Improvements in and relating to ultrasound probes - Patents.com

Abstract

A method and apparatus for providing inspection of a weld is disclosed, the method including the steps of preparing a weld inspection device in proximity to a weld on a substrate to be inspected, performing the inspection, where the substrate is subjected to an elevated temperature condition above ambient temperature by heating during the inspection, the performing the inspection including emitting ultrasonic waves within a volume of the substrate and the weld and receiving at least a portion of the ultrasonic waves from the substrate and the weld, thereby obtaining a plurality of signal sets, and processing one or more of the plurality of signal sets to provide weld inspection data, the processing including correcting for temperature distribution within the volume of the substrate and / or the weld at elevated temperatures.

Description

[Technical field]

[0001] The present disclosure relates to improvements in and relating to ultrasonic probes, their signal processing and compensation, and methods of use in non-destructive ultrasonic testing, particularly but not exclusively in high temperature configurations thereof. [Background technology]

[0002] Ultrasonic testing is used for non-destructive testing of a variety of objects. A transmitting transducer emits ultrasonic waves that enter the object, interact with the object and its sub-features, and then return to the receiving transducer. The temperature of the object affects the speed of sound within the object and can therefore adversely affect the quality of the test and imaging.

[0003] When testing welds, it is desirable to be able to perform ultrasonic testing after the weld is completed or after each individual weld pass, immediately after the weld or weld pass is made, thereby minimizing the time it takes to perform, test, and correct any corrections required for the weld or weld pass. However, this means that the ultrasonic inspection is performed while the object or substrate is hot and has temperature variations between portions of the object or substrate. Summary of the Invention

[0004] One potential objective of the present disclosure is to provide an ultrasonic probe and signal processing thereof that accounts for the effects of temperature and temperature variations with location within a substrate, providing accurate images over a wide range of temperatures and accounting for different temperatures at different locations, especially in high temperature situations.

[0005] According to a first aspect of the present disclosure, there is provided a method of providing inspection of a weld, comprising: (a) providing a weld inspection device proximate to a weld on a substrate to be inspected; (b) performing an inspection, the substrate being subjected to an elevated temperature above ambient temperature by heating during the inspection, the performing of the inspection comprising: a. emitting ultrasonic waves within a volume of the substrate and the weld; b. receiving at least a portion of the ultrasonic waves from the substrate and the weld, thereby obtaining a plurality of sets of signals; conducting an inspection, (c) processing one or more of the plurality of sets of signals to provide weld inspection data; Including, The process provides a method for providing inspection of a weld that includes compensation for temperatures within the substrate and / or weld volume at elevated temperatures.

[0006] The method may include an elevated temperature that is consistent across the substrate and / or includes a temperature distribution, such as a temperature gradient, in the substrate and / or the weld and / or the weld inspection apparatus.

[0007] The treatment involves the correction of temperature distributions, eg temperature gradients, within the volume of the substrate and / or weld at elevated temperatures.

[0008] The method may provide that the correcting includes correcting a path of at least a portion of the ultrasonic wave through a volume of the weld inspection device and / or the substrate and / or the weld to provide a corrected path.

[0009] The method may define a portion of an ultrasonic wave having a path characteristic in one or more elements through which the path passes. The method may define a portion of an ultrasonic wave in a weld inspection device having a path characteristic in one or more media elements. The method may define a portion of an ultrasonic wave in a substrate having a path characteristic in one or more substrate elements. The method may define a portion of an ultrasonic wave in a weld having a path characteristic in one or more weld elements. The method may define a portion of an ultrasonic wave at an interface between a weld inspection device and a substrate having a path characteristic in one or more interface elements.

[0010] The ultrasonic wave may have a path characteristic in each of a plurality of elements, for example, in each element adjacent to a preceding element and a succeeding element. The ultrasonic wave may have a path characteristic in each of one or more media elements. The ultrasonic wave may have a path characteristic in each of one or more interface elements. The ultrasonic wave may have a path characteristic in each of one or more substrate elements. The ultrasonic wave may have a path characteristic in each of one or more welding elements.

[0011] The ultrasonic wave may have a path characteristic in each of one or more media elements and then in each of one or more substrate elements. The ultrasonic wave may have a path characteristic in each of one or more interface elements, e.g., after a media element and / or before a substrate element. The ultrasonic wave may have a path characteristic in each of one or more welding elements, e.g., after a substrate element.

[0012] The ultrasonic wave may have a path characteristic in each of one or more substrate elements, for example on a return path, followed by a path characteristic in each of one or more media elements. The ultrasonic wave may have a path characteristic in each of one or more welding elements, for example before a substrate element. The ultrasonic wave may have a path characteristic in each of one or more interface elements, for example after a substrate element and / or before a media element.

[0013] The method may provide that a portion of the ultrasonic wave has path characteristics within an element of the weld inspection apparatus, e.g., a media element, and / or has path characteristics within an interface, e.g., an interface element, before the portion of the ultrasonic wave enters a first element of the substrate, e.g., a first substrate element. The method may provide that the first element has a temperature within a temperature distribution, such as a temperature gradient across a volume of the substrate and / or weld, with a temperature-compensated path characteristic determined for a portion of the ultrasonic wave within the first element, the temperature-compensated path characteristic being based on a temperature change between an element of the weld inspection apparatus, e.g., a media element and / or an interface element, and the first element of the substrate, e.g., a first substrate element.

[0014] The change in path characteristics can be calculated according to Snell's law of refraction.

[0015] The method may provide that a portion of the ultrasonic wave has a path characteristic within a first element of the substrate, e.g., the first substrate element, before the portion of the ultrasonic wave enters a second element of the substrate, e.g., the second substrate element. The method may further provide that the first element has a temperature within a temperature distribution, such as a temperature gradient across a volume of the substrate and / or weld, and the second element has a temperature within a temperature distribution, such as a temperature gradient across a volume of the substrate and / or weld, and a temperature corrected path characteristic is determined for a portion of the ultrasonic wave within the second element, e.g., the second substrate element, and the temperature corrected path characteristic is based on a change in temperature between the first element and the second element, e.g., between the first substrate element and the second substrate element.

[0016] The method may provide that a portion of the ultrasonic wave has a path characteristic within a first further element of the substrate, e.g., the first further substrate element, before the portion of the ultrasonic wave enters a second further element of the substrate, e.g., the second further substrate element. The method may further provide that the first further element has a temperature within a temperature distribution, such as a temperature gradient across a volume of the substrate and / or weld, and the second further element has a temperature within a temperature distribution, such as a temperature gradient across a volume of the substrate and / or weld, and a temperature corrected path characteristic is determined for a portion of the ultrasonic wave within the second further element, e.g., the second further substrate element, and the temperature corrected path characteristic is based on a change in temperature between the first further element and the second further element, e.g., between the first further substrate element and the second further substrate element.

[0017] The method may provide for determining a temperature compensated path characteristic for each element through which a portion of the ultrasonic wave passes within the weld inspection device and / or substrate and / or weld, e.g., each element through which a portion of the ultrasonic wave passes within the weld inspection device and / or substrate and / or weld.

[0018] The method may provide that the change in temperature is expressed as the change in speed of sound between the speed of sound of one element and the speed of sound of the next element.

[0019] The method may provide that multiple different portions of the ultrasonic wave within the weld inspection device and / or within the volume of the substrate and / or within the weld are compensated, for example, throughout the weld inspection device and / or substrate and / or weld, to provide a compensated path.

[0020] The method may provide multiple compensation passes, e.g., through the weld inspection device and / or the substrate and / or the weld, and / or back again, etc. The method may provide at least 5 compensation passes, in some cases at least 15 compensation passes, in some cases at least 25 compensation passes, and optionally at least 40 compensation passes, e.g., 64 compensation passes.

[0021] For the path of each beam of ultrasound emitted by the transducer, a corrected path may be provided, for example a 64 element phased array.

[0022] The method may provide that a region of interest is selected, the region of interest being within a volume of the substrate and weld through which ultrasonic waves have passed, the region of interest being subdivided into locations such as pixels. The method may apply signal correction to the locations such as pixels. The method may apply signal correction to each location such as pixel, beam path to a location such as pixel, and / or return beam path within the region of interest.

[0023] The method may provide that the signal correction is determined according to a relationship between a location, such as a pixel, and at least one, preferably at least one pair, of the corrected paths. The relationship may be a geometric relationship. The relationship may be a weighted correction based on a relative geometric position with respect to one or more corrected paths.

[0024] The method may provide that the signal correction is determined according to a relationship between a location, such as a pixel, and at least one, preferably at least one pair of positions on at least one, preferably at least one pair of corrected paths. The relationship may be a geometric relationship. The relationship may be a weighted correction based on a relative geometric position, such as a distance, between the location and at least one, preferably at least one pair of positions on at least one, preferably at least one pair of corrected paths.

[0025] The method may include signal correction for a corrected path for an emitted ultrasound beam that passes through one or more locations, such as pixels. The signal correction may be performed primarily or exclusively based on the corrected path for the emitted ultrasound beam that passes through the locations, such as pixels. The signal correction may be based on a relationship that is a weighted correction based on a relative geometric position, such as a distance between a location and at least one, preferably at least one pair of positions on the corrected path, and the signal correction may be primarily or exclusively based, possibly potentially, on the corrected path for the emitted ultrasound beam that passes through the locations, such as pixels.

[0026] The method may include signal correction for one or more locations, such as pixels not traversed by a corrected path for the emitted ultrasound beam. The method may include signal correction for locations, such as pixels not traversed by a corrected path for the emitted beam, based on calculated or observed corrections for locations traversed by the corrected path for the emitted beam. The signal correction may be performed based on calculated or observed positions of multiple locations, e.g., four locations, traversed by the emitted beam.

[0027] The signal correction may be performed based on calculated or observed positions of a plurality of locations on the first emission beam and a plurality of locations on the second emission beam. The first and second emission beams may be adjacent beams in a set of beams. The first beam may be on one side of the location requiring signal correction and the second beam may be on the other side of the location requiring signal correction.

[0028] The signal correction for a location may be a weighted combination of signal corrections for one or more other locations, such as one or more other locations through which the emitted beam has passed. The weighted combination may be based on four locations. The weighted combination may be based on two locations on one beam and two locations on another beam.

[0029] The signal correction for a location may be weighted according to the ratio of the distance between the location on the first beam and the location on the second beam where the location occurs. The signal correction for a location may be weighted according to the ratio of how far the location is from the location on the first beam and how far the location is from the location on the second beam, e.g., weighting the correction from those locations. The distance may be considered along an equidistant arc from the transducer that passes through the first beam, the location to be corrected, and the second beam.

[0030] The signal correction for a location may be weighted according to the ratio of the distance between a first location on the first beam where the location occurs and a second location on the first beam. The signal correction for a location may be weighted according to the ratio of how far the location is from a first location on the first beam and how far the location is from a second location on the first beam, e.g., weighting the correction from those locations. Distance may be considered along the first beam.

[0031] Distance values ​​for each location, such as a pixel, within a region of interest or within a substrate may be pre-calculated and stored.

[0032] The speed of sound at each location, such as a pixel, within a region of interest, or even within a substrate, may be calculated.

[0033] The beam path through each boundary between locations, such as pixels, within a region of interest or within a substrate may be calculated according to, for example, Snell's law of refraction.

[0034] Temperatures may be mapped to locations, such as pixels, within a region of interest or substrate. The region of interest or substrate may be assigned temperature contours, for example, perpendicular to the surface of the substrate.

[0035] The method may include determining a transit time through each location determined by, for example, distance and speed within the zone. The method may reveal a net transit time as the sum of travel times through all locations on the route.

[0036] The method may provide that the result set includes one or more measured representations of the shape of the substrate and / or the weld groove and / or the weld, and the method further includes comparing the measured shape representation with a modeled shape representation, and accepting imaging of the region of interest if the comparison of the measured shape representation with the modeled shape representation reveals that the measured shape representation sufficiently matches the modeled shape representation.

[0037] The method may provide that the result set includes one or more measured representations of the geometry of the substrate and / or the weld groove and / or the weld, and the method further includes comparing the measured representations of the geometry with a modeled representation, and if the comparison between the measured representations of the geometry and the modeled representation reveals that the measured representation of the geometry is an insufficient match to the representation of the modeled geometry, a temperature distribution is predetermined to be used for correcting the temperature distribution within the volume of the substrate and / or weld at elevated temperatures.

[0038] The method may include providing a thermal model and generating a modeled temperature distribution location at an elevated temperature for at least a portion of the substrate using the thermal model, the portion of the substrate including a volume. The method may include measuring temperature at a plurality of locations with the substrate at the elevated temperature to obtain a measured temperature distribution location, and may further include comparing the measured temperature distribution location to the modeled temperature distribution location.

[0039] The method may provide that if comparison of the measured and modeled temperature distribution locations reveals that the modeled temperature distribution does not adequately fit the measured temperature distribution, then modifying the thermal model and / or the modeled temperature distribution locations and then re-comparing.

[0040] The method may provide that if the comparison of the measured temperature distribution location with the modeled temperature distribution location reveals that the modeled temperature distribution is a sufficient match to the measured temperature distribution, calculating properties of ultrasonic waves emitted during passage through at least a portion of the substrate and / or weld.

[0041] This first aspect of the disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the disclosure.

[0042] According to a second aspect of the present disclosure, there is provided an apparatus for providing inspection of a weld, the apparatus comprising: (a) a weld inspection device that, in use, is located adjacent to a weld on a substrate to be inspected and that, during inspection, heats the substrate and thereby subjects it to an elevated temperature above ambient temperature; (b) Welding inspection equipment: (1) Welding by emitting ultrasonic waves into the volume of the substrate and the weld; (2) receiving at least a portion of the ultrasonic waves from the substrate and the weld, thereby obtaining a plurality of sets of signals; (3) a weld inspection apparatus including one or more processors, the processor including an input for one or more of the plurality of sets of signals, the processor providing a correction for the input one or more of the plurality of sets of signals to provide corrected weld inspection data, the processor applying a correction for a temperature within a volume of the substrate and / or weld at an elevated temperature; An apparatus is provided.

[0043] This second aspect of the disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the disclosure.

[0044] According to a third aspect of the present disclosure, there is provided a method for determining one or more dimensions of one or more substrates to be welded, comprising the steps of: (i) providing a sizing device that emits ultrasonic waves; (ii) introducing one or more substrates to be welded into a dimensioning device; (iii) performing one or more dimensional determinations on the one or more substrates using ultrasound; (iv) using the one or more dimensioning determinations in one or more further operations performed on the one or more substrates; A method is provided.

[0045] The method may be part of a welding method and may include the step of providing a welding apparatus. The method may be an arc welding method. The method may include introducing one or more substrates to be welded into the welding apparatus. The one or more further operations may be part of the welding process and / or may be part of a post-weld process.

[0046] The ultrasonic waves may be provided by one or more beams emitted from the device to one or more substrates. The ultrasonic waves may be provided by multiple beams. The ultrasonic waves may be provided by multiple beams that extend in an arc. The arc may extend over at least 5°, in some cases at least 10°, in some cases at least 15°. The arc may extend over less than 40°, for example less than 30°, or in some cases less than 25°.

[0047] The one or more dimensions of the one or more substrates may be a distance between a first surface and a second surface, such as a thickness. The thickness may be established in the direction of the one or more beams of ultrasound. The thickness may be established perpendicular to the one or more surfaces of the one or more substrates.

[0048] The sizing device may also be a weld inspection device. The sizing device and / or the weld inspection device may use ultrasonics.

[0049] The method may include using heating to raise the temperature of one or more substrates above ambient temperature.

[0050] A step of raising the temperature of the one or more substrates above ambient temperature using heating may be provided prior to performing one or more dimensioning on the one or more substrates. A step of raising the temperature of the one or more substrates above ambient temperature using heating may be provided after performing one or more dimensioning on the one or more substrates.

[0051] The method may include performing one or more dimensionings on one or more of the substrates at a plurality of locations on the substrate. The method may include moving, e.g., rolling, a dimensioning device between the plurality of locations on the substrate. The plurality of locations on the substrate may follow an intended trajectory of the weld and / or a formed trajectory of the weld.

[0052] One or more dimensions may be measured at an angle relative to the substrate. The angle relative to the substrate may be considered as an angle relative to the first surface and / or the second surface of the substrate. The angle may be 90°+ / -25°, such as 90°+ / -20°, or in some cases 90°+ / -15°.

[0053] One or more dimensions may be measured at a range of angles relative to the substrate. The range of angles relative to the substrate may be considered as a range of angles relative to the first surface and / or the second surface of the substrate. The range of angles may be 90°+ / -up to 25°, for example 90°+ / -up to 20°, possibly 90°+ / -up to 15°, particularly 90°+ / -up to 10°.

[0054] The one or more dimensions may be measured perpendicular to one or more of the first surface and / or second surface of the substrate.

[0055] The method may include correcting the thickness measurements to account for the effect on the ultrasonic waves of, for example, temperature and / or temperature gradients within the substrate in each of the ultrasonic beam paths.

[0056] The method may include correcting the thickness measurements to account for, for example, the effects of refraction of the ultrasound in each ultrasound beam path.

[0057] The method may include providing a weld inspection device.

[0058] The method may include inspecting the produced weld using a weld inspection device. The method may include inspection of the weld provided by the inspection device at an inspection location on the one or more substrates, the inspection location at an elevated temperature above ambient temperature.

[0059] The one or more dimensions may include dimensions of the substrate along one or more paths through the substrate that are inspected and / or being inspected by the weld inspection device.

[0060] A combined dimensioning and weld inspection system may perform a dimensioning of the substrate, followed by one or more weld inspections. A combined dimensioning and weld inspection system may perform one or more dimensioning, followed by one or more weld inspections, further dimensioning, and possibly further weld inspections. A cyclical approach of dimensioning, weld inspection, dimensioning, and weld determination may be used.

[0061] The method may include using the one or more dimension determinations in one or more further operations performed on the one or more substrates by using the one or more dimension determinations in adjusting and / or correcting the weld inspection. The method may include adjusting a depth of focus. The method may include adjusting an image analysis area, e.g. an image analysis area considered for inspection. The method may include using the one or more dimension determinations in adjusting and / or correcting a position determined by the weld inspection. The method may include using the one or more dimension determinations in adjusting and / or correcting a depth into the substrate determined by the weld inspection.

[0062] This third aspect of the disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the disclosure.

[0063] According to a fourth aspect of the present disclosure, there is provided an apparatus for determining dimensions in a weld as needed, the apparatus comprising: (i) Welding equipment as required; and (ii) a sizing device; and Equipped with The element of the dimensioning apparatus includes an ultrasonic transmitter and receiver, the element of the dimensioning apparatus being provided with a substrate contacting surface, the element of the dimensioning apparatus providing an output data signal to another element of the dimensioning apparatus including a determiner that determines one or more dimensions of the substrate, the determiner providing the one or more dimensions to another processor that uses the one or more dimensions in one or more further operations performed in relation to the one or more substrates.

[0064] The apparatus may further include a weld inspection apparatus.The apparatus may include a combined sizing apparatus and weld inspection apparatus.

[0065] The determiner and another processor may be part of a common data processing component that combines the dimensioning apparatus and the weld inspection apparatus.

[0066] This fourth aspect of the disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the disclosure.

[0067] According to a fifth aspect of the present invention there is provided a method of inspecting a weld, the method comprising: (a) providing a weld inspection device proximate to a weld on a substrate to be inspected; (b) emitting ultrasonic waves to the substrate and the weld and receiving returning ultrasonic waves from the substrate and the weld, thereby obtaining a plurality of signal sets including a first signal set and a second signal; (c) processing at least the first set of signals and the second set of signals to provide weld inspection data; the first signal set includes a first series of first data elements, the first series of first data elements spanning a first time period, each first data element including a first variable value of a variable and a first time value of the first variable value, the first time period including a first sub-period of time; the second signal set includes a second series of second data elements, the second series of second data elements spanning a second time period, each second data element including a second variable value of the variable and a second time value of the second variable value, the second time period including a sub-period of the second time; the first time sub-period and the second time sub-period span the same time values; Processing of the first and second signal sets determines a modified signal set that contributes to the weld inspection data, the modified signal set including modified time subperiods that span the same time values ​​as the first and second time subperiods, and including, for a time value, a representation of the first variable value for that time value and a representation of the second variable value for that time value if the first and second variable values ​​are within a predetermined relationship to each other, and / or, for a time value, one of the representations of the first variable value for that time value or a representation of the second variable value for that time value if the first and second variable values ​​are not within the predetermined relationship to each other.

[0068] The first variable value may for example be a representation of the amplitude, in particular the amplitude of the rectified signal set, at a particular time value.

[0069] The second variable value may for example be a representation of the amplitude, in particular the amplitude of the rectified signal set, at a given time value.

[0070] The first time period, the second time period, and / or the one or more additional time periods may be the same time period. The first time period, the second time period, and / or the one or more additional time periods may be + / - 25% of the same time period. The first time period, the second time period, and the one or more additional time periods may be the time period of a received signal for an ultrasound beam or wave.

[0071] The first time sub-period and / or the second time sub-period and / or the one or more further sub-periods may be individual time values. The first time sub-period and / or the second time sub-period and / or the one or more further sub-periods may be a small set of individual time values.

[0072] A first period may be considered as multiple distinct or partially overlapping first sub-periods. A second period may be considered as multiple distinct or partially overlapping second sub-periods. A further period may be considered as multiple distinct or partially overlapping further time sub-periods. Multiple time sub-periods may be paired with each other at the same time value.

[0073] The modified signal set may be adapted in accordance with the method taking into account several time sub-periods.

[0074] The method may provide that in the modified signal set, the representation of the first variable value for its time value and the representation of the second variable value for its time value may be an average of the first variable value and the second variable value. The method may provide that in the modified signal set, the representation of the first variable value for its time value and / or the representation of the second variable value for its time value may be an average of all variable values ​​in a predetermined relationship with further variable values.

[0075] The method may provide that the first and second variable values ​​are within a predetermined relationship to each other if the first and second variable values ​​are within a predetermined threshold. The nature and / or defining function and / or value of the predetermined threshold may be variable. The threshold may change, in particular increase, if the system gain increases and / or if the noise in the signal set increases. The threshold may change, in particular increase, if the interference level in the signal set increases. The nature and / or defining function and / or value of the threshold may be determined in a calibration method.

[0076] The method may provide that the first and second variable values ​​are not within a predetermined relationship to each other if the first and second variable values ​​are outside a further predetermined threshold, possibly the same threshold. The nature and / or defining function and / or value of the predetermined threshold may be variable. Furthermore, the predetermined threshold may change, in particular increase, if the system gain increases and / or if the noise in the signal set increases. Furthermore, the predetermined threshold may change, in particular increase, if the interference level in the signal set increases.

[0077] The method may further provide that if the first and second variable values ​​are not within a predetermined relationship to one another, then one included formula selected from the first and second will be the formula having the lower variable value for that time value. The method may further provide that if the variable value and the variable value being compared are not within a predetermined relationship to each other, then the one formula included will be the formula having the lower variable value for that time value. The lower limit variable value may be the lower limit variable value considered as an absolute value.

[0078] The predetermined threshold may be determined relative to the amplitude of the first and second variable values.

[0079] The amplitude may be a pre-determined amplitude value. The amplitude may be, for example, a predicted amplitude value relative to a predicted noise amplitude. The amplitude may be a defined percentage or multiple of the predicted noise amplitude.

[0080] The amplitude may be an observed amplitude.

[0081] The amplitude may relate to the maximum amplitude observed for a set of variable values ​​that is deemed to be free of interference.

[0082] The amplitude may relate to the minimum value observed at a certain time or period across the entire set of variable values ​​under consideration, for example across the first variable value, the second variable value, and one or more further variable values. The amplitude may be the minimum value plus a factor. This amplitude may be considered as the interference-free threshold.

[0083] The threshold value may be the difference between the observed value and the analyzed value. The observed value and the analyzed value may be amplitudes. The analyzed value may be taken from the minimum value observed at a time or period across the entire set of variable values ​​under consideration, for example across the first variable value, the second variable value, and one or more further variable values. The analyzed value may be the minimum value plus a coefficient.

[0084] The method may further include obtaining one or more additional signal sets as part of the plurality of signal sets.

[0085] The method may further include processing one or more additional sets of signals to provide weld inspection data.

[0086] The method may further include one or more or all of the further signal sets comprising a further series of further data elements, each of the further series of further data elements spanning a further period of time, each of the further data elements comprising a further variable value of the variable and a further time value of the further variable value, each of the further periods comprising a further time sub-period.

[0087] The method may further include one of the further time subperiods spanning the same time values ​​as at least one of the first time subperiod and / or the second time subperiod.

[0088] The method may further include processing one or more or all of the additional signal sets to determine a modified set of signals that contribute to the weld inspection data.

[0089] The method may further include a modified time subperiod that spans the same time values ​​as the further time subperiod and a modified signal set that spans one or both of the first time subperiod and the second time subperiod.

[0090] The method may further include, when the first variable value, the second variable value, and the further variable value are within a predetermined relationship to one another, a modified signal set including, for a time value, a representation of the first variable value for the time value, a representation of the second variable value for the time value, and a representation of the one further variable value for the time value.

[0091] The method may alternatively or additionally include, when the first, second and third variable values ​​are not within a predetermined relationship to one another, for a time value, a correction signal set including one of: a representation of the first variable value for that time value, or a representation of the second variable value for that time value, or a representation of a further variable value.

[0092] Alternatively or additionally, the method further includes, for a time value, a correction signal set comprising two of: a representation of a first variable value for that time value, a representation of a second variable value for that time value, and a representation of a further variable value for that time value, wherein the two variable values ​​have a predetermined relationship to each other.

[0093] The method may alternatively or additionally further include, for a time value, a correction signal set including only one of: a representation of a first variable value for that time value, a representation of a second variable value for that time value, and a representation of one of a further variable value for that time value, where the variable values ​​are not within a predetermined relationship to each other.

[0094] The method may further provide that the modified signal set is compiled from a plurality of modified time sub-periods, each ranging over the same time values ​​as the first time sub-period and / or the second time sub-period and / or the one or more further time sub-periods.

[0095] The fifth aspect of the invention may include any of the features, additions, or possibilities described elsewhere in this document, including other aspects of the disclosure.

[0096] According to a sixth aspect of the present invention there is provided an apparatus for inspecting a weld, the apparatus comprising: (a) A welding inspection device comprising: a. a transmitter of ultrasonic waves to a substrate and weld in use and a receiver of ultrasonic waves returning from the substrate and weld in use, the receiver being connected to a processor to provide the processor with a plurality of acquired sets of signals including a first set of signals and a second set of signals; b. a processor adapted to receive a plurality of acquired sets of signals, the plurality of sets of signals including a first set of signals and a second set of signals; A welding inspection device including: the first signal set includes a first succession of first data elements, the first succession of the first data elements spanning a first time period, each first data element including a first variable value of a variable and a first time value of the first variable value, the first time period including a first time sub-period; the second signal set includes a second succession of second data elements, the second succession of second data elements spanning a second time period, each second data element including a second variable value of a variable and a second time value of the second variable value, the second time period including a second time sub-period, the first time sub-period and the second time sub-period spanning the same time values; The processor is adapted to determine a modified set of signals that contribute to the weld inspection data using the first set of signals and the second set of signals; The processor is adapted to define that, for a modified time subperiod spanning the same time values ​​as the first time subperiod and the second time subperiod, the modified signal set includes, for a time value, a representation of the first variable value for that time value and a representation of the second variable value for that time value if the first variable value and the second variable value are within a predetermined relationship to each other, and / or, for a time value, one of the representations of the first variable value for that time value or a representation of the second variable value for that time value if the first variable value and the second variable value are not within the predetermined relationship to each other.

[0097] The sixth aspect of the invention may include any of the features, additions, or possibilities described elsewhere herein, including other aspects of the disclosure.

[0098] According to a seventh aspect of the present disclosure, there is provided a welding method, comprising: (a) providing a welding device; (b) providing a plurality of sensor types; and (c) defining a first set of welding conditions for the welding process; (d) introducing one or more substrates to be welded into a welding apparatus; (e) performing a weld of one or more substrates; and (f) acquiring data from multiple sensor types during welding; and (g) providing a correlation between data obtained from at least two selected sensor types of the plurality of sensor types; (h) synchronizing one or more data points in the data obtained from one of the at least two selected sensor types with one or more data points of another of the at least two selected sensor types; (i) comparing data obtained from a plurality of sensor types with reference data of one or more sensor types; (j) determining whether the weld is of acceptable or unacceptable quality based on the one or more comparisons; (k) if the quality of the weld is unacceptable, the method includes performing one or more actions; A welding method is provided.

[0099] A seventh aspect of the present disclosure may include any of the features, possibilities, or additions described in the first and / or second aspects of the present disclosure relating to inspection of welds.

[0100] The method may be arc welding.

[0101] The method may provide that the correlation is both a temporal correlation and a location correlation.

[0102] The method may provide that the correlation is a temporal correlation.

[0103] The method may provide that the data from the at least two selected sensor types includes a time of occurrence of one or more data points in the data. The method may further provide that the correlation is a match of a time of occurrence of a data point in the data from the at least two selected sensor types.

[0104] The method may provide that data from at least two selected sensor types includes a plurality of times at which a temporal correlation is provided.

[0105] The method may provide that the data from the at least two selected sensor types includes a time of occurrence of every data point in the data.The method may provide that the data from the at least two selected sensor types includes a time of occurrence of every data point in each range (such as a bandwidth) into which the data is subdivided.

[0106] The method may provide for obtaining the time of occurrence relative to a timestamp introduced into the data from the selected sensor type for each of two or more selected sensor types providing the correlation. The method may provide for the timestamp to be applied, for example, simultaneously with the time of occurrence, such that the time of occurrence is obtained directly from the timestamp at that time. The method may provide for the time of occurrence to be obtained by the elapsed time from a timestamp and the time of the data point, such as, for example, where the timestamp occurs and the time of occurrence is calculated from the elapsed time from the timestamp when the timestamp is reached.

[0107] The time stamp may be applied, for example, by a processor that includes a clock generator and / or is provided with a clock signal.

[0108] The method may provide that the correlation is a positional correlation.

[0109] The method may provide that the data from at least two selected sensor types includes a plurality of locations for which a positional correlation is provided.

[0110] The method may provide that the data from the at least two selected sensor types includes a location of occurrence of one or more data points within the data. The method may further provide that the correlation is a match of a location of occurrence of a data point within the data from the at least two selected sensor types.

[0111] The method may provide that the data from at least two selected sensor types includes the location of occurrence of all data points in the data.The method may provide that the data from at least two selected sensor types includes the location of occurrence of all data points within each range (such as a bandwidth) into which the data is subdivided.

[0112] The method may provide for obtaining the location of occurrence relative to a location stamp introduced into the data from the selected sensor type for each of two or more selected sensor types that provide the correlation. The method may provide for a location stamp to be applied, for example, simultaneously with the location of occurrence, such that the location of occurrence is obtained directly from the location stamp at that time. The method may provide for the location of occurrence to be obtained by the time elapsed from the location stamp and the time of the data point. For example, the location of occurrence is calculated from the time elapsed from the location stamp when the location stamp occurs and when the location of occurrence is reached, and / or from the distance elapsed from the location stamp when the location of occurrence is reached.

[0113] The positional correlation can be relative to a position with respect to one or more substrates being welded, e.g., a position on one or more substrates. Multiple positional correlations for one or more substrates can be provided simultaneously. The position can have a known relationship to where the weld is performed, e.g., relative to the start point of the weld and / or relative to the end point of the weld.

[0114] The positional correlation may be relative to a location relative to the welding device, such as a location on the welding device. Multiple positional correlations may be provided simultaneously for the welding device. One or more locations on the welding device may be proximate to the welding electrode. The location may have a known relationship, in some cases, to where the weld is performed, such as a known relationship to a start point of the weld and / or an end point of the weld, which may be a separation between the electrode and the substrate.

[0115] Both a position relative to one or more substrates to be welded, e.g., a positional correlation to a position on one or more substrates, and a position relative to a welding device, e.g., a positional correlation to a position on a welding device, may be provided.

[0116] The position stamp may be applied by a processor. The processor may receive one or more position signals. The position signals may be emitted by one or more position sensors and / or position indicators. The position sensor may be a vision-based sensor, such as a camera. The position indicator may be a motion detection-based position indicator, such as, for example, an incremental encoder, that provides, for example, the occurrence of motion and / or the direction of motion and / or the position of an encoder. The tracking signal-based position indicator may be a radio frequency identification-based system, possibly having one or more radio frequency identification tags on the substrate and / or welding apparatus.

[0117] The method may include providing at least one type of correlation between data obtained from at least three, possibly at least four, possibly at least five selected sensor types of the plurality of sensor types. The same type of correlation may be provided between at least two of the selected sensor types. The same type of correlation may be provided between all of the selected sensor types.

[0118] The method may include providing at least two different types of correlation between data obtained from at least three, possibly at least four, possibly at least five selected sensor types of the plurality of sensor types. The same different types of correlation may be provided between at least two of the selected sensor types. The same different types of correlation may be provided between all of the selected sensor types. The different types may include a temporal correlation and a positional correlation.

[0119] The method may include synchronizing one or more data points within the data obtained from each selected sensor type. The method may include synchronizing one or more data points within the data obtained from each of at least two, possibly at least three, possibly at least four selected sensor types using the same correlation.

[0120] The method may include synchronizing at least 10% of the data points in the data obtained from at least one of the selected sensor types with at least 10% of the data points in the data obtained from at least another of the selected sensor types.

[0121] The method may provide that at least one of the plurality of sensor types, and possibly at least one of the selected sensor types, is selected from a voltage sensor, a current sensor, a welding arc sound emission sensor, a welding topology sensor, a welding image sensor, and an ultrasonic image sensor.

[0122] The voltage sensor can be one of a plurality of sensor types. The voltage sensor can be one of a selected sensor type.

[0123] The current sensor can be one of a plurality of sensor types. The current sensor can be one of a selected sensor type.

[0124] The welding arc sound emission sensor may be one of a plurality of sensor types. The welding arc sound emission sensor may be one of a selected sensor type. The welding arc sound emission sensor may be an acoustic sensor such as a microphone. The welding arc sound emission sensor may be sensitive to emissions arising from the arc forming the weld, and / or emissions arising from the interaction of the arc with the substrate, and / or emissions arising from the interaction of the arc with the shielding gas. The welding arc sound emission sensor may provide data informing characteristics related to the welding speed and / or the weld location sidewall arc and / or the weld location sidewall melting and / or the shielding gas flow rate.

[0125] The weld topology sensor may be one of a plurality of sensor types. The weld topology sensor may be one of a selected sensor type. The weld topology sensor may be a radiation-based sensor. The weld topology sensor may be an image-based sensor. The weld topology sensor may be a laser sensor. The weld topology sensor may be sensitive to radiation applied to and reflected by the substrate and / or weld. The applied radiation and / or reflected radiation is focused by the weld topology sensor. The weld topology sensor may provide information regarding a characteristic related to a shape of the substrate and / or weld.

[0126] The weld imaging sensor can be one of a plurality of sensor types. The weld imaging sensor can be one of a selected sensor type. The weld imaging sensor can be a camera-based sensor. The weld imaging sensor can provide data informing characteristics regarding a size of a weld pool, a shape of a weld pool, a temperature of a weld pool, a presence of deposits on the weld and / or substrate, a pattern of deposits on the weld and / or substrate, a shape of deposits on the weld and / or substrate, a presence of anomalies on the weld and / or substrate.

[0127] The ultrasonic imaging sensor can be one of a plurality of sensor types. The ultrasonic imaging sensor can be one of a selected sensor type.

[0128] One or more, or all, of the sensor types may provide a raw data set. The method includes processing the raw data set to generate a processed data set.

[0129] The determination of an individual sensor type may consider raw data and / or processed data. The determination of an individual sensor type may consider how far a data point is from a known data point set (e.g., a known data point set that indicates an acceptable weld). The known data point set may be represented by a single element, such as, for example, an average of the known data point set.

[0130] The set of known data points may be represented by a principal component analysis (PCA) model of the sensor type and its signal and / or data values.

[0131] The distance may be the Mahalanobis distance.

[0132] The distance may be expressed as an outlier score.

[0133] The determination may be made based on the distance of the data point relative to a threshold, e.g., a threshold distance. The determination may be made based on whether the data point is within or outside the boundaries of a known data point set.

[0134] This determination may be made using a Mahalambis distance novelty detection model by comparing the data points or signal values ​​to a previously developed principal component analysis (PCA) model for the sensor type and its signal and / or data values.

[0135] The processing of the raw data set may include the application of a denoising algorithm. The raw data set may be processed by splitting it into a series of bandwidths.

[0136] The decision may be made based on subsequent data points and / or signals using a model, such as a PCA model, of acceptable performance.

[0137] As data points and / or data sets determined to be acceptable welds and / or data points and / or data sets determined to be unacceptable welds are added, the models and / or model averages may be recalculated.

[0138] One or more or all of the sensor types may provide a raw data set, and the method includes processing the raw data set to generate a processed data set, the processing using a model. The method may include using different variants of the model for different parts of the welding process. In particular, a variant of the model may be used for one or more weld passes, and a different model may be used for one or more additional weld passes in the overall weld production process. A different variant of the model may be used for each weld pass in the overall weld production process. One or more further different variants may be used for repair welds in the weld production process, such as when a portion of a weld is removed and then rewelded. Another further different variant of the model may be used for each weld pass or weld bead in the overall welding process that is subject to repair welding.

[0139] The model may be an acoustic signal processing model. The model may be a power processing model, such as voltage and / or current and / or output power.

[0140] A model can be composed of several different data sources or sub-models.

[0141] The method may provide that synchronization provides an overall data structure in which data points from at least two selected sensor types are aligned with respect to one another in time and / or position.

[0142] The method may provide that data points from at least two selected sensor types are displayed to a user aligned with respect to time of occurrence and / or location of occurrence.

[0143] The method may provide that data from two or more of the selected sensor types and correlated data from the at least two sensor types is displayed to a user and / or stored.

[0144] The method may provide that a determination that the weld is of acceptable or unacceptable quality and / or one or more actions taken are displayed and / or stored.

[0145] The method may provide for including providing a correlation between data obtained from at least four selected sensor types of the plurality of sensor types.

[0146] The method may include inspection of the entire weld. The method may include inspection of a multi-pass weld, for example, after each pass and before the next pass, before all passes are completed.

[0147] The welding device may be mounted on any autonomous automated positioning system. The welding device may be mounted on a guide rail system or a column and boom system. The welding device may be mounted on a robotic arm, such as a multi-axis arm. The welding device may be an arc welder.

[0148] The weld inspection device may be one of a plurality of sensor types. The weld inspection device may be one of a selected sensor type. The weld inspection device may use ultrasonic, for example, phased array ultrasonic.

[0149] The welding inspection device may be mounted on any autonomous automated positioning system. The welding device may be mounted on a guide rail system or a column and boom system. The welding device may be mounted on a robotic arm, such as a multi-axis arm. The welding inspection device may be a phased array ultrasonic transducer. The welding inspection device may include an ultrasonic transmitter and a receiver. The welding inspection device may include a substrate contacting surface. The welding inspection device may be in physical contact with the inspection location. The substrate contacting surface may be in physical contact with the inspection location.

[0150] The method may include rolling a weld inspection device, e.g., a substrate contacting surface thereof, over a surface of the substrate, e.g., to one side, but optionally parallel to the weld.

[0151] The weld inspection apparatus may be provided with an internal cooling system. The method may include providing a coolant to the weld inspection apparatus and / or removing a coolant from the weld inspection apparatus.

[0152] The method may further provide that at least one sensor type of the plurality of sensor types, possibly at least one of the selected sensor types, is part of a weld inspection apparatus, and the method may include inspecting the weld using the weld inspection apparatus. The method may include inspecting the weld using the weld inspection apparatus to determine the presence or absence of weld defects at a weld location, particularly at a succession of consecutive locations forming the weld. The method may include inspecting the weld using the weld inspection apparatus to determine one or more characteristics of the defects at a location. The method may provide that the characteristics may include one or more of a size, a location, a type of defect, a shape of the defect, or a location of the defect relative to a shape of the weld and / or a length of the weld.

[0153] The method may include determining the presence or absence of a weld defect at the weld location in real time. The method may include determining the presence or absence of a weld defect at the weld location within 100 milliseconds, in some cases within 50 milliseconds, and in some cases within 20 milliseconds of a welding device leaving the weld location. The method may further include comparing the one or more characteristics to one or more criteria, and may further include determining whether a defective weld at a location meets or does not meet the welding criteria.

[0154] The method may include determining in real time whether a weld having a defect at a location meets or does not meet the welding criteria. The method may include determining whether a weld having a defect at a location meets or does not meet the welding criteria within 100 milliseconds, in some cases within 50 milliseconds, and in some cases within 20 milliseconds of the welding device leaving the welding location.

[0155] The method may provide that if the weld meets the welding criteria, a record of the weld is created and stored and may include a location of the defect relative to a geometry of the weld and / or a location of the defect relative to a length of the weld. The method may further provide that the record includes data from one or more of a plurality of sensor types. The method may provide that if the weld does not meet the welding criteria, one or more repair steps are applied to the weld.

[0156] The method may provide that at least one sensor type of the plurality of sensor types, possibly at least one of the selected sensor types, is a weld condition sensor, and the method includes inspecting the weld condition using the weld condition sensor.

[0157] The method may provide for including inspecting the weld to determine one or more parameters of the weld as it is formed.

[0158] The method may provide for including a comparison of the one or more parameters to one or more control parameters, and further including a determination of whether a risk level for a weld defect has been exceeded.

[0159] The method may provide for including one or more actions of changing the welding conditions from a first set of welding conditions of the welding method. The method may provide for changing the welding conditions from the first set of welding conditions to include stopping welding and / or alerting an operator. The method may provide for changing the welding conditions from the first set of welding conditions to include changing the welding conditions back to the first set of welding conditions and / or changing the welding conditions to a second set of welding conditions.

[0160] The welding method may further include inspection of the weld formed by the welding method. The welding method is (i) providing a weld inspection device; (ii) using heat to elevate the temperature of one or more substrates above ambient temperature; (iii) performing welding of the one or more substrates at an elevated temperature above ambient temperature using a welding device; (iv) inspecting the produced weld using a weld inspection device; (v) inspecting the weld at an inspection location on the one or more substrates with an inspection device at an elevated temperature above ambient temperature at the inspection location; It may further include.

[0161] The seventh aspect of the present disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the present disclosure.

[0162] According to an eighth aspect of the present disclosure, there is provided an apparatus for monitoring welding, comprising: (a) multiple inputs for data from multiple sensor types; (b) one or more processors, wherein a processor of the one or more processors: a. Receives input; b. processing the data obtained from at least two selected sensor types from the plurality of sensor types and applying a correlation between the data obtained from the at least two selected sensor types from the plurality of sensor types; c. using the correlation to process the data obtained from at least two selected sensor types of the plurality of sensor types and synchronize one or more data points in the data obtained from one of the at least two selected sensor types with one or more data points in the data obtained from another one of the at least two selected sensor types; one or more processors; (c) one or more outputs for the processed data; and The present invention provides an apparatus comprising:

[0163] The processor may further provide a comparator for receiving and comparing data from one or more of the plurality of sensor types, such as one or more of the selected sensor types, with reference data for one or more of the sensor types.

[0164] The comparator may output a determination of whether the weld is of acceptable or unacceptable quality based on the compared data. If the weld quality is determined to be unacceptable, the apparatus may further provide a control signal to trigger one or more actions by the control unit.

[0165] The apparatus may further comprise a control unit for receiving a control signal, for example, for receiving a first set of welding conditions for the welding method and / or for triggering one or more actions by the control unit, such as modifying the first set of welding conditions or terminating welding.

[0166] It may be provided that the apparatus includes a weld inspection apparatus for determining one or more characteristics of the defect.

[0167] The device may provide input from at least one of a plurality of sensor types, possibly at least one of a selected sensor type, where the sensor type and / or the selected sensor type is selected from a voltage sensor, a current sensor, a welding arc sound emission sensor, a welding topology sensor, a welding imaging sensor, and an ultrasonic imaging sensor.

[0168] The voltage sensor can be one of a plurality of sensor types. The voltage sensor can be one of a selected sensor type.

[0169] The current sensor can be one of a plurality of sensor types. The current sensor can be one of a selected sensor type.

[0170] The welding arc sound emission sensor may be one of a number of sensor types. The welding arc sound emission sensor is one of the selected sensor types. The welding arc sound emission sensor may be an acoustic sensor such as a microphone. The welding arc sound emission sensor may be sensitive to emissions arising from the arc forming the weld and / or emissions arising from the interaction of the arc with the substrate and / or emissions arising from the interaction of the arc with the shielding gas. The welding arc sound emission sensor may provide data informing characteristics related to the weld speed and / or the weld location sidewall arc and / or the weld location sidewall melting and / or the shielding gas flow rate.

[0171] The weld topology sensor may be one of a plurality of sensor types. The weld topology sensor may be one of a selected sensor type. The weld topology sensor may be a radiation-based sensor. The weld topology sensor may be an image-based sensor. The weld topology sensor may be a laser sensor. The weld topology sensor may be sensitive to radiation irradiated onto and reflected by the substrate and / or weld. The irradiated radiation and / or reflected radiation may be focused by the weld topology sensor. The weld topology sensor may provide information regarding a property related to a shape of the substrate and / or weld.

[0172] The weld imaging sensor can be one of a plurality of sensor types. The weld imaging sensor can be one of a selected sensor type. The weld imaging sensor can be a camera-based sensor. The weld imaging sensor can provide data informing characteristics regarding a size of a weld pool, a shape of a weld pool, a temperature of a weld pool, a presence of deposits on the weld and / or substrate, a pattern of deposits on the weld and / or substrate, a shape of deposits on the weld and / or substrate, a presence of anomalies on the weld and / or substrate.

[0173] The ultrasonic imaging sensor can be one of a plurality of sensor types. The ultrasonic imaging sensor can be one of a selected sensor type.

[0174] The apparatus may further provide a comparator including a first comparator for receiving and comparing the one or more characteristics to one or more criteria, the first comparator further providing for outputting a first determination as to whether the defective weld meets or does not meet the welding criteria. The apparatus may further provide a comparator including a second comparator for receiving and comparing one or more parameters of the weld as the weld is formed to one or more control parameters, the second comparator further providing for outputting a second determination as to whether a risk level of a weld defect has been exceeded. The apparatus may provide such that if a risk level of a weld defect is exceeded, a control signal provided by the apparatus is sent to a controller to trigger one or more actions, the one or more actions being to change the welding conditions from a first set of welding conditions of the welding method, e.g., to stop welding, and / or to alert an operator.

[0175] The eighth aspect of the present disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the present disclosure.

[0176] According to a ninth aspect of the present disclosure, there is provided a welding method, comprising: (a) providing a welding device; (b) providing a plurality of sensor types; and (c) defining a first set of welding conditions for the welding process; (d) introducing one or more substrates to be welded into a welding apparatus; (e) performing a weld of one or more substrates; and (f) acquiring data from multiple sensor types during welding; and (g) comparing the data obtained from the plurality of sensor types to reference data of one or more sensor types; (h) determining whether the weld is of acceptable or unacceptable quality based on the one or more comparisons; and Including, (i) the acquired data is acquired by processing data from a plurality of sensor types, the acquired data for at least one of the plurality of sensor types is processed using a computer model, the method including using different variants of the model for different portions of the welding process; A welding method is provided.

[0177] In particular, a variant of the model may be used for one or more weld passes and another model may be used for one or more additional weld passes in the overall weld production process. A different variant of the model may be used for each weld pass in the overall weld production process. One or more further different variants may be used for repair welds in the weld production process, such as removing and then rewelding a portion of a weld. Another further different variant of the model may be used for each weld pass or weld bead in the overall welding process that is subject to repair welding.

[0178] The model may be an acoustic signal processing model. The model may be a power processing model, such as voltage and / or current and / or output power.

[0179] A model can be composed of several different data sources or sub-models.

[0180] The ninth aspect of the present disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the present disclosure.

[0181] According to a tenth aspect of the present disclosure, there is provided an apparatus for monitoring welding, comprising: (a) multiple inputs for data from multiple sensor types; (b) one or more processors, wherein a processor of the one or more processors: a. receiving input from at least one of a plurality of sensor types; b. processing the data to provide acquired data for at least one of a plurality of sensor types, the processor having access to a computer model, the method including using different variants of the model for different portions of the welding process; A processor; (c) one or more outputs for the processed data; and An apparatus comprising:

[0182] The tenth aspect of the present disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the present disclosure.

[0183] According to an eleventh aspect of the present disclosure, there is provided a welding method, comprising: (a) providing a welding device; (b) providing a plurality of sensor types; and (c) introducing one or more substrates to be welded into a welding apparatus; (d) performing a weld of one or more substrates; and (e) acquiring data from multiple sensor types during welding; and (f) comparing the data obtained from the plurality of sensor types to reference data of one or more sensor types; (g) determining whether the weld is of acceptable or unacceptable quality; (h) determining includes considering combined data from at least two of the plurality of sensor types; A welding method is provided.

[0184] The method may provide that combined data is formed and then a decision is made taking into account the combined data.

[0185] The method may provide for comparing data obtained from one or more of a plurality of sensor types to reference data from one or more of the sensor types to initially determine whether the weld is of acceptable or unacceptable quality.

[0186] The method may provide that determining whether the weld is of acceptable or unacceptable quality using the combined data is a second determination separate from the first determination. In determining whether the weld is of acceptable or unacceptable quality, the results of the second determination may take precedence over the first determination.

[0187] The combined data based decision may include a neural network based step. The integrated data based decision may include a decision engine supporting machine learning. The integrated data based decision may include the use of a model.

[0188] The combined data may be used to make individual determinations of acceptable and / or unacceptable welds versus one or more determinations made using data from only one sensor type in each determination.

[0189] A decision based on the combined data may provide a decision regarding an acceptable weld and / or an unacceptable weld where a decision made using only data from one sensor type in each decision would not be able to make a decision, and / or would not be able to make a decision with an acceptable level of associated error, and / or would result in a different decision regarding an acceptable weld and / or an unacceptable weld.

[0190] The combined data based decision may be used for data points and / or sets of data that are close to the threshold, close to the threshold distance, and / or close to the distance for an indication of an acceptable weld and / or an unacceptable weld. If a data point and / or set of data points is far from the threshold, threshold distance, or distance, the combined data based decision may not be used.

[0191] The combined data may be from one or more sensor types associated with welding conditions applied during welding. The combined data may be obtained from one or more sensor types associated with weld observations performed at a location less than one second after welding, such as less than 1 / 10th of a second after welding. The combined data may not include data from one or more sensor types associated with NDT of the weld and / or data associated with weld observations performed at a location more than one minute after welding.

[0192] The neural network may be trained by providing labeled data to the neural network. The labeled data may be a data point and / or a series of data points that may be indicative of whether or not a weld is consistent with an acceptable weld and / or an unacceptable weld. The labeled data may be a data point and / or a series of data points that may be indicative of whether or not a weld is consistent with an acceptable weld and / or an unacceptable weld with respect to a particular characteristic that may provide an acceptable or unacceptable weld.

[0193] The neural network may be trained by providing supervised learning through operator-based input, such as a label, for example, whether a data point and / or a series of data points is indicative of an acceptable weld and / or an unacceptable weld.

[0194] Training using labeled data and / or supervised learning may be provided during the calibration phase and / or during production welding.

[0195] A neural network can be trained from a library of existing data, particularly labeled data, which can be fed into a neural network classifier.

[0196] The labeled data and / or supervised learning and / or library may provide data from multiple sensor types.

[0197] The second method, a library, may be used as the starting labeled data set. The first method, an operator calling to perform a calibration or test run on the actual welding system, may be used as an alternative from the start. The first method may be used to add to the data set of the second method.

[0198] Over time and / or as the neural network learns, it may modify the definitions used to determine acceptable and / or unacceptable welds.

[0199] The neural network may be trained by unsupervised learning, which may be performed in a calibration phase.

[0200] Neural networks may provide clustering or grouping based processing. Neural networks may look for patterns in the data, such as similarities and / or anomalies in the data.

[0201] The neural network can be trained using only the first method, or using the first method followed by the second method, or using the first and second methods in parallel.

[0202] This eleventh aspect of the disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the disclosure.

[0203] According to a twelfth aspect of the present disclosure, there is provided an apparatus for monitoring welding, comprising: (a) multiple inputs for data from multiple sensor types; (b) one or more processors, wherein a processor of the one or more processors: a. receiving input from at least one of a plurality of sensor types; b. comparing the data obtained from the multiple sensor types to reference data of one or more sensor types; c. Determining whether the weld is of acceptable or unacceptable quality in the output determination; Steps and Equipped with The apparatus further comprises: (c) a processor of the one or more processors, (a) forming combined data from data from two or more sensors of a plurality of sensor types to provide the combined data; (b) making a second determination using the combined data, the second determination being whether the weld is of acceptable quality or of unacceptable quality; A processor; (d) one or more outputs for an output determination and a second determination output; An apparatus comprising:

[0204] This twelfth aspect of the disclosure may include any of the features, possibilities, or options described elsewhere in the document, including other aspects of the disclosure.

[0205] Various embodiments of the present disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which: [Brief description of the drawings]

[0206] [Figure 1a] 1 is a cross-sectional side view of a probe according to an embodiment useful in practicing the present disclosure. [Figure 1b] FIG. 1b is the same view as FIG. 1a, but illustrates additional features. [Diagram 2] Figure 2a is a schematic diagram showing the effect of temperature differentials on an ultrasonic beam within a substrate, and Figure 2b shows the propagation of shear wave ultrasonic waves through a substrate with a defect having two different homogeneous and one inhomogeneous temperature distributions. [Diagram 3]Figure 3a is a plot of a typical non-uniform temperature distribution along a weld cross section derived from a thermal model and verified by thermocouple measurements, and Figure 3b is a plot of an ultrasonic beam propagating through a discrete thermal boundary showing the planned focal spot and actual geometric location points for a substrate with a non-homogeneous temperature distribution. [Figure 4] FIG. 1 is a schematic diagram illustrating the scan conversion process and how a pixel's value is derived by weighted contributions from the nearest surrounding time samples on the nearest beam. [Diagram 5] FIG. 1 is a schematic diagram of the process flow of the modeling, ray tracing, and imaging stages. [Figure 6] 1 is a series of cross sections through the weld groove, along with the location and sequence of weld passes being added, and a representative ultrasonic image of the weld groove in that state. [Figure 7] FIG. 13 is an illustration of an approach to substrate thickness verification by reconstructing the beam produced by the same ultrasonic device used to image the weld. [Figure 8] 1 is a diagram of ultrasound echoes observed in a signal trace and interference bursts occurring in the same signal trace at different times. [Figure 9a] FIG. 13. Signal traces from three consecutive data acquisitions and indications of interference bursts therein. [Figure 9b] FIG. 13. Signal traces from three consecutive data acquisitions and indications of interference bursts therein. [Figure 9c] FIG. 13. Signal traces from three consecutive data acquisitions and indications of interference bursts therein. [Figure 9d] FIG. 9B is a processed signal trace based on an average of the signal traces of FIGS. 9a, 9b, and 9c. [Figure 10]Figure 10a is a graph of four separate signal traces superimposed: three are the rectified signal traces from Figures 9a, 9b, and 9c with interference bursts but occurring at different time positions, and the fourth is the rectified signal after processing to remove the interference while preserving the actual echo. Figure 10b is an illustration of the process rectified signal trace with four traces, showing the sample-by-sample selected contributions from each of the three original rectified traces and the fourth trace which is their average. [Figure 11] FIG. 13 processes an unrectified signal trace from three other traces and also displays selected contributions that each of the three signal traces makes to the unrectified trace. [Figure 12] FIG. 2 is a diagram of an unrectified waveform according to one embodiment. [Figure 13] FIG. 1 is a schematic flow diagram relating to a process for selecting an interference processing technique to be used. [Figure 14] A diagram of interference in different channels shown by scope display, image display, and amplitude + time of flight display. [Figure 15] FIG. 15 illustrates the effect of processing on the multiple channels shown in FIG. 14. [Figure 16] FIG. 13 illustrates a further approach to identifying interferences in sections of a trace, identifying subsections that will not be used in the processed trace, and the resulting processed trace. [Figure 17] FIG. 1 is a schematic diagram of an adaptive control function for a welding process implemented in accordance with the present disclosure. [Figure 18a] 1 is an image of a defect detected by an ultrasonic probe. [Figure 18b] FIG. 18b is an image of the ultrasound probe used for detection in FIG. 18a. [Figure 19] 1 is a plot of outlier scores obtained from acoustic signal v data points for various welding conditions. [Figure 20a] FIG. 2 is a perspective view of a profile sensing device relative to a substrate and a weld. [Figure 20b]FIG. 1 is a schematic diagram of the overall weld formation sequence and shape through a series of weld passes. [Figure 21] 1 is a sequence of camera images of a weld site during welding. [Figure 22] 1 is a plot of arc voltage and a plot of Gaussian amplitude x Gaussian center v time of a welding process. [Diagram 23] FIG. 13 is a diagram of combined data types displayed to a user. [Figure 24] FIG. 13 is a diagram of a second level of processing applied to data from multiple sensor types. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0207] Ultrasonic inspection is used for non-destructive testing of a variety of objects. A transmitting transducer emits ultrasonic waves which enter the object, interact with the object and its sub-features, and then return to the receiving transducer. The temperature of the object affects the speed of sound within the object, so using an inappropriate speed can adversely affect the quality of the inspection and images.

[0208] When testing welds, it is desirable to be able to perform ultrasonic testing immediately after the weld or weld pass is made in order to minimize the time it takes to perform, test, and make any necessary corrections to the weld or weld pass, or optimize the weld parameters after the weld is completed or after each weld pass, however this means performing the ultrasonic testing while the object or substrate is at an elevated temperature and there are temperature variations between parts and / or within parts of the object or substrate.

[0209] To accurately detect defects and more generally be able to image the weld site, the present disclosure provides a method to account for the temperature profile, account for the effect of that temperature profile on ultrasonic waves, and correct the location, thereby providing accurate imaging over a wide range of temperatures and accounting for different temperatures at different locations, especially in high temperature situations, which helps enable early inspection of the weld site without the need to cool the weld site to reduce the effect of temperature on the ultrasonic waves.

[0210] This disclosure relates in part to producing accurate ultrasonic images of welds in high temperature environments. Details of a probe device suitable for use in such environments are provided with reference to Figures 1a and 1b at the end of this document. However, it should be noted that the probe need not be a roller probe, but a wedge type probe that is positioned on a substrate and moves from location to location could also be used and still obtain the benefits of this disclosure.

[0211] The basic structure of the probe device includes an axis of rotation RR extending through the probe 13 on which is provided a second axis element 70 connected to the axis element 64 by a series of releasable fasteners 72. The second axis element 70 provides mounting for the transducer 52, an anti-echo block 74, and an ultrasound transmission block 76.

[0212] The coupling element 28 between the probe 13 and the substrate [not shown] is a unitary piece of compatible material, as will be described below. The coupling element 28 is provided with a generally right cylindrical body portion 54 and inwardly turned rims 56a, 56b at either end.

[0213] The transducer 52, anti-echo block 74, and transmission block 76, together with the second axis element 70, axis element 64, and first mounting location 20, do not rotate as the probe 13 rolls over the surface of the object. Thus, the transducer 52 and associated components are always maintained in the same sensing orientation facing the object.

[0214] The interior volume 42 is provided with a coolant that is actively introduced and removed from the probe to maintain the probe within an acceptable temperature range, possibly up to a level exceeding the maximum vertical extent 50 of the transducer 52 that surrounds the cooling components.

[0215] Also attached to the second shaft element 70 is a mounting element 82 carrying a thermistor 84 for sensing the temperature of the coolant at a location 86 proximate the portion of the inner surface 78 adjacent the weld location.

[0216] For ultrasonic transmission, the coupling element 28 is a high temperature silicone rubber. The material selected is capable of withstanding temperatures in excess of 350° C. for extended periods of time. Such a material may have an attenuation of 0.87 dB / mm at 5 MHz and an acoustic impedance of 1.12 MRayls, providing a good match with other materials used.

[0217] With regard to the thickness of the coupling element 28, a balance is struck between the increased insulation from the probe contents that would be obtained by increasing the thickness and the unfavorable increase in attenuation that would be obtained by increasing the thickness. For the operating conditions considered, a thickness of such material between 2 mm and 10 mm, for example between 4 mm and 8 mm, is suitable.

[0218] The material selected for the coupling element 28 also provides sufficient conformability to conform to the surface of the object under moderate applied force levels. High force levels are undesirable in view of the equipment required to generate them and move the device over the test piece. Surfaces of objects encountered in the real world are not highly finished or smooth, so a conforming material is necessary to obtain good contact for transmitting ultrasound without excessive loss.

[0219] With respect to the passage of ultrasound, the anti-echo block 74 plays an important role in preventing ultrasound from bouncing around within the probe and causing noise or other adverse effects on the probe. Hydrogenated nitrile rubber (HNBR) has been found to be a suitable material, especially for its N-filler form, as it provides an attenuation of 6.4 dB / mm at 5 MHz.

[0220] All these features contribute to successful acoustic coupling of the probe to the object through the surface that occurs in reality.

[0221] In terms of coolants, air offers insufficient heat capacity and thermal conductivity for active cooling. Water is also not optimal as it has an acoustic impedance of 1.5 MRayls and is not compatible with the other components. Water is not suitable for the temperatures encountered and also has a tendency to bond with the sliding surfaces rather than act as a lubricant. Providing the coolant in the form of a water-soluble oil, for example, with an acoustic impedance of 1.1 MRayls, would be more compatible with the acoustic impedance of the coupling elements [around 1.1 MRayls].

[0222] The transducer 52 provides a 5 MHz 64 element phased array and is mounted to generate ultrasonic waves at 55° to the object. A pitch of 0.5 mm and an elevation angle of 10 mm may be used. The angled beam has the advantage that the weld can be fully inspected from a distant lateral location. Often a distant lateral location will provide better contact between the probe and the object than the location where the weld is being made. For example, a multi-pass weld will have a large recess until the weld is complete, preventing good contact and ultrasonic propagation to the object. Also, the angled inspection angle better matches the orientation of typical defects, which is problematic with 0° or low angle based approaches.

[0223] This type of transducer and transmission block configuration can be used to provide a fan-shaped scan beam defined by an emitted top beam [angled away from normal to the transducer face] and a bottom beam [near normal to the transducer face]. Regarding the performance required of the probe with respect to high temperature performance, the present disclosure provides a probe capable of inspecting objects at approximately 300° C. for long periods of time.

[0224] The coupling is dry but achieves the level required to propagate ultrasound through the interface into and out of the object.

[0225] High temperature polymers used in the bonding components are capable of withstanding prolonged contact with objects at such temperatures and allowing successful propagation of ultrasonic waves to and from the interface.

[0226] The coolant and the coolant-filled gap allow for efficient propagation of ultrasonic waves to and from the transmission block.

[0227] Since the transmission block is only exposed to temperatures close to the ambient temperature, there is no need to select a high temperature resistant material with poor ultrasonic propagation characteristics, providing optimal propagation characteristics for the transport block.

[0228] The speed of sound in all components along the acoustic path, namely the ultrasonic transmission block 76 [wedge between the transducer 52 and the conforming material 28], the conforming material 28 [tire], and the substrates being welded, changes with temperature. This means that the angle of refraction changes with the material temperature at each interface. Furthermore, the beam within each material will bend if there is a temperature gradient within that material. For example, referring to FIG. 2a, the beam wavefront 100 travels slower in the hotter portion 102 of the substrate 104 than in the cooler portion 106 of the substrate 104. Thus, the beam 100 deforms from a straight beam 100' upon entering the substrate 104, as the portion of the beam in the cooler portion 106 travels ahead of the portion of the beam in the hotter portion 102. Thus, a non-uniform temperature distribution will cause the beam 100 to bend in a direction perpendicular to the beam 100'. def It bends like this.

[0229] This effect can also be seen in Figure 2b, where (a) a homogeneous 25°C plate is used to show a shear wave propagating through the plate as a substrate to a defect location, (b) a homogeneous 150°C plate and the angle of incidence adjusted so that the beam is pointing at the same target is used to show another shear wave propagating through the plate as a substrate to a defect location, and (c) a shear wave propagating through the plate as a substrate to a defect location in an inhomogeneous temperature state.

[0230] To be able to take into account possible beam bending, it is necessary to be able to determine as much as possible the temperature of the entire ultrasound path, especially in the high temperature and high potential variation sections. If the temperature distribution is known, the beam bending can be predicted and appropriate corrections can be made in the configuration and operation of the device [e.g. to deliver a focused beam to the desired location] and / or applied in the signal processing of the returned beam [e.g. to provide a more accurate image].

[0231] The first step in the determination is to perform thermal modeling of how the heat input from preheating and additional heat input from the welding itself are distributed throughout the metal volume of the base material as a function of time.

[0232] One way to validate the thermal modeling was to prepare a substrate in the form of a steel plate with numerous thermocouples distributed across it. Thermocouples were provided on both the top and bottom surfaces of the substrate. The substrate was then preheated to an optimal temperature for welding and various weld head passes were performed to obtain the thermocouple response. While it is possible to broaden the scope of the modeling by varying the level of preheating, in many cases samples of the same substrate, and even different substrates [thickness, profile, steel type, etc.], will be within a consistent and strict temperature range. Thus, little variation due to preheating is expected.

[0233] By performing such thermal modeling, the safe operating temperature of the probe (based on knowing the response of the probe over time when exposed to different temperatures) can be compared to temperatures at various locations away from the weld, allowing the modeling to determine the closest safe distance for the probe to trail behind the weld head.

[0234] Importantly for thermal modeling through the substrate depth, these studies reveal that after welding the temperature distribution quickly stabilizes to thermal contours perpendicular to the metal surface. This greatly simplifies the shape of the temperature distribution and the correction calculations, especially when corrections need to be made for the effect of multiple skips through the substrate. A skip is when an ultrasonic wave that enters the substrate through an interface is internally reflected off an interface on the other side of the substrate and returns to the entry interface.

[0235] With regard to thermal modeling of the spread from the hottest weld location to the substrate, the temperature distribution is often, but not always, symmetrical around the center of the weld. Therefore, the beam develops a complex S-shaped distortion as it passes through the weld area. This can be seen in the temperature distribution in Figure 3a and the ray tracing analysis in Figure 3b. The positions are shown in mm on either side of the weld location, which experiences the highest temperature [1000°C in this case]. In this example, an exaggerated temperature gradient is used to emphasize the magnitude of the beam bending.

[0236] Referring to FIG. 3b, the deviations caused by temperature distribution in a weld inspection are shown in more detail.

[0237] Series of planning focuses 110 1 ~110 i Plan Focus 110 1 The plan focus 110 extends along a plane 112 having a near-planar end on which the plan focus 110 is provided. The near-planar end faces the probe 13. i is provided at the far planar end, further away from probe 13. Plane 112 is parallel to the top surface of the substrate in the case of a plate, and therefore represents a certain depth into the substrate in such case.

[0238] Planned focus 110 1 ~110 i, the transducer 52, and the beam it produces, is angled relative to the top and bottom interfaces of the substrate. An angle of 55° may be adopted at this point, defining the centre line 114 of the scan arc. The control system of the transducer 52 allows the beam to be steered to various angles to achieve a sector scan. All elements of the array [64 in this example] are used for each transmission of the beam, but the excitation time of each element is different, steering and focusing the resulting beam where required. In effect, this defines the boundaries 116 and 118 of the scan arc, and the minimum and maximum angles at these boundaries, and thus the planned focal points 110 at both ends. 1 and 110 i is defined.

[0239] As revealed by thermal modeling, the temperature contours 120 are often perpendicular to the surface of the substrate, however, the method may work with other contour or boundary orientations and / or profiles to fully and comprehensively account for different temperature gradients in different directions within the substrate.

[0240] The more thermal boundaries used to divide the substrate, the more accurate the calculation of the temperature distribution will be, and the more accurate the calculation of the actual geometric location points 122 1 ~122 i Of course, the temperature change is continuous, but multiple step changes can be used to accurately represent the same effect.

[0241] Each time the beam crosses a thermal boundary, a correction may be applied to the path of the beam based on the temperature change that the boundary exhibits. This provides a new path for the beam, creating a new path length section before crossing the next thermal boundary. Further corrections create yet another new path, and so on. The result is an actual geometric location point 122, determined as the geometric location where the signal sample actually occurs. 1 ~122 i 3b. The actual geometric location point 122 1 ~122 iThe profile of is caused by the beam bending to a gradually shallower angle and decelerating as the wave front encounters a higher temperature toward the weld center, plane 124. The angle steepens again and the wave front accelerates again, bending in the opposite direction after passing the weld center, plane 124.

[0242] The actual geometric location of the signal origin point 122 1 ~122 i and 110 planned focal points 1 ~110 i Due to such differences between the focal points, the scan conversion algorithm must convert the observed signal to the corresponding planned focal point, e.g. 1 , but not the actual geometric location point, e.g. 110 1 , which needs to be plotted.

[0243] A similar process takes place on receive, with the signal from each channel undergoing a different time delay so that the sums from all channels add together coherently, increasing the net response from a given direction and range. A potential drawback here is that the steering angle and focus are fixed at the time of acquisition, so any errors in the thermal distribution cannot be corrected for afterwards. FIG. 4 further explains how the location of pixels located between the first beam i and the next adjacent beam i+1 are taken into account. In this figure, A ij is the amplitude of sample j in beam i. ·W ij is the weighting coefficient of the contribution from sample j in beam i. ·α is the fraction of the virtual beam distance where the pixel lies from beam i to beam i+1. · β is the fraction of the distance from sample j to sample j+1 along the virtual beam in which the pixel lies. thus, A pixel =(1-α).(1-β).A ij +(1-α).(β).A ij+1 +(α).(1-β).A i+1j +(α).(β).Ai+1j+1 Apixel=Wij.Aij+Wij +1 .Aij+1+Wi+1j.Ai+1j+Wi+1j+1.Ai+1j+1 It is.

[0244] The configuration does not change between acquisitions, so to speed up image generation, we use i+α, j+β, W ij , W ij+1 , W i+1j , Wi +1j+1 The values ​​of are pre-calculated for every pixel in the viewing area and stored in a look-up table. During imaging operations, the addresses of the beam and sample data contributing to each pixel (i, i+1, j, j+1) can be derived by simple truncation of the i+α and j+β values ​​in the look-up table. Producing overlays such as beam markers of selected A-scan data is a straightforward operation when the beam is straight, but becomes more complicated for curved beams with thermal gradients. This requires a reverse look-up process from pixel to beam and sample, and it is convenient to keep the i+α and j+β values ​​in a look-up table.

[0245] FIG. 5 shows the complete flow path of the process, including the modeling, ray tracing, and imaging stages.

[0246] In the modeling stage, a thermal model is generated that indicates the temperature of each data point. The initial thermal model may be based on previously stored information from previous actual welding operations and / or previous thermal models.

[0247] Next, actual thermal data of the welding environment is collected. This data may be obtained by inputs (e.g., preheat temperature and / or welding conditions provided) and / or from temperature sensors (e.g., thermistors associated with probes, other temperature sensors, infrared sensors, etc.) in the welding environment that provide information regarding the temperature and location of the data. The combined data is used to form an actual thermal data set.

[0248] The actual thermal data set is then compared to the thermal model to assess their agreement with respect to temperature at the location. If the discrepancy in agreement is too large, then the thermal model is updated. A new comparison is made and the cycle is completed until the actual thermal data set and the thermal model are determined to match. The ray tracing stage can now begin.

[0249] In the ray tracing stage, the thermal model approved in the modeling stage is converted into an acoustic velocity map. The substrate is known and the acoustic velocity versus temperature of the substrate is revealed through calibration tests. The calibration data is stored and accessible for use in generating the acoustic velocity map. The temperature of the data points is converted into velocity. A velocity contour interval is then set and contours that pass through data points with the same velocity are applied to the acoustic velocity map. In this way, the substrate is divided into different velocity zones.

[0250] The starting point of each ray [beam component] is known along with its initial path. Each time the projection of that path crosses a velocity boundary, a refraction expression is applied. The refraction expressions applied vary in extent depending on the velocity difference that the contour represents compared to the previous contour, and in direction depending on whether the velocity contour is increasing or decreasing compared to the previous contour.

[0251] The speed and the length of the path between the contours gives the time to pass through that speed zone.

[0252] This process is repeated for each contour intersection in the probe on its way into the substrate. This process may also be repeated for the echoes of each contour intersection on its way out of the substrate. Once the ray tracing steps for each ray are completed, the process may proceed to the imaging stage.

[0253] In the imaging phase, the first step is to generate the region of interest where the results need to be displayed, which may be influenced by the geometry of the substrate, the shape or configuration of the weld, the method of providing the weld, etc.

[0254] The method for generating a pixel's value is essentially a linear interpolation between the four nearest real samples. To avoid the time penalty of calculating this on the fly, the addresses of the four samples and their respective weighting values ​​are pre-calculated and stored in a lookup table. Each newly acquired data set is then scan converted and each pixel is processed in turn, accessing the four samples, multiplying them by their associated weighting functions and then summing them together to form the final pixel value, resulting in a linear display image.

[0255] Since there is existing information of the geometry regarding the position, shape and orientation of the weld groove, it can be matched with the geometry observed in the results. If they correlate, acquisition can proceed. If they do not correlate, the correction was not applied correctly and a check is needed on the temperature data that goes into the correction.

[0256] If the detected temperature does not match the thermal model, it needs to be corrected and the acoustic velocity map re-adjusted for the correction. The corrected acoustic velocity map can be used in the process steps above.

[0257] Figure 6 shows an example of the signal result, in this case the progressive filling of a weld groove by several welding passes. In the first subfigure, there is no weld in the groove, and the groove edge is detected in the ultrasound image. As the welding passes in numerical order gradually fill the groove from below, the extent of the exposed groove edge decreases, and the echo in each image becomes smaller than the previous one and finally disappears.

[0258] If a defect is found, it will be displayed on the right side of the image compared to the rest of the weld groove wall.

[0259] The implementation described above is based on using thermal contours to derive transit times and correct temperature distributions. Beam reflections are fully taken into account to determine the beam path.

[0260] In another implementation, Full Matrix Capture (FMC) may be used in combination with the Total Focusing Method (TFM) for the same purpose, but taking into account a larger and more complete data set.

[0261] FMC is a technique similar to holography in that raw data signals, containing time sequences of amplitude and phase, are collected from the combination of all transmit (Tx) and receive (Rx) elements. An image is reconstructed from this FMC data set by creating a virtual path from every element to every pixel and performing a differential delay as a post-processing operation. This means that every pixel is in ideal focus in both transmit and receive, and for this reason the approach can be called the Total Focusing Method (TFM). The same technique as above with the gradients approximated with multiple separate thermal zones is used to determine the thermal gradient corrected transit times from each element to each pixel, as required by the TFM reconstruction algorithm.

[0262] (Verification of substrate thickness) As noted above, the temperature profile within the substrate is a significant issue that needs to be addressed. Applicant has also discovered that there are material variations in the thickness of substrates such as boards, even when stated to be of a given thickness by the manufacturer. This also needs to be taken into consideration in order to optimize imaging.

[0263] Beam path length is affected by thickness and the variations observed give a 0-5mm variation in path length compared to that associated with a specified thickness value for normal material thickness. Note that the path length is multiplied by the number of half-skips performed. To properly identify the location of a defect, it is necessary to account for the actual distance of the path length to and from the defect, rather than the path length assumed based on the specified substrate value.

[0264] As a pre-weld inspection step, it is possible to move the probe over the substrate in the same path as during weld inspection. The ultrasonic beam can be used to reveal the thickness of the substrate. This approach is shown in Figure 7.

[0265] The roller probe 13 described above is only partially shown in FIG. 7 with respect to the coupling element 28 and the transmission block 76. The transducer 52 is not shown, but is located next to the angled face of the transmission block 76. Unlike the NDT beam 800, which is an angled shear wave sector scan of a weld for inspection purposes described above, the thickness measurement beam 802 operates at 0° (i.e. perpendicular to the metal surface) with respect to the substrate 804 or object. The control electronics are used to move the beam away from the position of the NDT beam 800 and to specify the position of the thickness measurement beam 802. To ensure that the 0° beam is included, a sector scan is used with multiple longitudinal beams centered at 0°.

[0266] The 0° beam is used to check the amplitude of the back wall echo from the substrate 804 as a way of monitoring the quality of the coupling between the probe 13 and the substrate 804, in this case the beam is used to measure the thickness as well as to confirm that a given echo intensity is present.

[0267] The strongest echo received at the detector of the probe 13 corresponds to a first interface 806 where the probe 13 meets the substrate 804. The second strongest echo corresponds to an opposing interface 808 of the substrate and is an internal reflection. The method uses the echo and its repetitions to measure the local thickness and correct the location of the region of interest.

[0268] By using a narrow sector beam rather than a single beam for the thickness measurement beam 802, the method can reveal which angles produce the strongest repetition, which can provide information about the taper compression of the bonded element 28, which in this example is a tire. Tapering can cause beam refraction, another issue that may need to be corrected for when imaging the weld in its correct shape.

[0269] The detected position may be corrected for temperature as described above and used to account for the actual distance along the length of the path. Temperature correction in the transfer block 76 is still required to achieve the 0° measurement direction, i.e. perpendicular to the substrate and minimum distance through the substrate or thickness. This may be used at the same position along the path on the substrate during inspection of the weld to correct the image and correct the location of any defects found.

[0270] Alternatively, the probe may rapidly switch between the weld inspection mode and the thickness determination mode and repeat the cycle as it advances along the weld path to determine thickness during actual weld inspection.

[0271] As a result of either approach, the actual thickness at any location along the weld path is known, as compared to the theoretical value, and can be used to correct the images for the location of features and / or defects.

[0272] (Handling RF interference) Significant interference bursts from robot motors for welding arms, probe arms, and other tasks were observed in the ultrasonic detection signal traces. These were found to be several times larger than ultrasonic echoes, even from planar interfaces. Other bursts are not as strong and are therefore difficult to distinguish as bursts. The planar interface echo itself is often much stronger than the defect echo. Other sources of interference exist and may be addressed by the techniques provided herein. For example, electrical interference is generated by various motors, such as motors used for pumps in cooling systems encountered in welding applications and their measurements. Other sources of interference may include elevator drive motors and power supplies. Therefore, a solution to address such interference is desirable to enable effective imaging.

[0273] The wideband nature of the burst means that bandpass filtering techniques have limited effectiveness in reducing the amplitude of the interference.

[0274] The smaller the return signal, the greater the problem, and in a hot probe environment, the distance from the transducer to the compliant material, the dry interface to the substrate, and similar factors in the return signal all attenuate the signal, making noise even more of a problem.

[0275] Furthermore, the timing of interference bursts is irregular in occurrence and duration depending on the motor task each arm is being asked to perform.

[0276] FIG. 8 is an illustration of a signal trace in which an ultrasonic echo of a defect can be seen towards the midpoint of the signal trace, and an interference burst can be seen towards the latter half of the trace.

[0277] However, the intermittent nature of the interference may contribute to the solution.

[0278] In a first embodiment, intended to account for interference bursts, a transducer is used to transmit a series of beams through the substrate, and the return signals are detected as before. In the illustrated example, Figures 9a, 9b, and 9c, three such detected signal traces are shown.

[0279] Interference from the robot motor occurs in short high frequency bursts separated by gaps significantly longer than the repetition period of the ultrasonic pulse. This time difference means that if a burst from the robot motor appears in one beam, it will not appear in the next, so interference-free data from this second signal can be used to replace the contaminated sample in the first beam. If there is a second independent robot, it acts as a second interference source, and its bursts are asynchronous with those from the first robot. If interference from this second robot occurs in an interference-free sample area of ​​the second signal, the sample from the third signal will be clear and can be used in place of the contaminated sample.

[0280] In each case, a defect is applied at the same location in all three traces, so the defect generates echoes in the traces at the same time. The three traces arise from the same beam entering and exiting the substrate along substantially the same path [because the probe does not move any significant distance along its path in that time frame], and encounter the same substrate, so the time at which the signal traces begin to generate does not affect the echoes.

[0281] An interference burst is also present in the later portion of the first signal trace in Figure 9a. In some cases, a similar interference burst is also present in the earlier portion of the trace in the second signal trace in Figure 9b. Further interference bursts are also present in Figure 9b, but not in Figure 9a. Two further potential interference bursts are present in Figure 9c.

[0282] Figure 9d shows a signal trace that is a sample-by-sample average of the three signal traces of Figures 9a, 9b, and 9c, used to improve the signal-to-noise ratio [SNR]. As a result, the echo portions of the signal trace are still prominent, but they are reduced because only one signal trace contributes to the average of interference bursts in any part of the signal trace. This averaging approach can tell in a basic way and in some circumstances which are echoes and which are interference bursts. This technique does not attempt to identify the presence or timing of interference bursts. Such bursts still contribute to the average signal trace of Figure 9d. However, the amplitude of the bursts is significantly reduced by averaging multiple beams. The more beams that are averaged, the greater the reduction.

[0283] However, simply lowering the amplitude may not be sufficient if the interference burst is comparable or greater in magnitude than the echo and / or the signal is noisy or complex. If the burst gives the result the impression of an echo, it may lead to a false negative determination of the weld, requiring unnecessary further investigation and / or remedial action.

[0284] The above technique can be improved by correctly identifying the interference bursts so that echoes occurring in the same part of the signal trace are detected in some way.

[0285] A more sophisticated approach to removing interference from the signal traces is adopted in Figure 10a, which shows an overlay of three separate rectified signal traces 0, 1, and 2 representing the same beam at three different transmissions.

[0286] The mechanism to determine the likelihood that a sample is from an interference burst is by comparing the magnitude of the rectified (envelope detected) samples from each of signals 0, 1 and 2. The signal with the largest magnitude sample is a candidate for being from interference, so the smallest value across the signals is selected and that is used in the final processed waveform. If none of the signal samples are from an interference burst, the amplitude difference between the signals will be very small due to random noise in the sensor or instrument. Comparing this amplitude range to a threshold is therefore a mechanism to indicate the absence of interference, and using an average of the samples across all signals improves the SNR. The threshold may be selected as a value related to the expected noise amplitude, or may be selected as a defined percentage of the signal amplitude.

[0287] Signal trace 0 contains a high amplitude section towards the end of the signal trace which is not present in 1 or 2 and is therefore deemed to be a burst from the RF in stream 0. Thus, when the portion of stream 0 which contributes to the average signal trace is formed, this section is omitted from its contribution. This can be seen in the RF0 stream at the bottom of Figure 10b, where trace 300 is omitted. Other sections 302 are also omitted from the RF0 contribution and these spikes are not seen in the other streams.

[0288] Averaging the contributions of traces that are allowed to contribute to the average signal trace means that there is an improvement in the signal to noise ratio from averaging of uncontaminated traces.

[0289] In a similar manner, returning to FIG. 10a, 1 contains two high amplitude sections towards the beginning of the signal trace that are not present in 0 or 2 and are therefore deemed to be bursts from the RF in stream 1. Thus, when the portion of stream 1 that contributes to the average signal trace is formed, this section is omitted from its contribution. This can be seen in the RF1 stream at the bottom of FIG. 10b, where these portions of trace 304 are omitted.

[0290] A similar approach is taken for the two sets of two signal trace spikes in the second half of the trace and their omission from the contribution of stream 2 (section 306).

[0291] With these sections removed, the sections that contribute to the average trace are displayed at the bottom of Figure 10b, which results in the signal trace plotted in Figure 10b, which shows not only the expected echo spike 310 clearly observed in Figure 10a, but also the other two echo spikes 312 of Figure 10a. The remaining signal section, the lowest trace in Figure 10b, is informative and provides a significant amount of time over which the defect signal is detected.

[0292] Up to this stage in the processing, there are sections of the average trace which represent samples where none of the 0s, 1s or 2s have been determined to be interfering and can be replaced with an average of 0s, 1s and 2s. See the lowest of the four traces in Figure 10b. This averaging has the benefit of reducing the signal to noise ratio. If these sections were not present, one or more of the 0s, 1s or 2s have been identified as being from an interfering burst and so an average of only the clear signal would be used. This results in some improvement in the signal to noise ratio through averaging, but the number of elements contributing to the average is reduced by the signals that have been excluded.

[0293] Using this minimum or threshold approach across all potential traces is a fast and highly efficient method of combining the elimination of possible interference with the substitution of alternative trace sections. However, this technique does not identify bursts; it assumes that high amplitude sections that are significantly larger than matching values ​​in other traces that do not suggest interference are due to bursts. If interference is only identified in one beam, averaging over the remaining beams will improve the SNR and improve signal fidelity compared to the all-streams minimum approach, so a technique that more clearly identifies bursts would be useful.

[0294] In yet another approach, the collision detection and replacement of trace sections are processed in separate steps. This further approach is described below in conjunction with FIG. 16 and is a sample-by-sample approach.

[0295] By providing a general approach that can identify burst regions in each trace, all sections of the trace that are free of contamination can be properly identified. The criteria used to determine the identification of a burst and the replacement method for replacing sections that contain bursts can each be independently controlled and the criteria can be modified as required.

[0296] In this general approach, the first step is to identify the suspected interfering sections on each beam stream or trace. This is achieved as follows: 1.1. If the input data traces / streams are complete RF waveforms, perform envelope detection on each trace / stream waveform and produce a rectified trace / stream for each beam. 1.2. For each beam, find the minimum of each rectified sample across all traces / streams. This will be a good approximation of the interference-free value for that time sample. 1.3. Calculate the difference between each trace / stream sample and this minimum value and mark it as interference if it exceeds a threshold. This threshold can be a constant or adaptive, such as the minimum value multiplied by a predefined factor.

[0297] In a second step, the method performs interference cancellation in the following manner. 2.1. Use the interference marker (derived in 1.3) to reject the input trace / stream sample for that beam. 2.2. Calculate the average of the remaining samples across the trace / stream. 2.3. This average value is used as the time sample of the processed beam.

[0298] Referring to Figure 16, the three traces above are the rectified data and the top, middle and bottom waveforms at the top of Figure 16 show streams 0, 1 and 2 of a single beam. The highlighted sections of these interference detection plots (one A towards the middle of stream 0, one B towards the end of stream 1, two C, D towards the start of stream 2) are all sections where there is interference. Peak e is the echo.

[0299] Applying the above algorithm, this interference is detected and marked in the interference processed plot, the bottom three traces 0'', 1'', 2'' in Figure 16, eliminating sections that need to be further considered to arrive at the processed trace. Each subsection [represented by individual bar style peaks] within the interference burst sections A, B, C, D is highlighted for discarding. These identified and highlighted subsections are excluded from forming the processed trace (middle trace P in Figure 16). This processed trace P shows a processed sample of the beam, demonstrating both the removal of the interference burst and the reduction of noise in the remainder of the waveform.

[0300] 11 and 12 show a further modification of the previous approach, applicable to non-rectified signals. This is also a sample-by-sample selection, so the general approach of the previous embodiment may be used here as well, with flexibility in the criteria used and improved signal-to-noise ratio.

[0301] Here, the approach of identifying samples without interference bursts by selecting the signal with the smallest amplitude does not work: instead, the signal is first rectified, and then the original approach is applied to these modified waveforms to determine whether the samples in the final waveform are from signals 0, 1, 2 or the average. This selector mechanism is then applied to the original unrectified signal to produce the unrectified processed signal.

[0302] In Figure 11, a similar situation to Figure 10a exists with streams 0, 1, and 2. Stream 0 has an RF burst later in the signal trace, stream 1 has a burst earlier in the signal trace, and stream 2 has two bursts later in the signal trace. If these contributions are removed to create the average trace at the bottom of Figure 11, the same type of missing section occurs.

[0303] However, with further processing, it is possible to reconstruct the missing section using sections taken from one or both of the other signal traces. Thus, for section 330 omitted by RF burst 332 in stream 1, section 334 from stream 0 or section 336 from stream 2 can be used to fill section 338 of the average trace. If both sections 334 and 336 are used to fill, then the average may be used.

[0304] The result of this filling is shown in Figure 12, where significant sections of the entire trace are taken from the average of three traces, some from the average of two traces, while other sections are filled using one of the other traces. Missing sections of stream 0 are filled using stream 1 and / or stream 2. Figure 12 shows a continuous trace with no missing echoes and interference fully accounted for, with no defects or missing echoes.

[0305] Although the interference of the robot motor is significant, the interference frequency of the welding arc is unrelated to the frequency of the ultrasonic signal of interest and therefore does not require special treatment.

[0306] The approach for unrectified signals is also directly applicable to full-matrix capture signals, where the SNR of each received signal is inherently lower due to it arising from a smaller excitation energy, so the combination of removing interference bursts and improving the SNR by averaging in the absence of bursts is particularly effective.

[0307] Minimizing the computation time to process FMC signals is especially important in FMC data sets because there are many signals to process. If all received signals of a particular element's transmission are acquired in parallel, the nature of electromagnetic interference means that all received signals will have bursts at the same sample. Therefore, multiple pulse sets from only one receive channel need to be analyzed to derive a sample selector, which can then be used across the parallel receive channels, thus reducing computation time.

[0308] Since the above techniques require different amounts of computation, it is useful to select the computationally intensive technique only when necessary and use the alternative technique when not. In this regard, Figure 13 represents a likelihood selection approach taking into account the potential selection. Here, the "average over samples" approach [described above with reference to Figures 9a-9d], or the "minimum sample value" approach [described above with reference to Figures 10a, 10b, 11, and 12], or in the case of full matrix capture, the "single channel rectification" approach, are considered. The single channel is rectified as described above for the more general rectification.

[0309] Thus, referring to Figure 13, the system is set up to scan and acquire 1300 multiple signal traces from multiple transmissions. The repetition rate used is adjusted 1302 to avoid aliasing patterns, after which acquisition 1304 of the data [signal traces] may proceed.

[0310] With respect to the acquired data 1304, which may be similar to the full matrix capture approach, the processing considers 1306 whether there is interference across all parallel receive (Rx) channels.

[0311] If all channels 1308 are free of interference, the signal traces of each of the multiple transmits (Tx) are considered 1310 in turn for each beam. The process then performs rectification 1312 using the approach outlined above. The minimum and maximum values ​​of the signal traces are obtained 1314 and compared to pre-determined thresholds.

[0312] If the value is below the threshold 1318, the approach continues by using an average of samples from each of the multiple transmit Tx streams to obtain a result 1320. This is the approach described above with reference to Figures 9a-9c.

[0313] If the value exceeds the threshold 1322, a more discriminatory analysis approach is required, which selects the minimum value of each sample to construct a rectified stream 1324. The same selector is used to construct a rectified trace from the raw signal traces 1326 to obtain the result 1320. This is the approach described above with reference to Figures 10a, 10b, 11 and 12.

[0314] Returning to the earlier stages of the process, it is considered whether interference is present on all channels 1306, and if so 1328, a different process approach is used. In that case, only one channel needs to be analyzed 1330, again performed by rectification 1332. In this way, minimum and maximum values ​​are found 1334, and a sample selector 1336 may be used to select the appropriate Tx stream for all parallel Rx channels. Again, the result 1320 is reached.

[0315] This approach of seeing interference in all channels is illustrated in FIG. 14 which shows that, prior to processing, there is material interference in the amplitude trace 1404 and time-of-flight map 1406 of a scope display 1400, an image 1402, and other scans.

[0316] The post-processing positions are shown in FIG. 15, with clear improvements seen in each scan, the processed scope display 1500, the processed image 1502, and also the amplitude trace 1504 and time-of-flight map 1506 of the other scans.

[0317] By using multiple transmissions and choosing the smallest commutation value for each sample, any interference can be properly taken into account.

[0318] Comparing the processed results to the acquired data stream, the signal trace, produces a selector waveform that indicates which samples came from which stream. The selector can be used to create a processed RF stream from the original RF stream, which can enable acquisition and processing of FMC data.

[0319] If no interference is detected (small range of values ​​across the stream), the selected processing only uses the average samples, resulting in a low computational load. At the same time, if interference is detected, the processing may select an alternative processing method that provides a minimum sample value approach, and thus better address the potential impact of interference.

[0320] Increasing the number of transmissions considered allows more asynchronous interference sources to be considered and addressed.

[0321] The various approaches provided by these embodiments are useful when interference prevention or suppression is insufficient.

[0322] (Further Disclosure) This further disclosure relates to improvements in welding, quality control in welding, and methods of use thereof based on the use of multiple sensor types and the combination of data therefrom.

[0323] During welding, there are many variables that affect the formation of a weld and the quality of the weld. Attempts have been made to monitor the individual variables during welding to detect deviations therebetween and the possible resulting degradation of weld quality and / or the occurrence of defects. A weld has various properties when the weld is formed and immediately after it is formed. Attempts have been made to determine the individual properties during welding and detect deviations therebetween and the possible resulting degradation of weld quality and / or the occurrence of defects.

[0324] One potential objective of the present disclosure is to provide a method of welding and quality verification that may benefit from greater utilization of monitoring variables and / or determining characteristics to mitigate the formation of weld defects or the occurrence of other quality issues. One potential objective of the present disclosure is to provide a method of welding and weld condition verification that seeks to synchronize and fuse data from multiple sensor types and / or sources to obtain more information than the sum of the parts. Additionally, potential advantages may be obtained by seeking to analyze each data type in a more informative manner.

[0325] The seventh to twelfth aspects of the invention were referenced above.

[0326] A description of these aspects is provided below.

[0327] It is known to perform non-destructive testing (or inspection), NDT, after a weld has been formed and cooled to ascertain whether defects are present. Any defects found can then be repaired. However, in conventional techniques, such NDT is performed long after the weld has been formed, providing information about the defects long after the event, too late to prevent the formation of defects or mitigate the formation of further defects.

[0328] During welding, there are many variables that affect the formation of the weld and the quality of the weld, which may include wire feed speed, voltage, current, welding speed, distance between the welding device and the weld location, shielding gas, shielding gas flow rate, substrate shape / profile / configuration / dimensions, substrate preheat temperature, weld groove shape / profile / configuration / dimensions, weld or weld pass shape / profile / configuration / dimensions.

[0329] Attempts have been made to monitor individual variables during welding to detect their deviations and the resulting degradation of weld quality and / or the possible occurrence of defects.

[0330] As a weld is formed and immediately after it is formed, the weld has a variety of characteristics, which may include acoustic emissions from the weld location and / or arc, the topology of the weld being formed, and the visual characteristics of the weld as it is formed and / or after it is formed.

[0331] Attempts have been made to determine individual characteristics during welding and to detect their deviations and, as a consequence, a decrease in weld quality and / or the possible occurrence of defects.

[0332] (Overall process) The present disclosure seeks to obtain increased benefits in welding through greater use of variable monitoring and / or characterization to mitigate the formation of weld defects or the occurrence of other quality issues. The present disclosure also seeks to synchronize and fuse data from multiple sources to obtain information that is greater than the sum of the parts. Additionally, the present disclosure seeks to analyze each data type in a more useful manner.

[0333] This disclosure improves initial weld quality and reduces the risk of defect formation by expanding the range of types of weld data that can be collected, analyzed, and used as the weld progresses.

[0334] The present disclosure is directed to processing data obtained from multiple sensors and / or sensor types in real time and using machine learning / artificial intelligence algorithms to derive near real-time decisions regarding welding process control, such that decisions benefit from the involvement of multiple different sensor types.

[0335] In the illustrated disclosure, a combination of four different sensor types may be used to collect data to provide adaptive control of the weld to mitigate the potential for defect formation, however, the method is suitable for processing many more sensor types or data types and for use with sensor types other than those illustrated.

[0336] Referring to FIG. 17, a schematic diagram of the adaptive control function is shown.

[0337] In a welding process, a set of variables that affect and control the weld are defined and controlled, which may include wire feed speed, voltage, current, welding speed, spacing between the welding device and the weld location, shielding gas, shielding gas flow rate, substrate shape / profile / configuration / dimensions, substrate preheat temperature, weld groove shape / profile / configuration / dimensions, and weld or weld pass shape / profile / configuration / dimensions.

[0338] In the welding process, on the left side of Figure 17, an NDT step [Post-weld High Temperature NDT] is performed using a high temperature capable roller type ultrasonic probe [more on this below], which provides ultrasonic data [NDT display] that feeds into an overall assessment of weld qualification marking the end of the process.

[0339] In the welding process, not only is the produced weld monitored and repair actions initiated thereon, but a series of parallel steps, starting with process monitoring, are carried out aiming to prevent or minimize the occurrence of defects.

[0340] In the process monitoring step, data is measured and collected regarding the time-varying changes in voltage and current applied to the welding device. In this embodiment, voltage and current sensing is used to assess correct weld formation, or problems in correct weld formation.

[0341] At this point, data from the other three types of sensors is fed into the data streams to form the sensor data: laser, vision, and acoustic data streams from the appropriate sensors of that type, as described in more detail below. In this embodiment, sensing of the weld profile using laser scanning, visual assessment of the weld, and acoustic sensing of sounds emanating from the weld are also used to address correct weld formation, or issues in correct weld formation.

[0342] The process sequence includes the collection of sensor data using laser, vision and acoustic data streams. Additionally, a process monitoring step collects data on the voltage and current applied to the welding device. Post-thermal welding NDT also provides a fifth data type, the NDT Indication. These form the complete data stream that is considered. Each of these sensor types and the data types they provide are now considered. The various sensor types are: Phased array ultrasonic sensor type - measuring weld quality. Acoustic sensor type - derives the performance of the welding process from the sound of the welding arc. Laser sensor type - to monitor the topology of the weld. · Visual monitoring type - Replaces the operator's vision in weld pool visualization. Voltage and current monitoring type - to measure the power to keep the welding process optimal.

[0343] (Ultrasonic sensor type) This sensor type is concerned with the outcome of the weld produced, whereas the other sensor types focus on the weld produced. Thus, while this sensor type provides information regarding the need for repair action, the later sensor types described below may be used to mitigate the occurrence of defects in the first place.

[0344] Ultrasonic sensor types include transducers that provide a 5Mhz 64 element phased array mounted to generate ultrasonic waves at 55° to the substrate. A 0.5mm pitch and 10mm elevation angle may be used. The angled beam is convenient in that it allows the weld to be fully inspected from laterally spaced locations. Often the laterally spaced locations provide better contact between the probe and the object than the location where the weld is being made. For example, a multi-pass weld creates a large recess that prevents good contact and propagation of ultrasonic waves to the object until the weld is complete. This is problematic with 0° or low angle based approaches.

[0345] This type of transducer and associated ultrasonic transmission block configuration can be used to provide a fan-shaped scan beam defined by an emitted upper beam (angled away from normal to the transducer face) and a lower beam (near normal to the transducer face).

[0346] Regarding the high temperature performance required of the probe, the probe will be capable of inspecting substrates at temperatures of approximately 300°C for long periods of time.

[0347] The coupling is dry but achieves the level required to propagate ultrasound through the interface into and out of the object.

[0348] The high temperature polymers used in the coupling components are capable of withstanding prolonged contact with objects at such temperatures and allowing successful propagation of ultrasonic waves to and from the interface.

[0349] The coolant and the gap between the coolant-filled sensor-type elements may also allow for effective propagation of ultrasonic waves to and from the transmission block.

[0350] Since the transmission block is only exposed to temperatures close to the ambient temperature, there is no need to select a high temperature resistant material with poor ultrasonic propagation characteristics, providing optimal propagation characteristics for the transport block.

[0351] The sensor types also include an integrated surface temperature measurement sensor type and a coolant temperature measurement sensor type.

[0352] (Conditions and effects of welding location) In arc welding, a power source is used to create a sufficient voltage difference between the electrode of the welding device and the substrates to be welded to create an arc, which results in an electric current. The arc heats the substrates to a molten state [potentially consuming the electrode]. Upon cooling, the molten metal solidifies and joins the two substrates.

[0353] The speed at which the welding device moves relative to the substrates affects the degree of melting, the shape of the weld pool, etc., which affects the quality of the weld.

[0354] The welding site is typically protected by a shielding gas to prevent atmospheric oxygen, water, or water vapor from reaching the weld site. Shielding gases are typically inert or semi-inert gases such as argon or helium. The flow rate of the shielding gas affects its ability to perform its function.

[0355] To ensure high weld quality, many operating variables must be carefully controlled within the welding process. These variables may be taken into account indirectly, as illustrated in the sections below. Although not directly sensed in the illustrated embodiment, other types of sensors may measure the speed of travel of the welding device, the flow of the shielding gas, and the flow rate of the shielding gas, and these sensor types and their data sets may be added to the processing.

[0356] (acoustic sensor type) The acoustic sensor type collects high frequency audio signals generated during welding. These signals arise from the arc forming the weld, the arc-substrate interaction, and the arc-shielding gas interaction. The detected audio signals were found to be sensitive to several key variables within the welding process. Figure 19 shows example values ​​of outlier scores for various weld characteristics.

[0357] The outlier score is obtained by a mathematical approach that considers how far a particular data value is from a set of known data values ​​that have been classified as acceptable values ​​for the characteristic and / or sensor type data being evaluated.

[0358] One such approach used in this disclosure is the use of a Mahalambis distance novelty detection model by comparing the incoming audio signal values ​​with previously developed principal component analysis (PCA) models of sensor types and their signal and / or data values.

[0359] Signal processing involves taking the audio signal and applying a denoising algorithm to the raw data. The audio signal is further processed using a short-time Fourier transform to convert the raw time series data into the frequency domain. Statistical features are then extracted from each of a series of bandwidths spanning the frequency range of interest. In this example, the bandwidth used is 39.1 kHz, generating 312 features that describe the acoustic signal of the particular signal instance. The entire feature set [312 features per signal instance] is then optimised by removing redundant features and standardised to improve robustness of features with small standard deviations.

[0360] In the initial setup of the PCA model, the signal resulting from the above processing, and therefore the remaining feature set, is set to be the acceptable performance of the weld with respect to that variable (in this case the acoustic signal). Thus, the remaining feature set may finally be reduced using PCA to obtain the principal components and model that define that model with respect to the acceptable performance of the weld.

[0361] The PCA model of acceptable performance may be used to consider subsequent signals. These subsequent signals are subject to the same noise cancellation and other steps defined above. For each feature, the location of the feature value relative to the distribution of feature values ​​for acceptable weld performance revealed, i.e., the feature value, is determined. The mean of the distribution is calculated and the distance of the feature value relative to the mean is measured. The distance provides a quantification of the outliers, shown in FIG. 19, revealing distance variations consistent with acceptable weld conditions versus distance variations that are more exceptional and indicative of poor weld performance.

[0362] This approach is useful for providing a unit-free, scale-invariant quantification that accounts for correlation of feature values ​​within a distribution. The mean can be recalculated each time an acceptable feature value is added to the distribution and / or PCA model, or it can be calculated based on a fixed set of existing acceptable feature values ​​used in the PCA model. Referring to sample results obtained using this process and displayed in Figure 19, the first set of data points indicate that good welding operation parameters have occurred. These parameters are verified individually as they are applied. As can be seen, they provide a nicely clustered set of data points A for the log-scale outlier score axis.

[0363] The second set of data points indicates that the welding speed is too fast, i.e. the welding device and substrate are moving too fast relative to each other. Two different welding speed deviations are shown, which are well-spaced with few outliers in two sets of data points B. 1 and B. 2 Gives.

[0364] The third set of data points shows sidewall arcing during the weld, i.e. the arc is shorting to the sidewall instead of to the intended weld location in the weld groove. Again, these give a well-consolidated data point set C with few outliers.

[0365] The fourth set of data points shows sidewall melting, i.e. the arc melts the sidewall and causes melting at the sidewall and not in the weld groove. The data points are well clustered with few outliers as data point D.

[0366] The fifth set of data points indicates that the shielding gas flow rate is too high, which could potentially lead to undesirable porosity issues. Like the other data point sets, this set E is well defined with few outliers. A similar location would be found if the shielding gas flow rate was too low.

[0367] When the aforementioned undesirable welding conditions are present, the outliers or feature values ​​are much higher than when good welding conditions are present. As a result, the threshold value Th acoustic A Mahalanobis distance selection value can be set and used to distinguish between data obtained from an acoustic sensor type that indicates good welding conditions and data obtained from an acoustic sensor type that indicates poor welding conditions. Thus, the acoustic data type is selected based on the threshold Th acoustic The system has the ability to alert the operation or trigger a stop of welding if the defect is breached or remains breached for a specified number of data points. This provides real-time acoustic signal based identification of defect creation. This analysis can be provided continuously and for each pass of a multi-pass welding approach.

[0368] Importantly, the algorithm used is more complex than simply using a single Principal Component Analysis (PCA) model previously developed for a sensor type and its signal and / or data values. As shown in Figure 20b, in multi-pass welding, the weld passes gradually fill the weld groove. This means that the weld groove depth, weld groove shape, and fill amount change from pass to pass. All these changes, and possibly other changes between passes, affect the acoustic signal emitted and detected. Therefore, this approach uses a separate model for each pass to most accurately evaluate the observed data against the predicted data for that pass.

[0369] The individual models can be obtained by a neural network approach that trains them based on existing paths, or they can learn to improve the model for a particular path in a sequence of paths as the paths are performed during operation and the number of each path increases.

[0370] The use of separate models for each pass also applies to the use of separate models for re-passes, for example as part of a weld repair operation: the geometry and therefore the acoustics of the weld groove in such a re-pass will be significantly different from a normal pass.

[0371] (Laser sensor type) The next sensor type used is visual and is intended to evaluate the geometry of the weld created.

[0372] FIG. 20a shows a cross section of a substrate 2 with a weld 20 formed. In this example, the weld 20 is linear, but other weld paths are conceivable as well. The visual sensor type device 22 includes a casing 24 with a light source 26 therein that can illuminate the substrate 2 and the weld 20 over an illumination width 28. The device 22 has a working range 30 that allows for accurate imaging. The light returns to the device 22 where a receiver 32 focuses the light onto a sensor matrix 34 and a signal is generated. In this example, a 2D laser profile scanner is used, but other types could be substituted.

[0373] A laser as light source 26 is used to reveal various details of the weld and its surroundings, such as the weld profile, the remaining groove profile, the weld bead width, and the material deposited outside the weld groove. The inspection may use a plane perpendicular to the substrate surface beside the weld groove and perpendicular to the longitudinal axis of the weld groove. Additionally, the weld bead profile along the weld groove may also be considered.

[0374] Figure 20b is a diagram of a typical weld groove over the course of a series of weld passes. Each weld pass adds welds to the welds already in the groove in a predetermined order [numbered] to build up the overall weld. As can be seen, the weld passes contribute to a predictable shape of the weld pass itself and a predictable change in the shape of the weld groove if the weld is progressing correctly.

[0375] The signal from device 22 may be used to form a profile image of the entire weld track at each position along the weld track. The actual profile may be compared to an expected profile and deviations may be recorded. The deviations may be measured using a threshold Th. profile Thus, the profile data can be compared to a threshold value Th profile has the ability to alert the operator and / or trigger a stop of welding if the profile signal is breached or remains breached for a specified number of data points. This can provide identification of defect creation based on real-time profile signals. This analysis can be provided continuously and for each pass of a multi-pass welding approach. Effective geometric validation is provided.

[0376] (visual sensor type) The next sensing area uses a high dynamic range camera to capture images of the weld location, including where the weld has not yet formed, where the weld is forming, and where the weld is solidifying and further cooling.

[0377] Figure 21 shows a series of images of this type collected from various weld locations. By processing these with a combination of artificial intelligence and traditional machine vision tools, the system can detect visual anomalies caused by changes in the images or defects associated with abnormal weld conditions. This can be accomplished by processing a single image or by combining and processing multiple images from previous data, both from recent images and from past parts and passes.

[0378] In this area, one or more variables may be considered. For example, the size of the weld pool [width, trailing length, leading length], the shape of the weld pool [oval, teardrop shaped, etc.], the temperature of the weld pool may be considered along with an analysis of the pattern, shape of the deposited material, and visible anomalies. A profile of known decision parameters for various types of defects may be used to compare the live output from each part of the algorithm to known good values. These values ​​may be a combination of the presence or absence of a particular visual feature, a numerical band or threshold, or a classification. If the image is determined to be beyond these parameters, the occurring situation for one or more or all may be compared to a desired position for one or more or all, and the deviations may again be used to trigger an alert or to stop the weld.

[0379] (Voltage and current sensor types) The applied voltage affects the formation of the arc and the current in the arc. This in turn affects the power and therefore the rate of dissolution of the base material [and electrode, if consumed]. These are important variables for the quality of the weld. These are the variables that affect, for example, the size of the weld pool.

[0380] The welding voltage also needs to be automatically and continuously adjusted to reflect the distance between the welding equipment and the substrate being welded, based on a known, fixed position of the substrate and a variable but known XYZ position of the robotic arm carrying the welding device, i.e. the separation of these two positions.

[0381] The sensing system for monitoring voltage and current is orders of magnitude faster than those built into existing automatic voltage control approaches. The power system is capable of handling 500A and is scalable up to 1000A and beyond, and also provides voltage and current monitoring at the nanosecond level between measurements. Thus, very detailed information of voltage and current is obtained and short-term variations can be taken into account.

[0382] To avoid the sensing / detection itself interfering with the output power performance, there is a tendency for the input voltage and current, and therefore the input power, to be detected rather than the output power.

[0383] The data and the approach used to process it is similar to that of the acoustic sensor type described above.

[0384] 22, the arc voltage is plotted against time as plot V. In this example, the welding process is shut down after 30 seconds and the voltage returns to zero.

[0385] Also plotted in Figure 22 is the Gaussian Amplitude x Gaussian Center value versus time. As can be seen, for an initial period of the first 2 or 3 seconds, the values ​​on this plot are above the acceptable threshold, indicating concern for the quality of the weld. After the first 2-3 seconds, and certainly after 8 seconds, the plot drops well below the threshold, indicating a high quality weld for this set of variables.

[0386] In addition to these variables, Figure 22 also includes an indication of when argon is scarce as a shielding gas. The indication shows the occurrence of argon deficiency by a plot point, and also the degree of argon deficiency by the vertical position of the plot point.

[0387] (Overall Process – continued) Returning to FIG. 17, now that the details of the operation of the sensor types and the consideration of their data have been clarified, the process and the data therein undergo a data reconciliation step. Data reconciliation is provided based on input provided through the user interface and / or historical data obtained from storage. The storage may contain data from earlier in the performance of this weld and / or data of a number of previous welds performed by the system, and / or data from other systems [e.g., data from calibration processes, etc.], all of which may contribute to the historical data and therefore to the data reconciliation.

[0388] In the live analysis step, two different levels of processing may be implemented.

[0389] The first level of processing provides synchronization of data types from different sensor types, not only those illustrated above, but also any number of other sensor types that may be introduced and used to measure key characteristics directly and / or indirectly related to the welding system and the occurring weld.

[0390] A programmable logic controller (PLC) is used to temporally synchronize the various data types by providing a master timestamp at the start of data collection from the sensor types in the system. The PLC continuously checks that each sensor of each sensor type in the system is continuously providing data, and that the provided data is being collected at a constant rate. Additional master timestamps may be applied periodically while the data collection process is ongoing.

[0391] A master time stamp can mean that data collected from a microphone serving as a sensor type for acoustic analysis, an input power analysis for draw measurement during welding, an area scan camera providing a vision system, and a laser profile scanner providing a 3D profile sensor type are all coordinated to represent data from the same time, i.e., the same location within the weld.

[0392] Importantly, the PLC also receives data from an incremental encoder that provides position data. The encoder may be attached to the substrate being welded and / or the welding equipment, indicating its physical location at that time. The encoder triggers data collection and provides consistent correlation of the sensor data with the weld position of the component. Again, the same master timestamp is applied to the data from the encoder. This means that the actual location of the same revealed position that can be applied simultaneously in a synchronized data arrangement is known. The data is synchronized in time and space.

[0393] This processing allows sensor data collected at various locations on the component to be post-processed into one cohesive data structure, allowing the raw sensor data, post-processed information, and analytical output to be correlated with the physical location of the weld. This data can be displayed to the operator in a user display, allowing them to adjust the welding process based on the system's output.

[0394] The second level of processing is implemented by data reduction and / or the application of machine learning. This second level of processing takes the correlated data and feeds it into a second level of processing and analysis (using machine learning or other suitable analytical methods) to correlate any number of features from both the raw data and the output of the sensor type level analysis to weld defects. In practice, a decision engine supporting machine learning is employed.

[0395] In this method, where a defect is suspected, a general assessment is made as to whether the defect is within acceptable characteristics, such as size, or exceeds acceptable characteristics, and thus a defect requiring recording or repair action. When NDT-type sensing is performed, direct measurement of the defect is made by the imaging performed. This directly reports the size and possibly other characteristics of the defect. However, when the consideration of defects is based on the welding conditions occurring during welding, the potential defect is being considered indirectly, and the question is being considered: "Are these conditions likely to lead to a defect?" A second level of processing may improve the determination of acceptable and / or unacceptable welds in this context.

[0396] 24, two different sensor types are considered: Type A on the left and Type B on the right. The two sensor types can be any of the sensor types described herein and / or other types of sensor types that provide data regarding the welding performance or results of the welding.

[0397] Referring to the Type A sensor type, within region 800 there is a series of data points that have been reliably established as indicative of an acceptable weld. These may be evident from test runs verified by other sensing and / or NDT, or may be modeled cases.

[0398] There are also data points from a weld, such as data point 802, that are clearly outside the acceptable range and are well on the unacceptable side of threshold 804 that may be applied in initially determining whether a data point represents an acceptable or unacceptable weld. Other data points, such as data point 806, exceed the threshold and are considered unacceptable under the definition of threshold 804.

[0399] Instances that are difficult to interpret are data points that are outside of region 800 but below threshold 804, such as data points 808 and 810. A decision based on a single sensor would determine that these welds are acceptable because they are below the threshold.

[0400] Gains are gained from a second level of processing that considers data points [802, 806, 808, 810] across multiple sensor types to make a complete determination. Referring to the right side and to the Type B sensor type, data point 802 again lies within the identified acceptable weld region 800. Similarly, data point 806 is well above threshold 804, again indicating an unacceptable weld from this sensor type alone.

[0401] Looking at data points 808 and 810, both are outside of region 800 but below threshold 804. In a single sensor type approach, these edge cases would again be considered acceptable welds, however, a second level of processing takes into account the location between multiple sensor types to gain additional information.

[0402] The use of neural networks can employ two approaches to considering location across multiple sensor types, which can be used as alternatives to each other, or one in parallel with the other, or even in series.

[0403] In the first approach, labeled data is provided and supervised learning takes place. The labeled data may come from either or both of two sources. First, especially in the early stages of processing, such as during calibration or early production weld runs, the labeled data may come from experimental results. So, continuing with the example of FIG. 24, the locations of both points 808 and 810 may be flagged to the operator to allow for evaluation by the operator as acceptable or unacceptable. The operator may not only be provided with data from a single sensor type to review and make a decision, but with flagged data from multiple different sensor types on that data [a data point or sequence of data points] and may make a decision based on more nuanced criteria. The results are used to label the data and thus are utilized within the data pool from which the neural network learns. Human knowledge and interpretations are input into the neural network by user decisions to provide supervision.

[0404] A second method of providing labeled data is to utilize a library of existing data. This data is also labeled according to the decision and imparts the necessary operator knowledge. The library is fed into a neural network classifier to reveal the processing location of the library data set. The library can span data from multiple sensor types and determine a second level of processing type, which is the location of the data point across the tests and results of multiple sensor types. The classifier can then determine a classifier score for any data point associated with the library data, and then quantify the likely error of the data point [or set of data] that is being examined to make a decision. The classifier can use a Bayes' Theorem based approach for classification and error quantification.

[0405] The second method, a library, may be used as the starting point for a labeled data set. The first method, an operator calling to perform a calibration or test run on the actual welding system, may be used as an alternative from the start. However, the first method may be used to add to the data set of the second method to progress the training of the neural network from more general welding system positions to positions tuned for that particular welding system.

[0406] Returning to FIG. 24, an operator may determine that data point 808 is acceptable because it is close enough to the tolerance region 800, but that data point 810 is unacceptable because it is too close to the threshold across multiple sensor types. This may cause the neural network to make small adjustments to the tolerance region 800 and / or threshold 804 [value or form]. Over time, repeated decisions of this type may lead to more pronounced and optimized modifications of the boundaries of the tolerance region 800 and / or threshold 804. For example, the tolerance region 800 may be expanded and / or the threshold may be tightened. The same results for data points 808 and 810 at a later, more advanced learning stage in the process may be referred to as acceptable and unacceptable because in the case of data point 808 it is within the revised tolerance range, and in the case of data point 810 it is beyond the revised threshold.

[0407] An example of such an analysis in a real-world scenario might be where the left side relates to a visual image sensor type, and one or more images in succession suggest there is a problem with the sidewall proximity of the weld. Considering an acoustic sensor type on the right side, a sidewall proximity problem might be suggested, thereby confirming the overall determination that the weld is not acceptable.

[0408] While the above example references the location of one type of sensor relative to the other type of sensor to determine acceptability or unacceptability, the determination may be more detailed than that and determine the nature of the problem with the welding conditions. Thus, data from two types of sensors may inform the nature of the problem even if data from one type of sensor alone only suggests a problem.

[0409] An issue to note is that importing historical data from other welding situations and environments is not a strong starting point for a library to determine acceptable weld performance. This is because other welding operations and other welding environments have different variables that can affect the data for those welding operations. For example, in the case of a visual imaging sensor, the lighting and lighting angle of the welding environment, the nature of the substrate, the angle and spacing of the welding torch, etc. can all affect the data, which may not match data retrieved in another environment where the lighting is different, for example.

[0410] As an alternative or in addition to this first approach using labeled data and supervised learning, the ability of neural networks to perform clustering or grouping-based processing [looking for similarities and / or anomalies in the data] may be utilized to reduce or avoid the need for library-based data and / or supervised learning. There are various techniques, such as K-means clustering, that can establish the centroid of a cluster and reveal distance certainty around it. Several other clustering techniques can also be applied. These can be used to reveal the location of acceptable regions 800 and / or thresholds 804, which can be modified with more data and learning. One can get started without needing a large library of labeled data.

[0411] As mentioned above, the first and second approaches may be used in combination, not just as alternatives. Thus, the first approach may be used to begin training the neural network, and the second approach may take over after training with the first approach has progressed. It is also possible for the neural network to simultaneously train from both the first and second approaches to maximize the data fed to it, especially if the library is being expanded from ongoing welds on other welding systems other than the welding system being considered. It is also possible to pool training from similarly configured and operated welding systems.

[0412] Over time, the amount of data, the accuracy of the evaluation, and the complexity of the various cases of data that can be successfully evaluated will increase, especially if the method is used in a production version with large volumes of welding.

[0413] All results obtained from live analysis of data and / or data reduction and / or machine learning may have a copy sent to storage after acquisition for future use or to contribute to the knowledge of the system and / or similar systems.

[0414] A key step in the analysis results is to compare the data value or position to one or more thresholds set for the data type (thresholding). Examples of thresholding approaches are provided in various sections for the specific sensor types illustrated, but are broadly applicable to each sensor type and the data type it produces. The position relative to the threshold may be considered indicative of a defect (defect indicator) and may be an important part of the weld quality result, displayed to the user via the user display, which also receives and displays the received sensor data.

[0415] 23 is an example of a user display. The user display may provide a real-time output of the analysis of all sensor types so that the welding process can be reviewed as it is running and the process can be adjusted if a poor quality weld is found in the output of the analysis. Additionally, it is also possible to review the entire weld data after each pass and examine the data to identify areas that need repair or adjustment before or during the next weld.

[0416] The weld quality results step may stop the welding process and / or issue a warning to the operator, such as via a user display, if an unacceptable weld is encountered.

[0417] Based on the position of the user display, the operator or the system itself may adjust one or more variable parameters used to control and conduct the welding process in real time.

[0418] The weld quality results are reflected in a Weld Qualification, which is provided for the entire weld. If a defect is identified by NDT, the size and location of the defect are checked against appropriate criteria to ensure that the level of the defect is acceptable. If not, remedial action is taken. If acceptable, 3D life record data is stored within the Weld Qualification and made available throughout the life of the weld and for any required decommissioning support.

[0419] (Data processing and storage) In addition to considering each data type individually, further benefits are gained by considering the data types as a combined data stream. To enable this, each data set is time-stamped, and synchronizing the timestamps synchronizes the time of all data sets.

[0420] This means that all data sets can be combined and saved as one data file, allowing for subsequent reuse.

Claims

1. 1. A method for providing inspection of a weld, comprising: (a) providing a weld inspection device proximate to a weld on a substrate to be inspected; (b) performing an inspection, wherein the substrate is subjected to an elevated temperature above ambient temperature by heating during the inspection, and performing the inspection comprises: a. emitting ultrasonic waves into the volume of the substrate and weld; b. receiving at least a portion of the ultrasonic waves from the substrate and the weld, thereby obtaining a plurality of sets of signals; conducting an inspection, (c) processing one or more of the plurality of signal sets to provide weld inspection data; Including, The treatment comprises correcting the temperature distribution within the volume of the substrate and / or weld at elevated temperatures; A method for providing inspection of welds.

2. The method of claim 1 , wherein the correcting comprises correcting a path of at least a portion of the ultrasonic wave into the volume of the substrate and weld to provide a corrected path.

3. 3. The method of claim 2, wherein a portion of the ultrasonic waves has path characteristics within a media element before the portion of the ultrasonic waves enters a first element of the substrate, the first element having a temperature within the temperature distribution of the volume of the substrate and / or weld, and a temperature-compensated path characteristic is determined for the portion of the ultrasonic waves within the first element, the temperature-compensated path characteristic being based on a temperature change between the media element and the first element.

4. 4. The method of claim 2 or 3, wherein a portion of the ultrasonic waves has path characteristics in a first element of the substrate, the first element having a temperature within a temperature distribution for a volume of the substrate and / or weld, the path characteristics directing a path of the portion of the ultrasonic waves to a second element, the second element having a temperature within a temperature distribution for the volume of the substrate and / or weld, and a temperature-compensated path characteristic is determined for the portion of the ultrasonic waves in the second element, the temperature-compensated path characteristic being based on a temperature change between the first element and the second element.

5. The method of claim 4 , wherein the temperature corrected path characteristics are determined for each element in the substrate and / or weld through which a portion of the ultrasonic waves passes.

6. The method of claim 3 , wherein the temperature change is expressed as a change in speed of sound between the speed of sound at one element and the speed of sound at the next element.

7. 4. The method of claim 1, wherein different portions of the ultrasonic waves into the volume of the substrate are compensated to provide compensated paths.

8. 4. The method of claim 1, wherein a region of interest is selected, the region of interest being within a volume of the substrate and weld through which the ultrasonic waves have passed, the region of interest being subdivided into locations such as pixels, and a correction is applied to the locations such as pixels.

9. 4. The method of claim 1, wherein the method includes correcting for one or more locations, such as pixels, through which the emitted ultrasound beam does not pass, based on calculated or observed corrections for locations through which the emitted beam passes.

10. 10. The method of claim 9, wherein the correction is based on calculated or observed positions for a plurality of locations on the first emitted beam and a plurality of locations on the second emitted beam.

11. 9. The method of claim 8, wherein the correction for the location is a weighted combination of corrections for one or more other locations, for example one or more other locations through which the emitted beam has passed.

12. 9. The method of claim 8, wherein the correction to the location is weighted according to the ratio of the distance the location lies between the location on the first beam and the location on the second beam.

13. 9. The method of claim 8, wherein the correction to the location is weighted according to a proportion of the distance the location lies between a first location on a first beam and a second location on a first beam.

14. 4. The method of claim 1, wherein the result set includes one or more measured representations of the shape of the substrate and / or weld groove and / or weld, and the method further includes comparing the measured shape representations with a modeled shape representation, and accepting imaging of the region of interest if the comparison between the measured shape representations and the modeled shape representations reveals that the measured shape representations sufficiently match the modeled shape representations.

15. 4. The method according to claim 1, wherein the result set comprises one or more measured representations of the shape of the substrate and / or weld groove and / or weld, and the method further comprises comparing the measured representations of the shapes with a modeled representation, and if the comparison of the measured representations of the shapes with the modeled representations reveals that the measured representations of the shapes are an insufficient match to the modeled representations of the shapes, redetermining a temperature distribution used to correct the temperature distribution within the volume of the substrate and / or weld at high temperatures.

16. 4. The method of claim 1, wherein the method comprises providing a thermal model and generating a modeled temperature distribution location at the high temperature state for at least a portion of the substrate, the portion of the substrate comprising the volume, using the thermal model.

17. 17. The method of claim 16, wherein the method includes measuring temperature at a plurality of locations with the substrate at an elevated temperature to obtain measured temperature distribution locations, and further includes comparing the measured temperature distribution locations to the modeled temperature distribution locations.

18. 17. The method of claim 16, further comprising modifying the thermal model and / or the modeled temperature distribution locations and re-comparing if a comparison of the measured temperature distribution locations with the modeled temperature distribution locations reveals that the modeled temperature distribution does not adequately fit the measured temperature distribution.

19. 17. The method of claim 16, further comprising: calculating characteristics of ultrasonic waves emitted during passage through at least a portion of the substrate and / or weld if comparison of the measured temperature distribution location with the modeled temperature distribution location reveals that the modeled temperature distribution is a sufficient match to the measured temperature distribution.

20. 1. An apparatus for providing inspection of a weld, comprising: (a) a weld inspection device that, in use, is positioned adjacent to a weld on a substrate to be inspected, and that heats the substrate during inspection, thereby subjecting the substrate to an elevated temperature above ambient temperature; (b) the welding inspection device, a. Welding by emitting ultrasonic waves into the volume of the substrate and weld; b. receiving at least a portion of the ultrasonic waves from the substrate and the weld, thereby obtaining a plurality of sets of signals; Fits The welding apparatus includes one or more processors, the processor having an input for one or more of the plurality of sets of signals, the processor providing a correction for the input one or more of the plurality of sets of signals to provide corrected weld inspection data, the processor applying a correction for temperatures within a volume of the substrate and / or weld at elevated temperatures.

21. A welding method comprising: (a) providing welding equipment; (b) providing a plurality of sensor types; (c) defining a first set of welding conditions for the welding method; (d) introducing one or more substrates to be welded into said welding apparatus; (e) performing welding of one or more substrates; (f) acquiring data from the plurality of sensor types during welding; (g) providing a correlation between data obtained from at least two selected sensor types of the plurality of sensor types; (h) synchronizing one or more data points in data acquired from one of the at least two selected sensor types with one or more data points of another of the at least two selected sensor types; (i) comparing data obtained from the plurality of sensor types with reference data for one or more of the sensor types; (j) determining whether the weld is of acceptable or unacceptable quality based on the one or more comparisons; and Including, (k) if the quality of the weld is unacceptable, the method includes performing one or more actions; Welding method.

22. The method of claim 21 , wherein the correlation is a temporal correlation and a location correlation.

23. The method of claim 21 , wherein the correlation is a temporal correlation.

24. 24. The method of any one of claims 20 to 23, wherein the data from at least two selected sensor types includes a time of occurrence of one or more data points in the data, and the correlation is a match of the time of occurrence of data points in the data from the at least two selected sensor types.

25. 25. The method of claim 24, wherein the time of occurrence is obtained for each of the two or more selected sensor types providing the correlation relative to a timestamp introduced into the data from the selected sensor type.

26. 26. The method of claim 25, wherein the time of occurrence is obtained directly from a time stamp at that time.

27. 26. The method of claim 25, wherein the time of occurrence is obtained by the time elapsed since the timestamp and the time of the data point.

28. The method of claim 21 , wherein the correlation is a positional correlation.

29. 24. The method of any one of claims 21 to 23, wherein the data from the at least two selected sensor types includes a location of occurrence of one or more data points within the data, and the correlation is a match of a location of occurrence of data points within the data from the at least two selected sensor types.

30. 25. The method of claim 24, wherein for each of the two or more selected sensor types providing the correlation, a location of occurrence is obtained relative to a location stamp introduced into the data from the selected sensor type.

31. 26. The method of claim 25, wherein the location of occurrence is obtained directly from the location stamp of the location.

32. The method of claim 25, wherein the location of occurrence is obtained by the time elapsed since the location stamp and the time of the data point.

33. 24. The method of any one of claims 21 to 23, wherein synchronization provides an overall data structure in which data points from the at least two selected sensor types are aligned with each other in time and / or position.

34. 24. The method of any one of claims 21 to 23, wherein data points from the at least two selected sensor types are displayed to the user aligned with respect to time and / or location of occurrence.

35. 24. The method of any one of claims 21 to 23, wherein the data from the two or more selected sensor types and the correlated data from the at least two selected sensor types is displayed to a user and / or stored.

36. 36. The method of claim 35, wherein the determination of whether the weld is of acceptable or unacceptable quality and / or the one or more actions taken are displayed and / or saved.

37. 24. The method of any one of claims 21 to 23, wherein the method includes providing a correlation between data obtained from at least four selected sensor types of the plurality of selected sensor types.

38. 24. The method of any one of claims 21 to 23, wherein at least one selected sensor type of a plurality of selected sensor types is part of a weld inspection device, and the method includes inspecting the weld using the weld inspection device.

39. 40. The method of claim 38, including inspecting the weld using the weld inspection device to determine one or more characteristics of a defect.

40. 40. The method of claim 39, wherein the characteristics include one or more of size, location, defect type, defect shape, or defect location relative to the shape of the weld and / or the length of the weld with reference to the shape of the weld and / or relative to the length of the weld.

41. 40. The method of claim 39, wherein the method further comprises comparing one or more characteristics to one or more criteria to determine whether the weld having the defect meets or fails to meet welding criteria.

42. 42. The method of claim 41, wherein if the weld meets the welding criteria, a record of the weld is created and saved.

43. 43. The method of claim 42, further comprising: the record including data from one or more of the plurality of sensor types.

44. 42. The method of claim 41, wherein if the weld does not meet the welding criteria, one or more repair steps are applied to the weld.

45. 24. The method of any one of claims 21 to 23, wherein the at least two selected sensor types of the plurality of selected sensor types are weld condition sensors, the method including inspecting the weld condition using the weld condition sensors.

46. 46. ​​The method of claim 45, including inspecting the weld to determine one or more parameters of the weld as it is formed.

47. 47. The method of claim 46, further comprising comparing the one or more parameters to one or more control parameters, including determining whether a risk level for a weld defect has been exceeded.

48. 24. The method of any one of claims 21 to 23, wherein the method further comprises one or more operations of changing the welding conditions from the first set of welding conditions of the welding method.

49. 49. The method of claim 48, wherein the change in the welding conditions from the first set of welding conditions is to stop welding and / or to alert an operator.

50. 49. The method of claim 48, wherein changing the welding conditions from the first set of welding conditions is changing the welding conditions back to the first set of welding conditions and / or changing the welding conditions to a second set of welding conditions.

51. 24. The method of any one of claims 21 to 23, wherein at least one of the at least two selected sensor types is selected from a voltage sensor, a current sensor, a welding arc sound emission sensor, a welding topology sensor, a welding image sensor, and an ultrasonic image sensor.

52. 1. An apparatus for monitoring welding, said apparatus comprising: (a) multiple inputs for data from multiple sensor types; (b) one or more processors, wherein a processor of the one or more processors: a. Receives input; b. processing the data obtained from at least two selected sensor types from the plurality of sensor types and applying a correlation between the data obtained from at least two selected sensor types from the plurality of sensor types; c) using the correlation to process the data acquired from at least two selected sensor types of the plurality of sensor types and synchronize one or more data points in the data acquired from one of the at least two selected sensor types with one or more data points in the data acquired from another one of the at least two selected sensor types; one or more processors; (c) one or more outputs for the processed data; and 1. An apparatus for monitoring welding, comprising:

53. 53. The apparatus of claim 52, wherein the processor is adapted to provide a temporal correlation, and the data output from the processor includes time information in the data from one of the at least two selected sensor types and time data in the data from another of the at least two selected sensor types.

54. 54. The apparatus of claim 52 or 53, wherein the processor is adapted to provide a positional correlation, and the data output from the processor includes position information in the data from one of the at least two selected sensor types and position data in the data from another of the at least two selected sensor types.

55. 54. The apparatus of claim 52 or 53, wherein the processor is adapted to provide a time of occurrence of one or more data points in the data from the at least two selected sensor types, and the correlation is a match of times of occurrence of data points in the data from the at least two selected sensor types.

56. 54. The apparatus of claim 52 or 53, wherein the processor is adapted to provide a position of occurrence of one or more data points in the data from the at least two selected sensor types, and the correlation is a match of a position of occurrence of data points in the data from the at least two selected sensor types.

57. 54. The apparatus of claim 52 or 53, wherein input is from at least one of the plurality of sensor types, possibly at least one of the selected sensor types, and wherein the sensor types and / or selected sensor types are selected from a voltage sensor, a current sensor, a welding arc sound emission sensor, a welding topology sensor, a welding image sensor, and an ultrasonic image sensor.

Images