Methods and processors for controlling an industrial system

Abstract

Methods and systems for controlling an industrial system are disclosed. The industrial system includes a plurality of sub-components. The method includes receiving time-series data representative of operations of at least one of the plurality of sub-components, applying a machine learning algorithm (MLA) configured to process the time-series data to generate: a root cause of an anomaly in operation of the at least one of the plurality of sub-components; and a remedial action parameter to bring operation of the at least one of the plurality of sub- components to specification. The method also includes transmitting instructions for triggering the remedial action.

Description

[0001] METHODS AND PROCESSORS FOR CONTROLLING AN INDUSTRIAL SYSTEM

[0002] FIELD OF THE TECHNOLOGY

[0003] The present technology is related to machine learning solution, and more specifically, to methods and processors for controlling industrial systems.

[0004] BACKGROUND

[0005] Industrial systems refer to systems that integrate machinery, processes, and resources within manufacturing and production domains. At the core of these systems are various components, including manufacturing equipment, which encompasses a spectrum of machinery and tools utilized in production tasks, from assembly lines to specialized tools tailored for specific operations and applications. Industrial systems often implement quality control measures to ensure adherence to standards through inspection, testing, and the mitigation of defects to uphold customer satisfaction. Some industrial systems use automation and robotics to optimize efficiency and reduce labor costs through automated assembly lines, robotic welding, and material handling systems. Some industrial systems implement energy management strategies, such as efficient equipment usage and renewable energy integration to minimize costs and / or environmental impact. Some industrial systems employ information processing systems to facilitate operations through computer systems configured to perform resource planning, real-time monitoring, data analytics, and predictive maintenance.

[0006] SUMMARY

[0007] It is an object of the present invention to ameliorate at least some of the inconveniences present in the prior art.

[0008] Developers of the present technology have realized that the integration of Artificial Intelligence (Al) in industrial systems can be used to ameliorate manufacturing and production processes, offering efficiency, precision, and adaptability. Al applications in industrial systems encompass various aspects, including predictive maintenance, quality control, process optimization, and autonomous operations.

[0009] In some embodiments of the present technology, Al may be leveraged by an industrial system for predictive maintenance, where machine learning algorithms (MLAs) analyze data from sensors and equipment to anticipate potential failures before or as they occur. This proactive approach minimizes downtime, reduces maintenance costs, and enhances operational efficiency. Quality control in the industrial process may also be ameliorated by using Al through advanced image recognition and data analytics techniques. Al-powered systems can detect defects with high accuracy and speed, ensuring that only products meeting quality standards are released. This results in improved product quality, reduced waste, and increased customer satisfaction.

[0010] In a first broad aspect of the present technology, there is provided a method of controlling an industrial system. The industrial system includes a plurality of sub-components. The method is executable by a supervisory computing system associated with the industrial system. The method comprises receiving time-series data representative of operations of at least one of the plurality of sub-components. The method comprises applying a machine learning algorithm (MLA) configured to process the time-series data to generate: a root cause of an anomaly in operation of the at least one of the plurality of sub-components; and a remedial action parameter to bring operation of the at least one of the plurality of sub-components to specification. The method comprises transmitting instructions for triggering the remedial action on the at least one of the plurality of sub-components.

[0011] In some embodiments of the method, the method further comprises applying a further MLA to generate a human-recognizable instruction representative of the remedial action parameter, and the transmitting further comprises transmitting the human-recognizable instruction to an operator of the industrial system.

[0012] In some embodiments of the method, the MLA being implemented as a Deep Neural Network.

[0013] In some embodiments of the method, the MLA is one of a Convolutional Neural Network, Recurrent Neural Network, and a Convolutional Recurrent Neural Network.

[0014] In some embodiments of the method, the remedial action parameter comprises a plurality of remedial scenarios.

[0015] In some embodiments of the method, the further MLA is implemented as a generative MLA.

[0016] In some embodiments of the method, the further MLA is implemented as a Large Language Model (LLM).

[0017] In some embodiments of the method, the industrial system is an injection molding system.

[0018] In some embodiments of the method, the industrial system is one of: an abrasive manufacturing system, an automobile electronics manufacturing system, a metal refining system, a heating and air-conditioning (HVAC) equipment manufacturing system, an engine manufacturing system, a lighting fixtures manufacturing system, a gas manufacturing system, a semiconductor machinery manufacturing system, a printing system, an aircraft manufacturing system, a battery manufacturing system, a food processing system, a paper manufacturing system, a fertilizer manufacturing system, and a glass manufacturing system.

[0019] In some embodiments of the method, the method further comprises, prior to processing by the MLA, applying one or more filtering algorithms onto the received time-series data.

[0020] In some embodiments of the method, the method comprises selecting the MLA amongst a plurality of MLAs, the selecting being based on a type of the anomaly to be detected.

[0021] In a second broad aspect of the present technology, there is provided a computer system for controlling an industrial system. The industrial system includes a plurality of sub-components. The computer system is associated with the industrial system. The computer system is configured to receive time-series data representative of operations of at least one of the plurality of subcomponents. The computer system is configured to apply a machine learning algorithm (MLA) configured to process the time-series data to generate: a root cause of an anomaly in operation of the at least one of the plurality of sub-components; and a remedial action parameter to bring operation of the at least one of the plurality of sub-components to specification. The computer system is configured to transmit instructions for triggering the remedial action on the at least one of the plurality of sub-components.

[0022] In some embodiments of the computer system, the computer system is further configured to apply a further MLA to generate a human-recognizable instruction representative of the remedial action parameter, and to transmitting further comprises the computer system to transmit the human- recognizable instruction to an operator of the industrial system.

[0023] In some embodiments of the computer system, the MLA being implemented as a Deep Neural Network.

[0024] In some embodiments of the computer system, the MLA is one of a Convolutional Neural Network, Recurrent Neural Network, and a Convolutional Recurrent Neural Network.

[0025] In some embodiments of the computer system, the remedial action parameter comprises a plurality of remedial scenarios.

[0026] In some embodiments of the computer system, the further MLA is implemented as a generative MLA. In some embodiments of the computer system, the further MLA is implemented as a Large Language Model (LLM).

[0027] In some embodiments of the computer system, the industrial system is an injection molding system.

[0028] In some embodiments of the computer system, the industrial system is one of: an abrasive manufacturing system, an automobile electronics manufacturing system, a metal refining system, a heating and air-conditioning (HVAC) equipment manufacturing system, an engine manufacturing system, a lighting fixtures manufacturing system, a gas manufacturing system, a semiconductor machinery manufacturing system, a printing system, an aircraft manufacturing system, a battery manufacturing system, a food processing system, a paper manufacturing system, a fertilizer manufacturing system, and a glass manufacturing system.

[0029] In some embodiments of the computer system, the computer system is further configured to, prior to processing by the MLA, apply one or more filtering algorithms onto the received time-series data.

[0030] In some embodiments of the computer system, the computer system is further configured to select the MLA amongst a plurality of MLAs, to select being based on a type of the anomaly to be detected.

[0031] In the context of the present specification, the words “first”, “second”, “third”, etc. have been used as adjectives only for the purpose of allowing for distinction between the nouns that they modify from one another, and not for the purpose of describing any particular relationship between those nouns. Further, as is discussed herein in other contexts, reference to a “first” element and a “second” element does not preclude the two elements from being the same actual real-world element.

[0032] These and other aspects and features of non-limiting embodiments of the present technology will now become apparent to those skilled in the art upon review of the following description of specific non-limiting embodiments of the technology in conjunction with the accompanying drawings.

[0033] Embodiments of the present technology each have at least one of the above-mentioned object and / or aspects, but do not necessarily have all of them. It should be understood that some aspects of the present technology that have resulted from attempting to attain the above-mentioned object may not satisfy this object and / or may satisfy other objects not specifically recited herein. Additional and / or alternative features, aspects and advantages of embodiments of the present technology will become apparent from the following description, the accompanying drawings and the appended claims.

[0034] BRIEF DESCRIPTION OF THE DRAWINGS

[0035] A better understanding of the embodiments of the present technology (including alternatives and / or variations thereof) may be obtained with reference to the detailed description of the nonlimiting embodiments along with the following drawings, in which:

[0036] FIG. 1 depicts a schematic diagram of an injection system including a controller implemented in accordance with non-limiting embodiments of the present technology;

[0037] FIG. 2 is a schematic view of components of a control system including a controller of the molding system of FIG. 1 ;

[0038] FIG. 3 is a schematic representation of an industrial system and a supervisory computing system in accordance with non-limiting embodiments of the present technology;

[0039] FIG. 4 is a scheme-block illustration of a method executable by a processor of a computer system of FIG 5; and

[0040] FIG. 5 is a schematic representation of a computer system in accordance with non-limiting embodiments of the present technology.

[0041] DETAILED DESCRIPTION

[0042] Reference will now be made in detail to various non-limiting embodiment(s) of a molding system and a related method for the operation thereof. It should be understood that other non-limiting embodiment(s), modifications and equivalents will be evident to one of ordinary skill in the art in view of the non-limiting embodiment(s) disclosed herein and that these variants should be considered to be within scope of the appended claims.

[0043] Furthermore, it will be recognized by one of ordinary skill in the art that certain structural and operational details of the non-limiting embodiment(s) discussed hereafter may be modified or omitted (i.e. non-essential) altogether. In other instances, well known methods, procedures, and components have not been described in detail. Developers of the present technology have devised methods and processors for controlling systems comprising industrial equipment. Broadly, industrial equipment encompasses a vast array of machinery and tools designed to facilitate manufacturing, processing, and other industrial operations. Such systems include machines that are typically robust, specialized, and engineered to perform specific tasks efficiently and reliably. From conveyor belts and assembly lines to heavy- duty machinery like drills, lathes, presses, molds, industrial equipment plays a role in various sectors such as automotive, aerospace, construction, and the like.

[0044] One non-limiting feature of industrial equipment is the ability to handle high workloads and operate under demanding conditions and extended periods of time. Whether it's operating in extreme temperatures, high-pressure environments, or corrosive settings, industrial machinery can be often constructed from durable materials like steel, cast iron, or specialized alloys to ensure longevity and resilience. It is contemplated that industrial equipment can be equipped or integrated with advanced automation and control systems, enhancing precision, productivity, and safety in manufacturing processes.

[0045] Developers have devised some methods for inter alia improving efficiency, reduce downtime, and enhance performance of industrial systems and / or individual components thereof. In at least some embodiments, one or more components of an industrial system may include smart sensors, predictive maintenance systems, and robotics to optimize operations and streamline production workflows.

[0046] In one non-limiting example, industrial systems can include assembly systems for automotive production lines, facilitating the integration and assembly of various vehicle components. These systems utilize robotic arms, conveyors, CNC machines, and specialized machinery to perform tasks such as welding, riveting, and fastening. By automating assembly processes, manufacturers can achieve higher production speeds, improve product quality, and optimize resource utilization.

[0047] In an other non-limiting example, industrial systems can include thermoforming systems used to manufacture packaging containers and trays from thermoplastic materials through a heat-forming process. These systems include thermoforming machines, mold sets, and trimming equipment for producing blister packs, clamshells, and trays for retail and food packaging applications. Thermoforming systems offer manufacturers cost-effective solutions for customizing packaging designs, enhancing product visibility, and extending shelf life.

[0048] In a further non-limiting example, the industrial system can include injection molding systems for producing produce plastic parts and components. Such systems can include injection molding machines such as a hopper for resin pellets, a barrel where the pellets are melted, an injection unit that injects the molten material into a mold cavity, and a clamping unit that holds the mold closed under pressure during the cooling and solidification process. Injection molding machines vary in size, capacity, and complexity, accommodating a wide range of production needs and part geometries. Such systems can include molds and tooling components for defining the shape, size, and features of the molded parts. These molds are typically made from steel or aluminum and consist of two halves, the cavity and the core, which create the desired part geometry. Injection molding tooling includes various features such as runners, gates, and cooling channels to facilitate the injection molding process and optimize part quality and cycle times. Such systems can include auxiliary components to enhance productivity, efficiency, and part quality. Auxiliary components may include hopper loaders, material dryers, and granulators for resin handling and preparation, mold temperature controllers, sprue pickers, and part conveyors to help streamline the injection molding process and optimize operational workflows. Such systems can include process control components to monitor and control parameters of the injection molding process to ensure consistent part quality and production efficiency. These components include sensors, actuators, and controllers to control variables such as temperature, pressure, injection speed, and cooling time. Process control components help optimize cycle times, reduce scrap rates, and minimize energy consumption, contributing to overall process reliability and profitability.

[0049] Injection Molding System

[0050] With reference to FIG. 1, there is depicted a non-limiting embodiment of the injection molding machine 100 which can be adapted to implement embodiments of the present technology. For illustration purposes only, it shall be assumed that the injection molding machine 100, which is part of a molding system 200, is configured to make preforms (not depicted) for subsequent blowmolding into a final shaped container. It should be noted that the molding machine 100 can be implemented as any suitable type of a molding machine, such as but not limited to, a hydraulic machine, an all-electric machine, and hybrids thereof.

[0051] It should be understood that in alternative non-limiting embodiments, the molding system 200 may include other types of molding systems, such as, but not limited to, compression molding systems, compression injection molding systems, transfer molding systems, and the like. Also, the molding system 200 could include additional machines such as a blow-molding machine (not depicted) disposed downstream from the molding machine 100 and / or a dryer disposed upstream from the molding machine 100. Alternatively, the blow-molding machine can be located separately from the molding system 200, such as being located in a different facility, be under control of a different entity, and the like.

[0052] In the non-limiting embodiment of FIG. 1 , the molding machine 100 includes a stationary platen 102 and a movable platen 104. The stationary platen 102 is also referred to as a fixed platen 102. In some embodiments of the present technology, the molding machine 100 may include a third non-movable platen (not depicted). Alternatively, or additionally, the molding machine 100 may include turret blocks, rotating cubes, turning tables and the like (all not depicted but known to those of skill in the art).

[0053] The injection molding machine 100 further includes an injection unit 106 for plasticizing and injection of the molding material. The injection unit 106 can be implemented as a single stage or a two-stage injection unit. In some alternative non-limiting embodiments of the present technology, the injection unit 106 can be implemented as one or more injection units (not depicted). It is however noted that the working principle of the molding machine 100 can be implemented in any known working principles.

[0054] In operation, the movable platen 104 is moved towards and away from the stationary platen 102 by means of stroke cylinders (not shown) or any other suitable means. Clamp force (also referred to as closure or mold closure tonnage) can be developed within the molding machine 100, for example, by using tie bars 108, 110 (typically, four tie bars 108, 110 are present in the molding machine 100) and a tie-bar clamping mechanism 112, as well as (typically) an associated hydraulic system (not depicted) that is usually associated with the tie-bar clamping mechanism 112. It will be appreciated that clamp tonnage can be generated using alternative means, such as, for example, using a column-based clamping mechanism, a toggle-clamp arrangement (not depicted) or the like.

[0055] The molding machine 100 has a mold 113 having a first mold half 114, and a second mold half 116. The first mold half 114 can be associated with the stationary platen 102 and the second mold half 116 can be associated with the movable platen 104. In the non-limiting embodiment of FIG. 2, the first mold half 114 defines the mold cavities 118. The mold cavities 118 generally form an array, such that there is a given number of columns and a given number or rows.

[0056] The second mold half 116 includes mold cores 120 complementary to the mold cavities 118, such that the mold cores 120 also generally form an array similar to the array formed by the mold cavities 118, where the array formed by the mold cores 120 has the same number of columns and rows as the array formed by the mold cavities 118. As will be appreciated by those of skill in the art, the mold cores 120 may be formed by using suitable mold inserts or any other suitable means. Even though not depicted in FIG. 1, the first mold half 114 may be further associated with a melt distribution network, commonly known as a hot runner, for distributing molding material from the injection unit 106 to each of the mold cavities 118.

[0057] FIG. 1 depicts the first mold half 114 and the second mold half 116 in a so-called “mold open position” where the movable platen 104 is positioned generally away from the stationary platen 102 and, accordingly, the first mold half 114 is positioned generally away from the second mold half 116. For example, in the mold open position, a molded article (not depicted) can be removed from the first mold half 114 and / or the second mold half 116. In a so-called “mold closed position” (not depicted), the first mold half 114 and the second mold half 116 are urged together (by means of movement of the movable platen 104 towards the stationary platen 102) and cooperate to define (at least in part) molding cavities into which the molten plastic (or other suitable molding material) can be injected, as is known to those of skill in the art.

[0058] The injection molding machine 100 can further include a robot 122, sometimes referred to as a retrieval device and / or a removal device, operatively coupled to the stationary platen 102. Those skilled in the art will readily appreciate how the robot 122 can be operatively coupled to the stationary platen 102 and, as such, it will not be described in detail herein. The robot 122 includes a mounting structure 124, an actuating arm 126 coupled to the mounting structure 124 and a takeoff plate 128 coupled to the actuating arm 126. The take-off plate 128 includes a plurality of molded article receptacles 130.

[0059] Generally speaking, the purpose of the plurality of molded article receptacles 130 is to remove molded articles (such as preforms, in this example) from the one or more mold cores 120 (or the one or more mold cavities 118) and / or to implement post mold cooling of the molded articles. In the non-limiting example illustrated herein, the plurality of molded article receptacles 130 includes a plurality of cooling tubes for receiving a plurality of molded preforms.

[0060] The molding machine 100 further includes a post-mold treatment device 132 operatively coupled to the movable platen 104. Those skilled in the art will readily appreciate how the post-mold treatment device 132 can be operatively coupled to the movable platen 104 and, as such, it will not be described here in any detail. The post-mold treatment device 132 includes a mounting structure 134 used for coupling the post-mold treatment device 132 to the movable platen 104. The post-mold treatment device 132 further includes a plenum 129 coupled to the mounting structure 134. Coupled to the plenum 129 is a plurality of treatment pins 133. The number of treatment pins within the plurality of treatment pins 133 generally corresponds to the number of receptacles within the plurality of molded article receptacles 130.

[0061] The molding machine 100 further includes a computing system 140, also referred to herein as a processor 140 or as a “controller”, configured to control one or more operations of the molding machine 100. As will be described below, the processor 140 is further configured to control one or more operations of the molding system 200 that the molding machine 100 is part of, and which will be described in greater detail below. As will be appreciated by those skilled in the art, the computing system 140 may include a plurality of processors or computer-implemented devices operatively connected together.

[0062] The processor 140 includes a human-machine interface (not separately numbered) or an HMI, for short. The HMI of the processor 140 can be implemented in any suitable interface. As an example, the HMI of the processor 140 can be implemented in a multi-functional touch screen. An example of the HMI that can be used for implementing non-limiting embodiments of the present technology is disclosed in co-owned United States Patent 6,684,264, content of which is incorporated herein by reference, in its entirety.

[0063] Those skilled in the art will appreciate that the processor 140 may be implemented using preprogrammed hardware or firmware elements (e.g., application specific integrated circuits (ASICs), electrically erasable programmable read-only memories (EEPROMs), etc.), or other related components. In other embodiments, the functionality of the processor 140 may be achieved using a processor that has access to a code memory (not shown) which stores computer-readable program code for operation of the computing system, in which case the computer-readable program code could be stored on a medium which is fixed, tangible and readable directly by the various network entities, or the computer-readable program code could be stored remotely but transmittable to the processor 140 via a modem or other interface device (e.g., a communications adapter) connected to a network (including, without limitation, the Internet) over a transmission medium, which may be either a non-wireless medium (e.g., optical or analog communications lines) or a wireless medium (e.g., micro wave, infrared or other transmission schemes) or a combination thereof.

[0064] In alternative non-limiting embodiments of the present technology, the HMI does not have to be physically attached to the processor 140. As a matter of fact, the HMI for the processor 140 can be implemented as a separate device. In some embodiments, the HMI can be implemented as a wireless communication device (such as a smartphone, for example) that is “paired” or otherwise communicatively coupled to the processor 140. The processor 140 can perform several functions including, but not limited to, receiving from an operator control instructions, controlling the molding machine 100 based on the operator control instructions or a pre-set control sequence stored within the processor 140 or elsewhere within the molding machine 100, acquiring one or more operational parameters associated with the molding system and the like (such as one or more of: a pump speed, accumulator charge, a fill rate, a hold time, a cooling time, a transfer speed, an injection start, a clamp tonnage, and a tonnage lock). According to non-limiting embodiments of the present technology, the processor 140 is further configured to process one or more of the acquired operational parameters associated with the molding system 200 and output information to the operator using the HMI and the like. It is contemplated that the processor 140 may be implemented in a distributed manner over a plurality of physical machines, without departing from the scope of the present technology.

[0065] The molding machine 100 further includes a number of monitoring devices (not depicted), the monitoring devices being configured to acquire various operational parameters associated with the performance of the molding machine 100. Generally speaking, these monitoring devices are known in the art and, as such, will not be described here at any length.

[0066] Just as an example, the injection molding machine 100 may include a counter to count mold opening and closing to determine the number of cycles over a period of time and / or the cycle time of each cycle. The injection molding machine 100 may also include a number of pressure gauges to measure pressure within various components of the injection molding machine 100 (such as hydraulic fluid pressure or molding material pressure).

[0067] According to non-limiting embodiments of the present technology, the processor 140 is configured to acquire a plurality of operational parameters associated with the molding machine 100. The nature of the so-acquired plurality of operational parameters can vary. How the processor 140 acquires the plurality of operational parameters will depend, of course, on the nature of the so- acquired plurality of operational parameters.

[0068] The processor 140 can acquire operational parameters (such as one or more of: a pump speed, accumulator charge, a fill rate, a hold time, a cooling time, a transfer speed, an injection start, a clamp tonnage, and a tonnage lock) by monitoring the operation of the molding machine 100. Just as an example, the processor 140 can acquire the cycle time by monitoring the performance of the molding machine 100. Naturally, the processor 140 can acquire some of the machine variables by either the operator entering them using the HMI or by reading a memory tag (not depicted) associated with the mold (i.e. the above described first mold half 114 and the second mold half 116) that is used in the molding machine 100. Various implementations of the memory tag (not depicted) are known in the art. Generally speaking, the memory tag (not depicted) may store information about the mold, the molded article to be produced, pre-defined control sequences, setup sequences and the like.

[0069] In some non-limiting embodiments of the present technology, the processor 140 can acquire the operational parameters by receiving an indication of those parameters from the operator. However, within some implementations of the molding machine 100, it is possible for the processor 140 to acquire some (or even all) of the operational parameters by monitoring performance of the molding machine 100. In some non-limiting embodiments of the present technology, this can be implemented by inspecting the molded articles and assessing compliance with specification, then accepting, rejecting articles on the basis of the assessment. This can be done by suitable devices (such as for example, an optical inspection device) or by an operator. Naturally, other ways for the processor 140 to acquire some or all of these or other operational parameters are possible, some of which will be described below.

[0070] In alternative non-limiting embodiments of the present technology, the processor 140 can acquire the operational parameters from other components of the molding system 200, such as but not limited to the aforementioned blow-molding machine (not depicted) disposed downstream from the molding machine 100 and / or a dryer disposed upstream from the molding machine 100. Additionally or alternatively, the operational parameters can be associated with a quality parameter of the blown or fdled container made from the molded article (this is particularly applicable in those embodiments, where the molded article is a preform for subsequent blow-molding into a final shaped container).

[0071] Additionally or alternatively, the operational parameters can be obtained from an inspection (which can be executed by an inspection system, not depicted, or an operator of the molding system 200). One example of such operational parameter that can be obtained through the inspection includes but is not limited to parameters of a molded article, such as physical attributes - dimensions, weight, etc. In other words, the inspection system can generate an indication of a current value of the operational parameter.

[0072] Additionally or alternatively, the inspection can identify impurities present in the molded articles or molding related defects (such as but not limited to short shots and the like).

[0073] Another example of a source of operational parameters can be a device for supplying material to the molding machine 100, such as a post consumer plastic recycling device (which can be implemented as a liquid state polycondensation device or the like), which can provide operational parameters that include but are not limited to resin parameters (such as viscosity, quality of flakes, moisture level, residence time, and the like). Broadly speaking, the operational parameters obtained by the controller 140 from such devices can include viscosity, temperature profile of the melt, status of the filter elements, quality of the incoming flakes (i.e., ground up old bottles or other types of recycled material), moisture of the incoming flakes, residence time in the hopper, etc.

[0074] In some embodiments of the present technology, the aforementioned inspection of the molded articles can provide operational parameters from which to define statistical ranges (of key dimensions). These statistical ranges can be used by the controller 140 to determine what operational parameters are within acceptable ranges for the optimal performance of the molding machine 100, based at least in part on the statistical variation.

[0075] Control system

[0076] With reference to FIG. 2, there is depicted a block diagram of a non-limiting embodiment of a control system 200 for operating the injection molding system 200. The control system 200 can be implemented, for example, in the computing system 140 (FIG. 1). As shown, the control system 200 is configured in layers. The tiers include an enterprise platform layer 202, a supervisory control layer 204 and a control layer 206.

[0077] The control layer 206 includes a plurality of control modules, each of which generally controls operation of a respective subsystem of the injection molding system 200. For example, the depicted control modules 207-1, 207-2, ...207-10 are responsible for controlling injection unit 106, the movable platen 104, the robot 122 and other components of the injection molding system 200. Each control module may include one or more programmable logic controllers (PLCs) coupled to individual actuators and sensors within the subsystem.

[0078] The supervisory control layer 204 includes a supervisory controller 205. The supervisory controller 205 interfaces with control modules of the control layer 206 in order to direct and coordinate molding operations, manage configuration of the subsystems and to direct production of articles. The supervisory control layer 204 is interconnected with enterprise platform layer 202 by way of a network. The network may be a local-area network (LAN) or a wide-area network such as the internet. The supervisory controller 205 is operable to receive and interpret operating information associated with the molding system 200, such as status messages, from each of the control modules 207 of the control layer 206. In an example, sending of messages may be initiated by the control modules. For example, messages may be sent in response to initiation or completion of a processing step, or on a periodic basis.

[0079] The supervisory controller 205 also implements the aforementioned human machine interface (HMI) or operator interface functionality, which may include a graphical user interface presented to the operator on one or more display panels, which may be touch sensitive. The HMI may also be equipped with hardware buttons or other manual controls for specific functions.

[0080] The enterprise platform layer 202 comprises one or more servers 203 and may serve as a data repository for operational data required for production of molded articles using the injection molding system 200.

[0081] In accordance with the non-limiting embodiments of the present technology, master data structures listing may be maintained as part of enterprise platform layer 202 and at any given time, only a subset of data may be copied to the supervisory control layer 204 and the control layer 206. For example, data may be stored in a master database at the enterprise layer 202 and subsets of data may be written to memory at the supervisory control layer 204 or at the control modules of control layer 206. The data copied may, for example, be only data relating to configurations that are possible with tooling physically available for installation.

[0082] The enterprise platform layer 202 and / or the supervisory layer 204 may be configured to monitor production / output of the injection molding system 200. Specifically, orders defining production requirements may be input or received at the enterprise platform layer 202 and / or the supervisory layer 204.

[0083] The production requirements may include, for example, types and quantities of the molded articles required, and time at which the articles are required. Based on the production requirements, the enterprise platform layer 202 and / or the supervisory layer 204 can schedule production of specific types and quantities, and send instructions to the supervisory control layer 204 to cause production according to the schedule. In some embodiments, enterprise platform layer 202 communicates with supervisory control layers 204 of multiple molding systems and can schedule and direct production by each of the molding systems. In some embodiments, the enterprise platform layer 202 is interconnected with the supervisory control layers 204 of a plurality of different molding systems (not depicted). In such embodiments, the enterprise platform layer 202 may request configuration information for each molding system, to determine which systems are capable of producing articles of the desired types and in the desired quantities. Production scheduling may involve distributing production instructions based on the capabilities and current production scheduled.

[0084] In some embodiments, the enterprise platform layer 202 and / or the supervisory control layer 204 provide an interface for outside users such as operators to input instructions and monitor production. For example, operators may access the enterprise platform layer 202 and / or the supervisory control layer 204 through the interface to place production orders, or to retrieve data on production progress, or the like. The interface may be provided to by way of a wide-area network (WAN) such as the internet. For example, users may interact with the enterprise platform layer 202 and / or the supervisory control layer 204 through a website or mobile application or by calls to one or more APIs.

[0085] Supervisory computing system

[0086] With reference to FIG. 3, there is depicted a supervisory computing system 350 in communication with an industrial system 300. Broadly the supervisory computing system 350 is configured to acquire data from one or more sub-components of the industrial system 300, process the acquired data, and in response, control one or more sub-components of the industrial system 300.

[0087] In one embodiment, the industrial system 300 may be implemented as the injection molding system 200 described above. However, this may not be the case in each and every embodiment of the present technology. In other non-limiting embodiments of the present technology, the industrial system 300 may be at least one of the following systems, but not limited to: abrasive manufacturing system, automobile electronics manufacturing system, metal refining system, heating and air- conditioning (HVAC) equipment manufacturing system, engine and turbine manufacturing system, lighting fixtures manufacturing system, gas manufacturing system, semiconductor manufacturing system, printing system, aircraft manufacturing system, battery manufacturing system, brewery and food processing system, fertilizer manufacturing system, glass manufacturing system.

[0088] In one example, the industrial system 300 may be employed for abrasive and sandpaper manufacturing with a range of specialized sub-components configured to each stage of the industrial process. The sub-components comprise mixers and / or kneaders for blending abrasive grains and bonding agents to form a homogeneous slurry. The sub-components comprise coating machines equipped with rollers and / or applicators to apply the abrasive coating onto backing materials, and curing ovens and / or drying chambers to solidify the bonding agents and remove excess moisture. The sub-components comprise cutting machines, slitters, and / or die-cutting equipment that are utilized to shape the coated abrasive material into sheets, rolls, or discs, ensuring uniformity and consistency in size and shape. The sub-components may also comprise inspection equipment such as optical scanners and / or dimensional gauges employed to assess the quality and integrity of the abrasive products before they undergo packaging by bundling machines or automated packaging lines. Throughout the manufacturing process, one or more subcomponents may be controlled to ensure efficiency, accuracy, and adherence to pre-determined quality standards, enabling the production of abrasive products for various industrial applications.

[0089] In an other example, the industrial system 300 may be employed for automobile electronics manufacturing with specialized sub-components configured for different stages of production. Initial stages involve circuit board fabrication, where sub-components such as automated solder paste printers, pick-and-place machines, and reflow ovens are utilized to assemble electronic components onto printed circuit boards (PCBs). The sub-components comprise Surface Mount Technology (SMT) and Through-Hole Technology (THT) machines employed to accommodate different component types and sizes, ensuring optimal placement accuracy and soldering quality. Following PCB assembly, sub-components such as automated testing and inspection equipment, including automated optical inspection (AOI) systems and in-circuit testers (ICT), can be used to verify the functionality and integrity of electronic assemblies, detecting defects and ensuring compliance with quality standards. Once validated, electronic assemblies undergo encapsulation and potting processes using sub-components such as injection molding machines or conformal coating systems to protect against environmental factors. Automated assembly lines equipped with robotic arms and conveyors can be used to integrate electronic components into automotive systems, such as engine control units (ECUs), sensors, and infotainment systems. Throughout the manufacturing process, one or more sub-components may be controlled to facilitate production of automobile electronics that meet industry standards and customer expectations for safety and efficiency.

[0090] In an additional example, the industrial system 300 may be employed for copper, zinc, and lead refining with sub-components configured to extract and purify metals from raw ores or recycled materials. Initially, the raw materials undergo crushing and grinding using sub-components such as jaw crushers, gyratory crushers, and / or ball mills to reduce particle size and increase surface area for subsequent processing. The crushed ore is then subjected to beneficiation processes such as flotation, gravity separation, or magnetic separation to concentrate the desired metal-bearing minerals from the ore gangue. Once concentrated, the ore undergoes smelting in sub-components such as furnaces or converters, where high temperatures and reducing agents are used to extract the metal from the ore concentrate. Smelting sub-components include blast furnaces, reverberatory furnaces, and electric arc furnaces, each of which can be optimized for specific ore types and process requirements. Following smelting, the crude metal undergoes refining processes to remove impurities and further purify the metal. Refining techniques utilized by one or more subcomponents include electrolysis, solvent extraction, and pyrometallurgical refining. Such subcomponents include electrolytic cells, leaching tanks, and refining kettles employed to achieve high purity levels. Finally, the refined metals are cast into ingots, bars, or other forms using casting machines and molds, ready for further processing or use in various industries such as construction, electronics, and manufacturing. Throughout the refining process, one or more sub-components may be controlled to ensure environmental safeguards with regulatory standards and production of high-quality metals.

[0091] In yet another example, the industrial system 300 may be employed for heating and air- conditioning (HVAC) equipment manufacturing and comprises sub-components configured for different stages of production. Initially, sheet metal fabrication sub-components such as shears, press brakes, and laser cutting machines are used to shape and form metal components, including ductwork, housings, and heat exchangers. Welding equipment, including MIG / TIG welders and spot- welding machines, is employed to join metal components together. In parallel, automated coil winding machines and core assembly sub-components are utilized to manufacture components such as electric motors and transformers used in HVAC systems. Following component fabrication, assembly lines equipped with conveyors, robotic arms, and assembly fixtures are utilized to integrate components into HVAC units, including furnaces, air handlers, heat pumps, and air conditioning units. Testing and quality control equipment such as pressure gauges, airflow meters, and thermal imaging cameras are employed to ensure the performance, efficiency, and safety of HVAC systems before they are packaged and shipped to customers. Throughout the manufacturing process, one or more sub-components may be control to adhere to industry standards, regulatory requirements, and manufacture energy-efficient HVAC equipment that meets customer needs and contributes to indoor comfort and air quality.

[0092] In a further example, the industrial system 300 may be employed for engine and turbine manufacturing with sub-components configured for different stages of production. Initially, machining centers equipped with Computer Numerical Control (CNC) technology are used to fabricate precision components such as engine blocks, cylinder heads, and turbine blades from raw materials such as metal alloys and / or composites. Cutting tools such as drills, end mills, and lathes are utilized to shape and finish components to tight tolerances and specifications. In parallel, casting and forging equipment such as die casting machines and forging presses are employed to produce complex engine and turbine components. Following component fabrication, assembly lines equipped with robotic arms, conveyors, and / or automated assembly fixtures integrate components into complete engines, turbines, or engine modules. Testing and quality control subcomponents such as dynamometers, vibration analyzers, and thermal imaging cameras are utilized to ensure the performance, reliability, and safety of engines and turbines before they are packaged and shipped to customers. Throughout the manufacturing process, one or more sub-components may be controlled to adhere to industry standards and regulatory requirements to ensure the production of high-quality engines and turbines that meet customer needs and performance specifications for applications such as aerospace, automotive, power generation, and marine propulsion.

[0093] In an other example, the industrial system 300 may be employed for lighting fixtures manufacturing with sub-components configured for different stages of production. Initially, sheet metal fabrication sub-components such as shears, press brakes, and stamping machines are used to shape and form metal components, including housings, reflectors, and mounting brackets. Injection molding machines and / or extrusion sub-components are employed to manufacture plastic components such as diffusers, lenses, and covers. Assembly lines equipped with conveyors, robotic arms, and assembly fixtures integrate components into complete lighting fixtures, including LED fixtures, fluorescent fixtures, and the like. Testing and quality control equipment such as photometers, spectrometers, and thermal chambers are utilized to ensure the performance, efficiency, and safety of lighting fixtures before they are packaged and shipped to customers. Throughout the manufacturing process, one or more sub-components may be controlled to produce energy-efficient lighting fixtures that meet customer needs and contribute to illumination and ambiance in various settings, including residential, commercial, and industrial spaces.

[0094] In an additional example, the industrial system 300 may be employed for gas manufacturing, such as oxygen and hydrogen, for example, comprises sub-components configured for different stages of production. Initially, raw materials such as air or water undergo separation processes to isolate oxygen and hydrogen molecules. For oxygen production, air separation units (ASUs) equipped with cryogenic distillation columns or pressure swing adsorption (PSA) systems are used to extract oxygen from atmospheric air, utilizing the differences in boiling points or adsorption properties of gases. Hydrogen production may include water electrolysis or steam methane reforming (SMR) processes, utilizing electrolyzers or reformer units to split water molecules or hydrocarbons into hydrogen and oxygen. Following gas separation or generation, purification sub-components such as molecular sieves, pressure filters, and catalytic converters remove impurities and contaminants to achieve high purity levels required for industrial applications. Finally, compression and storage equipment such as compressors, cylinders, and cryogenic tanks are utilized to package and distribute gases to customers, including industries such as metal fabrication, chemical processing, and healthcare.

[0095] In a further example, the industrial system 300 may be employed for semiconductor machine manufacturing with sub-components configured for different stages of production. Initially, precision machining centers equipped with CNC technology are used to fabricate complex components such as wafer handling systems, vacuum chambers, and lithography equipment from high-quality materials such as aluminum, stainless steel, and ceramics. Additionally, cleanroom facilities equipped with ultra-pure water systems, air filtration units, and environmental controls can be used to ensure a contamination-free environment for semiconductor equipment assembly. Assembly lines utilize robotic arms, conveyors, and precision alignment fixtures to integrate components into complete semiconductor machinery, including wafer fabrication equipment, etching systems, and inspection tools. Testing and calibration equipment such as laser interferometers, spectrometers, and particle counters are employed to ensure the precision, accuracy, and performance of semiconductor machinery before they are packaged and shipped.

[0096] In an other example, the industrial system 300 may be employed for printing, paper, food and textile applications. In the printing sector, specialized sub-components such as lithographic presses, digital printers, and flexographic machines are employed to produce various printed materials, including newspapers, packaging, and promotional materials. In paper manufacturing, sub-components such as pulp digesters, paper machines, and converting sub-components are used to process raw materials into various paper products, such as newsprint, cardboard, and tissue paper. In the food processing sector, sub-components such as mixers, ovens, extruders, and packaging equipment are utilized to produce a wide range of food products, including baked goods, snacks, beverages, and packaged foods. In the textile industry, sub-components such as spinning frames, weaving looms, knitting machines, and dyeing equipment are employed to process natural and synthetic fibers into textiles, fabrics, and garments.

[0097] In a further example, the industrial system 300 may be used for aircraft manufacturing. Initially, precision machining centers may use CNC machines to fabricate components such as fuselage sections, wings, engine casings, and landing gear from aerospace-grade materials such as aluminum alloys, titanium, and composites. Advanced manufacturing techniques such as additive manufacturing (3D printing) can also be utilized to produce intricate components with reduced weight and improved performance characteristics. Assembly lines equipped with robotics, automated guided vehicles (AGVs), and precision jigs and fixtures integrate components into complete aircraft structures, including airframes, propulsion systems, avionics, and interiors. Testing and quality control equipment such as non-destructive testing (NDT) tools, coordinate measuring machines (CMMs), and flight simulators are employed to ensure the performance, safety, and compliance of aircraft and components with regulatory standards and customer specifications. Furthermore, engine manufacturing process can be implemented using subcomponents such as turbine blade casting machines, engine assembly lines, and engine test cells to produce and validate the performance of aircraft engines. Parts manufacturing encompasses a wide range of specialized sub-components, including forging presses, composite layup machines, and surface treatment facilities, to produce various aircraft components such as fasteners, bearings, and electrical systems.

[0098] In another example, the industrial system 300 may be employed for battery manufacturing. Initially, raw materials such as lithium, cobalt, nickel, and graphite can be processed and prepared for battery production. A variety of sub-components may be used for mining, refining, and chemical processing operations to extract and purify the required materials. Next, the processed materials are combined to form battery components such as cathodes, anodes, electrolytes, and separators. Sub-components such as mixers, coating machines, and drying ovens are used to produce thin film coatings that enhance the performance and stability of battery components. Once the battery components are prepared, they are assembled into battery cells using automated assembly lines equipped with robotic arms, vacuum chambers, and precision welding equipment. The assembly process involves stacking layers of cathode, anode, and separator materials, encapsulating them in a protective casing, and filling them with electrolyte. Following cell assembly, the batteries undergo testing and quality control measures to ensure performance, safety, and reliability. Testing equipment such as battery cyclers, impedance analyzers, and thermal chambers are used to evaluate the capacity, voltage, temperature, and cycle life of batteries under various operating conditions.

[0099] In a further example, the industrial system 300 may be employed by brewery. Initially, raw materials such as malted barley, hops, water, and yeast are sourced and prepared for brewing. Malted barley is milled to expose starches, while hops are weighed and sometimes pelletized for ease of use. Water is treated to achieve the desired mineral composition and pH level, and yeast is cultured and propagated to ensure fermentation. Next, the milled barley is mashed in a mash tun with hot water to convert starches into fermentable sugars. This process, known as mashing, is facilitated by temperature-controlled sub-components such as mash tuns and lauter tuns. The resulting liquid, known as wort, is then separated from the spent grain through lautering and transferred to a brew kettle. In the brew kettle, the wort is boiled and hops are added at various stages to impart bitterness, flavor, and aroma to the beer. The boiling process also sterilizes the wort and concentrates its sugars. After boiling, the wort is rapidly cooled and transferred to fermentation vessels, where yeast is added to begin fermentation. Fermentation vessels are equipped with temperature control systems to maintain optimal fermentation conditions. Once fermentation is complete, the young beer is transferred to conditioning tanks or aging vessels for maturation and clarification. During this stage, the beer develops its flavor profile and undergoes conditioning to improve its clarity and stability. Depending on the beer style, additional ingredients such as fruit, spices, or wood may be added during conditioning. Finally, the conditioned beer is filtered, carbonated, and packaged into bottles, cans, or kegs using sub-components such as bottling lines, canning lines, and kegging systems.

[0100] In an additional example, the industrial system 300 may be employed for fertilizer manufacturing. Initially, raw materials such as nitrogen, phosphorus, and potassium sources are sourced and processed. Nitrogen may be obtained from ammonia or urea, phosphorus from phosphate rock or phosphoric acid, and potassium from potash or potassium chloride. Next, the raw materials undergo chemical reactions to produce fertilizers with specific nutrient compositions. For example, a variety of sub-components may be used for combining ammonia and nitric acid to produce ammonium nitrate, a nitrogen-rich fertilizer. Similarly, phosphate rock is reacted with sulfuric acid to produce phosphoric acid, which is then neutralized with ammonia to produce ammonium phosphate fertilizers. Following processing via sub-components facilitating chemical reactions, the fertilizers are granulated to improve handling, storage, and application characteristics. Granulation sub-components such as rotary drum granulators or fluidized bed granulators are used to agglomerate the fertilizers into spherical granules of uniform size and shape. Once granulated, the fertilizers may undergo additional processing steps such as coating or blending to enhance their performance and efficacy. Coating sub-components apply protective coatings to the fertilizer granules to control release rates, improve handling properties, and reduce nutrient losses through volatilization or leaching. Blending sub-components mix different fertilizers or additives to create custom fertilizer blends with specific nutrient ratios tailored to crop requirements.

[0101] In a further example, the industrial system 300 may be employed for glasses and contact lens manufacturing. Initially, raw materials such as silica sand, soda ash, limestone, and various additives are sourced and prepared for glass manufacturing. Specialized sub-components are used to melt these materials (e.g., in furnaces) at high temperatures to form molten glass, which is then shaped and formed into lenses through processes such as blowing, pressing, or casting. Precision molding sub-components, including molds and dies, are used to achieve the desired shape and curvature of the lenses. Once formed, the lenses undergo various finishing processes to enhance optical clarity and surface quality. This may include grinding, polishing, and coating processes performed using specialized sub-components such as lens edgers, polishers, and coating chambers. Additionally, lenses may be tinted, dyed, or treated with anti-reflective coatings to improve visual performance and aesthetics. In the case of contact lens manufacturing, raw materials such as hydrogel polymers or silicone hydrogels are processed and molded into lens blanks using injection molding or lathe-cutting equipment. The lens blanks are then subjected to lathing, polishing, and surface treatment processes to achieve the desired lens parameters, including curvature, thickness, and optical properties. Following finishing processes, lenses are inspected for quality and conformity to specifications using automated inspection systems and visual inspection stations. Quality control measures ensure that lenses meet regulatory standards and customer requirements for optical performance, durability, and safety.

[0102] Returning to the description of FIG. 3, the supervisory computing system 350 comprises a data acquisition module 302 configured to acquire data from one or more sub-components of the data acquisition module 302. In one embodiment, the data acquisition module 302 may acquire timeseries data from the one or more sub-components of the industrial system 300. Broadly, time-series data is any type of information presented as an ordered sequence. It can be represented as a collection of observations for a single subject, assembled over same or different, optionally equally spaced, time intervals. It can be said that data acquired by the data acquisition module 302 may comprise data collected at different points in time and organized chronologically.

[0103] In some non-limiting examples, the data acquisition module 302 may acquire time-series data from one or more sensors of the industrial system 300 such as, but not limited to: time-series data from pressure sensors monitoring pressure changes in pipelines or machinery, time series data from level sensors tracking fluid levels in tanks or reservoirs, time-series data from temperature sensors recording temperature variations in manufacturing processes, time-series data from pH sensors measuring acidity or alkalinity levels, time-series data from velocity sensors capturing movement speeds in conveyor belts or turbines, and the like. It is contemplated that time-series data may include one or more sensor readings associated with corresponding timestamps, creating a shared context for the data. In some embodiments of the present technology, the supervisory computing system 350 may employ the data acquisition module 302 to collect large amounts of data from a plethora of measurable points along an industrial process executed by the industrial system 300, enabling processing and decision-making operations. As it will be explained in greater details herein further below, the supervisory computing system 350 may employ the time-series data to determine actionable insights to drive efficiency, safety, and / or productivity of the industrial process performed by the industrial system 100.

[0104] In some embodiments, the data acquisition module 302 may be configured to apply one or more filtering algorithms for processing time-series data acquired from the one or more components of the industrial system 300. In some embodiments, the data acquisition module 302 interfaces with sensors, transducers, and other data-generating devices distributed throughout the industrial system 300. These sensors may capture real-time measurements of various parameters such as temperature, pressure, flow rates, vibration levels, and electrical signals. The acquisition module 302 may collect this raw sensor data, often in the form of analog signals, and digitizes it using analog-to-digital converters (ADCs), for example, for processing in a digital environment. Once the raw data is digitized, the data acquisition module 302 may apply filtering algorithms to preprocess the data before further analysis. Filtering algorithms can include techniques such as low-pass, high-pass, band-pass, or adaptive filters, depending on the specific characteristics of the acquired data and the desired outcome of the filtering process. For example, low-pass filters may be used to remove high-frequency noise from sensor signals, while high-pass filters may isolate specific frequency components relevant to detecting anomalies or faults in the industrial system 300. Additionally, the data acquisition module 302 may incorporate advanced signal processing techniques such as Fourier transforms, wavelet analysis, or Kalman filtering to extract useful information from the time-series data. These techniques enable the identification of patterns, trends, and anomalies in the data, facilitating predictive maintenance, fault detection, process optimization, and performance monitoring of industrial equipment and processes.

[0105] In additional embodiments, the acquisition module 302 may be configured to store one or more internal tables that identify pre-defined categories of time-series data to be monitored. These tables may be based on the customer preferences, operator defined priorities, and / or the vendor of the industrial system 300 pre-determined parameters, without departing from the scope of the present technology.

[0106] The supervisory computing system 350 can be configured to host a database system 312. Broadly, the database system 312 is a structured and organized collection of data that allows efficient storage, retrieval, and / or management. In this embodiment, the database system 312 serves as a central repository of the supervisory computing system 350 for storing information, such as records, documents, structured data, unstructured data, and the like. For example, the database system 312 include a data storage engine to persist data, the query processing module to translate queries into low-level operations, an indexing mechanism for data retrieval, an access control and security layer to manage permissions, transaction management module for maintaining data consistency, and a backup / recovery module to prevent data loss.

[0107] It is contemplated that the database system 312 may be specifically configured for managing storing, retrieving, and / or managing data associated with the operation of the industrial system 300. For example, scalability, flexibility, and performance metrics of the database system 312 may be selected for a specific application in the context of the industrial process performed by the industrial system 300.

[0108] In some embodiments, the database system 312 may be configured to store a plurality of data items 320. It is contemplated that the database system 312 may serve as a centralized repository, capable of efficiently storing, organizing, and retrieving vast amounts of data generated by various components and processes within the industrial system 300. The database system 312 may interface with the data acquisition modules 302, sensors, and data-generating devices distributed throughout the industrial system 300. These devices capture real-time measurements, status updates, event logs, and other data points related to equipment operation, process parameters, environmental conditions, and production metrics. The database system 312 may store this data in a structured format, ensuring consistency and integrity while accommodating the diverse range of data types and sources. The database system 312 may store the plurality of data items 320 in a relational and / or non-relational database management system (DBMS), organized into tables, documents, or other data structures optimized for efficient storage and retrieval. Optionally, each data item may be associated with relevant metadata, timestamps, and contextual information to provide a comprehensive understanding of its origin, relevance, and significance within the industrial system 300. In other embodiments, the database system 312 may be implemented with data organization, indexing, and query optimization algorithms to facilitate rapid access and retrieval of information. The database system 312 may be a distributed system and may store data locally, remotely, and / or a combination of both.

[0109] In some embodiments, the database system 312 may be configured to host a plurality of machine learning algorithms (MLAs) 330. In some embodiments, the plurality of MLAs 330 may beapriori trained using at least some of the plurality of data items 320. Machine learning algorithms

[0110] Broadly, the supervisory computing system 350 may be configured to employ MLAs for detecting operational anomalies with one or more sub-components of the industrial system 300 and provide instructions to operator(s) of the industrial system 300 which are indicative of potential remedial actions for remediating the operational anomalies. Generally, MLAs may be trained on data stored in the database system 312 to detect a root cause of an anomaly, determine a corresponding remedial action, and to generate human-understandable instructions for controlling operation of the one or more sub-components of the industrial system 300.

[0111] Training data may comprise a combination of one or more operational parameters of one or more operational components such as temperature, pressure, material characteristics, cycle times, and other relevant process variables. In some embodiments, training data may be generated based on historical time-series data provided by the industrial system 300.

[0112] One architecture that can be used for processing time-series data is Recurrent Neural Networks (RNNs). RNNs are well-suited for sequential data analysis due to their ability to capture temporal dependencies and recurrent connections between data points over time. As mentioned above, LSTM networks, a type of RNN, are effective for modeling long-term dependencies and handling time-series data with variable temporal dynamics. LSTM networks can leam from past observations to predict future trends, detect anomalies, and classify system states based on historical data.

[0113] An other architecture that can be used for processing time-series data is Convolutional Neural Networks (CNNs). While CNNs are traditionally used for image recognition tasks, they can also be adapted for one-dimensional data such as time-series signals. In the context of the industrial system 300, CNNs can leam spatial features from sensor data streams, for example, by applying convolutional filters across time-series sequences. This enables CNNs to detect localized patterns, extract relevant features, and classify system behavior based on sensor inputs or other types of time-series data sources within the industrial system 300.

[0114] Furthermore, hybrid architectures combining RNNs and CNNs, such as Convolutional Recurrent Neural Networks (CRNNs), offer time-series data analysis for industrial applications. CRNNs leverage the strengths of both RNNs and CNNs to capture both spatial and temporal dependencies in the data. For instance, CRNNs can leam hierarchical representations of complex time-series patterns by combining convolutional feature extraction with recurrent sequence modeling, enabling more accurate prediction and classification of system behavior. In addition to Neural Network architectures, ensemble methods such as random forests, gradient boosting machines (GBMs), and ensemble learning techniques offer alternative approaches for time-series data processing from various data sources within the industrial system 300. Ensemble methods combine multiple models to improve prediction accuracy, robustness, and generalization performance by aggregating predictions from diverse base learners.

[0115] The plurality of MLAs 330 comprises inter alia a first MLA 332 and a second MLA 334. In some embodiments, the first MLA 332 may be implemented via a Deep Neural Network (DNN) architecture, while the second MLA 334 may be implemented via a Large Language Model (LLM) architecture.

[0116] In one example, the DNN architecture may be an Attention-Based Deep Convolutional Autoencoding Prediction Network (AT-DCAEP), which is a DNN model designed for multivariate time series data anomaly detection. It combines a characterization network based on convolutional autoencoders with a prediction network that employs attention mechanisms. The AT-DCAEP may be trained for anomaly detection without requiring pre-labeled large-scale datasets. It is contemplated that the AT-DCAEP may be trained using one or more unsupervised training techniques without departing from the scope of the present technology. In an other example, the DNN architecture may be a CRNN, which is a non-linear forecasting algorithm used for correlated time-series data forecasting. In a further example, the DNN architecture may be a LSTM model configured to capture long-term dependencies in time-series data. Broadly, LLMs are advanced Al algorithms that can be trained on inter alia large text datasets for natural language processing tasks. LLMs can be used for text generation, translation, and summarization. A given LLM may comprise feedforward layers, embedding layers, and attention layers. In some embodiments of the present technology, the LLM can be used to generating Human-Understandable Instructions for detected anomalies and potential remedial action(s).

[0117] For example, when a given root cause of an anomaly is detected and a remedial action(s) is determined, the supervisory computing system 350 may use this data as input into the LLM configured to, in response, generates instructions for human operators for executing the remedial action(s). For example, instructions may include details about the root cause of the anomaly and recommended steps for performing one or more recommended remedial actions. The LLM may be trained to prioritize operator’s safety in the generated instructions. The supervisory computing system 350 may be configured to collect operator feedback and further train the LLM for improving quality of instructions. The supervisory computing system 350 may comprise an MLA selection module 304. The MLA selection module 304 is configured to automatically evaluate and compare the performance of multiple MLAs, selecting the most suitable algorithm for the specific task at hand based on predefined criteria and objectives.

[0118] Firstly, the MLA selection module 304 interfaces with a diverse range of MLAs stored in the database system 312, such as RNNs, CNNs, GBMs, and other models suitable for time-series data analysis. It should be noted that different architecture may have different advantages which depend on inter alia different types and combinations of industrial time-series data.

[0119] In some embodiments, the MLA selection module 304 may be pre-configured with a set of evaluation metrics and criteria to assess the performance of MLAs on given time-series data. These metrics may include, but are not limited to accuracy, precision, recall, Fl -score, mean squared error (MSE), root mean squared error (RMSE), and other relevant performance indicators tailored to the specific industrial application and objectives. Additionally, the MLA selection module 304 may consider computational efficiency, scalability, interpretability, and resource requirements when comparing potential use of one MLA against an other MLA for time-series data processing.

[0120] It is contemplated that the MLA selection module 302 may conduct experiments or simulations (off-line and / or in real-time) to benchmark the performance of one or more MLAs stored in the database system 312 on a representative dataset of time-series data from the industrial system 300. The representative dataset may include historical data, and / or real-time time-series data acquired (and potentially filtered) from one or more components of the industrial system 300. It is contemplated that such representative datasets may be partitioned into training, validation, and test datasets. Additionally, the MLA selection module 304 may be configured to implement cross- validation techniques, and conduct comparative analysis of MLA performance across different evaluation scenarios. Based on the results of the performance evaluation, the MLA selection module 304 may be configured to employ a decision-making algorithm or strategy to rank and prioritize MLAs according to their suitability for a specific task. This decision-making process may involve heuristic methods, statistical tests, ensemble techniques, or other machine learning algorithms trained on historical performance data to identify the MLA that best meets the predefined criteria and objectives. As it will be described a selected MLA may be deployed and / or used in real-time for processing time-series data from one or more sub-components of the industrial system, which has been acquired and potentially filtered by the data acquisition module 302, for leveraging its capabilities to analyze, predict, and / pr interpret complex temporal patterns and trends. The MLA selection module 304 may periodically re-evaluate MLA performance and adapt its selection criteria based on evolving data characteristics, system dynamics, and performance feedback, ensuring continuous optimization and improvement of data analysis processes in the supervisory control system 300.

[0121] It should be noted that selected MLA may be employed for performing an anomaly detection procedure 306 based on time-series data. Broadly, anomaly detection involves identifying unusual or unexpected patterns in the time-series data that deviate from normal operating conditions, signaling potential faults, failures, or abnormal behavior within one or more sub-components of the industrial system 300.

[0122] It should be noted that to utilize a given MLA for anomaly detection, the supervisory control system 300 may be configured to train the MLA using historical time-series data that represents normal operating conditions of one or more sub-components of the industrial system 300. This training dataset may comprise various sensor measurements, process variables, and operational parameters collected during typical system operation, providing the MLA with a in a sense an understanding of normal behavior. Once the MLA is trained, it can be deployed to monitor realtime and / or streaming time-series data from the one or more sub-components of the industrial system 300 acquired by the data acquisition module 302. The selected MLA may continuously analyze incoming time-series data streams, and compare current observations to learned patterns of normal behavior. By applying predictive modeling techniques, such as forecasting or modeling temporal dependencies, the MLA can generate expected values or confidence intervals for each data point based on historical trends and patterns. During operation, the MLA may be configured to evaluate deviations between observed data and expected values, and thereby identify instances where the data significantly diverge from the norm. These deviations, or anomalies, may manifest as sudden spikes, drops, oscillations, or irregular patterns in the time-series data that are not otherwise present in normal system variability.

[0123] In some embodiments, when an anomaly is detected, the MLA may be configured to trigger alerts, notifications, and / or alarms to notify operators or control systems of potential issues within the industrial system 300. Depending on the severity and criticality of the anomaly, the industrial system 300 may be controlled to initiate corrective actions, such as adjusting operating parameters, shutting down equipment, or activating maintenance protocols to prevent further damage or disruptions. How instructions are generated and communicated will be discussed in greater details below. Furthermore, the selected MLA for anomaly detection may incorporate feedback mechanisms to adapt and refine its anomaly detection capabilities over time. By continuously learning from new data and feedback from operators, the selected MLA can dynamically adjust its anomaly detection thresholds, update its models, and improve its accuracy and reliability in identifying abnormal behavior within the industrial system 300.

[0124] Use of recycled molding material optimization

[0125] In some embodiments, anomaly detection may be performed for controlling one or more properties of a material being processed by the industrial system 300. For example, in the case where the industrial system 300 is embodied as an injection molding system, one or more MLAs may be configured to ensure monitoring one or more characteristics of the resin material inputted into the injection molding system. The one or more MLAs may be trained to predict a change in one or more characteristics of the resin material based on time-series data acquired from the injection molding system.

[0126] Developers have realized that with the gain in popularity of recycling processes, for example, it is desirable to understand the effect of one or more sub-components of the injection molding system on the input resin and mitigate the risk of breaking molecular chains in the resin. It should be noted that maintaining the integrity of the resin molecular chain is important for maintaining material properties, avoiding undesirable by-products and ensuring the material is able to complete multiple recovery / reprocessing cycles. Keeping a given material “in the loop” multiple times is desirable with increasingly adopted recycled content requirements.

[0127] In some embodiments of the present technology, a training dataset may be generated for predicting a change in material property at a given point along the industrial process based on time-series data of one or more sub-components of the industrial system 300. A given MLA may be configured to in a sense “learn” how operational parameters of one or more sub-components of the industrial system 300 affect the material being processed. As a result, the supervisory computing system 350 may be configured to detect anomalies with physical operation of the one or more sub-components based on time-series data and anomalies with physical properties of material being processed based on time-series data. Therefore, the supervisory computing system 350 may be used to optimize and adapt the industrial process to ensure normal operation of the one or more sub-components of the industrial system 300 and / or to ensure maintaining one or more material properties of the material fed into the industrial system 300. It is contemplated that in some embodiments, the MLA selection module 304 may be configured to select a given MLA for detecting anomalies with operation of one or more sub-components of the industrial system 300, and another given MLA for detecting anomalies with one or more material properties of the material processed at a given stage of the industrial process performed by the one or more sub-components of the industrial system 300.

[0128] It should be noted that the supervisory control system 350 may be configured to employ a second MLA to generate human-recognizable instructions for communicating anomaly detection and root cause thereof to an operator of the industrial system 300. In one embodiment, the second MLA may be embodied as an LLM.

[0129] The LLM may be configured to produce human-recognizable instructions based on detected anomalies and their root causes to enhance operational efficiency and decision-making processes by the operator. As explained above, the first MLA may detect one or more anomalies within the with one or more sub-components of the industrial system 300 by analyzing deviations from normal operating conditions. A root cause analysis may be conducted to determine the underlying factors contributing to the detected anomaly, leveraging historical data and sensor readings. Once the root cause is identified, the LLM comes into play, generating clear and understandable instructions or recommendations to address the anomaly and / or the root cause thereof. Employing natural language generation techniques, the LLM converts the root cause analysis findings into coherent instructions, which can include specific steps for troubleshooting, adjusting system parameters, performing maintenance tasks, or initiating corrective actions. These instructions are then delivered to one or more operations through various communication channels, facilitating prompt and effective responses to abnormal events.

[0130] In some embodiments, iterative improvement processes driven by feedback from operators and maintenance personnel can be implemented by the supervisory computer system 350 to continuously ameliorate an instruction generation 308 procedure executed by the LLM, thereby enhancing the quality and relevance of the instructions generated in response to anomalies. It is contemplated that human-recognizable instructions generated by the second MLA make take different forms such as text, audio, video, and / or other visual indicators.

[0131] The supervisory computer system 350 comprises an interfacing module 310 configured to convey human-recognizable instructions to one or more operators of the industrial system 300. The interfacing module 310 is designed to deliver instructions in a format that is easily understandable and accessible to operators, leveraging various interface modalities such as screens, speakers, and other output devices.

[0132] The interfacing module 310 may be configured to interface with the output capabilities of the industrial system 300, including graphical user interfaces (GUIs), display screens, audio systems, and notification mechanisms. This allows the interfacing module 310 to deliver instructions through multiple channels, accommodating different preferences and operational requirements.

[0133] In some embodiments, for visual output, the interfacing module 310 can present instructions on display screens strategically placed within the industrial environment, such as control rooms, workstations, or production floors. Instructions may be displayed in a clear, legible format, accompanied by graphical elements, icons, or color-coded indicators to enhance comprehension and usability.

[0134] In other embodiments, the interfacing module 310 can leverage audio output capabilities, such as speakers or PA systems, to deliver spoken instructions or alerts to operators in real-time. This auditory feedback provides an additional layer of situational awareness, especially in noisy or high-activity environments where visual cues may be less effective.

[0135] In further embodiments, the interfacing module 310 can utilize notification mechanisms, such as email alerts, SMS messages, or mobile app notifications, to reach operators remotely or on-the- go. This ensures that operators receive critical instructions and updates wherever they are, enabling prompt response and action.

[0136] In additional embodiments, the interfacing module 310 may enable interactive communication, such as via touchscreens or voice-activated interfaces of the industrial system 300, to allow operators to acknowledge receipt of instructions, provide feedback, or initiate follow-up actions directly from the interface.

[0137] Computer system

[0138] With reference to FIG. 5, there is depicted a computer system 500 suitable for use with some implementations of the present technology. It is contemplated that the supervisory computing system 350 may comprise one or more components of the computer system 500 without departing from the scope of the present technology.

[0139] The computer system 500 comprises various hardware components including one or more single or multi-core processors collectively represented by processor 510, a graphics processing unit (GPU) 511, a solid-state drive 520, a random-access memory 530, a display interface 540, and an input / output interface 550.

[0140] Communication between the various components of the computer system 500 may be enabled by one or more internal and / or external buses 160 (e.g., a PCI bus, universal serial bus, IEEE 1394 “Firewire” bus, SCSI bus, Serial-ATA bus, etc.), to which the various hardware components are electronically coupled.

[0141] The input / output interface 550 may be coupled to a touchscreen 590 and / or to the one or more internal and / or external buses 560. The touchscreen 590 may be part of the display. In some embodiments, the touchscreen 590 is the display. In the embodiments illustrated in FIG. 5, the touchscreen 590 comprises touch hardware 594 (e.g., pressure-sensitive cells embedded in a layer of a display allowing detection of a physical interaction between a user and the display) and a touch input / output controller 592 allowing communication with the display interface 140 and / or the one or more internal and / or external buses 160. In some embodiments, the input / output interface 550 may be connected to a keyboard (not shown), a mouse (not shown) or a trackpad (not shown) allowing the user to interact with the computer system 500 in addition to or instead of the touchscreen 590. In some embodiments, the computer system 500 may comprise one or more microphones (not shown). The microphones may record audio, such as user utterances. The user utterances may be translated to commands for controlling the computer system 500.

[0142] It is noted some components of the computer system 500 can be omitted in some non-limiting embodiments of the present technology. For example, the touchscreen 590 can be omitted.

[0143] According to implementations of the present technology, the solid-state drive 520 stores program instructions suitable for being loaded into the random-access memory 130 and executed by the processor 510 and / or the GPU 511. For example, the program instructions may be part of a library or an application.

[0144] Computer-implemented methods

[0145] Given the architecture and examples provided herein, it is now possible to implement a method 400 for controlling operation of the industrial system 300. With reference to FIG. 4, there is depicted a flowchart diagram of the method 400, in accordance with certain non-limiting embodiments of the present technology, the method 400 can be executed by the processor 510 of the computer system 500. In some embodiments, it is contemplated that the supervisory computing system 350 may be implemented as the computing system 500. STEP 402: RECEIVING TIME-SERIES DATA REPRESENTATIVE OF OPERATIONS OF AT LEAST ONE OF THE PLURALITY OF SUB-COMPONENTS

[0146] The method 400 begins at step 402 with the supervisory computing system 350 configured to receive time-series data representative of operations of at least one of the plurality of subcomponents of an industrial system.

[0147] Broadly, time-series data is any type of information presented as an ordered sequence. It can be represented as a collection of observations for a single subject, assembled over same or different, optionally equally spaced, time intervals. It can be said that data acquired by the supervisory computing system 350 may comprise data collected by one or more sub-components of the industrial system 300, at different points in time and organized chronologically.

[0148] For example, time-series data may comprise time-series data from pressure sensors monitoring pressure changes in pipelines or machinery, time-series data from level sensors tracking fluid levels in tanks or reservoirs, time-series data from temperature sensors recording temperature variations in manufacturing processes, time-series data from pH sensors measuring acidity or alkalinity levels, time-series data from velocity sensors capturing movement speeds in conveyor belts or turbines, and the like, time-series data from one or more sensors configured to measure one or more material properties of a material being processed by the industrial system 300. It is contemplated that time-series data may include one or more sensor readings associated with corresponding timestamps, creating a shared context for the data.

[0149] In some embodiments of the present technology, the supervisory computing system 350 may employ the data acquisition module 302 to collect large amounts of data from a plethora of measurable points along one or more stages / steps of an industrial process executed by the industrial system 300.

[0150] In other embodiments of the present technology, the supervisory computing system 350 may be configured to acquire time-series data of a given type, corresponding to types of time-series data to be processed by one or more MLAs accessible by the supervisory computing system 350.

[0151] In some embodiments, the data acquisition module 302 may be configured to filter acquired / raw time-series data for further processing. The supervisory computing system 350 may collect this raw data, often in the form of analog signals, and digitizes it using analog-to-digital converters (ADCs), for example. The supervisory computing system 350 may apply filtering algorithms to preprocess the data before further analysis. Filtering algorithms can include techniques such as low-pass, high-pass, band-pass, or adaptive filters, depending on the specific characteristics of the acquired data and the desired outcome of the filtering process.

[0152] In additional embodiments, the acquisition module 302 may be configured to store one or more internal tables that identify pre-defined categories of time-series data to be monitored. For example, the internal tables may correspond to respective MLAs selectable for processing the time-series data. Additionally or alternatively, the internal tables may be based on the customer preferences, operator defined priorities, and / or the vendor of the industrial system 300 predetermined parameters, without departing from the scope of the present technology.

[0153] In some embodiments, the industrial system 300 may be implemented as a given injection molding system. It is contemplated that the industrial system may be implemented as one of an abrasive manufacturing system, an automobile electronics manufacturing system, a metal refining system, a heating and air-conditioning (HVAC) equipment manufacturing system, an engine manufacturing system, a lighting fixtures manufacturing system, a gas manufacturing system, a semiconductor machinery manufacturing system, a printing system, an aircraft manufacturing system, a battery manufacturing system, a food processing system, a paper manufacturing system, a fertilizer manufacturing system, and a glass manufacturing system, without departing from the scope of the present technology.

[0154] STEP 404: APPLYING A MACHINE LEARNING ALGORITHM (MLA) CONFIGURED TO PROCESS THE TIME-SERIES DATA TO GENERATE: A ROOT CAUSE OF AN ANOMALY IN OPERATION OF THE AT LEAST ONE OF THE PLURALITY OF SUBCOMPONENTS; AND A REMEDIAL ACTION PARAMETER TO BRING OPERATION OF THE AT LEAST ONE OF THE PLURALITY OF SUB-COMPONENTS TO SPECIFICATION

[0155] The method 400 continues to step 404 with the supervisory computing system 350 configured to apply a machine learning algorithm configured to process the time serries data to generate: a root cause of an anomaly in operation of the at least one of the plurality of sub-components, and a remedial action parameter to bring operation of the at least one of the plurality of sub-components to specification.

[0156] In some embodiments, the supervisory computing system 350 may select a given MLA from a plurality of MLAs to process the time-series data. It is contemplated that the supervisory computing system 350 may be configured to automatically evaluate and compare the performance of multiple MLAs, selecting the most suitable algorithm for the specific task at hand based on predefined criteria and objectives.

[0157] In some embodiments, the supervisory computing system 350 may select a given MLA amongst MLAs stored in the database system 312, such as RNNs, CNNs, GBMs, and other models suitable for time-series data analysis. It should be noted that different architectures may have different advantages which depend on inter alia different types and combinations of industrial time-series data.

[0158] In some embodiments, the selected MLA may be used for determining a root cause of an anomaly in operation of one or more sub-components. In other embodiments, the selected MLA may be used for determining a root cause of an anomaly with one or more properties of one or more materials processed by the industrial system 300. In some embodiments, the selected MLA may be used for determining one or more remedial action parameters to bring operation of the one or more sub-components to specification. In some embodiments, the selected MLA may be used for determining one or more remedial action parameters to bring operation of the one or more subcomponents so as to adjust the one or more properties of the one or more materials to specification.

[0159] In some embodiments, when an anomaly is detected, the supervisory computing system 350 may be configured to trigger alerts, notifications, and / or alarms to notify operators or control systems of potential issues within the industrial system 300. Depending on the severity and criticality of the anomaly, the industrial system 300 may be controlled to initiate corrective actions, such as adjusting operating parameters, shutting down equipment, or activating maintenance protocols to prevent further damage or disruptions. How instructions are generated and communicated will be discussed in greater details below.

[0160] In other embodiments, the supervisory computing system 350 may be configured to execute feedback mechanisms to adapt and refine its anomaly detection capabilities of one or more selected MLAs. By continuously learning from new data and feedback from operators, the supervisory computing system 350 may dynamically adjust anomaly detection thresholds, update its models, and improve its accuracy and reliability in identifying abnormal behavior within the industrial system 350 by generating a root cause of the anomaly and a remedial action to the anomaly. It is contemplated that the remedial action parameter may comprises one or more remedial scenarios to be executed by an operator of the industrial system 300.

[0161] It is contemplated that the MLA may be implemented as a Deep Neural Network. In some embodiments, the MLA may be one of a CNN, RNN, and CRNN. STEP 406: TRANSMITTING INSTRUCTIONS FOR TRIGGERING THE REMEDIAL ACTION ON THE AT LEAST ONE OF THE PLURALITY OF SUB-COMPONENTS OF THE INDUSTRIAL SYSTEM

[0162] The method 400 continues to step 406 with the supervisory computing system 350 configured to transmit instructions for triggering the remedial action on the at least one of the sub-components of the industrial system 300. In one embodiment, the supervisory computing system 350 may be configured to transmit operation control data to one or more sub-components of the industrial system 350 for automatically adjust their operation.

[0163] In other embodiments, the supervisory computing system 350 may be configured to apply a further MLA to generate a human-recognizable instruction representative of the remedial action parameter.

[0164] It should be noted that the supervisory control system 350 may be configured to employ a second MLA to generate human-recognizable instructions for communicating with an operator of the industrial system 300. In one embodiment, the second MLA may be embodied as a LLM. It is contemplated that the second MLA may be embodied as a generative MLA.

[0165] The LLM may be configured to produce human-recognizable instructions based on detected anomalies, their root causes, remedial actions, and / or remedial scenarios to enhance operational efficiency and decision-making processes of the operator. As explained above, the first MLA may detect one or more anomalies by analyzing deviations from normal operating conditions and / or material properties of the processed material. A root cause analysis may be conducted to determine the underlying factors contributing to the detected anomaly, leveraging historical data and sensor readings. Employing natural language generation techniques, the LLM converts the acquired data into coherent instructions, which can include specific steps for troubleshooting, adjusting system parameters, performing maintenance tasks, or initiating corrective actions. These instructions are then delivered to one or more operations through various communication channels, facilitating prompt and effective responses to abnormal events.

[0166] In some embodiments, iterative improvement processes driven by feedback from operators and maintenance personnel can be implemented by the supervisory computer system 350 to continuously ameliorate instruction generation by the second / further MLA, thereby enhancing the quality and relevance of the instructions generated in response to anomalies. It is contemplated that human-recognizable instructions generated by the second MLA make take different forms such as text, audio, video, and / or other visual indicators. In further embodiments, it is contemplated that the supervisory computing system 350 may be configured to transmit the human-recognizable instruction to an operation of the industrial system 300. The supervisory computing system 350 may be configured to transmit data to one or more operator-interface devices of the industrial system 300, including graphical user interfaces (GUIs), display screens, audio systems, and notification mechanisms. This allows the interfacing module 310 to deliver instructions through multiple channels, accommodating different preferences and operational requirements.

[0167] In some embodiments, for visual output, the supervisory computing system 350 can transmit instructions to displayed on display screens strategically placed within the industrial environment, such as control rooms, workstations, or production floors. Instructions may be displayed in a clear, legible format, accompanied by graphical elements, icons, or color-coded indicators to enhance comprehension and usability.

[0168] In other embodiments, the supervisory computing system 350 can leverage audio output capabilities of the industrial system 300, such as speakers or PA systems, to deliver spoken instructions or alerts to operators in real-time. This auditory feedback provides an additional layer of situational awareness, especially in noisy or high-activity environments where visual cues may be less effective.

[0169] In further embodiments, the supervisory computing system 350 can utilize notification mechanisms, such as email alerts, SMS messages, or mobile app notifications, to reach operators remotely or on-the-go. This ensures that operators receive critical instructions and updates wherever they are, enabling prompt response and action.

[0170] In additional embodiments, the supervisory computing system 350 may enable interactive communication, such as via touchscreens or voice-activated interfaces of the industrial system 300, to allow operators to acknowledge receipt of instructions, provide feedback, or initiate followup actions directly from the interface.

[0171] Modifications and improvements to the above-described embodiments of the present technology may become apparent to those skilled in the art. The foregoing description is intended to be exemplary rather than limiting. The scope of the present technology is therefore intended to be limited solely by the scope of the appended claims.

[0172] The description of the embodiments of the present technology provides only examples of the present technology, and these examples do not limit the scope of the present technology. It is to be expressly understood that the scope of the present technology is limited by the claims only. The concepts described above may be adapted for specific conditions and / or functions and may be further extended to a variety of other applications that are within the scope of the present technology. Having thus described the embodiments of the present technology, it will be apparent that modifications and enhancements are possible without departing from the concepts as described.

Claims

CLAIMS1. A method of controlling an industrial system, the industrial system including a plurality of sub-components, the method executable by a supervisory computing system associated with the industrial system, the method comprising: receiving time-series data representative of operations of at least one of the plurality of sub-components; applying a machine learning algorithm (MLA) configured to process the time-series data to generate: a root cause of an anomaly in operation of the at least one of the plurality of sub-components; and a remedial action parameter to bring operation of the at least one of the plurality of sub-components to specification; and transmitting instructions for triggering the remedial action on the at least one of the plurality of sub-components.

2. The method of claim 1, wherein the method further comprises applying a further MLA to generate a human-recognizable instruction representative of the remedial action parameter, and the transmitting further comprises transmitting the human-recognizable instruction to an operator of the industrial system.

3. The method of claim 1, wherein the MLA being implemented as a Deep Neural Network.

4. The method of claim 1, wherein the MLA is one of a Convolutional Neural Network, Recurrent Neural Network, and a Convolutional Recurrent Neural Network.

5. The method of claim 1, wherein the remedial action parameter comprises a plurality of remedial scenarios.

6. The method of claim 2, wherein the further MLA is implemented as a generative MLA.

7. The method of claim 2, wherein the further MLA is implemented as a Large Language Model (LLM).

8. The method of claim 1, wherein the industrial system is an injection molding system.

9. The method of claim 1, wherein the industrial system is one of: an abrasive manufacturing system, an automobile electronics manufacturing system, a metal refining system, a heating and air-conditioning (HVAC) equipment manufacturing system, an engine manufacturing system, a lighting fixtures manufacturing system, a gas manufacturing system, a semiconductor machinery manufacturing system, a printing system, an aircraft manufacturing system, a battery manufacturing system, a food processing system, a paper manufacturing system, a fertilizer manufacturing system, and a glass manufacturing system.

10. The method of claim 1, wherein the method further comprises, prior to processing by the MLA, applying one or more filtering algorithms onto the received time-series data.

11. The method of claim 1, wherein the method comprises selecting the MLA amongst a plurality of MLAs, the selecting being based on a type of the anomaly to be detected.

12. A computer system for controlling an industrial system, the industrial system including a plurality of sub-components, the computer system associated with the industrial system, the computer system being configured to: receive time-series data representative of operations of at least one of the plurality of sub-components; apply a machine learning algorithm (MLA) configured to process the time-series data to generate: a root cause of an anomaly in operation of the at least one of the plurality of sub-components; and a remedial action parameter to bring operation of the at least one of the plurality of sub-components to specification; and transmit instructions for triggering the remedial action on the at least one of the plurality of sub-components.

13. The computer system of claim 12, wherein the computer system is further configured to apply a further MLA to generate a human-recognizable instruction representative of the remedial action parameter, and to transmitting further comprises the computer system to transmit the human-recognizable instruction to an operator of the industrial system.

14. The computer system of claim 12, wherein the MLA being implemented as a Deep Neural Network.

15. The computer system of claim 12, wherein the MLA is one of a Convolutional Neural Network, Recurrent Neural Network, and a Convolutional Recurrent Neural Network.

16. The computer system of claim 12, wherein the remedial action parameter comprises a plurality of remedial scenarios.

17. The computer system of claim 13, wherein the further MLA is implemented as a generative MLA.

18. The computer system of claim 13, wherein the further MLA is implemented as a Large Language Model (LLM).

19. The computer system of claim 12, wherein the industrial system is an injection molding system.

20. The computer system of claim 12, wherein the industrial system is one of: an abrasive manufacturing system, an automobile electronics manufacturing system, a metal refining system, a heating and air-conditioning (HVAC) equipment manufacturing system, an engine manufacturing system, a lighting fixtures manufacturing system, a gas manufacturing system, a semiconductor machinery manufacturing system, a printing system, an aircraft manufacturing system, a battery manufacturing system, a food processing system, a paper manufacturing system, a fertilizer manufacturing system, and a glass manufacturing system.

21. The computer system of claim 12, wherein the computer system is further configured to, prior to processing by the MLA, apply one or more filtering algorithms onto the received time-series data.

2. The computer system of claim 12, wherein computer system is further configured to select the MLA amongst a plurality of MLAs, to select being based on a type of the anomaly to be detected.

Images