In-Situ Monitoring of Additive Manufacturing Layer-by-Layer Build Using Ultrasonic Spectroscopy

TECHNICAL FIELD

The present disclosure relates to monitoring of additive manufacturing (AM) layer-by-layer build using ultrasonic spectroscopy.

BACKGROUND

Additive manufacturing is the process of joining materials layer-by-layer to make components from three-dimensional (3D) model data. Additive manufacturing uses a stock material, such as a powder or filament that is formed into a layer according to the model data. Such layers are built-up, typically one on top of the last, until the entire component is formed, i.e., until the component is fully formed. Additive manufacturing allows complex parts to be built in various industries including consumer products, transportation, aerospace, robotics, medical, military, and academic research.

Conventional component inspection associated with additive manufacturing inspects the fully formed component for defects. If a defect is found, the component may be discarded and remade. Often, the defect may be introduced early in the additive manufacturing process, i.e., the defect may be introduced into one or more layers formed early in the additive manufacturing process; however, the defect is not caught when initially introduced because of the post fabrication inspection. As a result, the additive manufacturing continues to add layers to the defect until a defective component is fully formed. This results in wasted time and money. Moreover, the post fabrication inspection obviates any opportunity to employ a remedial or corrective measure to correct the defect at the time the defect is introduced.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of an example ultrasonic inspection system.

FIG. 2 is a block diagram of an example pulse generator of an ultrasonic transmitter of the ultrasonic inspection system.

FIG. 3 is an amplitude vs. time plot of an example amplified chirp pulse produced by the ultrasonic transmitter.

FIG. 4 shows an example frequency spectrum of the chirp pulse.

FIG. 5 shows different example frequency spectrums for different amplified chirp pulses generated by the ultrasonic transmitter.

FIG. 6 is a block diagram of an example signal processor of an ultrasonic receiver of the ultrasonic inspection system.

FIG. 7A shows an example first frequency spectrum produced by a frequency domain/spectral processor of the signal processor for a single graphite epoxy layer.

FIG. 7B shows an example second frequency spectrum (i.e., resonance spacing spectrum) produced by the spectral processor for the single graphite epoxy layer based on the first frequency spectrum.

FIG. 8A shows an example first frequency spectrum for a single silicon rubber layer.

FIG. 8B shows an example resonance spacing spectrum for the single silicon rubber layer, based on the first frequency spectrum of FIG. 8A.

FIG. 9A shows an example first frequency spectrum for a single graphite epoxy layer of different thickness than the layer for FIG. 7A.

FIG. 9B shows an example resonance spacing spectrum for the single graphite epoxy layer of different thickness, based on the first frequency spectrum of FIG. 9A.

FIG. 10A is an illustration of example multilayer test material in which the single layers examined in FIGS. 7A, 8A, and 9A are bonded together.

FIG. 10B shows an example first frequency spectrum for the multilayer test material.

FIG. 10C shows an example resonance spacing spectrum for the multilayer test material, based on the first frequency spectrum of FIG. 10B.

FIG. 11 is an illustration of an example matched filter in a time domain processor of the signal processor.

FIG. 12 is a plot of an example chirp pulse delivered to test material by the ultrasonic transmitter.

FIG. 13 is a plot of an example first chirp pulse reflection (of the delivered chirp pulse from FIG. 12) received from the test material by the ultrasonic receiver.

FIG. 14 is a plot of an example second chirp pulse reflection (of the delivered chirp pulse from FIG. 12) received from the test material by the ultrasonic receiver.

FIG. 15 is a plot of an example combined response of the first and second chirp pulses.

FIG. 16 is a plot of example pulse compression results produced by the time domain processor based on the combined response of FIG. 15.

FIG. 17 is an illustration of another example multilayer test material.

FIG. 18A shows an example first frequency spectrum for the multilayer test material of FIG. 17 in a case where the layers of the multilayer test material are fully bonded to each other.

FIG. 18B shows an example resonance spacing spectrum for the multilayer test material of FIG. 17, based on the first frequency spectrum of FIG. 18A.

FIG. 18C is an example time domain plot of compressed pulses produced by the time domain processor corresponding to the frequency spectrums of FIGS. 18A and 18B, for the fully bonded multilayer test material.

FIG. 19A shows an example first frequency spectrum for the multilayer test material of FIG. 17 in a case where a first pair of the layers of the multilayer test material are disbonded.

FIG. 19B shows an example resonance spacing spectrum for the multilayer test material of FIG. 17, based on the first frequency spectrum of FIG. 19A.

FIG. 19C is an example time domain plot of compressed pulses produced by the time domain processor corresponding to the frequency spectrums of FIGS. 19A and 19B, for the case where a first pair of layers of the multilayer test material are disbonded.

FIG. 20A shows an example first frequency spectrum for the multilayer test material of FIG. 17 in a case where a second pair of the layers of the multilayer test material are disbonded.

FIG. 20B shows an example resonance spacing spectrum for the multilayer test material of FIG. 17, based on the first frequency spectrum of FIG. 20A.

FIG. 20C is an example time domain plot of compressed pulses produced by the time domain processor corresponding to the frequency spectrums of FIGS. 20A and 20B, for the case where a second pair of layers of the multilayer test material are disbonded.

FIG. 21 is a flowchart of an example method performed by the ultrasonic inspection system.

FIG. 22 is a flowchart of another example method performed by the ultrasonic inspection system.

FIG. 23 is an illustration of an example computer display on which plots of a frequency spectrum, a resonance spacing spectrum, and a pulse compression time response are concurrently displayed.

FIG. 24 is a block diagram of an example controller of the ultrasound inspection system.

FIG. 25 is a block diagram of an example additive manufacturing (AM) environment including an AM apparatus (AMA) to construct a component using an AM process, and the ultrasonic inspection system (UIS) to inspect/monitor each layer of the component in-situ during the AM process using ultrasonic chirp pulses.

FIG. 26 is a flowchart of an example method of forming and inspecting layers of the component layer-by-layer, performed by the AMA and the UTA.

FIG. 27 is an illustration of an example of transducer pair placement for multi-position inspection in which multiple transducer pairs are in different positions beneath the component during additive manufacturing.

DESCRIPTION OF EXAMPLE EMBODIMENTS
Overview

An ultrasonic inspection system performs a method of inspecting a layer of a component formed on a build plate using an additive manufacturing process. The inspecting includes delivering to the layer, through the build plate and any intervening layers of the component previously formed on the build plate, an ultrasonic chirp pulse having a frequency that sweeps through a range of frequencies across a chirp bandwidth. The method also includes receiving from the layer ultrasonic energy including reflections of the ultrasonic chirp pulse delivered to the layer, and processing the ultrasonic energy to produce an ultrasonic signature indicative of a characteristic of the layer and the intervening layers. The inspecting is repeated for each of subsequent layers of the component formed on the build plate according to the additive manufacturing process, to inspect the component layer-by-layer as the component is built-up during the additive manufacturing process.

Example Embodiments

First, an ultrasonic inspection system is described below in connection with FIGS. 1-6 and 24. Then, frequency domain processing and time domain processing of ultrasonic energy from various test structures, and results of the processing, are described below in connection with FIGS. 7A-23. Finally, embodiments directed to additive manufacturing of a component and layer-by-layer ultrasonic inspection of the component in-situ during the additive manufacturing are described below in connection with FIGS. 25-27.

With reference to FIG. 1, there is shown a block diagram of an example ultrasonic spectroscopy inspection system 100 (referred to simply as an “ultrasonic inspection system 100”) including an ultrasonic transmitter (TX) 102 and an ultrasonic receiver (RX) 104 configured to interact with each other to perform ultrasonic (spectroscopy) inspections of test material 106 coupled to both the ultrasonic transmitter and the ultrasonic receiver. In general, test material 106 can be a single layer structure, such as a monolithic piece of metal, or a multi-layer structure having two or more layers, e.g., a sandwiched structured, in which adjacent layers are designed to be bonded to each other, and where the composition may vary from layer to layer. Ultrasonic inspection system 100 may also include a controller 105, coupled to transmitter 102 and receiver 104, to provide overall control of the system and implement various functions associated with transmitter 102 and receiver 104, as described more fully below in connection with FIG. 24.

FIG. 1 shows two arrangements by which a transmit transducer 126 of transmitter 102 and a receive transducer 130 of receiver 104 are coupled to test material 106. In a “through transmission” arrangement, transmit transducer 126 is arranged on one side of test material 106 and receive transducer 130 is arranged on the opposite side of test material 106 such that receive transducer 130 receives ultrasonic transmissions that travel all the way through test material from the transmit side to the opposite receive side. In the “single-sided reflection” arrangement, transmit transducer 126 and receive transducer 130 are arranged on the same side of test material 106 such that receive transducer 130 receives reflections of the ultrasonic chirped pulse instead of through transmissions. This arrangement enables the reflected ultrasonic chirped pulse to be evaluated in both the frequency domain and time domain, where individual peaks in the received time domain signal may be indicative of responses of individual layers within a multi-layer structure, the condition of inter-layer bonds, or the presence of defects within a single layer.

Transmitter 102 delivers an ultrasonic chirp pulse to test material 106, and receiver 104 detects resulting ultrasonic energy from the test material that is indicative of various defects and/or properties of the test material. Receiver 104 performs signal processing (referred to as ultrasonic sound analysis) on the detected ultrasonic energy to produce indications, including visual indications, of the various defects and/or properties. Depending on the transducer arrangement and the receiver capabilities, the indications may include, but are not limited to, frequency responses including material resonances, resonance spacings, and reflection time pulses, as discussed more fully below. The defects and properties indicated include, but are not limited to, a number of layers in test material 106, whether the layers are “disbonded,” layer thickness, porosity, and layer composition. The term “disbonded” means there is a bonding defect between layers. In the additive manufacturing process, this could mean large voids or porosity or an inadequate fusion between layers.

Transmitter 102 includes a pulse generator 110 to generate a chirp pulse 112 responsive to pulse generator control commands 114, a radio frequency (RF) power amplifier 120 to amplify the chirp pulse responsive to a gain control signal 122 to produce an amplified chirp pulse 124, and the ultrasonic transmit transducer (TT) 126 coupled to test material 106 so as to deliver an ultrasonic chirp pulse to the test material responsive to the amplified chirp pulse. Pulse generator 110 generates chirp pulse 112 as a sinusoidal voltage waveform, for example, that sweeps through a range of frequencies or a “swept frequency range” (also referred to as a “chirp bandwidth”) from a start frequency to a stop frequency during a time period equal to a pulse width of the chirp pulse. The start frequency and the stop frequency define a frequency position of the chirp bandwidth (also, a center frequency of the swept frequency range may represent the frequency position), while a difference between the start frequency and the stop frequency defines the chirp bandwidth.

Pulse generator 110 independently adjusts chirp pulse parameters, including the start frequency, the stop frequency, the pulse width, and an amplitude of chirp pulse 112 across the chirp bandwidth, responsive to pulse generator control commands 114. Thus, the chirp bandwidth may be adjusted over a range of chirp bandwidths from a narrowest chirp bandwidth to a widest chirp bandwidth, and the frequency position of the chirp bandwidth may be adjusted over a range of frequency positions of the chirp bandwidth from a lowest frequency position to a highest frequency position. In an example, pulse generator 110 may vary (i) the chirp bandwidth from a narrow bandwidth to a broader bandwidth (e.g., up to 20 MHz), (ii) the frequency position of the chirp bandwidth (e.g., up to 40 MHz), and (iii) the amplitude from +/−0.05 volts to +/−2.0 volts. Other ranges of the chirp pulse parameters are possible. Typically, pulse generator 110 adjusts the pulse width to be approximately 40 μs or greater to ensure sufficient ultrasonic energy for inspection analysis. In operation, pulse generator 110 adjusts the aforementioned chirp pulse parameters so that they are suited to inspect defects and/or material properties of interest of test material 106.

In one embodiment, pulse generator 110 provides chirp pulse 112 to RF power amplifier 120 as a single ended voltage, in which case the RF power amplifier includes a single ended input to receive the single ended voltage. In another embodiment, pulse generator 110 provides chirp pulse 112 to RF power amplifier 120 as a differential voltage, in which case the RF power amplifier includes a differential input to receive the differential voltage. RF power amplifier 120 amplifies chirp pulse 112 received at the input of the RF power amplifier according to a gain set by gain control signal 122 to produce amplified chirp pulse 124 at an output of the RF power amplifier, and provides the amplified chirp pulse to transmit transducer 126. RF power amplifier 120 provides amplified chirp pulse 124 to a drive input of transmit transducer 126 coupled to the output of the RF power amplifier. In response to amplified chirp pulse 124, transmit transducer 126 delivers an ultrasonic chirp pulse to test material 106. Transmit transducer 126 typically represents a capacitive load to the output of RF power amplifier 120, and the capacitance of the capacitive load may vary substantially across different types of transducers. An advantage of RF power amplifier 120 is its ability to drive a wide range of capacitances (capacitive loads) over a wide range of frequencies without any appreciable degradation of power amplifier gain or effect on a frequency spectrum of amplified chirp pulse 124.

RF power amplifier 120 may be any class of RF power amplifier, e.g., Class A, Class B, Class C, and so on, configured to provide a wide operating frequency range. The wide operating frequency range represents a frequency range over which the RF power amplifier 120 provides substantial RF gain. By way of a non-limiting example, the input voltage to RF power amplifier 120 can be on the order of millivolts or tens of millivolts (e.g., 80 millivolts), while the output of RF power amplifier 120 can be on the order of tens of volts (e.g., 50 volts), with gains on the order of tens of decibels (e.g., 40 dB). The operating frequency range is wider than the widest chirp bandwidth generated by pulse generator 110. Moreover, the gain of RF power amplifier 120 is approximately flat (i.e., the gain has an approximately flat frequency response) across the widest chirp bandwidth when the widest chirp bandwidth is positioned anywhere in the operating frequency range. An example operating frequency range of RF power amplifier 120 is from 100 kHz to 40 MHz.

With reference to FIG. 2, there is a block diagram of pulse generator 110, according to an embodiment. In the example of FIG. 2, pulse generator 110 includes a digital waveform generator 204, and a digital-to-analog converter (DAC) 206. Digital waveform generator 204 generates a digitized, frequency-swept waveform 210 responsive to pulse generator control commands 114, which also include a frequency vs. time sweep characteristic and an amplitude vs. time/frequency characteristic for the frequency-swept waveform. The frequency vs. time sweep characteristic may be in accordance with any desired frequency vs. time characteristic, e.g., the swept frequency may increase/decrease over time according to a linear function, a hyperbolic function, or any other type of function. DAC 206 converts digitized waveform 210 to a continuous-time, chirp pulse waveform 112. Digital waveform generator 204 may be implemented based on any presently known or hereafter developed digital waveform generator techniques.

With reference to FIG. 3, there is shown an amplitude vs. time plot of an idealized chirp pulse 300 delivered to test material 106 by transmit transducer 126 responsive to amplified chirp pulse 124. Chirp pulse 300 has a pulse width 302 over which a waveform 304 sweeps across a swept frequency range (i.e., chirp bandwidth) from a start frequency F1 at a start of the chirp pulse to a stop frequency F2 at an end of the chirp pulse. An amplitude of the idealized chirp pulse 300 is relatively flat or constant across the entire chirp bandwidth. In practice, to account for distortions caused by the frequency responses of transmit and receive transducers 126 and 130, a reference response waveform can be obtain by placing the transducers face to face and processing the received signal. The idealized chirp pulse 300 can than be adjusted on the transmit end (e.g., with amplitude modulation), essentially with the inverse of the distortions observed in this reference waveform, in order to compensate for the transducers' distortions. In this manner, a relatively flat frequency response can be ensured across the operating bandwidth notwithstanding the distortions introduced by the transducers.

With reference to FIG. 4, there is shown an idealized frequency spectrum 400 of chirp pulse 300. Again, in order realize this frequency response at the receiver, the transmit pulse must be modified to account for the distortions from the transducers determined from a reference waveform obtained by transmitting a pulse with the transducers face to face (no intervening material). In the example of FIG. 4, start frequency F1 is approximately 1 MHz, while stop frequency F2 is approximately 6 MHz. As can be seen in FIG. 3, the chirp pulse waveform tapers into and out of the frequency sweep, resulting the bandwidth appearing slightly wider than the 1 to 6 MHz range in FIG. 4. Frequency spectrum 400 is relatively flat as a result of the above-mentioned compensation in which the reference waveform resulting from the face-to-face response of the transducers is used to offset the transducer distortions by modifying the transmit waveform.

As mentioned above, pulse generator 110 may generate chirp pulse 112 with different chirp bandwidths and frequency positions (i.e., with different pairs of start and stop frequencies) responsive to control commands 114. As an example, FIG. 5 shows different frequency spectrums for different (normalized) chirp pulses delivered by transmit transducer 126 resulting from different chirp pulses (112) generated by signal generator 110. The frequency spectrums include a first frequency spectrum 502 having a first chirp bandwidth 502A (e.g., 450 kHz) positioned at first start and stop frequencies 502B, 502C (e.g., 50 kHz, 500 kHz), a second frequency spectrum 506 having a second chirp bandwidth 506A (e.g., 6 MHz) positioned at second start and stop frequencies 506B, 506C (e.g., 1 MHz, 7 MHz), and a third frequency spectrum 510 having a third chirp bandwidth 510A (e.g., 20 MHz) positioned at third start and stop frequencies 510B, 510C (e.g., 20 MHz, 40 MHz). Frequency spectrums 502-510 all fall within wide operating bandwidth 520 of RF power amplifier 120 and, therefore, benefit from the relatively high, relatively flat gain of the RF power amplifier across each of chirp bandwidths 502A, 506A, and 510A.

Pulse generator 110 controls an amplitude, e.g., peak-to-peak voltage, of chirp pulse 112 to avoid over driving RF power amplifier 120. When RF power amplifier 120 is over driven, the RF power amplifier clips the amplitude (e.g., sinewave clipping) of amplified chirp pulse 124. Therefore, pulse generator 110 generates chirp pulse 112 so that its amplitude remains just below an amplitude (referred to as a “limit amplitude”) that over drives RF power amplifier 120. Typically, the limit amplitude is frequency dependent, e.g., increases with frequency, across the operating frequency range of RF power amplifier 120. Thus, pulse generator 110 may control the amplitude of chirp pulse 112 to track the limit amplitude over frequency, e.g., to increase the amplitude of chirp pulse 112 with frequency in correspondence with an increase in the limit amplitude with frequency. The limit amplitude variations across the operating frequency range of RF power amplifier 120 may be determined empirically, and the amplitude of chirp pulse 112 may be adjusted to be just below the empirically determined limit amplitude based on control commands 114. In an example, the peak-to-peak voltage of chirp pulse 112 may be controlled to be in a range from +/−0.05 volts to +/−1.0 volts from a low end of the operating frequency range to a high end of the operating frequency range, to produce a relatively constant peak-to-peak voltage of amplified chirp pulse 124 of approximately +/−80 volts.

Additionally, as previously described, pulse generator 110 controls the amplitude of chirp pulse 112 to modify the amplitude of the ultrasonic energy delivered by transmit transducer 126 to test material 106 over the chirp bandwidth in order to compensate for transducer distortions determined from the reference waveform. While it is preferable to deliver a chirp pulse having a flat frequency spectrum to test material 106, a combined frequency response of RF power amplifier 120, transmit transducer 126, and receive transducer 130 may vary across a given chirp bandwidth. Accordingly, pulse generator 110 varies the amplitude of chirp pulse 112 across the chirp bandwidth to compensate for/cancel the way in which the combined frequency response varies across the chirp bandwidth, so that transducer 126 delivers the chirp pulse to test material 106 with a compensated or “normalized,” flat frequency spectrum. For example, pulse generator 110 increases or decreases the amplitude of chirp pulse 112 over the chirp bandwidth in a manner that essentially applies the inverse of the frequency distortions observed in the reference waveform at the receiver as a result of a test transmission with the transmit and receive transducer placed face to face with no intervening test material. This compensation provides a normalized/flat frequency spectrum (of the chirp pulse) at the receiver. Essentially, the use of the reference waveform enables the combined frequency response of RF power amplifier 120, transmit transducer 126, and receive transducer 130 over the chirp bandwidth may be characterized/measured. During regular operation, when pulse generator 110 generates chirp pulse 112 to inspect test material 106, the pulse generator controls (increases/decreases) the amplitude of chirp pulse 112 over the chirp bandwidth based on (i.e. to compensate for) the characterized/measured combined frequency response.

Referring again to FIG. 1, receiver 104 is now described with particular emphasis on the combination of both frequency domain and time domain process. It will be appreciated, however, that the above-described transmitter 102 and amplification scheme is suitable for operation in ultrasonic systems that do not employ all of the aspects of the described receiver 104. Receiver 104 includes an ultrasonic receive transducer 130 coupled to test material 106, a receive amplifier 132, an analog-to-digital converter (ADC) 134, a signal processor 136 (also referred to as an “ultrasonic sound analyzer 136”), and an output device 138, such as a computer display. When transmit transducer 126 delivers an ultrasonic chirp pulse to test material 106, receive transducer 130 detects from the test material an ultrasonic signal 140 (also referred to as ultrasonic energy 140) resulting from an interaction between the delivered ultrasonic chirp pulse and the test material, and provides the detected ultrasonic signal to receive amplifier 132.

Receive amplifier 132 amplifies ultrasonic signal 140 to produce an amplified ultrasonic signal 142, and provides the amplified ultrasonic signal to ADC 134. ADC 134 digitizes amplified ultrasonic signal 142 to produce a digitized ultrasonic signal 144 (representative of ultrasonic signal 140), and provides the digitized ultrasonic signal to signal processor 136. Signal processor 136 processes digitized ultrasonic signal 144 (also referred to as digitized ultrasonic energy 144) to produce processing results 146, stores the processing results, and provides the processing results to output device 138, e.g., for display. Processing results 146 provide visual indications or representations of defects and the various properties of interest of test material 106.

With reference to FIG. 6, there is a block diagram of signal processor 136, according to an embodiment. Signal processor 136 includes a frequency domain/spectral processor 602 to perform frequency domain or spectral processing on digitized ultrasonic signal 144, to produce spectral processing results 606 (included in results 146) for display. Signal processor 136 also includes a time domain processor 604 to perform time domain processing on digitized ultrasonic signal 144, to produce time domain processing results 608 (also included in results 146) for display. Frequency domain processor 602 and time domain processor 604 process digitized ultrasonic signal 144 in parallel and are thus able to (i) perform their respective spectral and time domain processing on the same ultrasonic signal, concurrently, and (ii) deliver their respective results 606, 608 to output device 138, concurrently. In other embodiments, processors 602 and 604 may perform their respective processing sequentially.

Spectral processor 602 performs Fourier transforms, e.g., Fast Fourier transforms (FFTs), on digitized ultrasonic signal 144 to generate frequency spectrums of the ultrasonic signal. The frequency spectrums show frequency resonances produced by the interaction of the chirp pulse delivered to test material 106 and one or more layers of the test material. The frequency resonances indicate various properties of the one or more layers. In an embodiment, spectral processor 602 performs (i) a first FFT on digitized ultrasonic signal 144 to produce a first frequency spectrum, and (ii) a second FFT on the first frequency spectrum to produce a second frequency spectrum, referred to as a “resonance spacing spectrum.” The second spectrum shows frequency spacings between frequency resonances of the first frequency spectrum, hence the name “resonance spacing spectrum.” Both the first and second frequency spectrums may be included in spectral processing results 606, and displayed on output device 138.

With reference to FIG. 7A, there is shown an example first frequency spectrum produced by spectral processor 602 (as described above) when transmitter 102 delivers an ultrasonic chirp pulse to a single graphite epoxy layer, 0.18″ thick, coupled to transmit and receive transducers 126 and 130. The chirp pulse has a chirp bandwidth of 6 MHz and is positioned in frequency at 0.5 MHz to 6.5 MHz. The first frequency spectrum prominently shows multiple frequency harmonics of a fundamental (resonance) frequency indicative of the graphite epoxy layer.

With reference to FIG. 7B, there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum from FIG. 7A. The second frequency spectrum prominently shows the fundamental (resonance) frequency of 0.33 MHz for the graphite epoxy layer (referred to as “Layer 1”).

With reference to FIG. 8A, there is shown an example first frequency spectrum produced by spectral processor 602 when transmitter 102 delivers the same chirp pulse used in connection with FIGS. 7A and 7B to a single silicon rubber layer, which may be used as an adhesive or bonding layer between other layers. The first frequency spectrum prominently shows multiple frequency harmonics of a fundamental (resonance) frequency indicative of the silicon rubber layer.

With reference to FIG. 8B, there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum from FIG. 8B. The second frequency spectrum prominently shows the fundamental (resonance) frequency of 0.64 MHz for the silicon rubber layer (referred to as “Layer 2”).

With reference to FIG. 9A, there is shown an example first frequency spectrum produced by spectral processor 602 when transmitter 102 delivers the same chirp pulse used in connection with FIGS. 7A and 8A to a single graphite epoxy layer 0.14″ thick. The first frequency spectrum prominently shows multiple frequency harmonics of a fundamental (resonance) frequency indicative of the graphite epoxy layer.

With reference to FIG. 9B, there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum from FIG. 9B. The second frequency spectrum prominently shows the fundamental (resonance) frequency of 0.42 MHz for the graphite epoxy layer (referred to as “Layer3”).

With reference to FIG. 10A, there is an illustration of test material 1002 having transmit transducer 126 coupled to an upper face of the test material and receive transducer 130 coupled to a bottom face of the test material that opposes the upper face. Test material 1002 is a multilayer structure that includes a Layer 1 (a graphite epoxy layer 0.18″ thick), a Layer 2 (a silicon rubber layer), and a Layer3 (a graphite epoxy layer 0.14″ thick) bonded together via the silicon rubber layer. The layers Layer 1, Layer 2, and Layer3 are the same as those discussed above in connection with FIGS. 7A, 7B), (8A, 8B), and (9A, 9B), respectively.

With reference to FIG. 10B, there is shown an example first frequency spectrum produced by spectral processor 602 when transmitter 102 delivers the chirp pulse used in connection with FIGS. 7A, 8A, and 9A to test material 1002. The first frequency spectrum shows a combination of frequency harmonics indicative of each of the 3 layers.

With reference to FIG. 10C there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum of FIG. 10B. The second frequency spectrum prominently shows fundamental frequencies for all three layers and from which the 3 layers can be identified.

As mentioned above, time domain processor 604 performs time domain processing on digitized ultrasonic signal 144 (which is representative of ultrasonic signal 140). Ultrasonic signal 140/144 includes reflections of the chirp pulse delivered to test material 106 from one or more layers of the test material. The reflections may be referred to as “reflected chirp pulses” or “chirp pulse reflections.” Because the layers are relatively thin, the chirp pulse reflections have reflection times between layers (e.g., 5 or 10 μs) that are much shorter than their pulse widths (e.g., 200 μs). This causes substantial time-overlapping of the chirp pulse reflections. Overlapping chirp pulse reflections tend to constructively and destructively interfere with each other. Consequently, ultrasonic signal 140/144 includes a series of time-overlapping (and interfering) chirp pulse reflections. The time-overlapping chirp pulse reflections subside over a finite time. The finite time during which the time-overlapping chirp pulse reflections persist is referred to as a “full multilayer response” of the test material 106. The full multilayer response is analogous to an impulse response of a system driven by an impulse, where the impulse is the chirp pulse delivered to test material 106 (representing the system).

Time domain processor 604 performs time domain processing on the series of time-overlapping chirp pulse reflections of the full multilayer response to compress each of the reflections into a respective compressed pulse (also referred to as a “reflection time peak”), such that the resulting compressed pulses are spaced apart from each other in time, i.e., they are not time-overlapping. In this way, the time-separated compressed pulses each indicate a distinct reflection or layer interface in test material 106. In an embodiment, time domain processor 604 includes a matched filter to filter the series of time-overlapping chirp pulse reflections included in ultrasonic signal 140/144. The matched filter is matched to the swept frequency characteristic of the chirp pulse, i.e., the time vs. frequency characteristic of the swept waveform generated by pulse generator 110.

With reference to FIG. 11, there is an illustration of a matched filter 1102 included in time domain processor 602. Matched filter 1102 receives a chirp pulse 1104 in ultrasonic signal 140/144 (e.g., which may be a reflection of the chirp pulse delivered to test material 106 from one of the layers of the test material). Matched filter 1102 filters/compresses chirp pulse 1104 (e.g., the reflection) to produce compressed pulse 1106 (e.g., as a “reflection time peak”), which has a pulse width that is substantially shorter than that of the received chirp pulse. The operations performed by time domain processor 602 and its outputs are described further in connection with FIGS. 12-16.

With reference to FIG. 12, there is a plot of an example chirp pulse delivered by transmitter 102 to test material 106 at a start time of 0 microseconds (μs). The chirp pulse has a pulse width of approximately 185 μs.

With reference to FIG. 13, there is a plot of a first reflection of the delivered chirp pulse (i.e., a first chirp pulse reflection, referred to as “Response 1”) received by receiver 102 from test material 106, a time delay of 5 μs. The first chirp pulse reflection may be a reflection of the delivered chirp pulse from a first layer of test material 106.

With reference to FIG. 14, there is a plot of a second reflection of the delivered chirp pulse (i.e., a second chirp pulse reflection, referred to as “Response 2”) received by receiver 102 from test material 106, after a time delay of 10 μs. The second chirp pulse reflection may be a reflection of the delivered chirp pulse from a second layer of test material 106. The first and second chirp pulse reflections are time-overlapped over most of their respective pulse widths and together form a combined response, shown in FIG. 15.

With reference to FIG. 15, there is a plot of the combined response (referred to as “combined response 1 & 2”) received from test material 106. The combined response exhibits both constructive and destructive interference between Response 1 and Response 2.

Time domain processor 602 performs pulse compression on the combined response to produce pulse compression results, as shown in FIG. 16. FIG. 16 is a time domain plot of the pulse compression results, which include (i) a first compressed pulse 1602 (i.e., a first reflection time peak 1602) representative of the first chirp pulse reflection Response 1, and (ii) a second compressed pulse 1604 (i.e., a second reflection time peak 1604) representative of the second chirp pulse reflection Response 2. Unlike the respective reflections from which they were derived, compressed pulses 1602 and 1604 are distinct because they are non-overlapped in time, i.e., are spaced-apart from each other. Accordingly, distinct compressed pulses 1602, 1604 clearly indicate respective distinct layers of test material 106.

With reference to FIG. 17, there is an illustration of multilayered test material 1702 having transmit transducer 126 and receive transducer 130 coupled to the same (upper) face of the test material. Multilayer test material 1702 includes a carbon phenolic layer (Layer 1) 0.39″ thick, a silicon rubber layer (Layer 2), and a Lucite layer (Layer 3) 0.48″ thick bonded together. With the arrangement shown in FIG. 17, transmitter 102 delivers a chirp pulse to multilayer test material 1702. The delivered chirp pulse has a pulse width of approximately 185 μs, a chirp bandwidth of approximately 1900 kHz, and is positioned in frequency at approximately 300 kHz to 2200 kHz.

FIGS. 18A-18C show inspection results for test material 1702 produced by receiver 104 responsive to the delivered chirp pulse as described in connection with FIG. 17.

With reference to FIG. 18A there is shown an example first frequency spectrum produced by spectral processor 602 (as described above).

With reference to FIG. 18B there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum from FIG. 18A. The second frequency spectrum prominently shows the resonance frequencies for the Lucite layer (Layer 3) and the carbon phenolic layer (Layer 1), but not the silicon rubber layer (Layer 2). The resonance frequency is too high to have multiple peaks in this frequency range.

With reference to FIG. 18C, there is shown an example time domain plot produced by time domain processor 604. The time domain plot includes a compressed pulse corresponding to a reflection from the carbon phenolic layer (Layer 1), a compressed pulse corresponding to a reflection from the silicon rubber layer (Layer 2), and a compressed pulse corresponding to a reflection from the Lucite layer (Layer 3). The aforementioned compressed pulses are non-overlapping in time. Thus, each compressed pulse is indicative of a corresponding one of the layers, Layer 1-3. While the resonance for the silicon rubber layer Layer 2 is missing from the second frequency spectrum of FIG. 18B, that layer is clearly indicated by a compressed pulse on the time domain plot of FIG. 18C. Thus, an advantage of performing both spectral processing and time domain processing (to perform pulse compression of the reflected chirp pulses) concurrently on the same received ultrasonic signal, and then displaying the respective processing results concurrently on one or more displays, is that the combination of techniques provides a more complete picture of the defects and/or properties of the test material. For example, when the frequency domain processing on the ultrasonic energy causes respective frequency resonance peaks corresponding to distinct layers of the multilayer structure having a same frequency resonance to overlap in a combined frequency resonance peak such that the distinct layers are not separately indicated on the frequency domain plot, but the time domain processing results in time-separated reflection time pulses that separately indicate the distinct layers on the time domain plot, then displaying results of both the frequency domain processing and the time domain processing allows comparison between the two for a more complete picture of the test material.

FIGS. 19A-19C show inspection results for test material 1702 produced by receiver 104 responsive to the chirp pulse delivered as described in connection with FIG. 17, but in a case where the silicon rubber layer (Layer 2) and the Lucite layer (Layer 3) are disbonded.

With reference to FIG. 19A, there is shown an example first frequency spectrum produced by spectral processor 602. In FIG. 19A, the first frequency spectrum from FIG. 18A (the bonded case) is shown in dotted line for ease of comparison.

With reference to FIG. 19B there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum from FIG. 19A. The second frequency spectrum prominently shows the resonance frequencies for the carbon phenolic layer (Layer 1), but not the Lucite layer (Layer 3) because of the disbonded layers. In FIG. 19B, the second frequency spectrum from FIG. 18B (the bonded case) is shown in dotted line for ease of comparison.

With reference to FIG. 19C, there is shown an example time domain plot produced by time domain processor 604. The time domain plot includes a compressed pulse (i.e., amplitude peak) corresponding to a reflection from the carbon phenolic layer (Layer 1) and a compressed pulse corresponding to a reflection from the silicon rubber layer (Layer 2), but no compressed pulse corresponding to the Lucite layer (Layer 3) because of the disbonded rubber and Lucite layers.

FIGS. 20A-20C show inspection results for test material 1702 produced by receiver 104 responsive to the delivered chirp pulse as described in connection with FIG. 17, but in a case where the carbon phenolic layer (Layer 1) and the silicon rubber layer (Layer 2) are disbonded. In FIG. 20A, the first frequency spectrum from FIG. 18A (the bonded case) is shown in dotted line for ease of comparison.

With reference to FIG. 20A, there is shown an example first frequency spectrum produced by spectral processor 602.

With reference to FIG. 20B there is shown an example second frequency spectrum produced by spectral processor 602, as an FFT of the first frequency spectrum from FIG. 20A. The second frequency spectrum prominently shows the resonance frequencies for the carbon phenolic layer (Layer 1), but not for the Lucite layer (Layer 3) because of the disbonded layers.

With reference to FIG. 20C, there is shown an example time domain plot produced by time domain processor 604. The time domain plot includes a first compressed pulse corresponding to a first chirp pulse reflection from the carbon phenolic layer (Layer 1), a second compressed pulse corresponding to a second chirp pulse reflection from the from the carbon phenolic layer (Layer 1), but no compressed pulse corresponding to either the Lucite layer (Layer 3) or the silicon rubber layer (Layer 2) due to the disbonded layers.

With reference to FIG. 21, there is a flowchart of an example method 2100 performed by ultrasonic inspection system 100. Method 2100 includes various ones of the operations described above.

At 2105, pulse generator 110 generates chirp pulse 112 having a chirp bandwidth. Pulse generator 110 is able to adjust the chirp bandwidth from a lowest chirp bandwidth to a highest chirp bandwidth, and a frequency position of the chirp bandwidth so that the adjusted chirp bandwidth and the frequency position of the chirp bandwidth are suited to inspecting defects or material properties of test material 106.

At 2110, RF power amplifier 120 amplifies chirp pulse 112 to produce amplified chirp pulse 124. RF power amplifier 120 has an operating frequency range greater than the highest chirp bandwidth and a gain that is relatively flat across the highest chirp bandwidth when the frequency position of the highest chirp bandwidth falls anywhere in the operating frequency range.

At 2115, transmit transducer 126 generates an ultrasonic pulse responsive to the amplified pulse and delivers the ultrasonic pulse to test material 106.

At 2120, ultrasonic receiver 104 receives an ultrasonic signal (140/144) from the test material that results from the ultrasonic pulse delivered to the test material, and processes the received ultrasonic signal to produce analysis results 146 indicative of test material defects or properties.

With reference to FIG. 22, there is a flowchart of another example method 2200 performed by ultrasonic inspection system 100. Method 2200 includes various ones of the operations described above.

At 2205, ultrasonic transmitter 102 delivers to a multilayer structure, e.g., test material 106, an ultrasonic pulse that sweeps through a chirp bandwidth.

At 2210, ultrasonic receiver 104 receives from the multilayer structure ultrasonic energy 140/144 including a series of time-overlapping reflections of the pulse delivered to the multilayer structure at 2205 from layers of the multilayer structure. The series of time-overlapping reflections may represent a full multilayer response of test material 106, as described above.

At 2215, spectral processor 602 performs frequency domain processing on the received ultrasonic energy (140/144) including the series of time-overlapping reflections to produce frequency resonance peaks respectively indicative of distinct layers of the multilayer structure.

At 2220, time domain processor 604 performs time domain processing on the received ultrasonic energy (140/144) to compress the series of time-overlapping reflections into respective time-separated compressed amplitude peaks/reflection time peaks. In an embodiment, spectral processor 602 and time domain processor 604 concurrently perform their respective processing on the (same) ultrasonic energy (140/144) including the series of time-overlapping reflections, so as to generate their respective processing results concurrently.

At 2225, output device 138, e.g., a computer display, displays the frequency resonance peaks on a frequency domain plot and displays the reflection time peaks on a time domain plot. The frequency and time domain plots may be displayed concurrently on one or more computer displays.

With reference to FIG. 23, there is an illustration of a computer display 2300 (e.g., output device 138) on which plots of a frequency spectrum 2305, a resonance spacing spectrum 2310, and a pulse compression time response 2315 generated by signal processor 136 as described above are concurrently displayed.

Reference is now made to FIG. 24, which shows an example block diagram of controller 105 of ultrasound inspection system 100, according to an embodiment. Controller 105 may provide overall control of inspection system, and may also incorporate components of transmitter 102 and ultrasonic receiver 104 described above. There are numerous possible configurations for controller 105 and FIG. 24 is meant to be an example. Controller 105 includes a processor 2444 and memory 2448.

Processor 2444 may include a collection of microcontrollers and/or microprocessors, for example, each configured to execute respective software instructions stored in the memory 2448. Processor 2444 may generate pulse generator control commands 114, gain control signal 122, and analysis results 146. Processor 2444 may also receive digitized received ultrasonic signal 144 from ADC 134. Portions of memory 2448 (and the instruction therein) may be integrated with processor 2444.

The memory 2448 may comprise read only memory (ROM), random access memory (RAM), magnetic disk storage media devices, optical storage media devices, flash memory devices, electrical, optical, or other physical/tangible (e.g., non-transitory) memory storage devices. Thus, in general, the memory 2448 may comprise one or more computer readable storage media (e.g., a memory device) encoded with software comprising computer executable instructions and when the software is executed (by the processor 2444) it is operable to perform the operations described herein. For example, the memory 2448 stores or is encoded with instructions for Control logic 2450 to perform overall control of ultrasonic inspection system 100 and operations described herein related to pulse generator 110 and signal analyzer 136. For example, Control logic 2450 may include frequency domain processor (FDP) logic 2452 to implement the functions of the FDP, time domain processor (TDP) logic 2456 to implement the functions of the TDP, and pulse generator (PG) logic 2458 to implement functions of the pulse generator described above. Control logic 2450 also includes logic to implement operations performed by ultrasonic inspection system 100 to inspect a component layer-by-layer during an additive manufacturing process used to form the component, as described below.

In addition, memory 2448 stores data 2480 used and generated by logic 2450-2458, including, but not limited to: normalization information used to normalize the amplitude of the chirp pulse delivered by transmit transducer 126, chirp pulse parameters (e.g., chirp bandwidth, amplitude, frequency position, and frequency sweep characteristic) used by pulse generator 110, samples of digitized ultrasonic receive signal 144 from ADC 134, analysis results, and information associated with inspection of the component during the additive manufacturing process, as described below.

Embodiments directed to additive manufacturing (AM) of a component (also referred to as an “AM build” of the component) and layer-by-layer ultrasonic inspection of the component during the AM build using ultrasonic chirp pulses are now described in connection with FIGS. 25-27. Such ultrasonic inspection may be used in many different applications to inspect multiple layer materials and perform bondline analysis, corrosion/defect detection, material characterization, and so on. Applications may include, but are not limited to, aircraft component production quality control, corrosion detection, porosity or internal defect analysis, and so on. As used herein, the term “defect,” with respect to a layer of a component (or multiple layers of the component), is construed broadly to mean any imperfection that may range from minor to serious in nature. The defect renders the layer imperfect or suboptimal with respect to a standard or expectation of quality, and may negatively impact the performance or integrity of the layer (and thus the component generally) compared to when the layer does not exhibit the defect, i.e., when the layer is not defective/suboptimal.

With reference to FIG. 25, there is a block diagram of an example AM environment or AM system (AMS) 2500 including an AM apparatus (AMA) 2502 to construct/build a component C layer-by-layer as a series of layers L(1)-L(N) using an AM process, and ultrasonic inspection system (UIS) 100 to inspect each layer of the component in-situ using ultrasonic chirp pulses as the layers are formed/added during the AM process. The components of AMS 2500 are not drawn to scale in FIG. 25. AMA 2502 includes: a support or build plate 2510 having a substantially planar top surface 2510T to support component C while the component is being formed, and a substantially planar bottom surface 2510B through which ultrasonic chirp pulses are delivered to inspect the component; a deposition device (Dep. Dev.) 2512 to form layers L(1)-L(N) (collectively referred to as “layers L”) of component C on the build plate (i.e. to build-up the component layer-by-layer); an optional deposition sensor 2514 to monitor conditions of the layers; and an AMA controller 2516 coupled to the deposition device and the deposition sensor to provide overall control of component layer formation via the AM process. AMA 2502 also includes a fabrication chamber 2520 to contain deposition device 2512, build plate 2510 carrying component C, and deposition sensor 2514.

Operations performed by AMA 2502 to build-up component C are described briefly. AMA controller 2516 receives in digital form a geometric model of a desired component to be built, and transforms the geometric model into thin, virtual, layer-by-layer cross-sections. Deposition device 2512 receives stock material, such as a metal, polymer powder, or filament, from a stock material reservoir (not shown). Controller 2516 controls deposition device 2512 to form sequential layers L(1)-L(N) of the stock material (on top surface 2510B of build plate 2510), one on top of the other, in accordance with the layer-by-layer cross-sections. As each of layers L is formed, it is formed on previous ones of the layers, creating a partially, in-process component. The layer forming is repeated, to build-up component C layer-by-layer. Any known or hereafter developed AM technique may be used to form layers L from the stock material.

Such AM techniques may employ selective laser sintering, direct metal laser sintering, selective heat sintering, electron beam freeform fabrication, electron beam melting, stereolithography, direct droplet deposition, additive friction stir deposition, and fused deposition modeling. For example, selective laser sintering uses a laser to selectively melt and fuse a current layer L(i) of powdered thermoplastic, ceramic, or metal stock material of component C by scanning cross-sections, derived from the geometric model, on the top surface of the component (i.e., on top of previously formed ones of layers L, e.g., layers L(i−1), L(i−2), and so on). After current layer L(i) is formed in this manner, a new or next layer L(i+1) of powdered stock material is applied on top of the current layer, and, in the absence of layer defects detected using UTA 100, as described below, the layer forming process is repeated until component C is fully formed in accordance with the geometric model. Deposition device 2512 may include a powder delivery mechanism or nozzle to deliver the powder to build plate 2510 (or to the top of partially built component C) and a laser to melt the powder, AMA controller 2516 may provide to UIS 100 an indication of when the layer has been completed to enable the UIS to perform inspection of the layer at that time. AMA controller 2516 may also provide to UIS 100 various information relating to the geometrical component model, such as expected component thickness at various stages of the AM build, and so on.

As AMA 2502 forms layers L of component C during the AM build, UIS 100 inspects the layers L, layer-by-layer, using ultrasonic chirp pulses. The ultrasonic inspection of component C takes measurements after each layer is formed, and uses one or more of a frequency signature, a resonance behavior, and a time domain analysis to evaluate a quality of a last build layer to generate meaningful quality control and feedback control during AM build. To this end, UIS 100 includes ultrasonic transmit transducer TT and ultrasonic receive transducer RR of UIS 100 fixed in contact with/coupled to bottom surface 2510B of build plate 2510 adjacent to each other at a position, generally indicated at P, directly beneath component C. An example separation distance between the adjacent pair of transducers may be approximately 0.1 of an inch, and position P may coincide with a midway point between the transducers. Thus, ultrasonic transmit and receive transducers TT and RT represent a co-located/adjacent pair of transmit and receive transducers coupled to bottom surface 2510B at position P. Position P is also referred to as a “generalized position P” because it places/refers to both the transmit and the receive transducers, generally.

At a high-level, AMA 2502 and UIS 100 form and inspect layers L of component C, individually, as follows. AMA 2502 forms a current layer on top of component C (which includes the current layer and any intervening layers previously formed on build plate 2510 beneath (i.e., underneath) the current layer, i.e., between the current layer and top surface 2510B of the build plate). Before AMA 2502 forms a next layer on top of the current layer, UIS 100 causes transmit transducer TT to deliver an ultrasonic chirp pulse T to component C through build plate 2510. In response, receive transducer RT receives ultrasonic energy including reflections R of the ultrasonic chirp pulse from the build plate and from the component, through the build plate. In the arrangement of FIG. 1, the ultrasonic chirp pulse T and reflections R travel to and from component C in opposite vertical (up and down) directions that are substantially perpendicular to horizontal directions of planar surfaces 2510T and 2510B of build plate 2510. Controller 105 processes the reflections using techniques described above in connection with FIGS. 1-23, and techniques described below, to produce an ultrasonic signature indicative of one or more characteristics of the current layer (and the layers of component C in general). Such characteristics include, but are not limited to, voids or porosity of one or more of layers L, a thickness of the component, or layer-to-layer fusion.

The ultrasonic signature may include spaced-apart reflection time peaks and/or frequency spectrums including frequency resonance peaks. Controller 105 may further process the ultrasonic signature to determine/compute the one or more characteristics. UIS controller 105 and AMA controller 2516 may communicate with each other over a communication link connecting the two controllers, in order to help control the layer forming and inspecting processes described herein. The layer forming and inspecting processes are described below in further detail in connection with FIGS. 26 and 27.

With reference to FIG. 26, there is a flowchart of an example method 2600 of forming and inspecting layers L of component C layer-by-layer (i.e., individually) during the AM process, performed by AMA 2502 and UIA 100.

At 2602, AMA 2502 forms a layer (also referred to as a “current layer”) of component C on build plate 2510 (which may be an initial/first layer, or a subsequent layer) using an AM technique, as described above.

After AMA 2502 has formed the layer at 2602, but before the AMA forms a next layer on top of the layer, UIA 100 inspects/monitors the layer as described in next operations 2604-2616. Layer inspection operations 2604-2616 may be performed as an independent process.

At 2604, UIA 100 generates an ultrasonic chirp pulse having a frequency that sweeps through a range of ultrasonic frequencies across a chirp bandwidth. In an example, the range of ultrasonic frequencies may be 6 to 13 MHz. Ultrasonic transmit transducer TT delivers the ultrasonic chirp pulse to the layers of component C through build plate 2510.

At 2606, UIA 100, via ultrasonic receive transducer RT, receives from build plate 2510 and component C ultrasonic energy including reflections of the ultrasonic chirp pulse. The reflections may include a series of time-overlapping reflections of the ultrasonic chirp pulse from the various structures through which the ultrasonic chirp pulse propagates, including build plate 2510, and layers of component C including the layer formed at 2602 and any intervening layers.

At 2607, UIA 100 (e.g., controller 105) processes the ultrasonic energy including the reflections to produce, and store in memory, processing results in the form of an ultrasonic signature indicative of one or more characteristics of the layer and the intervening layers of component C. The reflections may include time-overlapping reflections. UIA 100 performs time domain processing of the ultrasonic energy to compress the time-overlapping reflections into respective time-separated reflection time peaks that represent the aforementioned ultrasonic signature. The time domain processing includes filtering the series of time-overlapping reflections using a filter response matched to a chirped frequency characteristic of the ultrasonic chirp pulse. Time separation between various ones of the reflection time peaks indicate a thickness of build plate 210 and a thickness of component C, including all of the layers (including the current layer) of the component. Additionally, the number and shape of the reflection time peaks indicates other characteristics of the layers of component C, such as consistency of layer thickness and/or velocity. UIA 100 additionally, or alternatively, performs frequency domain processing of the ultrasonic energy to produce a frequency spectrum including frequency resonance peaks for component C that represent the ultrasonic signature. The frequency domain processing may include performing a first Fourier transform on the ultrasonic energy, and, optionally, performing a second Fourier transform on the first Fourier transform. Resonance behavior of component C as well as wavelength interaction in the frequency spectrum indicate characteristics of the layers of the component, such as property and defect state. UIA 100 may perform the time domain processing and the frequency domain processing sequentially or concurrently, to produce results of the time and frequency domain processing sequentially or concurrently, respectively.

At 2608, UIA 100 displays results of the time processing, e.g., reflection time peaks, on a time domain plot. UIA 100 also displays results of the frequency domain processing, e.g., frequency resonances, on a frequency domain plot.

At 2610, one or more characteristics of one or more of the layers of component C are determined based on the ultrasonic signature. In an example, the thickness of component C is computed based on the time separation between various ones of the reflection time peaks of the ultrasonic signature. UIA 100 may determine the one or more characteristics, or a user may determine the characteristics based on the results displayed at operation 2608, or as otherwise presented to the user. UIA 100 may display the determined characteristics, and may also communicate the determined characteristics to AMA 2502.

Layer defect detection and remediation may optionally be performed at next operations 2612-2616.

At 2612, it is determined whether the characteristic as determined at 2610 (i.e., one of the one or more characteristics as determined) represents a defect in the layer (or of component C, generally). For example, it is determined whether the characteristic as determined matches an expected/predetermined characteristic within a predetermined tolerance. To do this, the characteristic as determined may be compared against a predetermined expected (acceptable) range of the characteristic. If the comparison indicates that the characteristic as determined falls within the acceptable range, the layer is deemed not to be defective, i.e., is of acceptable quality. On the other hand, if the comparison indicates that the characteristic falls outside of the acceptable range, the layer is deemed to be defective, i.e., of unacceptable quality. For example, the determined ultrasonic signature of component C may be compared to a range of acceptable ultrasonic signatures expected at the current iteration of the AM processing (the current stage of layer-forming), where the range is in accordance with the current cross-section layer of the geometrical model that the layer represents. Such information may be provided from AMA controller 2516 to UIS controller 105, and then used by the UIS controller to determine whether there is a defect.

At 2614, if it is determined that the characteristic as determined represents a defect in the layer (i.e., a defective layer), corrective action may be taken. For example, UIS controller 105 provides feedback to AMA controller 2516 indicating the defect. Responsive to the feedback, AMA 2502 may alter layer forming settings (e.g., laser power) used by the additive manufacturing to form the subsequent layer(s), and form the subsequent layer(s) with the altered layer forming settings, to mitigate/remediate the defective layer. Detecting and correcting the defect early on when the layer is first formed (and then remediated) saves time and money that would have otherwise been wasted had a later inspection discovered the defect only after component C had been fully formed.

At 2616, if it is established that the determined characteristic does not represent a defect in the layer, no corrective action is taken, and the AM process continues to the next layer-forming iteration, i.e., to form the next layer on the (current) layer.

In operations 2614 and 2616, UIS 100 may display an indication of whether the layer is defective layer or whether the layer is not-defective, and may provide the indication to AMA controller 2516 of AMA 2502. Armed with the indication, AMA 2502 may then act accordingly under control of AMA controller 2516 or a user/operator, e.g., to take corrective action to remediate a defective layer, if necessary. In this way, UIS 100 may cause a defective layer to be remediated.

At 2618, the above-described layer forming operation 2602 and the individual layer inspection operations 2604-2616 are repeated for next/subsequent layers of component C in order to build-up, inspect, optionally remediate, and store an electronic signature history for the component layer-by-layer during the AM process, until the component is fully formed according to the component model.

Variations of method 2600 may be employed, as described below.

Method 2600 inspects component C with ultrasonic transducers TT and RT positioned generally at position P (as shown in FIG. 25) to produce an indication/measurement of one or more localized characteristics of component C that coincide with that position, because the ultrasonic energy delivered to and reflected from the component interacts with that portion of the component. Another embodiment, referred to as a “multi-position inspection” embodiment, inspects component C with multiple pairs of ultrasonic transducers (i.e., multiple pairs of transducers TT and RT) sequentially positioned at different spaced-apart (generalized) positions across component C, which produces sequential indications/measurements of the one or more characteristics at the different positions, as opposed to a single measurement. The different measurements of the one or more characteristics (e.g., thickness) may be compared to each other to reveal uniformity or non-uniformity of the characteristic across component C. Non-uniformity of characteristics (e.g., thickness) across component C may represent a defect that needs to be corrected.

Method 2600 may be modified as follows to implement the multi-position inspection embodiment (also referred to more simply as “multi-position inspection”). Multi-position inspection performs a first iteration, a second iteration, a third iteration, and so on, of operations 2604-2610 at sequential times t1, t2, t3, and so on, with transducer pairs at a first position P1, a second position P2, and a third position P3, and so on, beneath component C, to produce (e.g., determine) a first measurement of a characteristic (e.g., the ultrasonic signature) at the first position, a second measurement of the characteristic at the second position, a third measurement of the characteristic at the third position, and so on, respectively. The iterations of operations 2604-2610 may be performed concurrently rather than sequentially, for example, in an arrangement in which the multiple transducer pairs belong to multiple respective UISs, one per transducer pair. The multi-position inspection may then determine whether the first measurement, the second measurement, the third measurement, and so on, all match a predetermined (expected) acceptable range of the characteristic (e.g., the ultrasonic signature). If all of the measurements do not match, the multi-position inspection may declare a defective layer, and perform a remediation (e.g., adjust parameters to mitigate the defective layer). Otherwise, the multi-position inspection may declare that the layer is not defective, and remediation is not necessary.

With reference to FIG. 27, there is an illustration of an example of transducer pair placement on plate bottom 2510B of build plate 2510 in an example of multi-position inspection in which transducer pairs are positioned at P1 (center of component C), P2 (left edge of the component), and P3 (right edge of the component), to produce (by respective iterations of the set of operations 2604-2610) component measurements M1 (center), M2 (left), and M3 (right), at those positions, respectively.

Another arrangement referred to as “pitch with one, catch with many,” may include transmit transducer TT positioned at a transmit transducer position beneath component C, and multiple receive transducers RTs positioned at respective receive transducer positions near, but spaced-apart from, each other and the transmit transducer position. For example, the transmit transducer TT may be positioned centrally beneath component C, while the multiple receive transducers RTs may be positioned in equally spaced-apart positions surrounding the transmit transducer. In the pitch one, catch with many arrangement, transmit transducer TT transmits the chirp pulse. Then, multiple receive transducers RTs receive, at their respective spaced-apart positions, ultrasonic energy including reflections of the ultrasonic chirp pulse from build plate 2510 and component C.

In summary, in one form, a method is provided comprising: inspecting a layer of a component formed on a build plate using an additive manufacturing process, the inspecting including: delivering to the layer, through the build plate and any intervening layers of the component previously formed on the build plate, an ultrasonic chirp pulse having a frequency that sweeps through a range of frequencies across a chirp bandwidth; receiving from the layer ultrasonic energy including reflections of the ultrasonic chirp pulse delivered to the layer; and processing the ultrasonic energy to produce an ultrasonic signature indicative of a characteristic of the layer and the intervening layers; and repeating the inspecting for each of subsequent layers of the component formed on the build plate according to the additive manufacturing process, to inspect the component layer-by-layer as the component is built-up during the additive manufacturing process.

In summary, in another form, an ultrasonic inspection system (UIS) is provided comprising: an ultrasonic transmit transducer to deliver to a layer of a component formed on a build plate using an additive manufacturing process, through the build plate and any intervening layers of the component previously formed on the build plate, an ultrasonic chirp pulse having a frequency that sweeps through a range of frequencies across a chirp bandwidth; an ultrasonic receive transducer to receive from the layer ultrasonic energy including reflections of the ultrasonic chirp pulse delivered to the layer; and a controller to process the ultrasonic energy to produce an ultrasonic signatures indicative of a characteristic of the layer and the intervening layers; wherein the ultrasonic transmit transducer, the ultrasonic receive transducer, and the controller are configured to repeat respective operations for each of subsequent layers of the component formed on the build plate, to inspect the component layer-by-layer as the component is built up during the additive manufacturing process.

In summary, in yet another form, a non-transitory processor readable medium is provided. The processor readable medium stores instructions that, when executed by a processor, cause the processor to perform the methods described herein.

The above description is intended by way of example only. Various modifications and structural changes may be made therein without departing from the scope of the concepts described herein and within the scope and range of equivalents of the claims.

In-Situ Monitoring of Additive Manufacturing Layer-by-Layer Build Using Ultrasonic Spectroscopy

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

International Classifications

Abstract

Description

Claims