing defect based on the thermal energy density of each of the plurality of sample sections.
[0010]
[0010] An additive manufacturing method is disclosed that includes the steps of: generating multiple scans of an energy source across a build surface; determining a grid region including the multiple scans, the grid region characterized by a grid area; using an optical sensor to generate sensor readings during each of the multiple scans; using the sensor readings to determine a total amount of energy radiated from the build surface during the multiple scans; calculating a thermal energy density associated with the grid region based on the total amount of energy radiated and the grid area; determining that the thermal energy density associated with the grid region is characterized by a thermal energy density outside a range of thermal energy density values; and then adjusting the output of the energy source.
[0011]
[0011] An additive manufacturing method is disclosed that includes the steps of: defining a portion of a build surface as a grid including a plurality of grid areas, each having a grid area area; generating a plurality of scans of an energy source across the build surface; generating sensor readings during each of the plurality of scans using an optical sensor; for each of the plurality of scans, mapping a respective portion of the plurality of sensor readings to a respective one of the plurality of grid areas; for each of the plurality of grid areas: summing the sensor readings mapped to each grid area; calculating a grid-based thermal energy density based on the summed sensor readings and the grid area area; determining that the grid-based thermal energy density associated with one or more of the plurality of grid areas is characterized by a thermal energy density outside a range of thermal energy density values; and then adjusting the output of the energy source.
[0012]
[0012] The present disclosure will be readily understood by the following detailed description taken in conjunction with the accompanying drawings, in which like reference numerals indicate like structural elements, and in which: [Brief description of the drawings]
[0013] [Figure 1A] FIG. 1 is a schematic diagram of an optical sensing device for use in an additive manufacturing system with an energy source, which in this particular example is interpreted as a laser beam. [Figure 1B] FIG. 1 is a schematic diagram of an optical sensing device for use in an additive manufacturing system with an energy source, taken in this particular example to be an electron beam. [Diagram 2] FIG. 1 illustrates a sample scan pattern used in an additive manufacturing process. [Diagram 3] 1 is a flow chart illustrating a method for identifying portions of a part that are most likely to contain manufacturing defects. [Figure 4A]FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4B] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4C] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4D] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4E] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4F] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4G] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 4H] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Diagram 5] 1 is a flowchart detailing a method for completing an IPQM assessment using scanlet data separation. [Figure 6A] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 6B] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 6C]FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 6D] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 6E] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 6F] FIG. 1 illustrates data relating to a step-by-step process for using thermal energy density to identify portions of parts most likely to contain manufacturing defects. [Figure 7A] FIG. 13 shows test results comparing IPQM metrics with post-process metallography. [Figure 7B] FIG. 13 shows test results comparing IPQM metrics with post-process metallography. [Figure 7C] FIG. 13 shows test results comparing IPQM metrics with post-process metallography. [Figure 8] FIG. 13 illustrates an alternative process by which data recorded by an optical sensor, such as a non-imaging photodetector, can be processed to characterize an additive manufacturing build process. [Figure 9A] 1 is a visual depiction of how multiple scans can contribute to the power introduced in each grid area. [Figure 9B] 1 is a visual depiction of how multiple scans can contribute to the power introduced in each grid area. [Figure 9C] 1 is a visual depiction of how multiple scans can contribute to the power introduced in each grid area. [Figure 9D] 1 is a visual depiction of how multiple scans can contribute to the power introduced in each grid area. [Figure 10A] FIG. 1 illustrates an exemplary turbine blade suitable for use in the described embodiments. [Figure 10B] FIG. 10 illustrates an exemplary manufacturing configuration in which 25 turbine blades can be manufactured simultaneously on a build surface 1006. [Figure 10C] 10C is a cross-sectional view of different layers of the configuration depicted in FIG. 10B. [Figure 10D] FIG. 10C is another cross-sectional view of different layers of the configuration depicted in FIG. 10B. [Figure 11A] FIG. 2 is a cross-sectional view of a base portion of a turbine blade. [Figure 11B] FIG. 2 is another cross-sectional view of a base portion of a turbine blade. [Figure 11C] 1 is a photograph showing the difference in surface consistency between two different base portions. [Figure 12] FIG. 1 illustrates thermal energy density of a part associated with several different builds. [Figure 13] FIG. 1 illustrates an example of how thermal energy density can be used to control the operation of a component using in-situ measurements. [Figure 14A] FIG. 1 illustrates an example of how thermal energy density can be used to control the operation of a component using in-situ measurements. [Figure 14B] FIG. 1 illustrates an example of how thermal energy density can be used to control the operation of a component using in-situ measurements. [Figure 14C] 11 is another power density graph highlighting various physical effects resulting from setting the energy source too far from the process window. [Figure 14D] FIG. 13 illustrates how the size and shape of the melt pool can vary according to laser power and scan speed settings. [Figure 15A] FIG. 1 illustrates how grids can be dynamically generated to characterize and control additive manufacturing operations. [Figure 15B] FIG. 1 illustrates how grids can be dynamically generated to characterize and control additive manufacturing operations. [Figure 15C]FIG. 1 illustrates how grids can be dynamically generated to characterize and control additive manufacturing operations. [Figure 15D] FIG. 1 illustrates how grids can be dynamically generated to characterize and control additive manufacturing operations. [Figure 15E] FIG. 1 illustrates how grids can be dynamically generated to characterize and control additive manufacturing operations. [Figure 15F] FIG. 1 illustrates how grids can be dynamically generated to characterize and control additive manufacturing operations. [Figure 16] FIG. 16 illustrates an example control loop 1600 for establishing and maintaining feedback control of an additive manufacturing operation. [Figure 17A] FIG. 1 illustrates normal distribution of powder across a build plate. [Figure 17B] FIG. 13 illustrates how the thickness of the resulting powder layer can vary when an insufficient amount of powder is collected and spread across the build plate by the recoater arm. [Figure 17C] A black and white photo of the build plate where a short feed of powder only partially covers the nine workpieces placed on the build plate. [Figure 17D] FIG. 13 illustrates how the detected thermal energy density varies substantially when the energy source scans across all nine workpieces using the same input parameters. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0014]
[0038] FIG. 1A illustrates an embodiment of an additive manufacturing system that uses one or more optical sensing devices to determine thermal energy density. Thermal energy density is sensitive to changes in process parameters, such as energy source power, energy source speed, and hatch spacing. The additive manufacturing system of FIG. 1A uses a laser 100 as an energy source. The laser 100 emits a laser beam 101 that passes through a partially reflective mirror 102 and into a scanning and focusing system 103, which then projects the beam onto a small area 104 on a work platform 105. In some embodiments, the work platform is a powder bed. Due to the high material temperature, optical energy 106 is emitted from the small area 104.
[0015]
[0039] In some embodiments, the scanning and focusing system 103 may be configured to collect a portion of the optical energy 106 emitted from the beam interaction region 104. The partially reflective mirror 102 may reflect the optical energy 106 as shown by an optical signal 107. The optical signal 107 may be interrogated by multiple on-axis optical sensors 109, each receiving a portion of the optical signal 107 via a series of additional partially reflective mirrors 108. It should be noted that in some embodiments, the additive manufacturing system may include only one on-axis optical sensor 109 with a fully reflective mirror 108.
[0016]
[0040] It should be noted that the collected optical signal 107 may not have the same spectral content as the optical energy 106 emitted from the beam interaction region 104 due to some attenuation of the signal 107 after passing through multiple optical elements such as the partially reflective mirror 102, the scanning and focusing system 103, and a series of additional partially reflective mirrors 108. Each of these optical elements has its own transmission and absorption characteristics, which may result in varying amounts of attenuation and limit certain portions of the spectrum of the energy emitted from the beam interaction region 104. The data generated by the on-axis optical sensor 109 may correspond to the amount of energy imparted to the work platform.
[0017]
[0041] Examples of on-axis optical sensors 109 include, but are not limited to, light to electrical signal converters (i.e., photodetectors) such as pyrometers and photodiodes. Optical sensors can also include spectrometers, and slow or high speed cameras operating in the visible, ultraviolet, or infrared frequency spectrum. The on-axis optical sensor 109 is in a reference frame that moves with the beam, i.e., it sees all areas that the laser beam contacts and can collect optical signals 107 from all areas of the work platform 105 that the laser beam 101 contacts as it scans across the work platform 105. The optical energy 106 collected by the scanning and focusing system 103 travels a path that is approximately parallel to the laser beam, so the sensor 109 can be considered an on-axis sensor.
[0018]
[0042] In some embodiments, the additive manufacturing system can include off-axis sensors 110 that are in a stationary frame of reference relative to the laser beam 101. These off-axis sensors 110 can have a given field of view 111 that can be very narrow or can encompass the entire work platform 105. Examples of these sensors include, but are not limited to, pyrometers, photodiodes, spectrometers, high or low speed cameras operating in the visible, ultraviolet, or IR spectral ranges, etc. Off-axis sensors 110 that are not aligned with the energy source are considered off-axis sensors. The off-axis sensors 110 can be sensors that combine a set of physical measurement modalities, such as a laser ultrasonic sensor that actively excites or "pings" the deposit with one laser beam and uses a laser interferometer to measure the resulting ultrasonic waves or "ringing" of the structure in order to measure or predict the mechanical properties or mechanical integrity of the deposit during the build. The laser ultrasonic sensor / interferometer system can be used to measure the elastic properties of the material, which can, for example, provide insight into the porosity and other material properties of the material. Additionally, defect formation that causes vibrations in the material can be measured using a laser ultrasonic / sensor interferometer system.
[0019]
[0043] Additionally, the recoater arm 112, which is the mechanical device that spreads the powder, may have contact sensors 113. These sensors may be accelerometers, vibration sensors, etc. Finally, there may be other types of sensors 114. These may include contact sensors such as thermocouples to measure the macro-thermal field, or acoustic emission sensors that can detect cracks and other metallurgical phenomena that occur in the deposit during the build. These contact sensors may be utilized during the powder addition process to characterize the operation of the recoater arm 112. Using the data collected by the on-axis optical sensor 109 and the off-axis sensor 110, process parameters related to the recoater arm 112 can be detected. Thus, non-uniformities in the surface of the spread powder can be detected and addressed by the system. Rough surfaces resulting from variations in the powder spread process can be characterized by the contact sensors 113 to predict possible problem areas or non-uniformities in the resulting part.
[0020]
[0044] In some embodiments, peaks in the powder spread may be fused by the laser beam 101 resulting in a subsequent layer of powder with a corresponding peak. At some point, the peaks may contact the recoater arm 112 and potentially damage the recoater arm 112, resulting in further powder non-uniformities. Thus, embodiments of the present invention may detect non-uniformities in the spread powder before they cause non-uniformities in the build area on the work platform 105. One of ordinary skill in the art would recognize many variations, modifications, and alternatives.
[0021]
[0045] In some embodiments, the on-axis optical sensor 109, the off-axis sensor 110, the contact sensor 113, and the other sensors 114 may be configured to generate in-process raw sensor data. In other embodiments, the on-axis optical sensor 109, the off-axis optical sensor 110, the contact sensor 113, and the other sensors 114 may be configured to process the data and generate reduced dimensional sensor data.
[0022]
[0046] In some embodiments, a computer 116 including a processor 118, a computer readable medium 120, and an I / O interface 122 is provided and coupled to appropriate system components of the additive manufacturing system to collect data from various sensors. The data received by the computer 116 can include in-process raw sensor data and / or reduced dimensional sensor data. The processor 118 can use the in-process raw sensor data and / or reduced dimensional sensor data to determine the power of the laser 100 and control information including coordinates related to the work platform 105. In other embodiments, the computer 116 including the processor 118, a computer readable medium 120, and an I / O interface 122 can provide control of various system components. The computer 116 can send, receive, and monitor control information related to the laser 100, the work platform 105, and the recoater arm 112 to control and adjust the respective process parameters of each component.
[0023]
[0047] The processor 118 can be used to perform calculations using data collected by the various sensors to generate in-process quality metrics. In some embodiments, data generated by the on-axis optical sensor 109 and / or the off-axis optical sensor 110 can be used to determine thermal energy density during the build process. Control information related to the movement of the energy source across the build surface can be received by the processor. The processor can then use the control information to correlate data from the on-axis optical sensor 109 and / or the off-axis optical sensor 110 with corresponding positions. This correlated data can be combined to calculate thermal energy density. In some embodiments, the thermal energy density and / or other metrics can be used by the processor 118 to generate control signals for process parameters, such as laser power, laser speed, hatch spacing, and other process parameters, in response to the thermal energy density or other metric being out of a desired range. In this way, issues that could otherwise ruin a production part can be ameliorated. In embodiments where multiple parts are produced at once, the process parameters can be quickly modified in response to a metric being out of a desired range to prevent adjacent parts from receiving too much or too little energy from the energy source.
[0024]
[0048] In some embodiments, the I / O interface 122 can be configured to transmit collected data to a remote location. The I / O interface can be configured to receive data from a remote location. The received data can include a baseline data set, historical data, post-process inspection data, and classification data. The remote computing system can use the data transmitted from the additive manufacturing system to calculate in-process quality metrics. The remote computing system can transmit information to the I / O interface 122 depending on the particular in-process quality metric.
[0025]
[0049] For an electron beam system, FIG. 1B illustrates a possible configuration and arrangement of sensors. An electron beam gun 150 generates an electron beam 151 that is focused by an electromagnetic focusing system 152 and then deflected by an electromagnetic deflection system 153, resulting in a finely focused and targeted electron beam 154. The electron beam 154 generates a hot beam material interaction zone 155 on a workpiece 156. Light energy 158 emanates from the workpiece 156 and may be isolated locally to the interaction region 155 or collected by a series of optical sensors 159, each with its own field of view 160 that may encompass the entire workpiece 156. In addition, the optical sensors 159 may have their own tracking and scanning systems that can track the electron beam 154 as it moves across the workpiece 156.
[0026]
[0050] Whether or not the sensor 159 has optical tracking, the sensor 159 can be composed of pyrometers, photodiodes, spectrometers, and high-speed or low-speed cameras operating in the visible, UV, or IR spectral regions. The sensor 159 may be a sensor that combines a set of physical measurement modalities, such as a laser ultrasonic sensor that actively excites or "pings" the deposit with one laser beam and measures the resulting ultrasonic waves or "ringing" of the structure using a laser interferometer to measure or predict the mechanical properties or mechanical integrity of the deposit during the build. In addition, there may be contact sensors 113 on the recoater arm. These sensors may be accelerometers, vibration sensors, etc. Finally, there may be other types of sensors 114. These may include contact sensors such as thermocouples to measure the macro-thermal field, and acoustic emission sensors that can detect cracks and other metallurgical phenomena occurring in the deposit during the build. In some embodiments, one or more thermocouples may be used to calibrate the temperature data collected by the sensor 159. It should be noted that the sensors described in connection with Figures 1A and 1B can be used in the described manner to characterize the performance of an additive manufacturing process that involves continuous material accumulation.
[0027]
[0051] FIG. 2 shows possible hatch patterns for scanning an energy source across a powder bed. At 200, an area of a workpiece is treated by an energy source scanning along a long path length with alternating directions. In this embodiment, a hatch interval 204 is shown between a first scan 206 and a second scan 208. At 202, an area of a workpiece is divided into smaller checkerboards 214 that can be scanned by a first scan 210 and a second scan 212 successively from left to right and top to bottom. In other embodiments, the scan order of the individual checkerboards can be randomized. Several hatch patterns can be utilized in connection with the additive manufacturing processes disclosed herein. Those skilled in the art will recognize many variations, modifications, and alternatives.
[0028]
[0052] FIG. 3 shows a flow chart illustrating an example process 300 for using data generated by an additive manufacturing system to determine thermal energy density and identify portions of a part that are most likely to contain manufacturing defects. Data generated by the on-axis optical sensor 109 and the off-axis optical sensor 110 can be used alone or in combination to determine thermal energy density. At 302, a raw photodiode data trace is received. The raw photodiode data trace may be generated, for example, using voltage data generated by the sensor in response to detection of emitted thermal energy. At 304, a particular scan i In some embodiments, individual photodiode data traces may be separated from the remaining sensor readings by reference to energy source drive signal data (drive signals involved in the operation and actuation of the energy source). i Raw photodiode data trace (hereafter referred to as pdon i In some embodiments, the area under pdon is determined. i may represent the integrated photodiode voltage. In some embodiments, pdoni scan i represents the average reading of the photodiode during the scan. i The part identified in 308 is the part's associated area A p These two values can be expressed as pdon i can be determined by correlating with the energy source position data. The process can calculate a total scan count at 310. At 312, the scan i The length L associated with i can be determined. i can be calculated using equation (1), where x1 i ,y1 i and x2 i ,y2 i scan i Represents the start and end positions of:
number
[0029]
[0053] 314, the total length Lsum of all scans used to manufacture the part p can be determined. p is the length L of each scan associated with the part i The proportional area of the scan, A, can be determined by summing the i It is possible to determine i can be calculated using equation (2):
number
[0030]
[0054] 316, the proportional thermal energy density (TED) of the i-th scan, TED i TED iis an example set of reduced order process features. TED is calculated using the raw photodiode data. From this raw sensor data, the TED calculation extracts the reduced order process features from the raw sensor data. TED i is sensitive to all user-defined laser powder bed fusion process parameters, e.g. laser power, laser speed, hatch spacing, etc. i can be calculated using equation (3):
number
[0031]
[0055] For purposes of this description, "reduced order" refers to one or more aspects of: data compression, i.e., less feature data compared to the raw data; data reduction, i.e., the systematic analysis of raw data to yield process metrics or other figures of merit; data aggregation, i.e., the clustering of data into distinct groups and a smaller set of variables that characterize the clustering as opposed to the raw data itself; data transformation, i.e., the mathematical manipulation of data to linearly or nonlinearly map the raw data into another variable space of lower dimensionality using a transformation law or algorithm; or other related techniques having the net effect of reducing data density, reducing the number of dimensions of data, reducing data size, transforming data into another reduced space, or all of these occurring simultaneously.
[0032]
[0056] TED i can be used for analysis during the comparison of process quality metrics (IPQM) with the baseline data set. The resulting IPQM can be calculated for each scan. At 318, the IPQM quality baseline data set and the calculated TED ican be compared. In areas of the part where the difference between the calculated TED and the baseline data set exceeds a threshold, those areas can be identified as likely to contain one or more defects and / or further processing can be performed on the areas in near real time to remedy defects caused by variations in the TED from the baseline data set. In some embodiments, the portions of the part that may contain defects can be identified using a classifier. The classifier can group results as either nominal or non-nominal and can be represented in graphical and / or text-based media. The classifier can use multiple classification methods including, but not limited to: statistical classification, both single and multivariate; heuristic-based classifiers; expert system-based classifiers; look-up table-based classifiers; classifiers simply based on upper or lower control limits; classifiers working in conjunction with one or more statistical distributions that can establish nominal vs. non-nominal thresholds based on confidence intervals and / or degree of freedom considerations; or other classification schemes, whether implicit or explicit, that can identify whether a set of feature data is nominal or non-nominal. For purposes of this description, "nominal" means a set of process results that were within predefined specifications, and post-process measurement attributes of parts thus produced fall into a regime of values considered acceptable, or other quantitative, semi-quantitative, objective, or subjective methodology for establishing "acceptable" components. Additional description of IPQM classifications is provided in U.S. Patent Application Serial No. 15 / 282,822, filed September 30, 2016, the disclosure of which is incorporated herein by reference in its entirety for all purposes.
[0033]
[0057] It should be understood that the particular steps illustrated in FIG. 3 provide a particular method of collecting data and determining thermal energy density in accordance with an embodiment of the present invention. Other sequences of steps may be performed in accordance with alternative embodiments. Also, the individual steps illustrated in FIG. 3 may include multiple sub-steps that may be performed in various sequences, as appropriate for the individual step. Moreover, additional steps may be added or existing steps may be deleted, depending on the particular application. Those skilled in the art will recognize many variations, modifications, and alternatives.
[0034]
[0058] 4A-4H show steps used in process 300 to determine the TED and identify any portions of the part that are likely to contain defects. FIG. 4A corresponds to step 302 and shows the raw photodiode signal 402 for a given scan length. The x-axis 450 shows time in seconds and the y-axis 460 shows the photodiode voltage. In some embodiments, optical measurements can be made alternatively or in addition by a pyrometer. The signal 402 is the raw voltage of the photodiode. The rise 404 and fall 406 of the photodiode signal 402 can be clearly seen, as well as the scattering and fluctuations 408 of the signal while the laser is on. The data is collected at a given number of samples / second. The fluctuations 408 of the photodiode signal 402 can be caused by fluctuations of the powder melting on the powder bed. For example, one of the small valleys in the photodiode signal 402 can be caused by energy absorbed by larger particles in the particle bed transitioning from a solid to a liquid state. In general, the number of data points in a given segment of the photodiode signal between the rising and falling events can be related to the scan period and the sampling rate.
[0035]
[0059] FIG. 4B shows the raw photodiode signal 402 and the laser drive signal 410. The laser drive signal 410 shown in FIG. 4B may be generated using energy source drive signal data, in this case the laser drive signal 410, or a command signal that instructs the laser to be on and off for a particular scan length. The photodiode signal 402 is superimposed on the laser drive signal 410. The rising edges 412 and falling edges 414 of the laser drive signal 410 correspond to the rising edges 404 and falling edges 406 of the photodiode signal 402. The data shown in FIG. 4B may be used in step 304 to identify the portion of the raw photodiode signal 402 that corresponds to the scan. In some embodiments, the laser drive signal 410 is about 0V when the laser is off and about 5V when the laser is on. Step 304 may be accomplished by isolating all data associated with the photodiode signal where the laser drive signal 410 exceeds a particular threshold, for example 4.5V, and excluding from the analysis all data where the laser is below this threshold.
[0036]
[0060] 4C illustrates one embodiment of step 306, which includes determining the area 416 under the raw photodiode signal 402. In some embodiments, the area under the curve may be calculated using equation (4):
number
[0037]
[0061] Figure 4D shows the location of the scan 420 relative to the part and the total scan count 424. Both values can be used to determine the TED corresponding to the scan location on the part. Figure 4E shows a rendered area of a part of interest 426. Figure 4E also shows multiple additional parts 428 and a witness coupon 430. All of the parts in Figure 4E are depicted placed on a powder bed 432.
[0038]
[0062] FIG. 4F shows traces associated with a portion of the photodiode data and laser drive signal data corresponding to four scans that can be used with the remainder of the photodiode data to determine a total sample count 434. The total sample count is used to determine the total scan length of the part, LSum p can be calculated. The total sample count is determined by summing the laser on periods 436. In some embodiments, the sum of the laser on periods and the average speed of the scanning energy source during the laser on periods can be used to determine the total scan length.
[0039]
[0063] After the scan data is collected, the TED for each layer can be calculated from the TED associated with each laser scan and displayed in graph 440 shown in FIG. 4G. Graph 440 shows TED values that fall within nominal region 442 and non-nominal region 444. The TED regions are divided by baseline threshold 438. In this way, layers of the part that are likely to contain defects can be easily identified. Further analysis can then be focused on layers that have non-nominal TED values.
[0040]
[0064] 4H illustrates how the TED values for each scan can be displayed in three dimensions using point cloud 446. Point cloud 446 shows the location in three-dimensional space of the TED values from nominal regions 442 and non-nominal regions 444 by displaying the non-nominal values as a different color or intensity than the nominal values. The non-nominal values indicate portions of the part that are most likely to contain manufacturing defects, such as porosity due to keyhole formation or voids due to lack of fusion. In some embodiments, the system can generate and send control signals to alter one or more process parameters based on the TED.
[0041]
[0065] FIG. 5 shows a flow chart illustrating an example process 500 for determining thermal energy density using data generated by an additive manufacturing system to identify portions of a part that are most likely to contain manufacturing defects. Data generated by the on-axis optical sensor 109 and the off-axis optical sensor 110 can be used alone or in combination to determine thermal energy density. At 502, photodiode time series data can be collected. The photodiode time series data can be generated, for example, using voltage data associated with the sensors. At 504, laser drive time series data is collected. The laser drive time data may be associated with additional process parameters such as laser power, laser speed, hatch spacing, xy position, etc. A process at 506 can slice the photodiode time series data by dropping portions of the photodiode time series data that correspond to portions of the laser drive time series data that indicate a laser off condition. In some embodiments, the laser drive signal is about 0V when the laser is off and about 5V when the laser is on. A process at 506 can isolate all data where the laser drive signal is above a certain threshold, for example 4.5V, and exclude all data where the laser is below this threshold from the analysis. In some embodiments, the sample series data may include a photodiode signal that periodically drops to about 0.2V because the laser is just turning on and there are times when the laser is heating the material.
[0042]
[0066] The process at 506 outputs only the laser-on photodiode data 508. The process at 510 can use the laser-on photodiode data to convert the time series data to sample series data. The process at 510 divides the laser-on photodiode data into "N" sample sections. The use of 20 sample sections is intended to provide an example of one embodiment of the present invention. Any number of sample sections of varying accuracy / resolution can be used. In some embodiments, a set of sample sections can be referred to as a scanlet 520 since it typically requires multiple scanlets 520 to compose a single scan. The process at 512 can count (516) the number of samples. The process at 514 can render the area of the lasered part. In some embodiments, the number of pixels in the display associated with the lasered part can be used to determine the lasered part area 518. In other embodiments, the number of scans and data associated with the process parameters can be used to calculate the area. At 522, the process normalizes the scanlet data using the total sample count, the lasered part area 518, and the scanlet data 520. In the illustrated embodiment, the scanlet metric data 524 is the thermal energy density of the portion of the part associated with each scanlet. In some embodiments, the scan data may also be categorized by scan type. For example, an additive manufacturing machine may utilize scans with different characteristics. In particular, a contour scan, or a scan designed to finish the outer surface of a part, may be significantly more intense than a scan designed to sinter an interior region of the part. Thus, separating the data by scan type may provide more consistent results. In some embodiments, the identification of the scan type may be based on scan intensity, scan duration, and / or scan location. In some embodiments, the scan type may be identified by correlating the detected scan with scans dictated by a scan plan associated with the part under build.
[0043]
[0067] Process 528 then receives the baseline scanlet metric data and the thermal energy density and outputs an IPQM quality assessment 530. The IPQM quality assessment 530 can be used to identify the portions of the part that are most likely to contain manufacturing defects. Process 528 can include a classifier as previously described herein. In addition to the methods and systems described above, process 528 can use Mahalanobis distance to compare candidate data, such as scanlet metric data 524, to baseline scanlet metric data. In some embodiments, the Mahalanobis distance for each scanlet can be calculated using the baseline scanlet metric data. While the embodiment disclosed in connection with FIG. 5 discussed the use of a laser as an energy source, it will be apparent to one of ordinary skill in the art that many modifications and variations are possible in light of the above teachings, for example, the laser can be replaced with an electron beam or other suitable energy source.
[0044]
[0068] It should be understood that the specific steps illustrated in FIG. 5 provide a particular method for determining thermal energy density and identifying portions of a component that are most likely to contain manufacturing defects, according to another embodiment of the present invention. Other sequences of steps may be performed according to alternative embodiments. For example, alternative embodiments of the present invention may perform the steps listed above in a different order. Also, individual steps illustrated in FIG. 5 may include multiple sub-steps that may be performed in various orders, as appropriate for the individual step. Furthermore, additional steps may be added or removed depending on the particular application. Those skilled in the art will recognize many variations, modifications, and alternatives.
[0045]
[0069] FIG. 6A illustrates photodiode time series data 602. Photodiode time series data may be collected from various on-axis or off-axis sensors as shown in FIG. 1 and FIG. 2. The x-axis 602 illustrates time in seconds and the y-axis 604 illustrates the voltage generated by the sensor. The voltage generated by the sensor is associated with energy emitted from the build surface that may impinge on one or more sensors. Samples 606 are shown on the traces of the photodiode time series data 602. FIG. 4B illustrates the process at 506 where the photodiode data is associated with the laser drive signal.
[0046]
[0070] Figure 6B shows the laser on photodiode data. The x-axis 608 represents the number of samples and the y-axis 610 represents the voltage of the raw sensor data. The voltage drop 612 is included in the analysis because although the voltage is significantly lower, the laser still actively contributes to heating the material.
[0047]
[0071] FIG. 6C shows the sample sequence data of the laser on the photodiode described in relation to step 510. The 20 sample section 620 can have any size. 20 samples corresponds to a laser travel distance of about 400 μm with a laser travel speed of 1000 mm / s. The noise of the XY signal itself is about 150 μm. In some embodiments, for sections of less than 20 samples, for example a 2 sample section, the measured distance and the noise may be in such a ratio that the location of the point cannot be determined with confidence. In some embodiments, for a spatial resolution of less than 1 mm, a limit of 50 samples can be used. Thus, the number of samples into which the data is segmented should be in the range of 20≦N≦50 for 50 KHz data with a laser travel speed of 1000 mm / s. Many modifications and variations are possible in light of the above teachings, as will be apparent to those skilled in the art.
[0048]
[0072] 6D corresponds to process 522 and illustrates an embodiment in which an average value 618 for each scanlet is determined. In some embodiments, inputs to process 522 include total sample count 516, laser hit part area 518, and scanlet data 520. Using these inputs, the averages can be used to determine the area under the curve (AUC) as shown in equation (5):
number
[0049]
[0073] FIG. 6E shows the laser-hit part area A for each scanlet 622. i , and the laser hit part area for all scans 624. In addition to the area, the length of the scan L i and the L of the entire part i Sum of LSum p It is possible to calculate L i can be calculated using equation (6):
number
[0050]
[0074] 6F illustrates a laser-imposed part rendering area 626 associated with a layer in the build plane. i , area of part A p , the scan length L i , and the total length LSum p Once determined, TED can be calculated using equation (7):
number
[0051]
[0075] 7A-7C show post-process porosity measurements and corresponding normalized in-process TED measurements. These figures show that in-process TED measurements can be an accurate IPQM predictor of porosity and other manufacturing defects. FIG. 7A shows a comparison of TED metric data to a baseline data set. The plots show the values of each photodiode of the IPQM metrics, both individually and in combination. On-axis photodiode data 702 can be obtained from a sensor aligned with the energy source. Off-axis photodiode data 704 can be collected by a sensor not aligned with the energy source. A combination of on-axis and off-axis photodiode data 706 provides the highest sensitivity to changes in process parameters. The x-axis 708 shows the build surface layer of the part, and the y-axis 710 shows the Mahalanobis distance between the calculated TED and baseline metrics.
[0052]
[0076] The Mahalanobis distance can be used to normalize the TED data. The Mahalanobis distance indicates how many standard deviations each TED measurement is from a normal distribution of TED measurements. In this case, the Mahalanobis distance indicates how many standard deviations each TED measurement is from the average TED measurement collected during the build of control layers 526-600. The bottom chart in FIG. 7A also shows how TED varies with global energy density (GED) and porosity. Notably, this set of experiments shows that TED can be configured to predict the porosity of a part without destructive testing.
[0053]
[0077] In some embodiments, the performance of the additive manufacturing device can be further validated by comparing quantitative metallographic (e.g., size and shape of pores or intermetallic particles) and / or mechanical property (e.g., strength, toughness or fatigue) characteristics of metal parts produced during test runs. Generally, the presence of unmelted metal powder particles in a test part indicates that not enough energy has been applied, while test parts that receive too much energy tend to form internal voids that can compromise the integrity of the fabricated part. Porosity 714 can be representative of these defects.
[0054]
[0078] In some embodiments, the nominal value used to generate FIG. 7A is obtained from a preceding test. In some embodiments, the nominal value can also be obtained from a subsequent test, since no calculations need to be performed during additive manufacturing operations. For example, if one is trying to compare the performance of two additive manufacturing machines, a nominal value can be identified by performing tests using a first additive manufacturing machine. The performance of the second additive manufacturing machine can then be compared to the nominal value determined by the first additive manufacturing machine. In some embodiments, where the performance of the two additive manufacturing machines is within a predetermined threshold of five standard deviations, comparable performance can be expected from the two machines. In some embodiments, the predetermined threshold can be a 95% statistical confidence level derived from an inverse chi-square distribution. This type of testing methodology can also be utilized in identifying changes in performance over time. For example, after a machine is calibrated, the results of a test pattern can be recorded. After a certain number of manufacturing operations are performed by the machine, the additive manufacturing machine can be run again. The initial test pattern run after calibration can be used as a baseline to identify changes in the performance of the additive manufacturing machine over time. In some embodiments, the settings of the additive manufacturing device may be adjusted to return the additive manufacturing device to its post-calibration performance.
[0055]
[0079] FIG. 7B shows the post-process metallography of a part built using an additive manufacturing process. FIG. 7B shows part 718 and a corresponding cross-section 720 of the part. Sections 1-11 correspond to sections 1-11 of FIG. 7A. The changes in process parameters, and the resulting changes in porosity, can be seen in the cross-section 720 of the part. Notably, sections 2 and 3 have the highest porosity of 3.38% and 1.62%, respectively. Higher porosity is indicated in the cross-section by an increasing number of defect marks 722 in the sample part.
[0056]
[0080] Figure 7C shows the IPQM results on the corresponding cross sections determined during metallography. Each cross section has a J / mm 2 Included are the energy density 724 and porosity 714 in units. The sample with the greatest number of defect marks 722 corresponds to the TED measurement with the most normalized distance from the baseline. The plot shows that a low normalized distance can predict a high density and low porosity metallography, while a high normalized Malahabonis distance is highly correlated with high porosity and poor metallography. For example, a low power setting used in producing layers around layer 200 results in high porosity 714 and numerous defect marks 722. In contrast, using the middle of the load setting at or around the 600th layer results in no discernible defect marks 722 and the smallest porosity value recorded is 0.06%.
[0057]
[0081] FIG. 8 illustrates an alternative process in which data recorded by an optical sensor, such as a non-imaging photodetector, can be processed to characterize an additive manufacturing build process. At 802, raw sensor data is received that can include both build surface intensity data and an energy source drive signal that are correlated to each other. At 804, individual scans can be identified and located within the build surface by comparing the drive signal to the build surface intensity data. Generally, the energy source drive signal provides at least a start and end location from which a range that a scan can extend can be determined. At 806, the raw sensor data related to the intensity or power of each scan can be binned into corresponding X&Y grid areas. In some embodiments, the raw intensity or power data can be converted to energy units by correlating the dwell time of each scan in a particular grid area. In some embodiments, each grid area can represent one pixel of an optical sensor monitoring the build surface. It should be noted that different coordinate systems, such as polar coordinates, can be used to store the grid coordinates, and the storage of coordinates should not be limited to Cartesian coordinates. In some embodiments, different scan types can be binned separately such that analysis can be performed only on certain scan types. For example, an operator may want to focus on contour scans as these types of scans are most likely to contain unwanted variations. At 808, the energy input at each grid region may be summed so that the total amount of energy received at each grid region can be determined using equation (8).
number
[0058]
[0082] This summation can be performed just prior to adding a new layer of powder to the build surface, or the summation can be delayed until a predetermined number of layers of powder have been deposited. For example, the summation can be performed only after depositing and fusing five or ten different layers of powder during an additive manufacturing process. In some embodiments, a sintered layer of powder can add approximately 40 microns to the thickness of the part, although this thickness will vary depending on the type of powder used and the thickness of the powder layers.
[0059]
[0083] At 810, the standard deviation of the samples detected and associated with each grid region is determined. This allows for identification of grid regions with small or large variations in power readings. Variations in the standard deviation may indicate a problem with the sensor performance or where one or more scans are missing or have power levels well outside normal operating parameters. The standard deviation may be determined using equation (9).
number
[0060]
[0084] At 812, the total energy density received at each grid area may be determined by dividing the power reading by the total area of the grid area. In some embodiments, the grid area may have a square shape with a length of about 250 microns. The energy density of each grid area may be determined using equation (10).
number
[0061]
[0085] If multiple parts are being built, different grid regions can be associated with different parts at 814. In some embodiments, the system can include stored part boundaries that can be used to quickly associate each grid region and its associated energy density to each part using the coordinates of the grid regions and the boundaries associated with each part.
[0062]
[0086] At 816, the area of each layer of the part may be determined. If the layer contains voids or serves to define an internal cavity, a large portion of the layer may not receive energy. Thus, the affected area may be calculated by summing only those grid areas identified as receiving some energy from the energy source. At 818, the total amount of power received by the grid areas in the portion of the layer associated with the part may be summed and divided by the affected area to determine the energy density for that layer of the part. The area and energy density may be calculated using equations (11) and (12).
number
number
[0063]
[0087] At 820, the energy density of each layer can be summed to obtain a metric that indicates the total amount of energy received by the part. The overall energy density of the part can then be compared to the energy density of other similar parts on the build surface. At 822, the total energy from each part is summed. This allows a high level of comparison to be made between different builds. Comparing builds can help identify systematic differences such as powder variations or changes in overall power output. Finally, at 824, the summed energy values can be compared to other layers, parts, or build surfaces to determine the quality of the other layers, parts, or build surfaces.
[0064]
[0088] It should be understood that the particular steps illustrated in FIG. 8 provide a particular way of characterizing an additive manufacturing build process according to another embodiment of the present invention. Other orders of steps may be performed in accordance with alternative embodiments. For example, alternative embodiments of the present invention may perform the above steps in a different order. Also, individual steps illustrated in FIG. 8 may include multiple sub-steps that may be performed in various orders, as appropriate for the individual step. Furthermore, additional steps may be added or removed depending on the particular application. Those skilled in the art will recognize many variations, modifications, and alternatives.
[0065]
[0089] 9A-9D show visual depictions of how multiple scans can contribute to the power introduced at individual grid regions. FIG. 9A shows a grid pattern made up of multiple grid regions 902 distributed across a portion of a part being built by an additive manufacturing system. FIG. 9A also shows a first pattern of energy scans 904 extending diagonally across the grid regions 902. The first pattern of energy scans 902 can be applied by a laser or other intense source of thermal energy scanning across the grid 904. FIG. 9B shows how the energy introduced across the part is represented by a single grayscale color in each of the grid regions 902 representing the amount of energy received, with darker shades of gray corresponding to larger amounts of energy. Note that in some embodiments, the size of the grid regions 902 can be reduced to obtain higher resolution data. Alternatively, the size of the grid regions 902 can be increased to reduce memory and processing power usage.
[0066]
[0090] 9C shows a second pattern of energy scans 906 that overlaps at least a portion of the energy scans of the first pattern of energy scans. As explained in the text accompanying FIG. 8, when the first and second patterns of energy scans overlap, the grid area is displayed in a darker shade to show how the energy of both scans has increased the amount of energy received in the overlapping scan pattern. Obviously, this method is not limited to two overlapping scans, but can include many other additional scans that are added together to fully represent the energy received in each grid area.
[0067]
[0091] FIG 10A illustrates an exemplary turbine blade 1000 suitable for use with the described embodiments. The turbine blade 1000 includes multiple different surfaces and includes multiple different features that require many different types of complex scans to generate. In particular, the turbine blade 1000 includes a hollow blade portion 1002 and a tapered base portion 1004. FIG 10B illustrates an exemplary manufacturing setup in which 25 turbine blades 1000 can be manufactured simultaneously on a build surface 1006.
[0068]
[0092] 10C-10D show different cross-sectional views of different layers of the configuration shown in FIG. 10B with a gridded TED-based visualization layer. FIG. 10C shows layer 14 of turbine blade 1000, where the TED-based visualization layer shows how the lower ends of selected ones of the base portions 1004 may define multiple voids in turbine blades 1000-1, 1000-2, and 1000-3. Because the energy density data is associated with separate grid regions, these voids that may otherwise be completely hidden within the turbine blade are clearly visible in this gridded TED-based visualization. FIG. 10D shows how the upper ends of the base portions 1004 may define multiple hidden voids in turbine blades 1000-1, 1000-2, and 1000-3 that are clearly discernible from the gridded TED-based visualization layer shown in FIG. 10D.
[0069]
[0093] 11A-11B show cross-sectional views of base portions 1004-1 and 1004-2 of two different turbine blades 1000. FIG. 11A shows base portion 1004-1 manufactured using nominal manufacturing settings. The exterior surfaces 1102 and 1104 of base portion 1004-1 receive substantially more energy than the interior 1106 of base portion 1004-1. The increased energy input to the exterior surfaces provides a more uniform hardened surface, resulting in an annealing effect along the exterior surfaces 1102 and 1104. This additional energy may be introduced using a high energy contour scan targeting the exterior surfaces 1102 and 1104. FIG. 11B shows base portion 1004-2 produced with the same manufacturing settings as 1004-2, except omitting the contour scan. All scans utilized in the manufacturing operations to generate base portion 1004-2 were also included in the generation of base portion 1004-1, but by summing the energy density inputs of all scans covering each grid area, an operator can clearly see the differences between base portions 1004-1 and 1004-2.
[0070]
[0094] FIG. 11C illustrates the difference in surface consistency between base portions 1004-1 and 1004-2. Clearly, omitting the contour scan of base portion 1004-2 has a substantial impact on the overall exterior surface quality. The exterior surface of base portion 1004-1 is smoother, more consistent and less porous. The annealing effect on base portion 1004-1 should also impart substantially greater strength than base portion 1004-2.
[0071]
[0095] FIG. 12 shows the thermal energy density of parts associated with several different builds. Builds A-G each contain thermal emission data for approximately 50 different parts (represented by individual circles 1202) built during the same additive manufacturing operation. This figure illustrates how thermal emission data may be used to track variations between different builds. For example, lots A, B, and C all have similar TED distributions, but builds D, E, and F within tolerance have consistently lower thermal emission data. In some cases, these types of variations may be due to changes in powder lots. In this way, thermal emission data can be used to track systematic errors that may be negatively impacting the overall output quality. Note that while this chart shows average TED values based on gridded TED, similar charts can be created for scan-based TED methodologies.
[0072]
[0096] 13-14B show an example of how thermal energy density can be used to control part operation using in-situ measurements. FIG. 13 shows how parts produced using different combinations of energy source power and scan speed can be produced and then destructively analyzed to determine the resulting part density in grams per cubic centimeter as shown. In this trial, a part associated with part density 1302 and having a part density of 4.37 g / cc was produced using the manufacturer's recommended scan and laser power settings. Based on the resulting density readings, the location of dashed line 1304 can be determined. Line 1304 shows where the resulting drop in the amount of energy input caused some of the powder to not heat up enough to fuse together, causing the part density to fall below a threshold density level. Also, if too much energy is added to the system, the density of the part can be reduced, resulting in the formation of keyholes in the part as the powder vaporizes instead of melting during the manufacturing operation. Dashed line 1306 can be determined experimentally from the density data and is identified in this example by the part density being as low as 4.33 g / cc. Line 1308 represents the optimum energy density contour where energy density and part geometry remain substantially constant. Density tests show how the average density of parts made using settings distributed along line 1308 remains relatively consistent.
[0073]
[0097] 14A-14B show thermal energy density contours overlaid on a portion of a power density graph to illustrate how thermal density measurements collected during an additive manufacturing operation vary depending on the settings used by the energy source. As shown in FIGS. 14A-14B, darker shading indicates higher thermal energy density and lighter shading indicates lower thermal energy density. From part density testing and thermal energy density contours, control limits for a particular part can be determined. In this case, the control limits shown by ellipse 1402 have been determined, and the power and scan speed parameters can be varied from the settings used to produce part 1302 up to 3σ along line 1308 and 1σ along a line perpendicular to line 1308. In some embodiments, allowable variations in power and speed allow in-process adjustments to be made to maintain the desired thermal energy density during a production run. Ellipse 1404 shows the overall process window that can accommodate further variations outside the control limits. In some embodiments, this process window can be used to identify variations where the resulting part can be validated using only in-process data. It should be noted that while the depicted control and process windows are shown as ovals, other process window shapes and sizes are possible and may be more complex depending, for example, on the part geometry and material eccentricity. During manufacturing operations, optical sensor readings measuring thermal radiation can be used to determine thermal energy density in situ. If the thermal energy density falls outside of expected ranges when maintaining laser power and scan speed within the control limits shown by oval 1402, a warning can be issued that the portion of the part having an abnormal thermal energy density reading may be defective.
[0074]
[0098] In some embodiments, the process window can be incorporated into a modeling and simulation program that models one or more optical sensors that collect sensor readings used to determine thermal energy density. Once the modeling and simulation system has been iterated to a first order approximation of the instruction set for the workpiece, the expected thermal energy density can be output to an additive manufacturing machine for further testing. If the additive manufacturing machine includes optical sensors and computing equipment configured to measure thermal energy density, the modeled thermal energy density data can be used for further testing and validation on the additive manufacturing machine. A comparison of the modeled and measured thermal energy densities can be used to ascertain how well the performance of the instruction set matches the expected results in situ.
[0075]
[0099] FIG. 14C is another power density graph highlighting various physical effects resulting from the energy source settings being too far from the process window 1404. For example, the power density graph shows that adding a large amount of laser power at a low scan speed results in the formation of a keyhole in the workpiece. The formation of the keyhole occurs because some of the powder material evaporates from receiving too much energy. In addition, low power combined with a high scan speed can cause the powder to not fuse. Finally, high power levels combined with a high scan speed can result in ball-up of molten metal during the build operation. It should be noted that varying the thickness of the deposited powder layer can result in a shift in the line separating the conduction mode zone from the keyhole formation zone and a lack of a fusion zone. For example, increasing the thickness of the powder layer has the effect of increasing the slope of the line separating the keyhole from the conduction mode zone and a lack of a fusion zone, since thicker layers generally require more energy to undergo liquefaction.
[0076]
[0100] FIG. 14C also illustrates how power and scan speed settings corresponding to the conduction mode zone generally do not result in any of the critical defects discussed above, but by maintaining the laser settings corresponding to values within the process window 1404, the resulting grain structure and / or density of the part can be optimized. Another benefit of maintaining the energy source settings within the process window 1404 is that these settings should maintain the thermal energy density readings within a narrow range. If the thermal energy density values are outside of the predetermined range during a manufacturing run, it may indicate a problem with the manufacturing process. These problems can be addressed by moving the settings closer to the central region of the process window or by updating the process window for subsequent parts. In some embodiments, the manufacturer may discover that the failure was caused by a defective powder or other rare aberrations that should not be accounted for in the subsequent manufacturing process. It should be understood that the depicted process window is not the same for all parts of the part and may vary significantly depending on which area or specific layer of the part was being worked on at a particular time. The size and / or shape of the process window 1404 may also vary according to other factors such as hatch spacing, scan length and scan direction in some embodiments.
[0077]
[0101] FIG. 14D illustrates how the size and shape of the melt pool can vary according to the laser power and scan speed settings. Exemplary melt pools 1406-1418 show exemplary melt pool sizes and shapes for various laser power and speed settings. It can be seen that, in general, larger melt pools result from higher amounts of laser power and lower scan speeds. However, in this particular configuration, the size of the melt pool is more dependent on the laser power than the speed.
[0078]
[0102] 15A-15F illustrate how grids can be dynamically created to characterize and control additive manufacturing operations. FIG. 15A illustrates a top view of a cylindrical workpiece 1502 positioned on a portion of a build surface 1504. The workpiece 1502 is shown undergoing additive manufacturing operations. FIG. 15B illustrates how the cylindrical workpiece 1502 is divided into multiple tracks 1506 along which an energy source can melt powder distributed on the top surface of the cylindrical workpiece 1502. In some embodiments, the energy source can alternate directions 1506 as shown, while in other embodiments, the energy source can always move in one direction. Additionally, the direction of the tracks 1506 can change from layer to layer to further randomize the direction of the scans used to build the workpiece 1502.
[0079]
[0103] FIG. 15C illustrates an exemplary scanning pattern of the energy source as it forms a portion of the workpiece 1502. The direction of movement of the exemplary energy source across the workpiece 1502 is diagonal, as indicated by the arrow 1508. Each scan 1510 of the energy source is oriented in a direction perpendicular to the direction of movement of the energy source along the track 1506 and can completely traverse the track 1506. The energy source can be turned off for a short period of time between successive individual scans 1510. In some embodiments, the duty cycle of the energy source can be about 90% when traversing each of the tracks 1506. By employing this type of scanning strategy, the energy source can cover the width of the track 1506 as it traverses the workpiece 1502. In some embodiments, the swath 1510 can have a width of about 5 mm. In some embodiments, the width of the melt pool generated by the energy source can be on the order of about 80 microns, which can substantially reduce the number of tracks required to form the workpiece 1502.
[0080]
[0104] 15D-15E show how grid regions 1512 are dynamically generated along each track 1506 and sized to correspond to the width of each individual scan 1510. The system can predict the exact location of the subsequent scan by reference to the energy source drive signal directed to the energy source. In some embodiments, the width of the grid 1512 can match the length of the individual scan 1512 or be within 10% or 20% of the length of the individual scan 1512. Again, the scan length of the individual scan 1512 can be predicted by reference to the energy source drive signal. In some embodiments, the grid regions 1512 can be square or rectangular in shape. As the energy source continues along the track 1506, a thermal energy density can be determined for each of the grid regions 1512. In some embodiments, the thermal energy density reading in grid region 1512-1 can be used to adjust the power of the energy source in the next grid region, in this case grid region 1512-2. For example, if the thermal energy density readings generated by the individual scans 1510 in grid region 1512-1 are significantly higher than expected, the power output of the energy source can be reduced, the speed of the energy source scanning across the individual scans 1510 can be increased, and / or the spacing of the individual scans 1510 can be increased within grid region 1512-2. These adjustments can be made as part of a closed loop control system. While only five individual scans 1510 are shown within each region, this is done for illustrative purposes only, and the actual number of individual scans within grid region 1512 can be substantially higher. For example, if the width of the melt zone generated by the energy source is approximately 80 microns, approximately 60 individual scans 1510 may be required for all powder within a 5 mm square grid region 1512 to be within the melt zone.
[0081]
[0105] FIG. 15F illustrates the edge region of the workpiece 1502 after the energy source has finished traversing the pattern of the track 1506. In some embodiments, the energy source can continue to add energy to the workpiece 1502 after the majority of the powder has melted and resolidified. For example, a contour scan 1514 can be traced along the periphery 1516 of the workpiece 1502 to provide a surface finish to the workpiece 1502. It should be appreciated that the contour scan 1514 shown is substantially shorter than the individual scans 1510. As such, the grid area 1518 can be much narrower than the grid area 1512. It should also be noted that the grid area 1518 is not purely rectangular in shape since, in this case, it follows the contour of the periphery of the workpiece 1502. As another example of a possible difference in scan length, a workpiece may include walls of varying thickness. Varying wall thicknesses can result in varying scan lengths within a single grid area. In such cases, the area of each grid region can be kept consistent by increasing the length of the grid region while decreasing its width to accommodate the changing length of the individual scans.
[0082]
[0106] FIG. 16 illustrates a closed loop control example showing a feedback control loop 1600 for establishing and maintaining feedback control of an additive manufacturing operation. At block 1602, a baseline thermal energy density of the next grid area that the energy source is about to traverse is input into the control loop. This baseline thermal energy density reading can be established from a modeling and simulation program and / or from previously performed experiments / test runs. In some embodiments, this baseline thermal energy density data can be adjusted by an energy density bias block 1604 that includes energy density readings of various grid areas recorded during the previous layer. The energy density bias block 1604 can include adjustments to the baseline energy density block if the previous layer received too much or too little energy. For example, if the optical sensor reading indicates a lower than nominal thermal energy density in one area of the workpiece, the energy density bias value can increase the baseline energy density value of the grid area that overlaps with the grid area that is below the nominal thermal energy density reading. In this way, the energy source can fuse additional powder that was not fully fused during the previous layer or layers.
[0083]
[0107] FIG. 16 also illustrates how inputs from blocks 1602 and 1604 cooperate to generate an energy density control signal that is received by a controller 1606. The controller 1606 is configured to receive the energy density control signal and generate heat source input parameters configured to generate a desired thermal energy density within the current grid region. The input parameters can include power, scan speed, hatch spacing, scan direction, and scan time. The input parameters are then received by the energy source 1608, and changes to the input parameters are adopted by the energy source 1608 for the current grid region. Once the optical sensor measures the scans of the energy source 1608 that make up the current grid region, at block 1610, the thermal energy density of the current grid region is calculated and compared to the energy density control signal. If the two values are the same, the energy density control signal is not changed based on the optical sensor data. However, if the two values are different, the difference is added or subtracted to the energy density control signal for the scan made in the next grid region.
[0084]
[0108] In some embodiments, the grid areas of the current layer and all previous layers may be dynamically generated grid areas oriented according to the path and scan length / width of the scan performed by the energy source. In this type of configuration, both the baseline energy density and the energy density bias may be based on the dynamically generated grid areas. In other embodiments, the grid area of the current layer may be dynamically generated and the energy density bias data 1604 may be based on energy density readings associated with static grid areas defined before the additive manufacturing operation begins, so that the static grid areas remain fixed throughout the part and do not change position, size, or shape. The grid may be uniformly shaped and spaced if a Cartesian grid system is required, but may also take the form of grid areas that make up a polar grid system. In other embodiments, the grid area of the current layer is statically generated before the build operation is performed and the energy density bias data is also statically generated and can share the same grid used for the current layer.
[0085]
[0109] In some embodiments, thermal radiation density can be used instead of thermal energy density in the control loop 1600. Thermal radiation density may refer to other factors in addition to thermal energy density. For example, thermal radiation density can be a weighted average of multiple features including thermal energy density along with one or more other features such as peak temperature, minimum temperature, heating rate, cooling rate, average temperature, standard deviation from the average temperature, rate of change of average temperature over time, etc. In other embodiments, one or more other features can be used to verify that the scans making up each grid area have reached the desired temperature, heating rate or cooling rate. In such embodiments, the verification feature can be used as a flag indicating that the input parameters of the energy source may need to be adjusted within a defined control window to achieve the desired temperature, heating rate or cooling rate. For example, if the peak temperature in a grid area is too low, the power can be increased or the scan speed can be decreased. It should be noted that the above description of the control loop 1600 in relation to various types of grid TEDs allows those skilled in the art to understand that scan TED metrics can also be used in similar loop configurations. TED analysis of recoater arm short feed
[0086]
[0110] FIG. 17A shows a normal distribution of powder 1702 across the build plate 1704. This depiction shows how the powder is spread evenly without any height variations. In contrast, FIG. 17B shows how the thickness of the powder layer 1702 can vary greatly when an insufficient amount of powder 1702 is retrieved by the recoater arm. As the recoater arm begins to run out of powder 1702, the thickness of the layer of powder 1702 gradually tapers off until a portion of the build plate 1704 is completely bare of powder 1702. Early detection of this type of phenomenon is important for accurate defect detection, as this type of error can be very detrimental to the overall quality of the part.
[0087]
[0111] 17C is a black and white photograph of the build plate where nine workpieces positioned on the build surface are only partially covered due to a short feed of powder 1702. In particular, the three workpieces on the right side of the photograph are completely covered, while the three workpieces on the left side are nearly completely uncovered.
[0088]
[0112] FIG. 17D shows how the detected thermal energy density differs substantially when the energy source scans across all nine workpieces using the same input parameters. Region 1706 produces a substantially higher thermal energy density reading because the emissivity of the powder is substantially higher and the thermal conductivity of the build plate or solidified powder material is higher than the effective thermal conductivity of the powder. The higher thermal conductivity reduces the amount of energy available to radiate back to the optical sensor, thereby reducing the detected thermal radiation. Additionally, the lower emissivity of the material itself reduces the amount of radiated thermal energy. Thermal Energy Density and Global Energy Density
[0089]
[0113] When power supplied by an energy source is provided to a workpiece, it melts the material that makes up the workpiece, but that power can also be dissipated by several other heat and mass transfer processes during the additive manufacturing process. The following equation describes the various processes that can absorb the emitted power if the energy source is a laser scanning across a powder bed: P TOTAL LASER POWER =P OPTICAL LOSSES AT THE LASER +P ABSORPTION BY CHAMBER GAS +P REFLECTION +P PARTICLE AND PLUME INTERACTIONS +P POWER NEEDED TO SUSTAIN MELT POOL +P CONDUCTION LOSSES +P RADIATION LOSSES +P CONVECTION LOSSES +P VAPORIZATION LOSSES Formula (13)
[0090]
[0114] OPTICAL LOSSES AT THE LASER refers to power losses due to imperfections in the optical system that transmits and focuses the laser light at the build surface. Imperfections result in absorption and reflection losses of the emitted laser within the optical system. ABSORPTION BY CHAMBER GAS refers to power losses due to gases within the build chamber of an additive manufacturing system that absorb a small portion of the laser power. The effect of this power loss depends on the absorptivity of the gas at the wavelength of the laser. REFLECTION LOSSES refers to power lost due to light escaping the laser light that is never absorbed by the powder bed. PARTICLE AND PLUME INTERACTIONS refers to interactions between the laser and the plume, and / or particles that are released during the deposition process. Power losses due to these effects can be ameliorated by shielding gas circulating through the build chamber, but a small amount of power loss usually cannot be completely avoided. POWER NEEDED TO SUSTAIN MELT POOL refers to the portion of the laser power absorbed by the work material to melt the powder and ultimately heat it up to the temperature ultimately achieved by the melt pool. CONDUCTION LOSSES refers to the portion of the power absorbed by conduction to the solidified metal below the powder and to the powder bed itself. In this way, the powder bed and solidified material that make up the part conduct heat away from the melt pool. This conductive transfer of thermal energy is the primary form of energy loss from the melt pool. RADIATION LOSSES refers to the portion of the laser power emitted from the melt pool and the material surrounding the melt pool that is hot enough to emit thermal radiation. CONVECTION LOSSES refers to the loss caused by the transfer of thermal energy to the gas circulating in the build chamber. Finally, VAPORIZATION LOSSES refers to the small portion of the powder material that evaporates under laser irradiation.Since the latent heat of vaporization is very large, this is a powerful cooling effect on the molten pool and can cause significant energy losses at high total laser beam powers.
[0091]
[0115] The Thermal Energy Density (TED) metric is based on the optical measurement of light that is the result of the emission of light from a heated area, the transmission of this light through an optical system, the collection of the light by a detector, and the conversion of this light into an electrical signal. The equation governing blackbody radiation over all possible frequencies is given by the Stefan-Boltzmann equation shown below in Equation (14):
number
[0092]
[0116] The variables in equation (14) are shown in Table (1) below. JPEG0007576873000014.jpg58146
[0093]
[0117] There are additional intervening factors that affect the emitted light before it is collected by the sensor and therefore becomes the voltage used to calculate the TED metric, which are summarized below in equation (15). V VOLTAGE USED BY TED ={P RADIATED -P VIEW FACTOR -P OPTICAL LOSSES AT RADIATED WAVELENGTHS -P SENSOR LOSS FACTOR}*(SENSOR SCALING FACTOR) formula (15)
[0094]
[0118] These various terms from equation (15) are explained in Table (2) below. JPEG0007576873000015.jpg74143
[0095]
[0119] A figure of merit often used in additive manufacturing is the global energy density (GED), which is a parameter that combines the various PROCESS INPUTs as shown in equation (16) below: GED=(BEAM POWER) / {(TRAVEL SPEED)*(HATCH SPACING)} Equation (16)
[0096]
[0120] The GED units of energy per unit area are: (JOULES / SEC) / {(CM / SEC)*(CM)}=JOULES / CM 2 However, it should be noted that while GED may have the same units as TED, GED and TED are generally not equivalent. As an example, TED is obtained by dividing the radiant power from the hot area by the area, while GED is a measure of the input power. As described herein, TED relates to the RESPONSE or PROCESS OUTPUT, and GED relates to the PROCESS INPUT. The inventors believe that, as a result, TED and GED are different measurements from one another. In some embodiments, the area utilized in determining TED is different from the melt pool area. As a result, some embodiments do not have a direct correlation between TED and melt pool area. Advantageously, TED is sensitive to a variety of factors that directly affect the additive manufacturing process.
[0097]
[0121] Although the embodiments described herein have determined thermal energy density using data generated by optical sensors, the embodiments described herein may be implemented using data generated by sensors that measure other indicia of in-process physical variables. Sensors that measure indicia of in-process physical variables include, for example, force and vibration sensors, contact thermal sensors, non-contact thermal sensors, ultrasonic sensors, and eddy current sensors. It will be apparent to one of ordinary skill in the art that many modifications and variations are possible in light of the above teachings.
[0098]
[0122] Various aspects, embodiments, implementations or features of the described embodiments may be used separately or in any combination. Various aspects of the described embodiments may be implemented by software, hardware, or a combination of hardware and software. Also, the described embodiments may be realized as computer readable code on a computer readable medium for controlling a manufacturing operation or as computer readable code on a computer readable medium for controlling a manufacturing line. A computer readable medium is any data storage device that can store data that can then be read by a computer system. Examples of computer readable media include read-only memory, random access memory, CD-ROMs, HDDs, DVDs, magnetic tapes, and optical data storage devices. Also, the computer readable medium may be distributed over network-coupled computer systems, such that the computer readable code is stored and executed in a distributed manner.
[0099]
[0123] The foregoing description, for purposes of explanation, used specific terminology to provide a thorough understanding of the described embodiments. However, it will be apparent to one skilled in the art that specific details are not required to practice the described embodiments. Thus, the foregoing descriptions of specific embodiments have been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the described embodiments to the precise forms disclosed. It will be apparent to one skilled in the art that many modifications and variations are possible in light of the above teachings.
Claims
1. depositing a layer of build material on the build surface; dividing at least a portion of the build surface into a plurality of grid regions, each grid region having a respective area; generating a plurality of scans across the plurality of grid areas using an energy source to fuse the layer of build material within the grid areas; using a sensor to detect energy emitted from the build surface while the energy source fuses the layers of build material; determining a thermal energy density of the grid areas of the plurality of grid areas from the energy detected by the sensor and the respective areas of the grid areas; An additive manufacturing method comprising:
2. 2. The additive manufacturing method of claim 1, wherein the thermal energy density is determined from the energy released from the grid areas of the plurality of grid areas during the fusing and the area of the grid areas.
3. The additive manufacturing method of claim 1 , wherein the sensor is a photodiode.
4. The additive manufacturing method of claim 1 , wherein the build material comprises a metal powder.
5. The additive manufacturing method of claim 1 , wherein the energy source comprises a laser.
6. 2. The additive manufacturing method of claim 1, wherein the determined thermal energy density is compared to a threshold and the grid region is identified as potentially defective when the determined thermal energy density exceeds the threshold.
7. depositing a layer of build material on the build surface; dividing at least a portion of the build surface into a plurality of grid regions, each grid region having a respective area; fusing the build material within each of a plurality of grid regions using an energy source; using a sensor to detect energy released while the energy source fuses the build material; determining a thermal energy density for each of the plurality of grid regions based on the detected energy and the respective area of each grid region; An additive manufacturing method comprising:
8. 8. The additive manufacturing method of claim 7, wherein the thermal energy density determined for each grid area of the plurality of grid areas is determined from energy released from each grid area during the fusing divided by the respective area of each grid area.
9. The additive manufacturing method of claim 7 , wherein the sensor is a photodiode.
10. The additive manufacturing method of claim 7 , wherein the layer of build material comprises a metal powder.
11. The additive manufacturing method of claim 7 , wherein the fusing of the build material comprises melting the build material with the energy source.
12. The additive manufacturing method of claim 7 , wherein the thermal energy density determined for each grid region of the plurality of grid regions is compared to a threshold value.
13. 8. The additive manufacturing method of claim 7, wherein the step of depositing a layer is performed by a recoater arm that spreads a layer of powder.
14. a build material disposed across the build surface; an energy source configured to fuse at least a portion of the build material; a sensor configured to detect energy emitted from the build surface; 1. A processor comprising: Dividing at least a portion of the build surface into a plurality of grid regions; receiving data from the sensor while the build material is being fused by the energy source; determining an area of each grid area of the plurality of grid areas; a processor configured to calculate a thermal energy density for each grid area of the plurality of grid areas, the thermal energy density being calculated based on the data received from the sensor and the area of each grid area of the plurality of grid areas; Additive manufacturing systems, including:
15. the energy source fuses the build material within each grid area of the plurality of grid areas by generating a plurality of scans within each grid area of the plurality of grid areas; 15. The additive manufacturing system of claim 14, wherein each scan of the plurality of scans includes turning on the energy source, moving the energy source, and turning off the energy source.
16. 16. The additive manufacturing system of claim 15, wherein the processor is further configured to sum data generated by the sensor for each scan of the plurality of scans within each grid area of the plurality of grid areas.
17. The additive manufacturing system of claim 14 , wherein the build material comprises a metal powder.
18. The additive manufacturing system of claim 14 , wherein the energy source comprises a laser.
19. 15. The additive manufacturing system of claim 14, wherein the processor is further configured to compare the calculated thermal energy density of each grid area of the plurality of grid areas to a threshold value.
Citation Information
Patent Citations
Three-dimensional product manufacturing device and manufacturing method
JP2006513055A
Method and corresponding apparatus for monitoring laser beam energy density by image analysis
JP2016540895A
Systems and methods for additive manufacturing operations
US20170090462A1
Filtering temperature distribution data of build material
US20180281286A1
Controlling heating of a surface
WO2016050319A1
