WO2023003512A2

WO2023003512A2 - Additive manufacturing method and apparatus

Info

Publication number: WO2023003512A2
Application number: PCT/SG2022/050511
Authority: WO
Inventors: Sudharshan RAMAN
Original assignee: Phasio Pte. Ltd.
Priority date: 2021-07-21
Filing date: 2022-07-19
Publication date: 2023-01-26
Also published as: WO2023003512A3

Abstract

An additive manufacturing method and apparatus, wherein the method comprising: receiving print data of an object printable by an additive manufacturing (AM) machine; obtaining expected print data of the printing of the object; obtaining actual print data of the printing of the object provided by the AM machine; comparing the expected print data with the actual print data to obtain a comparison result; determining the printed object as a failed product if the comparison result is out of bounds of predetermined upper serviceable limit (USL) and/or lower serviceable limit (LSL); and certifying or approving the printed object to proceed with further testing if the comparison result is within the limits of the predetermined USL and/or LSL.

Description

ADDITIVE MANUFACTURING METHOD AND APPARATUS

Technical Field

The present invention relates to the field of Additive Manufacturing (AM) and involves an additive manufacturing method and apparatus. In particular, it relates to the use of real-time or non-real-time data from in-process monitoring systems on Extrusion-based deposition systems to perform qualification and certification of printed parts of extrusion-based deposition systems.

Backqround

Additive Manufacturing (AM) is the field of manufacturing through deposition, sintering, welding, or otherwise forming material into a desired shape layer-by-layer. The advantages of AM over traditional manufacturing approaches are many-fold, including but not limited to Just in-Time production of parts, manufacturing of specialised alloys, formation of complex topologies and localised manufacturing. Many use cases of manufacturing involve the development of a digital inventory, wherein digital representations of parts are stored and kept ready for print. The digital storage of the part incurs minimal cost versus the physical storage of the actual part, yet the part can be printed quickly enough to satisfy customer demands.

Historically, the number of parts printed is always the same as the number of parts requested by the customer. The printing of additional parts for certifications such as ISO-9001 results in additional cost to the producer, along with the environmental repercussions of wasted material and energy. Development of a mechanism to eliminate this qualification wastage is therefore of interest for economic, environmental, and efficiency reasons.

Summary

According to an example of the present disclosure, there is provided an additive manufacturing method and an additive manufacturing apparatus as claimed in the independent claims. Some optional features are defined in the dependent claims.

Brief Description of Drawinqs Examples in the present disclosure will be better understood and readily apparent to one skilled in the art from the following written description, by way of example only and in conjunc tion with the drawings, in which:

FIG. 1 is a flowchart illustrating a certification and/or quality approval process for a printed product according to an example of the present disclosure.

FIG. 2 shows an example of an architecture of a processing unit or processor described in the present disclosure.

FIG. 3 is a flowchart illustrating an example of a workflow according to an example of the present disclosure.

FIG. 4 shows a graph of bed temperature of a sensor in an AM machine versus time comprising a range of acceptable values that is shaded and a graph plot of actual print data of the sensor.

FIG. 5 shows a graph of extruder temperature of a sensor in an AM machine versus time comprising a range of acceptable values that is shaded and a graph plot of actual print data of the sensor.

Description of the Invention

Examples of the present disclosure relate generally to quality assurance and certification systems for additive manufacturing. Additive manufacturing is the process of creating an object by building it one layer at a time.

A preferred example of the present disclosure includes a printer which is an extrusion-based deposition system or other additive manufacturing system which produces parts through a deposition process and relies on computer numerical control for instructions to build the part.

In additive manufacturing, a file, usually referred to as a build file, a GCODE file, or a print file is provided to the printer as a list of instructions for the machine to execute in order to produce the desired part.

For each GCODE file, a series of Upper and Lower Serviceable Limits (USL/LSL) shall be defined, with the goal of setting safe operating zones within which a part can be considered of good quality. These limits are defined as:

USL, LSL = max(f(x_\) g(E(x_t))), min(f(x_t) - g(E(x_t))) Where x_t is the sensed value at time / during the print, (x_t) is the expected value of x_t derived from the print file, (x) is a function, linear or nonlinear, which maps x_t into a ‘scoring’ space. Likewise, ((x_t)) is a function, linear or nonlinear, which maps (x_t) into a ‘scoring’ space.

The functions defining USL and LSL for different sensors may vary because a limit for temperature cannot be applied. For example, XYZ sensors may be evaluated through comparison through the expected values in the GCODE file, whereas the temperature and other physical sensors may be evaluated through analysis of previously completed prints, or simulation of the nominal thermal, stress, or harmonic environment.

The monitoring system is a real-time data ingestion system which records data from the printer on a real-time basis. This system could be connected to the printer via serial, WiFi, Bluetooth, ethernet or any other means through which real time data can be transmitted.

Before the print file is sent to the printer, a series of commands are added to the file to instruct the printer to produce verbose sensor updates. This step is required in many printers but not all printers, and can be omitted if it is not required. In the case of many extrusion-based deposition systems printers, this command is the M114 GCODE command, which provides position updates.

This technique also applies to other mechanisms of requesting verbose sensor updates from the printer, including polling and non-polling methods.

The modified print file is submitted to the printer and the print is allowed to start. Throughout the print, the monitoring system will record data inclusive of but not limited to print nozzle position(s), temperature sensors, pressure sensors, gas or other environmental sensors. The monitoring system may or may not record images of the part at key points throughout the print. The sensor (measurement) updates shall occur with a sampling rate, ideally within the range 1 Hz to 50Hz.

After the print, the print file (or GCODE file) is analysed to determine the expected values of each sensor measurement throughout the print. These expected values may include print nozzle position(s), temperature sensors, pressure sensors, gas or other environmental sensors. The monitoring system now, having access to both the true data from the sensors and also the expected data from the print file, compares the sensors one-by-one using the scoring mechanism described above.

This comparison from the monitoring system may take place in real-time, or retroactively after the print is completed.

Violations of the USL and LSL shall be classified as either critical or non-critical. If a violation of the service or serviceable limit exceeds the critical threshold, this print shall be considered to have failed the qualification. Otherwise, this print shall be considered to have passed the process qualification.

It is also possible to extend the above process by making use of auxiliary data such as cameras and thermal images. In this case, the process described in the above steps shall be extended to include an extra step. In this step, parts passing the criteria already described above will undergo an additional review prior to qualification wherein an operation or computer vision algorithm inspects the images for defects including but not limited to delamination, lack of forming, impact damage to the print surface or warping.

The steps for obtaining a Digital DNA of an AM printed part (or object or component) using a system or an apparatus of an example of the present disclosure are as follows with reference to FIG. 1.

A user logs in to a webapp (a web application called “Phasio”) so as to be able to check the status of 3D printing/Additive manufacturing (AM) machines in a facility. This can be done by a computer (e.g. a desktop computer) or a mobile device. The mobile device can include a tablet device, a smartphone, a portable laptop or notebook, and the like. With reference to the example of FIG. 1, after logging in, the user can prepare for a print job at a step 102. The apparatus and/or the system comprising the software and hardware that enable the webapp to operate can be regarded as a monitoring system 104. The monitoring system 104 may include a processing unit (or processor) configured to execute instructions in a memory to operate the monitoring system 104 to perform its functions.

In the present disclosure, a 3D printer is an AM machine and these terms are used interchangeably. An AM machine can refer to a broader class of machines for AM, which includes the 3D printer. The user then chooses a 3D printer/AM machine in which he/she is going to print a part (component).

The user uploads a file in an acceptable file format, which is compatible with the respective printer. With reference to FIG. 1, the file will be submitted to the 3D printer/AM machine at a step 126.

The uploaded file contains the design of the part to be manufactured which has been approved by the user or the user’s organisation to be deployed to the 3D printer/AM machine.

This uploaded file contains the process parameter (or parameters), which is set by the user for the specific 3D printer/AM machine, and consumable (or consumables). This process parameter has already been verified to help produce a certified component. Although the terms “parameter” and “consumable” written here are singular, it is understood that more than one parameter and consumable can be involved during actual operation.

In the example illustrated by FIG. 1 , the monitoring system 104 determines expected print data in two ways. In a first way, the monitoring system 104 determines expected values from a GCODE file related to the part to be printed at a step 106 and build a model of expected print parameters at a step 108. The expected values can include XYZ values relating to coordinates in an XYZ coordinate system (i.e. the Cartesian system). In 3D printing, these coordinates can determine a position of an impact point of a laser, an electron beam, a hot end nozzle or the like, which may be moved around by different rails and driving systems in the 3D printer/AM machine. The expected values can also include sensor readings. For example, the monitoring system 104 can determine the expected XYZ values and expected sensor readings from the GCODE file.

Expected print data 112 will be produced from the model of expected print parameters after iterating over expected measurements at a step 110. The step 110 involves going through data of expected print parameters of previously completed prints (in particular, prior successful prints) of the part to identify a range of acceptable print parameters with upper serviceable limit (USL) and lower serviceable limit (LSL). The term “serviceable” means a state in which the part (or object or component) is still serviceable i.e. still deemed to be a useful product. The data of this range of acceptable print parameters and the upper and lower serviceable limits constitute the expected print data 112. In a second way, separately, expected values of physical sensor measurements are determined statistically or computationally at a step 130 by the monitoring system 104. The physical sensor measurements are iterated over expected measurements at a step 132 to obtain the expected values of the physical measurements 134 The step 132 involves going through expected measurements of previously completed prints (in particular, prior successful prints) of the part to identify a range of acceptable measurements with upper serviceable limit (USL) and lower serviceable limit (LSL). The data of this range of acceptable measurements and the upper and lower serviceable limits constitute the expected physical measurements 134.

Although the present example has 2 ways of determining expected print data, it would be appreciated by a skilled person that in another example, it is possible to only implement one of these 2 ways.

Information of the consumable is documented by the user into the webapp. This information should include the source of the consumable (a.k.a. where it is purchased from), the industry name of the consumable if available or an identifiable name of the consumable, and/or the tested nominal chemical composition of the consumable.

The consumable should meet the criteria for an ideal consumable to be used in a 3D printer/AM machine. This is to ensure the quality of the data being streamed from the 3D printer/AM machine to the webapp.

The user does a check to see if the 3D printer/AM machine is of proper working condition. It is to ensure the 3D printer/AM machine working condition does not affect the data quality being streamed to the webapp.

The user does a check to see if the 3D printer/AM machine is in a condition to be used. This is to ensure there are no remnants of an earlier process affecting the quality of a scheduled printing.

The user makes sure that all conditions are met before the start of the 3D printer/AM machine process. This is to ensure that Standard Operating Conditions (SOPs) are in place to produce process reliability.

The user starts the scheduled printing process, which triggers a stream of data which is sent from the 3D printer/AM machine to the webapp. In the example of FIG. 1 , such stream of data includes print data 128 produced by the 3D printer/AM machine based on the file submitted to the 3D printer/AM machine at the step 126.

In the example of FIG. 1, (1) the expected data produced by the monitoring system 104, (2) the actual print data 128 produced by the 3D printer/AM machine, and (3) the expected physical measurements 134 produced by the monitoring system 104 are streamed to the webapp for processing. Based on (1), (2) and (3), an error signal 114 can be produced if the print data 128 vary from data determined from the expected data 104 and the expected physical measurements 134. The print data 128 can include actual physical measurements and/or actual print parameters determined by sensors of the 3D printer/AM machine. (1) and (3) are essentially expected or reference print data for comparison with (2). (1) and (3) can be predetermined before they are compared with (2).

The data streamed from the 3D printer/AM machine to the webapp is collected and stored either in a cloud or on a server at users’ site.

The collected data is displayed in a dashboard in the webapp, which can be accessed by a computer (e.g. a desktop computer) or mobile device. The mobile device can include a tablet device, a smartphone, a portable laptop or notebook, and the like.

The collected data (e.g. the print data 128 of FIG. 1) can be checked in real time or non-real time to see if they fall within the respective upper serviceable limit (USL) and the respective lower serviceable limit (LSL). With reference to the example of FIG. 1, the error signal 114 produced is compared against the USL and/or LSL at a step 116.

The USL and LSL is set for all the sensor readings which can be exposed from the 3D printer/AM machine. Examples of such sensors are Laser Power, Base plate temperature, etc. For example, the error signal 114 of FIG. 1 can refer to an error signal of each of a plurality of such sensors, wherein specific USL and/or LSL are predetermined for each sensors.

As mentioned earlier, the USL and the LSL are determined with the help of prior successful jobs i.e. the process data generated, streamed and collected from an object (i.e. the part or component) which has been 3D printed/AM machine produced and certified by conventional certification technique. For example, if the error signal 114 does not exceed the USL and/or LSL (“No” in FIG. 1), the object printed by the 3D printed/AM machine, the object will be regarded as a good part and be sent for further testing to certify or approve the object at a step 120. When the streamed data from the 3D printing/AM process is out of bounds of the USL and LSL, the webapp makes a record of this and alerts the user with the help of a notification. If possible, the webapp also takes an image with the help of an onboard camera or thermal sensor which is connected to the webapp. The image can also be collected with the help of an external recording device if the user has set up such a capability in their facility. For example, if the error signal 114 exceeds the USL and/or LSL (“Yes” in FIG. 1), the object printed by the 3D printed/AM machine will be regarded as a failed part 118.

The webapp collects all the instances of Out of bound readings which are determined by the USL and LSL from the sensors during the 3D printing/AM process and documents it for the user.

The webapp then notifies the user of all such instances of out of bound sensor behaviour and displays it in a dashboard which can be viewed from a computer or mobile device.

The data collected from a part produced by 3D printing/AM process and has passed certification by conventional certification technique for 3D printed/AM process part will be called the Ground Truth (GR).

The data from the 3D printing/AM process for the part will now be analysed and compared against the GR.

If the part whose data is collected and displayed in the webapp falls within the acceptable range of the GR. The part is digitally certified for the user by a 3^rd party certification body. This example involving the GR illustrates one method of how the certification or approval at the step 120 in FIG. 1 can be performed.

If the part does not meet the minimum standards set by the GR, it is tested using traditional certification techniques to make sure the part meets the minimum criteria for use in the real world. The further testing described with reference to step 120 of FIG. 1 can be such traditional certification techniques.

This comparison based on the GR can be used to accelerate the part qualification of 3D printer/AM manufactured parts. FIG. 3 illustrates a workflow of a system or an apparatus of an example of the present disclosure. The steps of FIG. 3 provide more details on the steps 130, 132 and 134 of FIG. 1. The system or apparatus operating the workflow of FIG. 3 comprises an example of the monitoring unit 104 of FIG. 1 described earlier and the reference numeral “104” is reused. In the present example, the monitoring unit 104 comprises one or more servers running web applications developed with a set of Application Programming Interfaces (APIs). The monitoring unit 104 is configured to communicate with an Additive Manufacturing (AM) machine. The monitoring unit 104 is configured to work with JSON, which is an open standard file format and data interchange format that uses human-readable text to store and transmit data objects consisting of attribute-value pairs and arrays (or other serializable values). The monitoring unit 104 is also configured to operate 3 services, namely an incoming message transformer 304, a confidence interval service 310 and an outgoing message transformer 324.

In a first step 302 of the workflow, a user creates a JSON Request through a user interface running the APIs. It should be appreciated that in other examples, the request can be made in other formats such as an XML request. In the present example, the JSON request comprises past AM information of a part (or object or component) to be printed (e.g. dimensions of the part or component), and information of one or more sensors of the AM machine for which the upper and lower serviceable limits, USL and LSL, described earlier have to be calculated and obtained. Specifically, the JSON request comprises corresponding sensor readings of all prior successful printing jobs of the part or component. Each of the sensor readings comprise a value (e.g. the sensor measurement or reading) that can be either integer, decimal, string, or Boolean, along with the corresponding timestamp of the value. In the present example, the prior successful printing jobs and corresponding sensor readings are collated from prior jobs performed by the same or similar type of the AM machine.

The data in the JSON request may be obtained from a database and may be in a format that is not suitable for further processing. Flence, the JSON request is inputted to the Incoming Message Transformer 304, which transforms the JSON request into an Incoming Service Limits Message at a step 306. The Incoming Service Limits Message comprises an internal domain model, which is a proprietary means of representing and organising data to facilitate the processing of the data. In the present example, the internal domain model comprises one or more tables containing a list of different prior successful printing jobs, the plurality of sensors of each job, the corresponding sensor readings of the plurality of sensors, and the timestamps of the corresponding sensor readings. The internal domain may be sorted by time, grouped by jobs, and in each job group, sub-grouped by sensors and their readings. After the Incoming Service Limits Message is generated by the Incoming Message Transformer 304, at a step 308, the internal domain model is sent to the Confidence Interval Service layer 310 to calculate confidence intervals for each sensor.

At the Confidence Interval Service layer 310, as the different prior successful printing jobs will have run at different times, their sensor readings need to be normalised against the same time scale at a step 312. This is done by treating the start time of each job as the zeroth second and treating the timestamps after the zeroth second as time elapsed since the start time. The format of the internal domain model is organised to facilitate such time normalisation.

After the time normalisation at step 312, tables containing sensor readings of different jobs (i.e. map of job with normalised readings 314) are merged using the elapsed time as an index and the sensor readings or values are filled into a merged table (or dataframe) or tables (or dataframes) using time-interpolation. That is, the time series of readings for each sensor are parameterise at a step 316. In this manner, conceptually, the sensor readings of each job are mapped onto a same time scale of a graph through interpolation, and the graph provides the range of acceptable sensor readings. Merged table (or dataframe) or tables (or dataframes) 318 with time and the job specific sensor readings as columns can be generated to contain the data of the graph. The table or tables are considered as “merged” because they contain time normalised data of a plurality of prior successful jobs. For example, in 2 different prior successful jobs A and B, they both have the same type of sensor. Job A has a start time of ST 1 and the corresponding sensor has a first reading V1 that is 5 minutes after the start time ST1 , and job B has a start time of ST2 and the corresponding sensor has a first reading V2 that is 5.2 minutes after the start time ST2. During the time normalisation, ST1 and ST2 are set to zero, and during the time-interpolation, the merged table for the same sensor would be recorded with the readings V1 and V2 at the 5 and 5.2 minutes respectively. The gaps in the sensor readings between the 5 and 5.2 minutes (i.e. the gap of 0.2 minutes) in the graph can be filled in with sensor readings obtained through interpolation.

At a step 320, for each row of each merged table stating the readings of prior successful prints of a sensor i.e. at each time instance or time step, confidence and LSL and USL are calculated using the below formula: q = 1 - [(100 - confidence/ 100)/2] tValue PP(q, n_j0bs 1)

Cl = std(readina \ , readina2. reading N) c tValue Jⁿjobs LSL = ^{reading† , reading 2, .... , readingN) - Cl USL = [Jiireading l , reading2, .... , readingN) + Cl where - confidence = 95% by default (but can be changed by the user),

PPF = Percent Point Function, n_j0bs = number of jobs, std = standard deviation, m = mean, and reading 1, reading2, ... , readingN are the sensor readings of job1 , job2, ... , jobN at each time step.

After calculation at step 320, the one or more merged tables (dataframe) with calculated LSL and USL 322 is generated. The one or more merged tables (dataframes) with LSL and USL is transformed into another internal domain model, Outgoing Service Limits Message, at a step 326.

Thereafter, the Outgoing Service Limits Message is further transformed into a JSON response using the APIs into a preferred format for user consumption. This JSON response will have the serviceable limits for a specific sensor of a specific part that can be aligned with any future job’s timestamps and form a permissible limits band or range (expected print data range) around live or actual print data.

FIG. 4 and FIG. 5 illustrate the permissible limits band or range (expected print data range) around actual print data of an object being printed. FIG. 4 shows a graph of bed temperature of the bed temperature sensor (in degrees Celsius) versus printing time of the object. FIG. 5 shows a graph of extruder temperature of the extruder temperature sensor (in degrees Celsius) versus printing time of the object.

Specifically, FIG. 4 shows a range of acceptable (expected) sensor measurements of a bed temperature sensor and FIG. 5 shows a range of acceptable (expected) sensor measurements of an extruder temperature sensor. The shaded regions 402 and 502 in FIG. 4 and 5 respectively comprise the range of acceptable sensor measurements and their top and bottom boundaries represent the USL and LSL respectively. The range of acceptable sensor measurements with the USL and LSL are obtained from prior successful prints of the object.

FIG. 4 also shows a graph plot of actual print data (i.e. sensor readings) from a corresponding bed temperature sensor and FIG. 5 also shows a graph plot of actual print data (i.e. sensor readings) from a corresponding extruder temperature sensor of a specific part or component being printed by a printer. The solid lines 404 and 504 in FIG. 4 and 5 respectively are the actual print data of the sensor obtained from the printer when the object is being printed. The actual print data can be streamed live i.e. in real time while the object is being printed for comparison with the range of acceptable sensor measurements or transmitted after the object is printed for comparison with the range of acceptable sensor measurements.

In FIG. 4 and FIG. 5, the graph plots of the actual print data shown exceed the boundaries of the USL and the LSL. Flence, in the examples of FIG. 4 and FIG. 5, it can be determined that the manufacturing process is flawed and the object being printed may be flawed, or determined that the object being printed should be subject to further testing to determine whether it should be certified or approved as a good product. In the examples of FIG. 4 and FIG. 5, the printer is an “ANYCUBIC i3 mega S” printer. Other printers may also be used in other examples.

FIG. 2 shows in more detail an example of the processing unit (or processor) of the monitoring unit 104 of FIG. 1, or a processing unit (or processor) of the computer and the mobile device described earlier in the steps for obtaining a Digital DNA of an AM printed part.

The processing unit of the monitoring unit 104 may comprise a processing unit 1002 for processing software including one or more computer programs for running one or more computer/server applications to enable a backend logic flow or the method or methods for carrying out the steps as described with reference to the earlier Figures. The processing unit of the monitoring unit 104 can be a server in a network (e.g. internet). It may also be a processing unit of an AM machine (e.g. 3D printer).

Furthermore, the processing unit 1002 may include user input modules such as a computer mouse 1036, keyboard/keypad 1004, and/or a plurality of output devices such as a display device 1008. The display device 1008 may be a touch screen capable of receiving user input as well. A virtual keypad and/or keyboard can be provided by the display device 1008.

The processing unit 1002 may be connected to a computer network 1012 via a suitable transceiver device 1014 (i.e. a network interface), to enable access to e.g. the Internet or other network systems such as a wired Local Area Network (LAN) or Wide Area Network (WAN). The processing unit 1002 may also be connected to one or more external wireless communication enabled devices 1034 (For example, another apparatus 100, another user interface 200) via a suitable wireless transceiver device 1032, e.g. a WiFi transceiver, Bluetooth module, Mobile telecommunication transceiver suitable for Global System for Mobile Communication (GSM), 3G, 3.5G, 4G telecommunication systems, and the like. Through the computer network 1012, the processing unit 1002 can gain access to one or more storages i.e. data storages, databases, data servers and the like connectable to the computer network 1012 to retrieve and/or store data in the one or more storages.

The processing unit 1002 may include a processor 1018, a Random Access Memory (RAM) 1020 and a Read Only Memory (ROM) 1022. The processing unit 1002 may also include a number of Input/Output (I/O) interfaces, for example I/O interface 1038 to the computer mouse 1036, a memory card slot 1016, I/O interface 1024 to the display device 1008, and I/O interface 1026 to the keyboard/keypad 1004.

The components of the processing unit 1002 typically communicate via an interconnected bus 1028 and in a manner known to the person skilled in the relevant art.

The computer programs may be supplied to the user of the processing unit 1002, or the processor (not shown) of one of the one or more external wireless communication enabled devices 1034, encoded on a data storage medium such as a CD-ROM, on a flash memory carrier or a Hard Disk Drive, and are to be read using a corresponding data storage medium drive of a data storage device 1030. Such computer or application programs may also be downloaded from the computer network 1012. The application programs are read and controlled in its execution by the processor 1018. Intermediate storage of program data may be accomplished using RAM 1020.

In more detail, one or more of the computer or application programs may be stored on any non-transitory machine- or computer- readable medium. The machine- or computer- readable medium may include storage devices such as magnetic or optical disks, memory chips, or other storage devices suitable for interfacing with a general purpose computer. The machine- or computer- readable medium may also include a hard-wired medium such as that exemplified in the Internet system, or wireless medium such as that exemplified in the Wireless LAN (WLAN) system and the like. The computer program when loaded and executed on such a general-purpose computer effectively results in an apparatus that implements the steps of the computing methods in examples herein described.

In summary, examples of the present disclosure may have the following features.

An additive manufacturing method comprising: receiving print data of an object printable by an additive manufacturing (AM) machine (e.g. step 102); obtaining expected print data (e.g. 112, 134) of the printing of the object; obtaining actual print data (e.g. 128) of the printing of the object provided by the AM machine (Note: This can be done in real-time or non-real-time); comparing the expected print data with the actual print data to obtain a comparison result (e.g. step 116); determining the printed object as a failed product if the comparison result is out of bounds of predetermined upper serviceable limit (USL) and/or lower serviceable limit (LSL) (e.g. step 118); and certifying or approving the printed object to proceed with further testing if the comparison result is within the limits of the predetermined USL and/or LSL (e.g. step 120).

The further testing may comprise: certifying that the printed object meets predetermined standards; and storing the actual print data of the certified printed object as data relating to ground truth, wherein the data relating to ground truth is used to obtain the expected print data.

The further testing may comprise: inspecting images of the printed object for defects using an operation or computer vision algorithm to certify or approve the printed object or determine the printed object as a failed product.

A model of expected print parameters may be derived from the received print data (e.g. step 108), and the expected print data may comprise expected print parameters determined from print parameters of previously completed prints of the object.

The expected print data may comprise expected physical measurements of the printed object determined from expected values of physical sensor measurements.

The predetermined USL and/or LSL may be specific for each of one or more sensors of the AM machine.

The method may comprise: receiving updates of actual print data from each of the one or more sensors of the AM machine to perform the comparison and determine whether the comparison result specific to each sensor is out of bounds of the USL and/or LSL specific to each sensor. The method may comprise: when the comparison result is out of bounds of the predetermined USL and/or LSL, recording the comparison result and sending an alert to notify a user.

The method may further comprise: capturing an image of the printed object having the comparison result that is out of bounds of the USL an/or LSL by an onboard camera or a thermal sensor of the AM machine.

The predetermined USL and/or LSL may be classified as critical or non-critical, and the printed object may be deemed as a failed product only if the comparison result is out of bounds of the USL and/or LSL that is classified as critical.

The USL or LSL may be defined according to the formula:

USL, LSL = max(f(x_\) g(E(x_t))), min(f(x_t) - g(E(x_t))), wherein x is a sensed value at a time t during the printing of the object, (x_t) is an expected value of x_t derived from the inputted print data, f(x) is a function, linear or nonlinear, which maps x_t into a scoring space, and (E(x_t)) is a function, linear or nonlinear, which maps E(x_t) into a scoring space.

The expected print data may be determined from sensor readings of prior successful prints of the object by: normalising start time of the prior successful prints to map the sensor readings of the prior successful prints onto a common time scale (e.g. step 312); and time interpolating the sensor readings of the prior successful prints according to their time elapsed from the normalised start time to obtain the sensor readings of the prior successful prints at each time instance in the common time scale (e.g. step 316).

An additive manufacturing apparatus comprising: a processing unit (e.g. 1002) configured to execute instructions to operate the additive manufacturing apparatus to: receive print data of an object printable by an additive manufacturing (AM) machine; obtain expected print data of the printing of the object; obtain actual print data of the printing of the object provided by the AM machine (Note: This can be done in real-time or non-real-time); compare the expected print data with the actual print data to obtain a comparison result; determine the printed object as a failed product if the comparison result is out of bounds of predetermined upper serviceable limit (USL) and/or lower serviceable limit (LSL); and certify or approve the printed object to proceed with further testing if the comparison result is within the limits of the predetermined USL and/or LSL.

The further testing may be to certify that the printed object meets predetermined standards, and the apparatus may comprise a database to store the actual print data of the certified printed object as data relating to ground truth, wherein the data relating to ground truth is used to obtain the expected print data.

The apparatus may be operable to: inspect images of the printed object for defects using an operation or computer vision algorithm to certify or approve the printed object or determine the printed object as a failed product.

A model of expected print parameters may be derived from the received print data, and the expected print data comprises expected print parameters determined from print parameters of previously completed prints of the object.

The apparatus may be operable to: receive updates of actual print data from each of the one or more sensors of the AM machine to perform the comparison and determine whether the comparison result specific to each sensor is out of bounds of the USL and/or LSL specific to each sensor.

The apparatus may be operable to: when the comparison result is out of bounds of the predetermined USL and/or LSL, record the comparison result and send an alert to notify a user. The apparatus may be operable to: instruct an onboard camera or a thermal sensor of the AM machine to capture an image of a printed object having the comparison result that is out of bounds of the USL an/or LSL.

The USL or LSL may be defined according to the formula:

The expected print data may be determined from sensor readings of prior successful prints of the object, wherein start time of the prior successful prints are normalised and the sensor readings of the prior successful prints are mapped onto a common time scale, and wherein the sensor readings of the prior successful prints at each time instance in the common time scale are obtained by time interpolating the sensor readings of the prior successful prints according to their time elapsed from the normalised start time.

In the specification and claims, unless the context clearly indicates otherwise, the term “comprising” has the non-exclusive meaning of the word, in the sense of “including at least” rather than the exclusive meaning in the sense of “consisting only of”. The same applies with corresponding grammatical changes to other forms of the word such as “comprise”, “comprises” and so on.

While the invention has been described in the present disclosure in connection with a number of embodiments and implementations, the invention is not so limited but covers various obvious modifications and equivalent arrangements, which fall within the purview of the appended claims. Although features of the invention are expressed in certain combinations among the claims, it is contemplated that these features can be arranged in any combination and order.

Claims

1. An additive manufacturing method comprising: receiving print data of an object printable by an additive manufacturing (AM) machine; obtaining expected print data of the printing of the object; obtaining actual print data of the printing of the object provided by the AM machine; comparing the expected print data with the actual print data to obtain a comparison result; determining the printed object as a failed product if the comparison result is out of bounds of predetermined upper serviceable limit (USL) and/or lower serviceable limit (LSL); and certifying or approving the printed object to proceed with further testing if the comparison result is within the limits of the predetermined USL and/or LSL.

2. The additive manufacturing method as claimed in claim 1, wherein the further testing comprises: certifying that the printed object meets predetermined standards; and storing the actual print data of the certified printed object as data relating to ground truth, wherein the data relating to ground truth is used to obtain the expected print data.

3. The additive manufacturing method as claimed in claim 1 or 2, wherein the further testing comprises: inspecting images of the printed object for defects using an operation or computer vision algorithm to certify or approve the printed object or determine the printed object as a failed product.

4. The additive manufacturing method as claimed in any one of the preceding claims, wherein a model of expected print parameters is derived from the received print data, and the expected print data comprises expected print parameters determined from print parameters of previously completed prints of the object.

5. The additive manufacturing method as claimed in any one of the preceding claims, wherein the expected print data comprises expected physical measurements of the printed object determined from expected values of physical sensor measurements.

6. The additive manufacturing method as claimed in any one of the preceding claims, wherein the predetermined USL and/or LSL is specific for each of one or more sensors of the AM machine.

7. The additive manufacturing method as claimed in claim 6, wherein the method comprises: receiving updates of actual print data from each of the one or more sensors of the AM machine to perform the comparison and determine whether the comparison result specific to each sensor is out of bounds of the USL and/or LSL specific to each sensor.

8. The additive manufacturing method as claimed in any one of the preceding claims, wherein the method comprises: when the comparison result is out of bounds of the predetermined USL and/or LSL, recording the comparison result and sending an alert to notify a user.

9. The additive manufacturing method as claimed in any one of the preceding claims, wherein the method further comprises: capturing an image of the printed object having the comparison result that is out of bounds of the USL an/or LSL by an onboard camera or a thermal sensor of the AM machine.

10. The additive manufacturing method as claimed in any one of the preceding claims, wherein the predetermined USL and/or LSL are classified as critical or non-critical, and the printed object is deemed as a failed product only if the comparison result is out of bounds of the USL and/or LSL that is classified as critical.

11. The additive manufacturing method as claimed in any one of the preceding claims, wherein the USL or LSL is defined according to the formula:

USL, LSL = max(f(x_\) - g(E(x_t))), min(f(x_t) - g(E(x_t))), wherein x is a sensed value at a time t during the printing of the object, (x_t) is an expected value of x_t derived from the inputted print data, f(x) is a function, linear or nonlinear, which maps x_t into a scoring space, and (E(x_t)) is a function, linear or nonlinear, which maps E(x_t) into a scoring space.

12. The additive manufacturing method as claimed in any one of the preceding claims, wherein the expected print data is determined from sensor readings of prior successful prints of the object by: normalising start time of the prior successful prints to map the sensor readings of the prior successful prints onto a common time scale; and time interpolating the sensor readings of the prior successful prints according to their time elapsed from the normalised start time to obtain the sensor readings of the prior successful prints at each time instance in the common time scale.

13. An additive manufacturing apparatus comprising: a processing unit configured to execute instructions to operate the additive manufacturing apparatus to: receive print data of an object printable by an additive manufacturing (AM) machine; obtain expected print data of the printing of the object; obtain actual print data of the printing of the object provided by the AM machine; compare the expected print data with the actual print data to obtain a comparison result; determine the printed object as a failed product if the comparison result is out of bounds of predetermined upper serviceable limit (USL) and/or lower serviceable limit (LSL); and certify or approve the printed object to proceed with further testing if the comparison result is within the limits of the predetermined USL and/or LSL.

14. The additive manufacturing apparatus as claimed in claim 13, wherein the further testing is to certify that the printed object meets predetermined standards, and the apparatus comprises a database to store the actual print data of the certified printed object as data relating to ground truth, wherein the data relating to ground truth is used to obtain the expected print data.

15. The additive manufacturing apparatus as claimed in claim 13 or 14, wherein the apparatus is operable to: inspect images of the printed object for defects using an operation or computer vision algorithm to certify or approve the printed object or determine the printed object as a failed product.

16. The additive manufacturing apparatus as claimed in any one of claims 13 to 15, wherein a model of expected print parameters is derived from the received print data, and the expected print data comprises expected print parameters determined from print parameters of previously completed prints of the object.

17. The additive manufacturing apparatus as claimed in any one of the claims 13 to 16, wherein the expected print data comprises expected physical measurements of the printed object determined from expected values of physical sensor measurements.

18. The additive manufacturing apparatus as claimed in any one of the claims 13 to 17, wherein the predetermined USL and/or LSL is specific for each of one or more sensors of the AM machine.

19. The additive manufacturing apparatus as claimed in claim 18, wherein the apparatus is operable to: receive updates of actual print data from each of the one or more sensors of the AM machine to perform the comparison and determine whether the comparison result specific to each sensor is out of bounds of the USL and/or LSL specific to each sensor.

20. The additive manufacturing apparatus as claimed in any one of the claims 13 to 19, wherein the apparatus is operable to: when the comparison result is out of bounds of the predetermined USL and/or LSL, record the comparison result and send an alert to notify a user.

21. The additive manufacturing apparatus as claimed in any one of the claims 13 to 20, wherein the apparatus is operable to: instruct an onboard camera or a thermal sensor of the AM machine to capture an image of a printed object having the comparison result that is out of bounds of the USL an/or LSL.

22. The additive manufacturing apparatus as claimed in any one of the claims 13 to 21, wherein the predetermined USL and/or LSL are classified as critical or non-critical, and the printed object is deemed as a failed product only if the comparison result is out of bounds of the USL and/or LSL that is classified as critical.

23. The additive manufacturing apparatus as claimed in any one of the claims 13 to 22, wherein the USL or LSL is defined according to the formula:

24. The additive manufacturing apparatus as claims in any one of claims 13 to 23, wherein the expected print data is determined from sensor readings of prior successful prints of the object, wherein start time of the prior successful prints are normalised and the sensor readings of the prior successful prints are mapped onto a common time scale, and wherein the sensor readings of the prior successful prints at each time instance in the common time scale are obtained by time interpolating the sensor readings of the prior successful prints according to their time elapsed from the normalised start time.