Systems, methods, and media for manufacturing processes

Abstract

A manufacturing system is disclosed herein. The manufacturing system includes one or more stations, a monitoring platform, and a control module. Each of the one or more stations is configured to perform at least one step in a multi-step manufacturing process for a component. The monitoring platform is configured to monitor the progress of the component throughout the multi-step manufacturing process. The control module is configured to dynamically adjust processing parameters for each step of the multi-step manufacturing process to achieve a desired final quality metric for the component.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to U.S. Provisional Application Serial No. 62 / 979,639, filed on February 21, 2020, which is incorporated herein by reference in its entirety. Technical Field

[0003] The present disclosure generally relates to systems, methods, and media for use in manufacturing processes. Background Art

[0004] Continuous monitoring and adjustment of the manufacturing process is often required to manufacture samples that consistently meet the intended design specifications safely, on time, and with minimal waste. Summary of the Invention

[0005] In some embodiments, a manufacturing system is disclosed herein. The manufacturing system includes one or more stations, a monitoring platform, and a control module. Each of the one or more stations is configured to perform at least one step in a multi-step manufacturing process for a component. The monitoring platform is configured to monitor the progress of the component throughout the multi-step manufacturing process. The control module is configured to dynamically adjust processing parameters for each step of the multi-step manufacturing process to achieve a desired final quality metric for the component. The control module is configured to perform operations. These operations include receiving input related to the component at a step in the multi-step manufacturing process from the monitoring platform. The operations also include generating, by the control module, a final quality metric prediction based on an image of a sample. The operations also include determining, by the control module, that the final quality metric prediction value is not within a range of acceptable values. The operations also include adjusting, by the control module, control logic for at least a subsequent station based on the determination. Adjusting the control logic includes applying corrective actions to be performed by the subsequent station.

[0006] In some embodiments, a multi-step manufacturing method is disclosed herein. A computing system receives an image of a sample at one or more stations of a manufacturing system from a monitoring platform. Each station is configured to perform a step in the multi-step manufacturing process. The computing system generates a final quality metric prediction based on the image of the sample. The computing system determines that the final quality metric prediction is not within a range of acceptable values. Based on the determination, the computing system adjusts control logic of at least a subsequent station, wherein the adjustment includes applying a corrective action to be performed by the subsequent station.

[0007] In some embodiments, a three-dimensional printing system is disclosed herein. The system includes a processing station, a monitoring platform, and a control module. The processing station is configured to deposit multiple layers to form a sample. The monitoring platform is configured to monitor the progress of the sample throughout the deposition process. The control module is configured to dynamically adjust the processing parameters of each of the multiple layers to achieve a desired final quality metric for the component. The control module is configured to perform an operation. The operation includes receiving an image of the sample after the layer is deposited from the monitoring platform. The operation also includes generating, by the control module, a final quality metric prediction value based on the image of the sample. The operation also includes determining, by the control module, that the final quality metric prediction value is not within a range of acceptable values. The operation also includes adjusting, by the control module, the control logic for at least a subsequent layer to be deposited based on the determination, wherein the adjustment includes applying a corrective action performed by the deposition of the subsequent layer. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to enable the above-mentioned features of the present disclosure to be understood in detail, the present disclosure may be described in more detail with reference to embodiments, some of which are illustrated in the accompanying drawings. It should be noted, however, that the accompanying drawings illustrate only typical embodiments of the present disclosure and are therefore not to be considered limiting of its scope, as the present disclosure may admit to other equally effective embodiments.

[0009] Figure 1 is a block diagram illustrating a manufacturing environment according to an example embodiment.

[0010] Figure 2 is a block diagram illustrating a prediction engine according to an example embodiment.

[0011] Figure 3 is a block diagram illustrating an exemplary architecture of a clustering module according to an example embodiment.

[0012] Figure 4 is a block diagram illustrating the architecture of a machine learning module according to an example embodiment.

[0013] Figure 5 is a block diagram illustrating the architecture of a machine learning module according to an example embodiment.

[0014] Figure 6 is a flow chart illustrating a corrective method of performing a multi-step manufacturing process according to an example embodiment.

[0015] Figure 7A A system bus computing system architecture is shown according to an example embodiment.

[0016] Figure 7B A computer system having a chipset architecture according to an example embodiment is shown.

[0017] To facilitate understanding, identical reference numerals have been used, where possible, to denote identical elements common to the figures. It is contemplated that elements disclosed in one embodiment may be used in other embodiments without specific description. DETAILED DESCRIPTION

[0018] One or more of the techniques described herein are generally directed to a monitoring platform configured to monitor each step of a multi-step manufacturing process. For each step of the multi-step manufacturing process, the monitoring platform can monitor the progress of a sample and determine how the current state of the sample affects a final quality metric associated with the final sample. Typically, a final quality metric is a metric that cannot be measured at each step of the multi-step manufacturing process. Exemplary final quality metrics may include, but are not limited to, tensile strength, hardness, thermal properties, etc. of the final sample. For certain final quality metrics, such as tensile strength, destructive testing is used to measure such metrics.

[0019] One or more techniques described herein can use one or more artificial intelligence techniques to predict final quality metrics at each step of a multi-step manufacturing process. For example, one or more techniques described herein can leverage unsupervised K-means clustering and a deep learning network to learn cluster features. High-fidelity labels can be created for unreliable feedforward setpoints. The method can then be generalized using a regression deep learning network to relabel all images. Using the relabeled images, two networks can be trained to predict sample quality metrics at specific points in the manufacturing process.

[0020] The manufacturing process can be complex, involving raw materials being processed by various processing stations (or "stations") until a final sample is produced. In some embodiments, each processing station receives input for processing and may output an intermediate output, which may be passed to a subsequent (downstream) processing station for further processing. In some embodiments, a final processing station may receive input for processing and may output a final sample, or more generally, a final output.

[0021] In some embodiments, each station may include one or more tools / devices that can perform a set of processing steps. Exemplary processing stations may include, but are not limited to, conveyor belts, injection molding machines, cutting machines, molding machines, extruders, computer numerical control (CNC) milling machines, grinders, assembly stations, 3D printers, quality control stations, verification stations, and the like.

[0022] In some embodiments, the operation of each processing station can be controlled by one or more process controllers. In some embodiments, each processing station can include one or more process controllers that can be programmed to control the operation of the processing station. In some embodiments, an operator or a control algorithm can provide a station controller set point to the station controller, which set point can represent a desired value or range of values ​​for each control value. In some embodiments, the value used for feedback or feedforward in the manufacturing process can be referred to as a control value. Example control values ​​can include, but are not limited to: speed, temperature, pressure, vacuum, rotation, current, voltage, power, viscosity, materials / resources used by the station, production rate, downtime, toxic fumes, etc.

[0023] In some embodiments, a sample may refer to an output of a manufacturing process. For example, the output of a manufacturing process may be a circuit board that is part of a mobile device, a screen that is part of a mobile device, and / or a complete mobile device.

[0024] Figure 1 1 is a block diagram illustrating a manufacturing environment 100 according to an example embodiment. Manufacturing environment 100 may include a manufacturing system 102, a monitoring platform 104, and a control module 106. Manufacturing system 102 may broadly represent a multi-step manufacturing system. In some embodiments, manufacturing system 102 may represent a manufacturing system for additive manufacturing (e.g., a three-dimensional (3D) printing system). In some embodiments, manufacturing system 102 may represent a manufacturing system for subtractive manufacturing (e.g., CNC machining). In some embodiments, manufacturing system 102 may represent a manufacturing system for a combination of additive and subtractive manufacturing. More generally, in some embodiments, manufacturing system 102 may represent a manufacturing system for a general manufacturing process.

[0025] Manufacturing system 102 may include one or more stations 1081-108 n (generally referred to as "station 108"). Each station 108 can represent a step and / or station in a multi-step manufacturing process. For example, each station 108 can represent a layer deposition operation in a 3D printing process (e.g., station 1081 can correspond to layer 1, station 1082 can correspond to layer 2, etc.). In another example, each station 108 can correspond to a specific processing station. In some embodiments, the manufacturing process of the sample may include multiple steps. In some embodiments, the multiple steps may include an ordered sequence of steps. In some embodiments, the multiple steps may include an unordered (e.g., random or pseudo-random) sequence of steps.

[0026] Each station 108 may include a process controller 114 and control logic 116. Each process controller 1141-114 nThe control logic 116 may be programmed to control the operation of each corresponding station 108. In some embodiments, the control module 106 may provide a station controller setpoint for each process controller 114, which may represent a desired value or range of values ​​for each control value. The control logic 116 may reference attributes / parameters associated with the process steps of the station 108. In operation, the control logic 116 of each station 108 may be dynamically updated by the control module 106 throughout the manufacturing process, depending on the current trajectory of the final quality metric.

[0027] The monitoring platform 104 can be configured to monitor each station 108 of the manufacturing system 102. In some embodiments, the monitoring platform 104 can be a component of the manufacturing system 102. For example, the monitoring platform 104 can be a component of a 3D printing system. In some embodiments, the monitoring platform 104 can be independent of the manufacturing system 102. For example, the monitoring platform 104 can be installed on an existing manufacturing system 102. In some embodiments, the monitoring platform 104 can represent an imaging device configured to capture an image of a sample at each step of a multi-step process. For example, the monitoring platform 104 can be configured to capture an image of the sample at each station 108. Generally, the monitoring platform 104 can be configured to capture information related to sample production (e.g., images, voltage readings, speed readings, etc.) and provide the information as input to the control module 106 for evaluation.

[0028] The control module 106 can communicate with the manufacturing system 102 and the monitoring platform 104 through one or more communication channels. In some embodiments, the one or more communication channels can represent a single connection via the Internet, such as a cellular or wireless fidelity (Wi-Fi) network. In some embodiments, the one or more communication channels can connect the terminal, server, and mobile device using a direct connection, such as radio frequency identification (RFID), near-field communication (NFC), Bluetooth, or the like. TM , low-energy Bluetooth (BLE) TM , Wi-Fi TM , ZigBee TM , ambient backscatter communication (ABC) protocol, Universal Serial Bus (USB), Wide Area Network (WAN) or Local Area Network (LAN).

[0029] The control module 106 may be configured to control each process controller of the manufacturing system 102. For example, the control module 106 may be configured to adjust process controls associated with a particular station 108 or process step based on information captured by the monitoring platform 104. In some embodiments, the control module 106 may be configured to adjust process controls at a particular station 108 or process step based on a predicted final quality metric.

[0030] The control module 106 may include a prediction engine 112. The prediction engine 112 may represent one or more trained machine learning modules to predict a sample's final quality metric based on data measured at each individual step of a multi-step manufacturing process. In operation, the control module 106 may receive input from the monitoring platform 104. In some embodiments, such input may take the form of an image of the sample's current state after a step in the multi-step manufacturing process. Based on the input, the control module 106 may predict a sample's final quality metric. Based on the sample's predicted final quality metric, the control module 106 may determine one or more actions to be taken in subsequent manufacturing steps. For example, if the predicted final quality metric falls outside a range of acceptable values, the control module 106 may take one or more actions to correct the manufacturing process. In some embodiments, the control module 106 may interface with station controllers in subsequent stations 108 to adjust their respective control and / or station parameters. These adjustments may help correct the manufacturing process so that the final quality metric falls within an acceptable quality indicator range.

[0031] Figure 2is a block diagram illustrating a prediction engine 112 according to an exemplary embodiment. As shown, the prediction engine 112 may include at least a clustering module 202 and a machine learning module 204. Each of the clustering module 202 and the machine learning module 204 may include one or more software modules. One or more software modules may be a collection of codes or instructions stored on a medium (e.g., a memory of a computing system associated with the control module 106), the collection representing a series of machine instructions (e.g., program code) that implement one or more algorithm steps. Such machine instructions may be actual computer code that a processor interprets to implement the instructions or, alternatively, may be a higher-level encoding of instructions that is interpreted to obtain the actual computer code. One or more software modules may also include one or more hardware components. One or more aspects of the example algorithm may be performed by the hardware component (e.g., circuit) itself, rather than as a result of the instructions. In addition, in some embodiments, each of the clustering module 202 and the machine learning module 204 may be configured to transmit one or more signals between the components. In these embodiments, these signals may not be limited to machine instructions executed by a computing device.

[0032] In some embodiments, clustering module 202 and machine learning module 204 can communicate via one or more local networks 205. Network 205 can be of any suitable type, including a single connection via the Internet, such as a cellular or Wi-Fi network. In some embodiments, network 205 can connect terminals, services, and mobile devices using direct connections, such as radio frequency identification (RFID), near field communication (NFC), Bluetooth, or other similar communication methods. TM , Bluetooth Low Energy TM (BLE), Wi-Fi TM 、ZigBee TM , Ambient Backscatter Communication (ABC) protocol, USB, WAN, or LAN. Because the information being transmitted may be personal or confidential, one or more of these types of connections may be encrypted or otherwise protected for security reasons. However, in some embodiments, the information being transmitted may not be very personal, and therefore, a network connection may be selected for convenience rather than security.

[0033] The clustering module 202 may be configured to assign labels to images of samples during a manufacturing process. For example, multiple dimensions of a manufacturing process may vary. In particular, for additive manufacturing, several dimensions of a printed product may vary. The station 108 may receive one or more parameters for instructions regarding a manufacturing step in a manufacturing process. For example, in additive manufacturing, the manufacturing system 102 may utilize code (e.g., G-code) provided by the control module 106, which may include one or more parameters for the jth instruction on the ith layer, x i，j ,y i，j , z i，j , e i，j , and f i，j , where x i，j ,y i，j ,y i，j , z i，j is the position set point, e i，j can represent the length of filament extruded for a specific printing move, and f i，j It can indicate the speed at which the print head moves. In some embodiments, the code can also include meta-instructions for temperature control or axis homing.

[0034] In some embodiments, such as those utilizing additive manufacturing, to correlate a final quality metric (e.g., tensile strength) to a layer image, a deviation coefficient γe,i,j for each layer may be established such that:

[0035]

[0036] in, It can represent the apparent extrusion caused by printing movement. Each extrusion value e i，j Can have result modifiers γ e，i，j For N print moves in a layer, this resulting modifier can be abstractly averaged across the layer, defined as:

[0037]

[0038] Therefore, the average deviation coefficient γ can be measured e，i To evaluate the extrusion deviation of the entire layer. For a perfect performance of the manufacturing system 102 or station 108, γ e，i = 1.0; however, naturally occurring variations may ensure that this is not the case. In some embodiments, in order to correlate natural variations with the final quality metric (e.g., tensile strength), these set points may be artificially perturbed, assuming their labels are unreliable.

[0039] In some embodiments, the squeezing coefficient γ e，iCan be a parameter that directly affects the quality outcome of parts manufactured in additive manufacturing. In some embodiments, the oxygen / carbon dioxide content of the exhaust gas stream from a chemical manufacturing process can be used instead of the extrusion coefficient. In some embodiments, ellipsometry measurements of thin film deposition can be used. The extrusion coefficient vector Γ can be established e =[γ e，0 ,...,γ e，N ] T In some embodiments, the velocity coefficient vector, which is the deviation from the normal baseline, can be used in a similar manner, Γ f =[γ f，0 ,...,γ f，N ] T These vectors can be used as parameters of a function of the selected quality metric. In some embodiments where the selected quality metric is tensile strength, t s :

[0040] t s =f(Γ e , Γ f ) where the exact formula of f(·) may be unknown and an approximation will be learned by the clustering module 202.

[0041] In some embodiments, the result of directly measuring the selected set point, such as γ e，i , which may not be feasible. Therefore, clustering module 202 can be configured to estimate the bias coefficient γ of station 108 e，i Such estimates can be designed as:

[0042]

[0043] in, Can represent γ e，i The estimated value of I i can represent the top-down image of the i-th layer, and θ h can represent the learning parameters of the mapping function h(·).

[0044] Clustering module 202 can utilize an end-to-end clustering method that can be trained to simultaneously learn the parameters of the neural network that generates the feature vectors and the cluster assignments of the resulting feature vectors. In some embodiments, a K-means clustering technique can be used to partition the data points into K groups of clusters, where each data point can be assigned to the cluster with the closest mean, thereby enabling unsupervised, automatic labeling. In some embodiments, convolutional and pooling layers can be used sequentially in the network to extract features from the image, followed by one or more fully connected network layers that can use backpropagation. In some embodiments, the K-means cluster assignments on the feature vectors can be used as labels for computing gradients to update the neural network.

[0045] Figure 3 2 is a block diagram illustrating an exemplary architecture of the clustering module 202 according to an exemplary embodiment. As shown, the clustering module 202 may include a convolutional neural network 302, a classification module 304, and a regression module 306.

[0046] Convolutional neural network 302 can receive one or more input images from image collection 308 as input. In some embodiments, the one or more input images used for training can represent actual images captured by monitoring platform 104 and / or synthetically generated images. Convolutional neural network 302 can be trained to extract a feature vector from each of the one or more input images. For example, convolutional neural network 302 can be trained to minimize the variance within the resulting K clusters, given a feature vector generated by convolutional neural network 302.

[0047] The classification module 304 can be configured to receive as input one or more feature vectors extracted by the convolutional neural network 302. The classification module 304 can be configured to apply K-means clustering to the one or more feature vectors to classify the features. In some embodiments, the gradient of the backpropagation can be calculated using the fitting of the clusters. In some embodiments, the clusters can be calculated using K-means clustering of the classification module 304 and the forward propagation of the convolutional neural network 302. In some embodiments, the accuracy of the convolutional neural network 302 in predicting the cluster labels of a specified cluster for all images can be calculated. The group of clustered images with the highest accuracy in the Z iteration can be selected for further use.

[0048] The regression module 306 can be configured to generate a deviation score for an image. For example, the regression module 306 can use the input of the feature vector generated by the convolutional neural network 302 to generate a deviation score for the image. In some embodiments, the regression module 306 can be trained by collecting labels through a voting process implemented by K-means clustering. In some embodiments, the clustering dataset can use images from a process artificially perturbed with a known deviation coefficient, but the unreliability of the process increases the uncertainty of these images as pure labels. For each cluster, the modal value of the known deviation coefficient can be applied as a label to the group. The regression module 306 can then be trained using the input of the feature vector generated by the convolutional neural network 302 and the label of the modal cluster value associated with the feature vector. In this way, the clustering module 202 can input the image into the convolutional neural network 302, use the output feature vector as the input to the regression module 306, and then the regression module 306 can output a deviation score for the image.

[0049] In some embodiments, the regression module 306 may include multiple fully connected layers that may utilize linear activation functions.

[0050] Using the clustering module 202, a new set of data pairs can be generated for each image. The new label assignment using visual features can help alleviate the problems found in the traditional use of unreliable set points, which have been shown to be insufficient for supervised learning. The mapping function can be used with weights θ h The deep neural network is used to approximate, and the trained network h(I i ,θ m ) to estimate the new extrusion label set In some embodiments, high-fidelity labels can be used for further predictive training.

[0051] Back to Figure 2 , the machine learning module 204 can be configured to predict a final quality metric of a sample in the manufacturing process based on an image of the sample at the station 108 and one or more labels associated with the image. In some embodiments, the setpoint label can be used to train the machine learning module 204. In some embodiments, the machine learning module 204 can represent a fully connected neural network. In some embodiments, the machine learning module 204 can represent a gated recurrent unit with an internal attention mechanism. The machine learning module 204 can be configured to use High-fidelity label mapping of s =f(Γ e , Γ f ). This can be rewritten as:

[0052]

[0053] in, can represent the predicted tensile strength, and θ t can represent the learning parameters of the prediction function g(·).

[0054] Once sufficiently trained, the prediction engine 112 can make predictions related to the final quality metric of the sample using the machine learning module 204. When an end user or administrator wants to retrain the prediction engine 112, the clustering module 202 can be used.

[0055] Figure 4 is a block diagram illustrating the architecture of the machine learning module 204 according to an example embodiment. As shown and previously discussed, the machine learning module 204 can represent a fully connected neural network.

[0056] As shown, the machine learning module 204 may include a branch-merge architecture 400 utilizing a fully connected layer with Rectified Linear Units (ReLU) activation. and Γ fBefore passing it to a series of fully connected layers, the architecture 400 can utilize and Γ f are converted to higher dimensions, where a series of fully connected layers compress the measured tensile strength t s The predicted value of training on the label

[0057] As provided, the architecture 400 may include a first branch 402 and a second branch 404. The first branch 402 may receive As input. Input may be provided to a first fully connected layer 406. The output from the fully connected layer 406 may be provided to a rectified linear activation function (ReLU) 408. The output from the ReLU 408 may be passed to a second fully connected layer 410, followed by a second ReLU 412 and a dropout layer 414.

[0058] The second branch 404 may receive Γ f As input. Similarly, input Γ f may be provided to a first fully connected layer 416. The output from the fully connected layer 416 may be provided to a ReLU 418. The output from the ReLU 418 may be passed to a second fully connected layer 420, followed by a second ReLU 422 and a dropout layer 424.

[0059] The output from each branch 402 and 404 may be combined and provided as input to a fully connected layer 430. The output from the fully connected layer 430 may be provided to a dropout layer 432. The output from the ReLU 432 may be passed to a fully connected regressor 434 followed by a hyperbolic tangent (tanh) activation function 436. The output from the machine learning module 204 may be a predicted value.

[0060] Figure 5 is a block diagram illustrating the architecture of the machine learning module 204 according to an example embodiment. As shown and previously discussed, the machine learning module 204 can represent a gated recurrent unit with an internal attention mechanism.

[0061] As shown, the machine learning module 204 may include a gated recurrent unit architecture 500. The gated recurrent unit may be configured to predict a given and Γ f Final quality metric (e.g., tensile strength ). Architecture 500 may be used because it handles sequence data well and is able to preserve information about the fault layer through its predicted values.

[0062] Input 502 may include information on key layers of a sample or key process steps of a sample. and Γ f For example, in a manufacturing process, there may be some process steps that contribute more to the final quality metric than others. The identification of these steps can be done through correlation or theoretical analysis. In the case of additive manufacturing, the layers with the smallest surface area will have the highest stress during stretching. Therefore, layer regions can be defined that can meet the definition under normal deviations from extrusion. These layers or steps can be called "critical layers" or "critical steps". In some embodiments, the input labels used are estimates from the output of the clustering module 202 In some embodiments, the data sequence of each sample can be divided or distributed based on the number of gated cycle unit blocks used. Figure 5 As shown in the embodiment of FIG, input 502 can be divided into four groups, corresponding to four gated recurrent unit blocks 504. Each group of data can be provided to a respective gated recurrent unit block 504.

[0063] In some embodiments, the output from each gated recurrent unit block 504 can be concatenated and passed to a fully connected layer 506 and a regressor output layer 508 to estimate the final quality metric (e.g., tensile strength). ) continuous values.

[0064] Figure 6 is a flow chart illustrating a method 600 for performing a correction of a multi-step manufacturing process according to an example embodiment. The method 600 may begin at step 602.

[0065] At step 602, a canonical instruction set may be provided to manufacturing system 102. The canonical instruction set may represent a set of instructions for a manufacturing process. In some embodiments, the canonical instruction set may be provided to each station 108. In these embodiments, each canonical instruction set may specify processing parameters corresponding to a particular manufacturing step at each station 108.

[0066] At step 604, the monitoring platform 104 may capture images of the sample at the conclusion of the processing steps of the multi-step process. For example, the monitoring platform 104 may be configured to capture images of the sample after processing at each station 108. In some embodiments, the monitoring platform 104 may be configured to capture information related to sample production (e.g., images, voltage readings, speed readings, etc.) and provide the information as input to the control module 106 for evaluation.

[0067] At step 606, the prediction engine 112 may examine one or more images of the sample. For example, the prediction engine 112 may receive input (e.g., one or more images) of the sample from the monitoring platform 104 at the end of a particular processing step. Using the input, the prediction engine 112 may generate a predicted final quality metric for the sample.

[0068] Corrective actions may be applied to subsequent manufacturing steps at step 608. For example, based on the predicted final quality metric, the control module 106 may instruct a given station 108 to adjust one or more processing parameters corresponding to the corrective actions to be applied.

[0069] Figure 7A A system bus computing system architecture 700 according to an example embodiment is illustrated. One or more components of system 700 can electrically communicate with each other using a bus 705. System 700 can include a processor (e.g., one or more central processing units (CPUs), graphics processing units (GPUs), or other types of processors) 710 and a system bus 705 that couples various system components, including system memory 715 (e.g., read-only memory (ROM) 720 and random access memory (RAM) 725), to processor 710. System 700 can include a cache of high-speed memory directly connected to, near, or integrated as part of processor 710. System 700 can copy data from memory 715 and / or storage devices 730 to cache 712 for faster access by processor 710. In this way, cache 712 can provide a performance boost by avoiding delays incurred by processor 710 while waiting for data. These and other modules can control or be configured to control the processor 710 to perform various actions. Other system memory 715 may also be used. The memory 715 may include a variety of different types of memory with different performance characteristics. The processor 710 may represent a single processor or multiple processors. The processor 710 may include a general-purpose processor or hardware modules or software modules configured to control the processor 710 (such as service 1 732, service 2 734, and service 3 736 stored in the storage device 730), as well as one or more special-purpose processors that incorporate software instructions into the actual processor design. The processor 710 can essentially be a completely self-contained computing system, containing multiple cores or processors, buses, memory controllers, caches, etc. Multi-core processors can be symmetric or asymmetric.

[0070] To enable user interaction with computing device 700, input device 745 can be any number of input mechanisms, such as a microphone for voice, a touch screen for gesture or graphic input, a keyboard, a mouse, motion input, voice, etc. Output device 735 can also be one or more of the many output mechanisms known to those skilled in the art. In some instances, a multimodal system can enable a user to provide multiple types of input to communicate with computing device 700. Communication interface 740 can generally control and manage user input and system output. There is no limitation on the operation of any particular hardware configuration, so when developed, improved hardware or firmware configurations can easily replace the basic features herein.

[0071] The storage device 730 may be a non-volatile memory and may be a hard disk or other type of computer-readable medium that can store computer-accessible data, such as a magnetic cassette, a flash memory card, a solid-state storage device, a digital versatile disk, a cassette tape, random access memory (RAM) 725, read-only memory (ROM) 720, and combinations thereof.

[0072] The storage device 730 may include services 732, 734, and 736 for controlling the processor 710. Other hardware or software modules are contemplated. The storage device 730 may be connected to the system bus 705. In one aspect, a hardware module that performs a particular function may include software components stored in a computer-readable medium that are associated with the hardware components (e.g., the processor 710, the bus 705, the display 735, etc.) required to perform the function.

[0073] Figure 7BA computer system 750 with a chipset architecture according to an example embodiment is shown. Computer system 750 may be an example of computer hardware, software, and firmware that may be used to implement the disclosed technology. System 750 may include one or more processors 755, representing any number of physically and / or logically distinct resources capable of executing software, firmware, and hardware configured to perform recognition computations. One or more processors 755 may communicate with a chipset 760, which may control the input and output of one or more processors 755. In the illustrated example, chipset 760 outputs information to an output 765, such as a display, and may read and write information to a storage device 770, which may include, for example, magnetic media and solid-state media. Chipset 760 may also read data from and write data to RAM 775. A bridge 780 may be provided for connecting to various user interface components 785 for interfacing with chipset 760. Such user interface components 785 may include a keyboard, a microphone, touch detection and processing circuitry, a pointing device such as a mouse, and the like. In general, input to system 750 may come from any of a variety of sources that may be machine-generated and / or human-generated.

[0074] The chipset 760 may also be connected to one or more communication interfaces 790 having different physical interfaces. Such communication interfaces may include interfaces for wired and wireless local area networks, broadband wireless networks, and personal area networks. Some applications of the methods disclosed herein for generating, displaying, and using a graphical user interface (GUI) may include receiving an ordered data set through a physical interface, or generating the GUI by the machine itself by analyzing data stored in memory 770 or 775 through one or more processors 755. In addition, the machine may receive input from a user through the user interface component 785 and perform appropriate functions, such as browsing functions, by interpreting these inputs using one or more processors 755.

[0075] It will be appreciated that the example systems 700 and 750 may have multiple processors 710 or be part of a group or cluster of computing devices networked together to provide greater processing power.

[0076] Although the foregoing is directed to the embodiments described herein, other and further embodiments may be designed without departing from the basic scope thereof. For example, various aspects of the present disclosure may be implemented in hardware or software or a combination of hardware and software. One embodiment described herein may be implemented as a program product for use with a computer system. The program of the program product defines the functions of the embodiment (including the methods described herein) and may be contained on various computer-readable storage media. Illustrative computer-readable storage media include, but are not limited to: (i) non-writable storage media for permanent storage of information (e.g., a read-only memory (ROM) device such as a CD-ROM disk, flash memory, ROM chip, or any type of solid-state non-volatile memory that can be read by a read-only optical disc read-only memory (CD-ROM) drive within a computer); and (ii) writable storage media for storing variable information (e.g., a floppy disk in a floppy disk drive or hard drive or any type of solid-state random access memory). When carrying computer-readable instructions for the functions of the disclosed embodiments, such a computer-readable storage medium is an embodiment of the present disclosure.

[0077] Those skilled in the art will appreciate that the foregoing examples are illustrative and not restrictive. All permutations, enhancements, equivalents, and modifications thereof will be apparent to those skilled in the art upon reading the specification and studying the drawings, and are encompassed within the true spirit and scope of this disclosure. Therefore, the following claims are intended to encompass all such modifications, combinations, and equivalents within the true spirit and scope of these teachings.

Claims

1. A manufacturing system comprising: one or more stations, each station configured to perform at least one step in a multi-step fabrication process of the sample; a monitoring platform configured to monitor the progress of the sample throughout the multi-step manufacturing process; as well as a control module configured to dynamically adjust processing parameters of each step in the multi-step manufacturing process to achieve a desired final quality metric for the sample, the control module configured to perform operations comprising: receiving, from the monitoring platform, an image of the sample at a step in the multi-step manufacturing process; generating, by the control module, a predicted final quality metric based on the image of the sample; wherein the final quality metric is a metric associated with a property of the finished sample that cannot be measured before each step in the multi-step manufacturing process is completed; determining, by the control module, that the final quality metric prediction value is not within a range of acceptable values; and Based on the determination, control logic of at least a subsequent station is adjusted by the control module, wherein the adjustment includes applying a corrective action performed by the subsequent station.

2. The manufacturing system according to claim 1, wherein: The operations further include: A neural network is trained to generate the final quality metric prediction value based on the plurality of images of the plurality of samples.

3. The manufacturing system according to claim 1, wherein: The operations further include: A gated recurrent unit is trained to generate the final quality metric prediction value based on a plurality of images of a plurality of samples.

4. The manufacturing system according to claim 1, wherein: The operations further include: A clustering module is trained to label a plurality of images of a plurality of samples in order to train a machine learning module configured to generate the final quality metric prediction value based on the plurality of images.

5. The manufacturing system according to claim 1, wherein: The control module adjusts the control logic of at least the subsequent station, including: Yet another control logic is adjusted for yet another subsequent station.

6. The manufacturing system according to claim 1, wherein: Each of the one or more stations corresponds to layer deposition in a three-dimensional printing process.

7. A multi-step manufacturing method comprising: receiving, by the computing system, from a monitoring platform of the manufacturing system, an image of the sample at one or more stations, each station configured to perform a step in a multi-step manufacturing process; generating, by the computing system, a final quality metric prediction based on the image of the sample; wherein the final quality metric is a metric associated with a property of the finished sample that cannot be measured until each step in the multi-step manufacturing process is completed; determining, by the computing system, that the final quality metric prediction value is not within a range of acceptable values; and Based on the determination, control logic of at least a subsequent station is adjusted by the computing system, wherein the adjustment includes a corrective action to be performed by the subsequent station.

8. The method according to claim 7, further comprising: A neural network is trained to generate the final quality metric prediction value based on the plurality of images of the plurality of samples.

9. The method according to claim 7, further comprising: A gated recurrent unit is trained to generate the final quality metric prediction value based on a plurality of images of a plurality of samples.

10. The method according to claim 7, further comprising: A clustering module is trained to label a plurality of images of a plurality of samples in order to train a machine learning module configured to generate the final quality metric prediction value based on the plurality of images.

11. The method according to claim 7, wherein: The computing system adjusts the control logic of at least a subsequent station, including: Yet another control logic is adjusted for yet another subsequent station.

12. The method according to claim 7, wherein: Each of the one or more processing stations corresponds to layer deposition in a 3D printing process.

13. A three-dimensional printing system comprising: a processing station configured to deposit a plurality of layers to form a sample; a monitoring platform for monitoring the progress of the sample throughout the deposition process; as well as A control module configured to dynamically adjust processing parameters of each of the plurality of layers to achieve a desired final quality metric for the sample, the control module configured to perform operations comprising: receiving an image of the sample after the layer is deposited from the monitoring platform; generating, by the control module, a predicted final quality metric based on the image of the sample; wherein the final quality metric is a metric associated with a property of the finished sample that cannot be measured until each step in the multi-step manufacturing process is completed; determining, by the control module, that the final quality metric prediction value is not within a range of acceptable values; and Based on the determination, control logic of at least a subsequent layer to be deposited is adjusted by the control module, wherein the adjustment includes corrective action performed by deposition of the subsequent layer.

14. The system according to claim 13, wherein: The operations further include: A neural network is trained to generate the final quality metric prediction value based on the plurality of images of the plurality of samples.

15. The system according to claim 13, wherein: The operations further include: A gated recurrent unit is trained to generate the final quality metric prediction value based on a plurality of images of a plurality of samples.

16. The system of claim 13, wherein: The operations further include: A clustering module is trained to label a plurality of images of a plurality of samples in order to train a machine learning module configured to generate the final quality metric prediction value based on the plurality of images.

17. The system of claim 13, wherein: The control module adjusts the control logic of at least the subsequent layer, including: Adjusting yet another control logic of yet another subsequent layer.

Images