Defective cell detection system and method

Abstract

Methods and systems for identifying defective sample clips in real time during manufacturing are provided herein. Such a system uses monitored operational parameters associated with the operation of a manufacturing apparatus that manufactures a sample clip. One or more operational parameter datasets are obtained from existing device control and / or other sensors and compared to corresponding operational parameters associated with acceptable clips. The defect detection unit may be configured to compare the operational parameters and may select algorithms and / or models, such as machine learning models, from which defective clips may be identified. The comparing step may include determining whether the monitored operating parameter is within a predefined acceptable value range and / or deviates from a characteristic linetype of the operating parameter, thereby indicating that the card is defective. The detection of defective clips can be integrated into automated control, so that defective clips can be detected in real time and automatically discarded during manufacturing.

Description

Technical Field

[0001] This application claims the benefit of priority of U.S. Provisional Application No. 63 / 374,312, filed on September 1, 2022, which is incorporated herein by reference.

[0002] This application is generally related to a concurrently filed U.S. non - provisional application entitled "Seal Failure Detection Systems and Methods" [Attorney Docket No. 085430 - 1406031 - 018210US] and a U.S. provisional application entitled "Transfer Learning Methods and Models for Facilitating Defect Detection" [Attorney Docket No. 085430 - 1406033 - 018310US], the entire contents of which are incorporated herein by reference for all purposes.

[0003] The present invention generally relates to the field of manufacturing, and more particularly to identifying defective units, such as sample cartridges for analyzing fluid samples. Background Art

[0004] In recent years, significant developments have been made in the field of biodiagnostic devices, which help to manipulate liquid samples within sample cartridges to prepare samples for biodiagnostic testing by polymerase chain reaction (PCR). A notable development in this field is the Cepheid GeneXpert sample cartridge. The configuration and operation of these types of cartridges can be further understood by reference to U.S. Patent No. 6,374,684 entitled "Fluid Control and Handling System" and U.S. Patent No. 8,048,386 entitled "Fluid Handling and Control". While these sample cartridges represent a significant advancement at the beginning of the field, as with any precision instrument, the manufacture of sample cartridges presents certain challenges, especially when the assembly of multiple parts occasionally results in defects that cause the sample cartridge to leak or fail to maintain the internal pressure required for successful operation.

[0005] Traditional systems for manufacturing sample cartridges utilize a series of manufacturing processes and steps, some of which may introduce defects to the sample cartridge, such as defects that cause the sample cartridge to fail to maintain internal pressure, commonly referred to as seal failures. Certain steps, such as welding a lid device to the cartridge body and sealing reagents in the cartridge through a membrane, occasionally introduce seal defects that are difficult to detect. Existing detection methods include various seal test methods and visual inspections. However, these methods typically use destructive methods and / or occasionally fail to identify all defects. Existing test procedures require randomly selecting a majority of units (e.g., 200 cartridges) produced from the assembly / production line and conducting specialized seal tests (e.g., Seal Test Failure; STF tests). When more than a preset number of units (e.g., 10 cartridges) fail the test, this results in an entire batch consisting of a large number of units (e.g., 1500 to 1600) being discarded as defective. When these defective units are not detected during this quality inspection process, this allows the defective units to reach the customers.

[0006] Accordingly, there is a need to improve methods for detecting defective cartridges to avoid unnecessarily wasting entire batches and to prevent defective units from reaching the customers. In addition, there is a need for detection methods for defective cartridges that are non-destructive, do not require extensive testing after manufacturing, and are not prone to human error. Summary of the Invention

[0007] In one aspect, the present invention relates to a method for detecting defective sample cartridges. Such methods may include the steps of: obtaining one or more data sets of one or more monitored operating parameters during the manufacturing process of the sample cartridge; comparing the one or more data sets with a baseline or standard data set of the manufacturing process associated with acceptable sample cartridges; and identifying defective sample cartridges based on variations of the one or more data sets from the baseline or standard data set. In some embodiments, identifying defective sample cartridges is based on the monitored operating parameters exceeding the acceptable operating value range of the parameters associated with approved sample cartridges. The range of acceptable values may vary over time depending on the manufacturing process. In some embodiments, identifying defective sample cartridges is based on the monitored operating parameters deviating from the standard characteristic line type of the operating parameters associated with approved sample cartridges.

[0008] In some embodiments, the manufacturing process includes welding a clip part by a welder that forcibly engages with the clip part (such as a lid and a clip body) and applies ultrasonic energy to form a weld seam that seals the parts together. The operating parameters can include any of the following parameters: power, stroke distance, force, amplitude, frequency, or any combination thereof. In some embodiments, the operating parameter includes the power supplied to the welder during welding. In some embodiments, the operating parameter includes the stroke distance of the welding head of the welder during welding. In some embodiments, the operating parameter includes the force applied by the welding head during welding. In some embodiments, the operating parameter includes the amplitude of the ultrasonic wave applied during welding. In some embodiments, the operating parameter includes the frequency of the ultrasonic wave applied by the ultrasonic welding head during welding. In some embodiments, the manufacturing process includes heat-sealing a clip lid by a heat-sealing mechanism that presses a film against the lid and applies heat, thereby heat-sealing an opening in the lid to seal a reagent in the clip. In some embodiments, identifying a defective clip includes automatically identifying a defective clip based on one or more data sets obtained from an automatic control unit that controls the operation of the manufacturing equipment, and the method may further include automatically discarding any defective clips.

[0009] In another aspect, the present invention relates to a system configured to detect defective sample clips. Such a system includes one or more sensors communicatively coupled to a manufacturing device, wherein the sensors monitor operating parameters associated with the operation of the manufacturing device that performs a manufacturing process on a sample clip; and a processing unit communicatively coupled to the one or more sensors. Instructions for performing automatic detection of defective sample clips are recorded on the processing unit, and the instructions include the following steps: in the manufacturing process of the sample clip, obtaining one or more data sets of one or more monitored operating parameters from the one or more sensors; comparing the one or more data sets with a baseline data set of the manufacturing process associated with an acceptable sample clip; and identifying a defective sample clip based on the difference between the one or more data sets and the baseline data set. In some embodiments, the one or more sensors are integrated within a control unit that operates the manufacturing device. In some embodiments, the system is integrated within an automation software that controls the manufacturing process and the associated production line, and the system is configured to automatically discard any defective clips during the manufacturing process. In some embodiments, the processing unit is configured to identify a defective sample clip based on the monitored operating parameters exceeding an acceptable operating value range of the parameters associated with an approved sample clip. The range of acceptable values can vary over time. In some embodiments, the processing unit is configured to identify a defective sample clip based on the monitored operating parameters deviating from a characteristic line pattern of the operating parameters associated with an approved sample clip.

[0010] In some embodiments, the manufacturing process includes welding a clamp part by a welder that forcibly engages the clamp part, such as a lid and a clamp body, and applies ultrasonic energy to form a weld seam that seals the parts together. In some embodiments, the operating parameters include any of the following parameters: power, stroke distance, force, amplitude, frequency, or any combination thereof. In some embodiments, the operating parameter includes the power supplied to the welder during welding. In some embodiments, the operating parameter is the stroke distance of the welding head of the welder during the welding process. In some embodiments, the operating parameter includes the force applied by the welding head during the welding process. In some embodiments, the operating parameter includes the amplitude of the ultrasonic waves applied during the welding process. In some embodiments, the operating parameter includes the frequency of the ultrasonic waves applied by the ultrasonic welding head during the welding process. In some embodiments, the manufacturing process includes heat-sealing a clamp lid by a heat-sealing mechanism that presses a film against the lid and applies heat to heat-seal the film to the lid. In some embodiments, the processing unit is configured to determine defect detection in real time during the manufacturing process of a sample clamp. In some embodiments, the processing unit is configured to command the system to discard a defective clamp from the production line once it is identified. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1A is a flowchart of defective clamp detection according to some embodiments that utilizes inputs of monitored operating parameters of the manufacturing process compared to the corresponding parameters of acceptable clamps to identify defective clamps.

[0012] Figure 1B is a flowchart according to some embodiments showing another method of defective clamp detection that utilizes inputs of operating parameters of the manufacturing process in a machine learning model that analyzes the parameters relative to the corresponding parameters in acceptable clamps in order to identify defective clamps.

[0013] Figure 2A illustrates an exemplary sample clamp that has a welded lid device and a film seal provided to the user, with the lid on the top cover opened to receive a liquid sample. Figure 2B According to some embodiments, an exploded view of the sample clamp is shown, illustrating its main components, including a lid device, a multi-chamber, a reaction vessel, a valve assembly, and a base. Figures 2C to 2D Shows a detailed view of the lid device. Figure 2E Shows the lid device placed on top of the sample clamp body before ultrasonic welding by a welding head. Figure 2F Shows a schematic view of a part of the production line process and related operating parameters for automatically detecting defective clamps according to the methods described herein.

[0014] Figures 3A to 3BShows a manufacturing process flow chart and examples that illustrate various sources of clip defects in an exemplary manufacturing method of sample clips that may result in defective clips as specified herein.

[0015] Figure 4 Illustrates a data acquisition flow chart in a manufacturing process according to some embodiments, including operating parameters from a welding machine operation, which can be used to automatically detect defective clips.

[0016] Figures 5A to 9B Illustrates monitoring operating parameters in a manufacturing process of welding a lid device to a clip body according to some embodiments, with each figure showing a comparison of the operating parameters of three defective clips on the left side with the standard or baseline operating parameters of acceptable clips on the right side.

[0017] Figures 10 to 11 Shows a method for detecting defective clips according to some embodiments. Detailed Description

[0018] The present invention generally relates to the detection of manufacturing defects, particularly the identification of defective sample clips in manufacturing. In some embodiments, the method and system provide for the real-time automatic identification of defective clips during manufacturing. A flow chart of such an automatic defective clip identification method is as Figures 1A to 1B shown and will be discussed in further detail below. I. System Overview

[0019] In one aspect, the present invention relates to an automatic detection system for identifying defective sample clips. An exemplary sample clip for testing a target analyte is as Figure 2A shown. The sample clip includes a lid device 100 sealed on top of a clip body 200, and the clip body 200 contains reagents and a liquid sample. The lid device 100 includes a bottom lid portion sealed on the clip body and an open top lid portion, as shown, to allow a user to deposit a liquid sample in the clip. When the sample clip is provided to the user, the reagents are already set in selected chambers and are sealed within the clip by a film 110 sealed on top of the bottom lid. The film includes a central opening for a syringe and an opening for inserting a liquid sample.

[0020] Figure 2B Depicts an exemplary clip suitable for performing multi-target combinatorial detection as described herein. The illustrated clip is based on Cartridge (Cepheid, Inc., Sunnyvale, Calif.). The cartridge 100 includes a cartridge body 200 having a plurality of chambers 208 defined therein for receiving various reagents and / or buffers. The chambers are disposed around a central injection barrel 209 that is in fluid communication with a valve body 210 via a valve injection tube 211 extending into the injection barrel 209. The valve body 210 is interfaced within the cartridge body and is supported on a cartridge base 210. The cartridge typically includes one or more channels or cavities that may contain filter material (e.g., glass filter columns) that can bind and elute nucleic acids. In various embodiments, the cartridge further includes one or more temperature-controlled channels or chambers that can serve as thermal cycling chambers in some embodiments. A “plunger” (not shown) can be operated to draw fluid into the injection barrel 209, and rotation of the valve body / injection tube provides selective fluid communication between the various reagent chambers and channels and the reaction chamber. Thus, the various reagent chambers, reaction chambers, matrix materials, and channels are selectively in fluid communication via valves and rotation of the plunger, and reagent movement (e.g., chamber loading or unloading) is effected by the “injecting” action of the plunger. An attached reaction vessel 216 (“PCR tube”) provides an optical window for real-time detection of, e.g., amplification products, base identification, during sequencing operations by operating modules within the systems described herein. It should be understood that such reaction vessels can include a variety of different chambers, conduits, or micro-well arrays for detecting target analytes. The sample cartridge can be equipped with means for preparing a biological fluid sample prior to transport to the reaction vessel. Any chemical reagents required for viral or cell lysis, or devices for binding or detecting target analytes (e.g., reagent beads) can be housed in one or more chambers of the sample cartridge and thus be available for sample preparation.

[0021] The exemplary use of such a sample cartridge with reaction vessels for analyzing a biological fluid sample is described in U.S. Patent Application No. 6,818,185, entitled "Cartridge for Conducting Chemical Reactions," filed May 30, 2000, commonly assigned, the entire content of which is incorporated herein by reference for all purposes. Exemplary sample cartridges and associated instrument modules are shown and described in U.S. Patent No. 6,374,684, entitled "Fluid Control and Handling System," filed Aug. 25, 2000, and U.S. Patent No. 8,048,386, entitled "Fluid Handling and Control," filed Feb. 25, 2002, the entire content of which is incorporated herein by reference for all purposes. Various aspects of the sample cartridge can be further understood by reference to U.S. Patent No. 6,374,684, which more particularly describes certain aspects of the sample cartridge. Such a sample cartridge can include a fluid control mechanism connected to the chambers of the sample cartridge, such as a rotary fluid control valve. Rotation of the rotary fluid control valve allows fluid communication between the chamber and the valve body, thereby controlling the flow of a biological fluid sample stored in the cartridge into different chambers, where various reagents can be provided as needed according to a specific protocol to prepare the biological fluid sample for analysis. To operate the rotary valve, the cartridge handling module includes a motor, such as a stepper motor, which is typically coupled to a drive system that is coupled to a feature of the valve body to control the movement of the valve body in coordination with the movement of a syringe, thereby moving the fluid sample according to the desired sample preparation protocol. The fluid metering and dispensing function of the rotary valve according to a specific sample preparation protocol is demonstrated in U.S. Patent No. 6,374,684.

[0022] Figures 2C to 2DShows a detailed view of an exemplary cap device 100, which includes a central opening for the passage of a syringe / plunger, which affects the movement of liquid between chambers, and the central opening is surrounded by a plurality of shafts 102 (with channels), which project into an opening 104 in the top cap. Thus, the cap device 100 includes a substantially uniform bottom surface 106, and thus the internal welding pattern shown is not coextensive with any wall extending from the bottom surface 106. The chambers of the liquid container device disclosed herein may contain one or more reagents for a variety of purposes. These reagents may exist in a variety of forms. Non-limiting exemplary reagent forms may include solutions, dry powders, or lyophilized beads. The reagents may be used for different purposes, including but not limited to chemical and / or enzymatic reactions, sample preparation, and / or detection. Non-limiting example purposes may include lysing cells or microorganisms, purifying or isolating target analytes (e.g., specific cell populations, nucleic acids, or proteins), digesting or modifying nucleic acids or proteins, amplifying nucleic acids, and / or detecting target analytes. Other details of the cap device can be found in U.S. Patent No. 10,273,062, the entire content of which is incorporated herein by reference for all purposes.

[0023] Figure 2C Shows a top view of the underside of the bottom cap and the bottom surface of the top cap portion. The underside of the bottom cap includes a plurality of shafts 102 that project upward from the top surface of the bottom cap portion and are received in corresponding holes 104 in the top cap portion. The plurality of shafts 102 and the opening 104 surround a central opening 103, through which an injection instrument of a module projects during the operation of the sample cartridge to facilitate the flow of fluid between chambers by the movement of the valve body. Figure 2D Shows a bottom view of the underside of the bottom cap of the cap device 100, which includes the lower main surface and the top side of the top cap portion. The bottom surface of the bottom cap portion is welded to the top edge of the cartridge body. To facilitate welding, a raised welding ridge 101 is circumferentially provided along the periphery of the bottom cap between the edge alignment feature 107 and the outermost wall. When in place in an appropriate manner, the edge alignment feature 107 and the outermost wall can prevent the bottom cap from rotating excessively relative to the fluid container 200, thereby aligning the raised welding ridge 101 of the bottom cap with the weldable feature of the cartridge body (e.g., the top edge of the wall). A plurality of walls 108 extend from the central portion of the lower main surface. The walls are arranged in a petal shape around the central opening 103. Here, the walls form six petals. There is a raised welding pattern at the top edge of the walls. The raised welding pattern is connected to the welding ridge 101. In this way, a fluid zone is created outside the petals. When the liquid container and the bottom cap are welded via the raised welding pattern and the welding ridge, the sub-containers within the bottom container are fluidly isolated from each other (at least at the interface between the liquid container and the bottom cap).

[0024] Figure 2E Shows the relationship between the lid device 100 and the clip body 200. The clip body 200 includes a plurality of chambers that can be fluidly coupled or uncoupled depending on the position of an internal valve assembly. The chambers are defined by walls that extend to the top of the clip body 200. The molten interface between the lid device 100 and the clip body 200 is created by sealing the chambers to each other through a raised weld pattern 160, a weld ridge 156, and a weld interface between the chambers of the container 200. The lid device 100 is welded to the fluid container by an ultrasonic welding head 1901 that engages a lid interface while the lid sits on the container 200. The welding head 1901 typically includes a metal cylinder that is shaped to abut and surround a raised platform. The welding head is part of a larger welding device (not shown) that supplies energy to the welding head. In an exemplary welding process, the lid device is placed on top of the clip body such that the weld ridge is aligned with the top edge of the clip body. Next, the ultrasonic welding head is pressed against the lid and sufficient force is applied to increase the welding contact force, and ultrasonic energy is applied to facilitate ultrasonic welding. The operating parameters associated with this welding process include the power (W) supplied to the welder, the distance (mm) the ultrasonic welding head moves during the welding process, the force (N) applied by the ultrasonic welding head, the amplitude (%) of the ultrasonic energy applied, and the frequency (Hz) of the ultrasonic energy applied by the ultrasonic welding head. It will be understood that in some embodiments, the detection method may utilize only one or a subset of these monitoring parameters. In some embodiments, these parameters are already monitored by the equipment controller in its standard operation, so these parameters can be used to detect defective clips without any additional sensors. If any of these parameters have not been monitored, one or more additional sensors may be included to monitor these or other operating parameters. Commercially available ultrasonic welding equipment can be used in this process, which is available from manufacturers such as Hermann Ultrasonics, Bartlett, Ill. 60103 or Branson Ultrasonics (a division of Emerson Industrial Automation; located in Eden Prairie, Minn. 55344). The above welding operations are typically performed at a welding station along the manufacturing / production line of the sample clips.

[0025] Figure 2FShows a schematic diagram of a part of a cartridge manufacturing / assembly line 2100, which includes a welding station 2010, a reagent filling station 2020, and a film sealing station 2030. At the welding station, an automatic alignment fixture places the lid device 100 on top of the cartridge body 200, presses down the automatic ultrasonic welding head, and applies ultrasonic energy for at least 3 to 5 seconds to weld the lid to the cartridge body by ultrasonic welding. As described above, the shape and design of the welding ridges on the underside of the bottom lid are for sealing and welding to the top edge of the cartridge body chamber. After welding, the cartridge moves automatically in sequence to the reagent filling station, where one or more reagents (such as beads, liquids, powders, etc.) or other process materials (such as buffers, desiccants) are stored in selected chambers through the shaft openings in the lid. After the reagent or other process materials are placed in the chambers, a film seal is applied to the top surface of the bottom lid to seal the shaft passage openings, thereby sealing the reagent and process materials inside the cartridge. The automated equipment places a film layer on the bottom lid (the lid is an open structure) and heats it to seal the film to the bottom lid part. The film includes a central hole (such as a transverse cut) to allow the syringe of the instrument module to pass through the film, and another opening in the lower right corner allows the user to inject a liquid sample into the sample chamber. However, the remaining openings on the lid, including the shaft openings leading to the chambers containing the reagent and process materials, are sealed by the film.

[0026] As further described below, the automatic defective cartridge detection system obtains one or more operating parameters 2011 from the welder and then inputs them into the defective cartridge detection unit 2040, which compares the parameters with the corresponding operating parameters from acceptable cartridges (for example, several cartridges that have passed the applicable performance tests). Although these concepts are described for the welding manufacturing process, it can be understood that these same concepts can be applied to various other manufacturing processes, including but not limited to the thermal film sealing of reagents in the cartridge body at the film sealing station. Similarly, a second data set of operating parameters 2031 from the film station can be input into the defect detection unit to identify defective cartridges due to film sealing defects. The identification results from the two inputs can be fed back to the production line to discard the defective cartridges before labeling and completing the cartridges for shipment to consumers. II. Example of Detection Method A. Overview

[0027] In one aspect, the systems and methods described herein utilize one or more monitored operating parameters associated with performing one or more manufacturing processes along a manufacturing production line to identify defective cartridges in real-time during manufacturing. The monitored parameters can be obtained from existing control units that control the operation of manufacturing equipment using sensors integrated into the equipment, or through additional sensors. In some embodiments, the corresponding parameters from acceptable cartridges can be used to determine a baseline range or profile for a given parameter. In other embodiments, more complex relationships between multiple parameters can be examined by using specially developed algorithms. Given the complexity of the data and the relatively small amount of data associated with defects, it may be advantageous to utilize a machine learning (ML) model to determine the relationship between one or more monitored parameters and defective cartridges. By using an ML model, subtle variations in the monitored parameters that result in defective cartridges can be examined. Advantageously, such automated defect detection can identify defective cartridges in real-time during manufacturing, and thus defective cartridges can be removed during manufacturing. In some embodiments, the automated defect detection obviates the need for post-manufacture destructive testing of selected cartridges in each batch (especially destructive testing), thereby avoiding waste and reducing costs while significantly improving defect detection. B. Example of Defective Cartridge Detection

[0028] In one aspect, an object of the present invention is to develop a supervised model based on power (P), travel (T), force (F), amplitude (A), and frequency (Y) data that are reported from automated welder data (such as Hermann Welder data) during the welding process of an automated production line (such as a Reagents On-Board Assembly Line; ROBAL). In traditional systems, the automated welder data reports these five parameters at a resolution of 1 millisecond. This data shows how much power is supplied to the welder, how much the cartridge lid moves, and how much force is applied by the ultrasonic welder (such as a sonotrode or welding head). When the reported distance (i.e., travel) is insufficient, this may indicate improper welding, and discrete movements in force or amplitude may also indicate improper welding, which may result in a Seal Test Failure (STF) associated with a defective cartridge. As previously mentioned, existing procedures require discarding an entire batch when too many units fail the STF test, and more importantly, allow defective units to be delivered to our customers. Therefore, by developing a supervised model, defective cartridges can be automatically detected, thereby detecting defective units in real-time to improve product quality and avoid waste associated with discarding an entire batch.

[0029] In some embodiments, the automatic defect detection unit compares the monitored operating parameters with the corresponding operating parameters of acceptable cartridges. In some embodiments, the defect detection unit may use an algorithm that is determined by analyzing the monitored parameters from a sufficient number of acceptable cartridges until the values of the parameters converge to an identifiable range or characteristic profile. In some embodiments, an ML model can be used to associate one or more parameters of the manufacturing process with specific defects that result in cartridge defects. In some embodiments, the defects are associated with weld seals (e.g., over-welding, incomplete welding, shaft rupture) and / or membrane seals (e.g., incomplete seal, shaft melting). The parameter or characteristic value can include any attribute associated with the manufacturing process. Advantageously, this method allows for the detection of defects in the manufacturing process in real time so that defective cartridges can be removed during the process.

[0030] As Figure 1A shown, schematic diagram 1000 depicts an automated method that obtains one or more inputs of any number of parameters associated with a manufacturing process. In this embodiment, the parameters can be any parameters associated with the operation of equipment in the manufacturing process, such as power, stroke, force, amplitude, frequency, temperature, current, voltage. These parameters are fed as input values into a comparator that compares these parameters with the corresponding parameters of acceptable cartridges. The comparator can utilize a model or algorithm developed specifically for defective cartridge detection, which is based on the correlation between the parameters and cartridge or process defects. Preferably, the parameter comparison is performed in real time during manufacturing so that defective cartridges determined to be such can be removed from the production line. The automatic defective cartridge identification method described herein can supplement or replace standard testing and inspection by personnel.

[0031] As Figure 1B shown, schematic diagram 1100 depicts an automated method of feeding parameter inputs into an ML model, which may include supervised learning and / or unsupervised learning. This method can identify complex relationships between the parameters associated with the defects that result in defective cartridges. Given the complexity of the data and the examination of various parameters, this method can understand the relationships between the various parameters and identify defective cartridges that may not be understood otherwise or pointed out by standard testing and inspection procedures.

[0032] Figure 3AShows the cartridge assembly workflow 300, highlighting various sources of cartridge defects that lead to cartridge defects. An exemplary manufacturing process flow is shown on the left. The process flow includes lid welding and film sealing steps, which are typically associated with defects that cause seal test failures, and these defects are associated with defective cartridges. For example, 80% of seal test failures (STF) can be traced back to the lid welding step (e.g., insufficient welding or over-welding), and approximately 20% of seal test failures can be traced back to the film sealing step. Currently, defects caused by these steps can only be detected after sample cartridges are labeled and unloaded as finished products, whether through visual inspection by staff or in STF or functional tests, which may result in partial or complete batch rejection. Therefore, the automatic defective cartridge detection system and method described herein can detect defective cartridges in real time before product manufacturing is completed by analyzing the operating parameters associated with these manufacturing process steps, thus avoiding such waste. Therefore, the method described herein can improve defect detection with minimal or no adjustment to the existing production line. Figure 3B Shows examples of cartridge defects (e.g., liquid leakage, crystal leakage), cartridges, and manufacturing equipment.

[0033] Figure 4 Shows a flowchart 400 that details a method of integrating existing controller parameters for automatically detecting defective cartridges. Figure 4 Shows how programmable logic controller (PLC) data is matched with welder data. In conventional systems, there is usually no unique identifier for matching automation data (e.g., Rockwell database) and welder data (e.g., Hermann welder data). During the welding process, the welder creates and assigns a process ID for each weld. By combining the part ID and batch status data, a unique process ID can be created in the PLC and written to the welder, thereby enabling the matching of automation data and welder data. It is worth mentioning that this is just one method of implementing the automatic defective cartridge detection method in the existing manufacturing process data stream, and it is understood that various other methods can be used.

[0034] Figures 5 to 9 illustrate the operating parameters monitored in the welding manufacturing process of welding the lid device to the cartridge body. The figures shown on the left ( Figure 5A 、 Figure 6A 、 Figure 7A 、 Figure 8A and Figure 9A ) are data observed from three units that failed the STF test, and the figures shown on the right ( Figure 5B 、 Figure 6B 、 Figure 7B 、 Figure 8B andFigure 9B ) are the units that have passed the STF test. Each figure shows the upper and lower limits of the parameters (in red) and the average line type (in blue). For example, the two red lines can indicate the 99.5% confidence interval of the associated parameter. Figure 5 shows the variation of the power (W) supplied to the welder during the welding process. Figure 6 shows the variation of the welder travel distance (mm) during the welding process. Figure 7 shows the variation of the force (N) applied by the welder during the welding process. Figure 8 shows the variation of the amplitude (%) of the ultrasonic wave applied by the welder during the welding process. Figure 9 shows the frequency (Hz) of the ultrasonic wave applied by the welder during the welding process. Inconsistencies in the welding operation (such as misalignment, excessive lid) or product materials (such as lid breakage / melting, clip damage) can cause one or more of these parameters to deviate from the baseline or standard line type and / or range, such that if the monitored parameter exceeds the standard value range and / or line type of the parameter associated with the acceptable clip, a defective clip can be identified. It should be understood that this is not always the case, and developing a specialized algorithm can further improve the analysis of the variation of the monitored parameters associated with defective clips. In some embodiments, a model can be developed to identify defective clips from the monitored parameters associated with the manufacture of acceptable clips and the parameters associated with defective clips.

[0035] Based on the available parameter data, a proposed model can be developed in the following order: (1) Continue to collect data n until the variations of the power (P), travel (T), force (F), amplitude (A), and frequency (Y) data converge. (2) Develop the empirical distribution functions of the normal P, T, F, A, and Y, NP, NT, NF, NA, and NY respectively. (3) Compare the normal arcs with the real-time P, T, F, A, and Y, and calculate the difference i for each clip serial number: DPi, DTi, DFi, DAi, DYi (4) Find the distributions of DPi, DTi, DFi, DAi, DYi and calculate their p-values: PPi, PTi, PFi, PAi, PYi; (5) Create a vector of size 5, where these five p-values are associated with each clip (6) Run a supervised model using the labeled data based on these features.

[0036] Based on the available parameter data, another proposed model can be developed in the following order: (1) Generate a baseline vector from the welding data associated with the welder used to attach the lid to the body; (2) Collect the attachment data associated with the welder during the attachment of the lid to the body; (3) Convert the attachment data into a container vector; (4) Determine the manufacturing status by comparing the container vector with the baseline vector; (5) Transmit the manufacturing status to an output.

[0038] One or more methods may be performed via a non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform the method. The method can be used to detect manufacturing defects in containers, cartridges, or storage vessels in real time. Additionally, the output conveys status, results, information, instructions, or a combination thereof to a user or device. The output can be auditory, visual, tactile, or a combination thereof. The output can be, for example, a screen, a speaker, an interface, etc.

[0039] Existing automatic feature generation models typically do not consider the shape of time series models but generate thousands of features that may not have physical meaning and may lead to over-fitting of the model. The proposed model takes into account the physical meaning of the data and creates custom features before developing a supervised model. Based on the custom features, drift in welding parameters can also be monitored and considered when identifying defective cartridges. It should be noted that not all abnormal patterns associated with power (P), travel (T), force (F), amplitude (A), and frequency (Y) parameters will result in a defective cartridge (e.g., a failed seal test), and other relationships between the parameters can be examined. It should be understood that the above method is just an example of developing an applicable parameter model, and various other methods can be implemented. In some embodiments, various parameters can be input into an ML model to develop algorithms to illustrate the relationship between one or more parameters and defective cartridges, as well as the relationship between multiple different parameters and defective cartridges to further improve the detection of defective cartridges.

[0037] Figures 10 to 11 An exemplary method for detecting defective cartridges according to some embodiments is shown. Figure 10 A method is shown by which the concepts herein can be applied to detect defective products, which can include sample cartridges and various other manufactured products. The method includes the steps of: obtaining one or more operating parameters associated with the operation of a manufacturing device performing a product manufacturing process; comparing the one or more operating parameters with corresponding baseline or standard parameters associated with an acceptable product; and identifying a defective product based on the difference between the monitored operating parameters and the baseline or standard parameters. Figure 11 A method for detecting a defective sample cartridge as described herein is shown. The method includes the steps of: obtaining one or more operating parameters (e.g., power, travel, force, amplitude, frequency) associated with a welding operation between a lid and a body of the sample cartridge; comparing the one or more operating parameters with corresponding baseline or standard parameters associated with an acceptable sample cartridge (e.g., by using an algorithm or a model); and identifying a defective sample cartridge based on the comparison of the monitored operating parameters with the baseline or standard parameters.

[0038] In the foregoing specification, the invention has been described with reference to specific embodiments thereof, but those skilled in the art will recognize that the invention is not so limited. The various features, embodiments, and aspects of the above invention may be used singly or in combination. In addition, the invention can be used in any number of environments and applications beyond those described herein without departing from the broader spirit and scope of the specification. Accordingly, the specification and drawings are to be regarded as illustrative rather than restrictive. It should be recognized that the terms "comprising," "including," and "having" as used herein are specifically intended to be open-ended terms of the art. Any reference to a publication, patent, or patent application is hereby incorporated by reference in its entirety for all purposes.

Claims

1. A method for detecting a defective sample holder, characterized in that, The method includes: Obtaining one or more data sets of one or more monitored operating parameters in the manufacturing process of a sample cartridge, wherein the operating parameters are associated with the operation of a manufacturing device that performs the manufacturing process; Comparing the one or more data sets with a baseline data set of the corresponding parameters of the manufacturing process associated with an acceptable sample cartridge; and Identifying a defective sample cartridge based on the difference between the one or more data sets of the monitored operating parameters and the baseline data set.

2. The method according to claim 1, characterized in that, Identifying a defective sample cartridge is based on the monitored operating parameters exceeding the acceptable operating value range of the parameters associated with an approved sample cartridge.

3. The method according to claim 2, wherein The range of the acceptable values varies over time in the manufacturing process.

4. The method according to claim 1, characterized in that, Identifying a defective sample cartridge is based on the monitored operating parameters deviating from the characteristic line type of the operating parameters associated with an approved sample cartridge.

5. The method according to claim 1, characterized in that, The manufacturing process is welding of a cartridge part by a welder, which forcibly couples against the cartridge part and applies ultrasonic energy to form a weld seam that seals the parts together.

6. The method according to claim 5, wherein The manufacturing process is welding a lid device to a cartridge body to form a weld seam that seals the lid device to the cartridge body.

7. The method according to claim 5, characterized in that, The operating parameters include any one of the following: power, stroke, distance, force, amplitude, frequency, or any combination thereof.

8. The method according to claim 5, wherein The operating parameter includes the power supplied to the welder during the welding process.

9. The method according to claim 5, characterized in that, The operating parameter includes the travel distance of the welder during the welding process.

10. The method according to claim 5, characterized in that The operating parameter includes the force applied by the welder during the welding process.

11. The method according to claim 1, wherein The operating parameter includes the amplitude of the ultrasonic wave applied during the welding process.

12. The method according to claim 1, characterized in that, The operating parameter includes the frequency of the ultrasonic wave applied during the welding process.

13. The method according to claim 1, wherein The manufacturing process is heat-sealing a lid of a cartridge by a heat-sealing mechanism, which presses a film against the lid and applies heat to heat-seal the film on the lid.

14. The method according to claim 13, characterized in that, Identifying a defective cartridge includes automatically identifying a defective cartridge based on the one or more data sets obtained from an automatic control unit that controls the operation of the manufacturing device, and the method further includes automatically discarding any identified defective cartridges.

15. The method according to claim 1, characterized in that The comparison is performed using an algorithm developed from a supervised model of the corresponding operating parameters associated with multiple acceptable sample cartridges and the corresponding operating parameters associated with multiple defective sample cartridges.

16. A system for detecting defective sample holders, characterized in that, The system includes: One or more sensors that communicate with a manufacturing device that performs a manufacturing process on a sample cartridge, wherein the one or more sensors monitor operating parameters associated with the operation of the manufacturing device that performs the manufacturing process; and A processing unit that is communicatively coupled to the one or more sensors, wherein the processing unit records instructions thereon for performing automatic detection of defective sample cartridges, and the instructions include the following steps: Obtaining one or more data sets of one or more monitored operating parameters from the one or more sensors in the manufacturing process; Comparing the one or more data sets with a baseline data set of the corresponding operating parameters of the manufacturing process associated with an acceptable sample cartridge; and Identify defective sample cartridges based on variations between one or more data sets of the monitored operating parameters and the baseline data set.

17. The system according to claim 16, wherein The one or more sensors are integrated within a control unit operating the manufacturing equipment.

18. The system according to claim 16, wherein The system is integrated into an automation software controlling the manufacturing equipment and associated production lines.

19. The system according to claim 16, wherein, The processing unit is configured to identify defective sample cartridges based on the monitored operating parameters exceeding an acceptable operating value range of parameters associated with acceptable sample cartridges.

20. The system according to claim 19, wherein, The range of acceptable values varies over time in the manufacturing process.

21. The system according to claim 16, wherein The processing unit is configured to identify defective sample cartridges based on the monitored operating parameters deviating from a characteristic line profile of the operating parameters associated with acceptable sample cartridges.

22. The system according to claim 16, wherein The manufacturing process is welding of cartridge parts by a welder that forcibly couples a lid of the cartridge against a cartridge body of the sample cartridge and applies ultrasonic energy to form a weld seam sealing the sample cartridge together.

23. The system according to claim 22, wherein The operating parameters include any one of the following: power, stroke distance, force, amplitude, frequency, or any combination thereof.

24. The system according to claim 22, characterized in that, The operating parameter includes the power supplied to the welder during welding.

25. The system according to claim 22, wherein The operating parameter includes the stroke distance of the welder during the welding process.

26. The system according to claim 22, wherein The operating parameter includes the force applied by the welder during the welding process.

27. The system according to claim 22, wherein, The operating parameter includes the amplitude of the ultrasonic waves applied during the welding process.

28. The system according to claim 22, wherein The operating parameter includes the frequency of the ultrasonic waves applied by the welder during the welding process.

29. The system according to claim 16, wherein The manufacturing process is heat-sealing of a lid of the cartridge by a heat-sealing mechanism that presses a film against the lid and applies heat to heat-seal the film to the lid.

30. The system according to claim 16, wherein The processing unit is configured to determine in real time the identification of defective cartridges during the manufacture of sample cartridges.

31. The system according to claim 16, wherein The processing unit is configured to command the system to discard the defective cartridge from the production line upon identification of a defective cartridge.

32. The system according to claim 16, wherein The processing unit includes a supervised model of the respective operating parameters associated with a plurality of acceptable sample cartridges and the respective operating parameters associated with a plurality of defective cartridges, through which the monitored operating parameters can be compared with the baseline data set.

33. A method for real-time detecting manufacturing defects of a container having a lid and a body, characterized in that, The method includes: Generating a baseline vector from welding data associated with a welder for attaching the lid to the body; Collecting attachment data associated with the welder during attachment of the lid to the body; Converting the attachment data into a container vector; Determining a manufacturing status by comparing the container vector with the baseline vector; and Transmitting the manufacturing status to an output.

34. The method according to claim 33, wherein The welding data includes at least two data functions, the functions being power (P), stroke (T), force (F), amplitude (A), and frequency (Y).

35. The method according to claim 34, wherein The method further includes collecting the welding data until each of at least two of the data functions converges with an associated data function arc.

36. The method according to claim 35, characterized in that, The method further includes developing an empirical data distribution function based on the welding data.

37. The method according to claim 33, wherein The determining step includes a sub-step of quantifying the difference between the baseline vector and the container vector.

38. The method according to claim 37, wherein The determining step includes a sub-step of calculating a p-value based on the difference between the baseline vector and the container vector.

39. The method according to claim 38, wherein The method further includes generating a final vector based on the p-value.

40. The method according to claim 38, wherein, The method further includes starting a supervised model based on the p-value, the p-value being associated with labeled data.

41. A non-transitory computer-readable medium, characterized in that, Instructions are stored on the non-transitory computer-readable medium that, when executed by a processor, cause the processor to perform a method for real-time detection of manufacturing defects of a container, the method including: generating a baseline vector from a distribution of welding data associated with a welding machine for attaching a lid to a cartridge body; collecting attachment data associated with the welding machine during attaching the lid to the cartridge body, the attachment data including at least two parameters; converting the at least two parameters into a container vector; determining a manufacturing status by comparing the container vector with the baseline vector; and transmitting the manufacturing status to an output.

Images