Method and apparatus for implementing predictive analysis for a continuous process

Abstract

A method and apparatus for implementing predictive analytics for a continuous process is disclosed. An exemplary apparatus includes a virtual batch unit controller for implementing a sampled batch on a virtual batch unit. The sampled batch corresponds to a discrete time period of a continuous control system process. The virtual batch unit includes input and output parameters corresponding to parameters associated with the continuous control system process. The exemplary apparatus also includes a sampled batch analyzer for generating predictive analytics information indicating a predicted quality of an output of the continuous control system process at the end of the discrete time period based on an analysis of the sampled batch relative to an analytics model.

Description

Technical Field

[0001] The present disclosure relates generally to process control systems and, more particularly, to methods and apparatus for implementing predictive analytics for continuous processes. Background Art

[0002] Process control systems can be implemented as either batch process control or continuous process control. Batch process control involves the processing of specific batches, which are associated with a specific amount of material that is processed over a specific duration to produce a final output product. Thus, a batch process has a defined start and end (discrete time period) corresponding to the time required to process the material fed into the system at the beginning and produce the final output product. In contrast, continuous process control involves processing materials to produce an output product continuously. Thus, the duration of a continuous process can extend to any specified period of time (theoretically indefinitely), and as the duration increases, the amount of output product produced will continue to increase. Summary of the invention

[0003] An exemplary method and apparatus for implementing predictive analytics for a continuous process are disclosed. The exemplary apparatus includes a virtual batch unit controller for implementing a sampling batch on a virtual batch unit. The sampling batch corresponds to a discrete time period of a continuous control system process. The virtual batch unit includes input and output parameters corresponding to parameters associated with the continuous control system process. The exemplary apparatus also includes a sampling batch analyzer for generating predictive analytics information indicating a predicted quality of an output of the continuous control system process at the end of the discrete time period based on an analysis of the sampling batch relative to an analytics model.

[0004] An exemplary non-transitory computer readable medium includes instructions that, when executed, cause a machine to implement at least a sample batch on a virtual batch unit. The sample batch corresponds to a discrete time period of a continuous control system process. The virtual batch unit includes input and output parameters corresponding to parameters associated with the continuous control system process. The exemplary instructions further cause the machine to generate predictive analysis information indicating a predicted quality of an output of the continuous control system process at the end of the discrete time period based on an analysis of the sample batch relative to an analysis model.

[0005] An exemplary method includes implementing a sample batch on a virtual batch unit. The sample batch corresponds to a discrete time period of a continuous control system process. The virtual batch unit includes input and output parameters corresponding to parameters associated with the continuous control system process. The exemplary method also includes generating predictive analysis information indicating a predicted quality of an output of the continuous control system process at an end of the discrete time period based on an analysis of the sample batch relative to an analytical model. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 is a schematic diagram of an exemplary process control system.

[0007] Figure 2 Demonstrates the implementation Figure 1 An exemplary approach to an exemplary predictive analytics engine.

[0008] Figure 3 Included are graphs illustrating exemplary quality parameters of historical process data associated with an exemplary continuous control system process and dynamic behavior of process parameters.

[0009] Figure 4 yes Figure 3 An enlarged view of an exemplary graph of FIG.

[0010] Figure 5 is a schematic diagram of an exemplary virtual batch unit implemented according to the teachings disclosed herein.

[0011] Figure 6 is a graph illustrating the dynamic behavior of an exemplary process parameter of an exemplary continuous control system process in operation.

[0012] Figure 7 is an exemplary quality prediction interface that may be presented in accordance with the teachings disclosed herein.

[0013] Figure 8-Figure 10 It means that it can be executed to achieve Figure 1 and / or Figure 2 Flowchart of an exemplary method of an exemplary predictive analysis engine.

[0014] Fig.11 is a schematic diagram of an exemplary processor platform that may be used and / or programmed to perform Figure 8-Figure 10 An example method of, and / or more generally, implementing Figure 1 and / or Figure 2 An exemplary predictive analytics engine for .

[0015] The drawings are not drawn to scale. Generally, the same reference numbers will be used throughout the drawings and accompanying written description to refer to the same or like parts.

[0016] When identifying multiple elements or components that can be mentioned separately, descriptors "first", "second", "third", etc. are used in this article. Unless otherwise specified or understood based on the context of use, such descriptors are not intended to give any meaning of priority, physical order or arrangement in a list, or chronological order, but are only used as labels for individually quoting multiple elements or components to facilitate understanding of the disclosed examples. In some examples, the descriptor "first" can be used to refer to an element in the detailed description, while different descriptors such as "second" or "third" can be used in the claims to refer to the same element. In such a case, it should be understood that such descriptors are used only for ease of quoting multiple elements or components. DETAILED DESCRIPTION

[0017] Typically, batch control analysis is implemented during execution of a batch process to improve and / or maintain the safety, performance, and / or efficiency of the batch process by enabling fault detection and quality prediction of the process in substantially real time. More particularly, in some cases, batch control analysis involves multivariate analysis of the batch process relative to a model of the batch process generated from previously executed (e.g., historically archived) batch processes. Common tools used to create statistical models for batch analysis include principal component analysis (PCA) and latent structure projection (PLS) (also known as partial least squares).

[0018] While batch analysis enables fault detection and quality prediction of output products during batch processes, similar analysis is not available for continuous process control. Among other things, PCA and PLS (used in batch analysis) cannot account for the dynamic behavior of a continuous system that deviates from a steady state and then returns to a steady state. In particular, in a continuous process system, deviations from a steady state are detected and corrected based on changes in process parameters over time. Therefore, the response of a continuous process to a deviation is instantaneous or time-based. However, PCA and PLS do not contain a time dimension and are therefore not suitable for direct application to continuous steady-state process control systems. A potential approach to providing predictive analysis to a continuous process control system is to develop analytical models based on the physical mechanism of the process and time-dependent analysis based on, for example, differential equations. However, due to the challenges of training such models, this approach is unlikely to produce reliable models due to the nonlinearity and sensitivity of the dynamic behavior in such differential equations.

[0019] The examples disclosed herein enable predictive analysis of continuous process control systems using batch-like analysis techniques. More specifically, the examples disclosed herein specify continuous and discrete time portions of a continuous process, which can be simulated as a series or campaign of a continuous batch process. Each discrete time portion of a continuous process is referred to herein as a sampling batch to distinguish it from a typical batch associated with an actual batch process. According to the teachings disclosed herein, by dividing a continuous process into a plurality of discrete segments, the discrete segments can be considered as separate batches, thereby enabling the application of batch-like analysis techniques to generate predictive analysis information for a continuous process.

[0020] The International Society of Automation (ISA) adopted a framework for standard batch process control in 1995, called ISA-88. The ISA-88 standard defines a procedural control framework for batch processes in the context of a recipe that includes one or more unit procedures, which in turn may include an ordered sequence of operations, which in turn may include an ordered set of phases. Standard continuous process control systems are not subject to the ISA-88 standard and are therefore not typically defined in terms of procedural units, operations, and phases. However, in some examples disclosed herein, virtual batch units are generated for continuous process systems to run in parallel with the continuous process to provide predictive analysis of the process. In other words, in some examples, a continuous process system is controlled using standard continuous process control techniques, while a virtual batch unit that reflects the continuous process but is structured to mimic a typical batch process is implemented in parallel to run a sampled batch of the continuous system (corresponding to a discrete time portion of the process described above) so as to provide predictive analysis in a manner similar to batch analysis.

[0021] More specifically, in some examples, a simulated batch recipe can be defined to run continuous sampling batches on a virtual batch unit, thereby emulating the activities of a continuous batch (consistent with ISA-88) while also reflecting an actual continuous process. The implementation of the virtual batch process unit can reflect an actual continuous process control system because the virtual batch unit is defined to include input and output parameters corresponding to the relevant inputs, outputs, and / or process parameters of the continuous process to be monitored and analyzed in conjunction with each sampling batch. In addition, in some examples, the virtual batch unit is defined to include initial conditions (separate from the input parameters) that correspond to the values ​​of the process parameters when a new sampling batch is started.

[0022] By breaking down the continuous process into discrete parts implemented in conjunction with the virtual batch units defined above, batch-like analysis techniques can be used to analyze the individual parts (e.g., sampling batches) to provide fault detection and / or quality prediction of the current individual parts of the continuous process being analyzed. In addition, by repeating the analysis for each continuous sampling batch (e.g., discrete part) defined in the continuous process, continuous updating of the predictive analysis can be provided over the course of the entire continuous process.

[0023] In some examples, the duration of each sampling batch of a continuous process is set to a fixed period. In such examples, the analysis of the continuous process is performed without dynamic time warping (DTW), as is usually achieved with conventional batch analysis. In some examples, the fixed period or duration defined for the sampling batch corresponds to the residence time of the continuous process system. As used herein, the residence time of a process system (also referred to as the retention time) refers to the duration of time that a material is processed (e.g., resides therein) in the system. In other words, the residence time corresponds to the duration between when a particular material is initially introduced into the process as an input and when the particular material has been processed into a final output product. Typically, due to the reflux and / or mixing of materials, the residence time of a continuous process control system is not constant for materials introduced into the system at different time points. Therefore, the residence time can be represented as a distribution in which some materials stay longer in the process system than other materials. Therefore, in some examples, the residence time for defining the length of each sampling batch is approximated based on the average residence time of the continuous processing system. In some examples, the duration of each sampling batch is equal to the approximate value of the residence time. In other examples, the duration of each sampling batch can be greater than the residence time (e.g., up to 4 times the approximate value of the residence time). A longer duration of a sampling batch can facilitate the capture of dynamic process behavior in a process control system. On the other hand, a longer duration of a sampling batch can result in less frequent predictive analytics information.

[0024] As described above, conventional batch analysis (based on PCA and PLS) cannot be directly applied to continuous steady-state process systems because process behavior is dynamic (e.g., the deviation of the system and its correction are time-dependent). The examples disclosed herein illustrate dynamic (time-based) changes in process behavior in continuous process systems by developing analytical models using sampled batches (e.g., discrete portions) of historical process data of an associated continuous process as training data, wherein the historical sampled batches of the historical process data are selected to correspond to specific times when the process undergoes dynamic changes. In standard batch analysis, an analytical model is generated based on training data corresponding to historical process data associated with multiple batches. In other words, all data from the beginning to the end of the training batch is used to generate the model. In contrast, only isolated portions of the historical process data from the continuous process are used to define historical sampled batches, which are used as the basis for generating analytical models that can analyze sampled batches that are being implemented in real time. In addition, in some examples, the historical process data used to define one historical sampled batch may overlap in time with the historical process data used to define a second historical sampled batch.

[0025] For example, assume that a continuous processing system is currently in a non-steady state, such as during system startup. During startup, before the system reaches a steady state, process parameters and / or quality parameters (measured at the output) may change significantly. Using a single historical sampling batch corresponding to the historical process data associated with the entire startup period will not be able to account for the dynamic behavior that occurs during the period. Therefore, in some examples, multiple historical sampling batches can be extracted from the historical data corresponding to the startup period of a single continuous process, wherein the start time and end time of each historical sampling batch are slightly offset from each other. As a result, different historical sampling batches cover overlapping time periods. In some examples, the time offset of continuous historical sampling batches is defined to provide sufficient granularity to accurately capture the dynamic behavior of process parameters and / or quality parameters in the system. Therefore, the amount of overlapping data between different historical sampling batches depends on the amount of dynamic behavior (e.g., volatility) in the process parameters and / or quality parameters at the time point of interest. Typically, the startup period and shutdown period of a continuous process exhibit greater dynamic behavior than during the steady-state period. Therefore, in some examples, the historical sampling batches extracted from the historical process data of a continuous process overlap more densely around the startup and shutdown periods than the historical sampling batches extracted at the time corresponding to the system steady-state period. In some examples, the historical sampling batches corresponding to steady-state periods may not overlap at all. Furthermore, while the above examples are described with respect to historical data associated with a single continuous process having a single startup and a single shutdown, historical sampling batches may be extracted from multiple similar continuous processes and / or during multiple startups and shutdowns and associated steady-state periods therebetween.

[0026] Typically, continuous process control systems operate in steady state far more often than in startup or shutdown periods. Therefore, in some examples, historical sampling batches corresponding to steady state periods are used for model generation in greater numbers than historical sampling batches corresponding to startup or shutdown periods. However, a complete historical sampling batch training set includes at least some sampling batches corresponding to startup periods, some batches corresponding to steady state periods, and some batches corresponding to shutdown periods. Using the identified training set of historical sampling batches, an analysis model can be generated or trained based on the historical sampling batches in the same manner as a traditional batch analysis model is generated from historical batch data. Using the generated analysis model, a real-time sampling batch corresponding to the current continuous process can be analyzed to generate predictive analysis information indicating fault detection and / or quality prediction of the output product of the process.

[0027] Figure 1 is a schematic diagram of an example process control system 100 that can be implemented according to the teachings disclosed herein. In this example, the process control system 100 is implementing a continuous process. Figure 1 The exemplary process control system 100 includes one or more process controllers, one of which is indicated by reference numeral 102, one or more operator stations, one of which is indicated by reference numeral 104, and one or more workstations, one of which is indicated by reference numeral 106. The exemplary process controller 102, the exemplary operator station 104, and the exemplary workstation 106 are communicatively coupled via a bus and / or a local area network (LAN) 108, which is generally referred to as an application control network (ACN).

[0028] Figure 1 The example operator station 104 allows an operator to view and / or operate one or more operator display screens and / or applications that enable the operator to view process control system parameters, states, conditions, alarms, etc., and / or change process control system settings (e.g., set points, operating states, etc.). The example operator station 104 includes and / or implements an example predictive analysis engine 105 to generate predictive analysis for continuous processes being implemented within the process control system 100. Figure 2 Describe the implementation Figure 1 An exemplary manner of an exemplary predictive analysis engine 105.

[0029] In some examples, the predictive analysis engine 105 executes a virtual batch unit associated with the simulated recipe overview program to implement a specific sampled batch of a continuous process. As described above and below, a sampled batch is a discrete time portion of a continuous process that is implemented on a virtual batch unit to be analyzed independently of the actual control of the continuous process. In some examples, once the current sampled batch process ends, the predictive analysis engine 105 opens a new sampled batch on the virtual batch unit so that the continuous sampled batch forms a chain or activity of continuous batches corresponding to and running in parallel with the continuous process. During the execution of the virtual batch unit for each sampled batch, the predictive analysis engine 105 monitors the process parameters and / or other inputs and outputs associated with the current sampled batch substantially in real time. In addition, the exemplary predictive analysis engine 105 applies the analysis model to the current sampled batch substantially in real time to detect faults and / or generate a prediction related to the quality of the output of the continuous process at a future time period corresponding to the time when the current sampled batch will end. By performing this analysis on continuous sampled batches, the predictive analysis engine 105 can provide substantially real-time fault detection and quality prediction for the continuous process in a continuous manner. Such predictive analysis information may increase the operator's understanding of the current state of the process control system 100 and expected changes to the current state, thereby enabling the operator to respond more quickly and effectively to unexpected deviations in the process.

[0030] Figure 1 The exemplary workstation 106 can be configured as an application station to execute one or more information technology applications, user interaction applications, and / or communication applications. For example, the application station 106 can be configured to primarily execute applications related to process control, while another application station (not shown) can be configured to primarily execute communication applications that enable the process control system 100 to communicate with other devices or systems using any desired communication medium (e.g., wireless, hardwired, etc.) and protocol (e.g., HTTP, SOAP, etc.). One or more workstations and / or any other suitable computer systems and / or processing systems can be used to implement Figure 1 1. An exemplary operator station 104 and an exemplary workstation 106 are shown. For example, the operator station 104 and / or the workstation 106 may be implemented using a single-processor personal computer, a single-processor or multi-processor workstation, or the like.

[0031] Any desired communication medium and protocol may be used to implement Figure 1 108. For example, the LAN 108 may be based on a hardwired and / or wireless Ethernet communication scheme. However, any other suitable communication medium and / or protocol may be used. In addition, although Figure 1 A single LAN 108 is illustrated in FIG. 1 , however, Figure 1As shown, more than one LAN and / or other alternative communication hardware may be used in Figure 1 Provide redundant communication paths between exemplary systems.

[0032] Figure 1 The exemplary controller 102 of FIG. 1 is coupled to a plurality of smart field devices 110, 112, and 114 via a digital data bus 116 and an input / output (I / O) gateway 118. The smart field devices 110, 112, and 114 may be fieldbus-compatible valves, actuators, sensors, and the like, in which case the smart field devices 110, 112, and 114 communicate via the digital data bus 116 using the well-known Foundation fieldbus protocol. Of course, other types of smart field devices and communication protocols may be used instead. For example, the smart field devices 110, 112, and 114 may alternatively be Profibus and / or HART-compatible devices that communicate via the data bus 116 using the well-known Profibus and HART communication protocols. Additional I / O devices (similar to and / or identical to the I / O gateway 118) may be coupled to the controller 102 to enable additional groups of smart field devices (which may be Foundation fieldbus devices, HART devices, and the like) to communicate with the controller 102.

[0033] In addition to the example smart field devices 110 , 112 , and 114 , one or more non-smart field devices 120 and 122 may be communicatively coupled to the example controller 102 . Figure 1 The exemplary non-intelligent field devices 120 and 122 in can be, for example, conventional 4-20 milliamp (mA) or 0-10 volt direct current (VDC) devices that communicate with the controller 102 via corresponding hardwired links.

[0034] Figure 1 An exemplary controller 102 may be, for example, a DeltaV sold by Fisher-Rosemount Systems, Inc., an Emerson Process Management company. TM controller. However, any other controller may be used instead. In addition, although Figure 1 Only one controller 102 is shown in the example, but additional controllers and / or process control platforms of any desired type and / or combination of types may be coupled to the LAN 108. In any case, the example controller 102 executes one or more process control routines associated with the process control system 100 that have been generated by a system engineer and / or other system operator using an operator station 104 and downloaded to and / or instantiated in the controller 102.

[0035] although Figure 1 An exemplary process control system 100 is illustrated in which the methods and apparatus for performing predictive analysis on a continuous process control system described in more detail below may be advantageously employed, but one of ordinary skill in the art will readily appreciate that the methods and apparatus for controlling information presented to an operator and / or other personnel described herein may be advantageously used with other systems, if desired. Figure 1 Examples of the invention are compared to other process plants and / or process control systems having higher or lower complexity (eg, having more than one controller, spanning more than one geographic location, etc.).

[0036] Figure 2 Demonstrates the implementation Figure 1 An exemplary manner of an exemplary predictive analysis engine 105. Figure 2 The exemplary predictive analysis engine 105 includes an exemplary communication interface 202, an exemplary residence time analyzer 204, an exemplary historical sampling batch generator 206, an exemplary batch model generator 208, an exemplary virtual batch unit controller 210, an exemplary sampling batch analyzer 212, an exemplary database 214, and an exemplary user interface 216.

[0037] Figure 2 The exemplary predictive analysis engine 105 includes an exemplary communication interface 202 for communicating with other components in the exemplary process control system 100. In some examples, the communication interface 202 receives process control data (indicative of values ​​of process parameters) from the field devices 110, 112, 114, 120, 122 via the controller 102 in substantially real time. In some examples, such process control data is sent to a continuous historian database 218 for storage. In some such examples, the communication interface 202 obtains historical process data from the continuous historian database 218. In some examples, the continuous historian database 218 is implemented by the exemplary database 214 of the predictive analysis engine 105. In other examples (as shown in the illustrated example), the continuous historian database 218 is implemented separately from the predictive analysis engine 105 and / or separately from the operator station 104.

[0038] Figure 2 The exemplary predictive analysis engine 105 of the embodiment includes an exemplary residence time analyzer 204 for analyzing historical process data associated with a continuous control system process to approximate the residence time of the continuous process. As described above, the residence time corresponds to the duration of time that a material spends passing through or being processed by a continuous system. In some examples, the residence time analyzer 204 is based on the following examples: Figure 3The residence time is approximated by the duration between the time of first activation of the continuous system shown and the earliest response of the system at the outlet.

[0039] Figure 3 A first graph 300 showing changes in quality parameters 302, 304, 306 at the outlet of an exemplary continuous system over time and a second graph 308 showing changes in process parameters and / or other input parameters 310, 312, 314 (collectively referred to herein as process parameters for brevity) of the continuous system over the same time period. The quality parameters 302, 304, 306 may or may not correspond to the process parameters 310, 312, 314. Furthermore, although three quality parameters and three process parameters are shown for purposes of illustration, there may be any number of quality parameters and any number of process parameters.

[0040] In the illustrated example, the first graph 300 and the second graph 308 are aligned in time with a common time scale to enable comparison of the quality parameters 302, 304, 306 and the process parameters 310, 312, 314 at any given point in time during the process. As shown in the illustrated example, the continuous control system process can be divided into three general stages or periods, including a startup period 316, a steady-state period 318, and a shutdown period 320. The steady-state period 318 corresponds to when the continuous process has reached a substantially steady state, wherein the process parameters 310, 312, 314 have reached and remained at desired values ​​(e.g., set points) and / or within acceptable thresholds of such values, and the quality parameters 302, 304, 306 are similarly maintained at a substantially steady state (e.g., within acceptable thresholds of desired values). As the names imply, the startup period 316 and the shutdown period 320 correspond to time periods when the process first starts before the steady-state period 318 and when the process is shutting down after the steady-state period 318.

[0041] For a continuous control system, the steady-state period 318 may be a significantly longer period (e.g., multiple orders of magnitude) than the startup period 316 or the shutdown period 320. Figure 3In the example shown, the graphs 300, 308 include interruptions 322 within the steady-state period 318 to shorten the length of the period shown, which can be extended for any duration up to time S. Based on the time measurements shown in the example shown, the startup period 316 ends and the steady-state period 318 begins approximately 30 minutes after the process is first started. Although the process parameters 310, 312, 314 reached a steady state 20 minutes ago, the steady-state period 318 does not begin until 30 minutes because it takes that long for the second quality parameter 304 to reach a steady state. There is a lag in the response of the quality parameters 302, 304, 306 relative to the process parameters 310, 312, 314 because the quality parameters are measured on the output side of the process, while the process parameters (at least the inlet parameters) are measured on the input side of the process. This lag is the basis for the example residence time analyzer 204 to be able to approximate the residence time of a continuous process. More specifically, in some examples, the residence time analyzer 204 determines an approximate residence time for the process based on the duration between when the continuous system is first started (at time 0) and the earliest response of the system on the outlet side. Figure 3 In the example shown in FIG. 1 , the dwell time (identified by reference numeral 324) is approximately 10 minutes because that is when the first quality parameter 302 and the third quality parameter 306 first begin to respond. In this example, the fact that the second quality parameter 304 does not begin to respond until several minutes later is irrelevant to the determination of the dwell time 324. In some examples, the dwell time may be determined based on the earliest response of a particular quality parameter, regardless of whether other parameters may have responded before the particular parameter. For example, if the second quality parameter 304 were to be used, the dwell time would be determined to be closer to 13 minutes because that is when the second quality parameter 304 began to respond. Figure 3 The time when the second quality parameter first starts to change in the example shown in FIG. For the purpose of illustration, the residence time of the example shown is assumed to be 10 minutes.

[0042] return Figure 2 The exemplary predictive analysis engine 105 includes a historical sampling batch generator 206 for generating a historical sampling batch for generating or training an analysis model for generating predictive analysis information for a continuous process. As used herein, a historical sampling batch includes historical process data associated with a continuous control system process over a discrete time period (e.g., a specified start and end separated by a fixed duration). The historical process data corresponds to a process parameter of the control system, such as Figure 3 In some examples, the duration or length of the historical sampling batch is defined based on an approximation of the residence time 324 determined by the residence time analyzer 204. In other words, in some examples, the historical sampling batch is defined using a duration corresponding to the residence time of the continuous process system.

[0043] In some examples, the historical sample batch generator 206 generates a sample based on the value of the quality parameter at a particular time point of interest (in Figure 3 Specific start and end times of different historical sampling batches are determined using the first graph 300 of FIG. 302 . More specifically, as described above, the quality parameters 302 , 304 , 306 are measures of the quality of the output of the continuous process system. Therefore, in some examples, the quality parameter at a specific point in time is identified to specify the end time of the historical sampling batch. In addition, the value of the quality parameter at a specific quality reading point is used as the result or output of the associated historical sampling batch that ends at the same point in time. The start of a historical sampling batch is determined by backtracking in time by a fixed length of the historical sampling batch (corresponding to the residence time as described above). As a specific example, Figure 3 A quality reading point 326 selected in the first graph 300 is shown at approximately 22.5 minutes into the process. Following the arrow downward to the second graph 308, the quality reading point 326 defines an end time 328 for a historical sampling batch that includes historical process data (e.g., values ​​of process parameters 310, 312, 314) during the period framed by box 330. The duration (e.g., width) of box 330 corresponds to the dwell time 324 (e.g., 10 minutes in this example), thereby defining a start time 332 for the historical sampling batch at approximately 12.5 minutes into the continuous process. The example historical sampling batch generator 206 generates a historical sampling batch for the time period identified by box 330 by acquiring historical process control data (e.g., values ​​of process parameters 310, 312, 314) during the period bounded by box 330. The example historical sampling batch generator 206 also sets initial conditions for the historical sampling batch. In some examples, the initial conditions correspond to the values ​​of the process parameters 310 , 312 , 314 at the start time 332 of the historical sampling batch.

[0044] In some examples, the number of historical sampling batches generated by the historical sampling batch generator 206 and the time intervals between these historical sampling batches are based on the degree to which the historical process data and / or quality parameters change dynamically over time. Figure 3As shown in the first graph 300 of FIG. 300 , the first quality parameter 302 drops sharply shortly after a residence time 324 (10 minutes) and reaches a steady state before 15 minutes into the process. Therefore, between 10 minutes and 15 minutes into the process, the dynamic changes of the first quality parameter 302 are significant. The second quality parameter 304 also begins to decline shortly after the first residence time 324, but the decline is gentler than the first quality parameter 302 and takes a longer period of time. The third quality parameter 306 also begins to change around 10 minutes (a residence time) and fluctuates greatly in a period of time before reaching a steady state around 25 minutes into the process. The relatively gentle slope of the second quality parameter 304 causes the second quality parameter 304 to take a longer time to reach a steady state than the first quality parameter 302 or the third quality parameter 306.

[0045] In order to properly capture the relatively high degree of dynamic change in the first quality parameter 302 and the third quality parameter 306 during the startup period 316, the example historical sampling batch generator 206 may select or identify a plurality of quality reading points spaced at relatively small time intervals. Figure 3 25 minutes after the exemplary process represented by the graphs 300, 308 in FIG. 25, when the first quality parameter 302 and the third quality parameter 306 reach a steady state, the exemplary historical sampling batch generator 206 may select or identify additional quality reading points that are spaced at larger intervals because the degree of dynamic change in the second quality parameter 304 is less. After all three quality parameters 302, 304, 306 have reached a steady state (e.g., when the process enters a steady state period 318), the intervals between the quality reading points may be spaced further apart. Figure 4 This is demonstrated graphically in the example shown, which contains Figure 3 Magnified portions of graphs 300, 308.

[0046] exist Figure 4 In the example shown in FIG. 4 , 20 different quality reading points 402 are identified and labeled from Q1 to Q20. Figure 4As shown, the first quality reading point 402 (Q1) is located at the 10 minute mark corresponding to one dwell time 324. The first quality reading point 402 (Q1) is set at one dwell time so that the first portion of the historical process control data (identified by box 404) can define a complete historical sampling batch, which has a fixed length corresponding to the dwell time 324, as described above and shown in the example shown. In the example shown, the next 15 quality reading points 402 (Q2-Q16) are spaced at relatively small intervals to capture the dynamic changes of the quality parameters 302, 304, 306 from the 10 minute mark to the 25 minute mark (when the third quality parameter 306 reaches a steady state). For clarity, the first to sixteenth quality reading points 402 (Q1-Q16) are spaced at one minute intervals. However, in some examples, due to the relatively high degree of dynamic changes in the quality parameters during this time period, the time intervals of the quality reading points 402 may be greatly reduced (e.g., 1 second, 5 seconds, 10 seconds, 15 seconds, etc.). Alternatively, in the case where the dynamic changes in the quality parameters are less, the time intervals of the quality reading points 402 may be increased. For example, as described above, the slope of the second quality parameter 304 changes more gradually than the first parameter 302 and the third quality parameter 306. As a result, in the example shown, the quality reading points 402 starting at 25 minutes (Q16) and extending to the end of the startup period 316 at 30 minutes (Q18) are set at intervals of 2.5 minutes. After entering the steady-state period 318, since there are substantially no dynamic changes in the quality parameters 302, 304, 306, the quality reading points 402 are spaced further apart (e.g., every five minutes in the example shown).

[0047] Although the time intervals of the quality reading points 402 have been described as being based on the dynamic changes of the quality parameters 302, 304, 306, in some examples, the time intervals may be additionally or alternatively based on the dynamic changes of the process parameters 310, 312, 314. More specifically, in some examples, the time intervals of the quality reading points 402 are based on which parameter exhibits the greatest degree of dynamic change at the relevant time point. In addition, although the time intervals of the quality reading points 402 are shown at relatively fixed intervals, in some examples, different quality reading points 402 may be located at irregular intervals suitable for the variability of the parameters at the corresponding time points. In some examples, the timing of the quality reading points 402 is randomly selected within a suitable range for the entire duration of the continuous process. For example, in some examples, the historical sampling batch generator 206 may determine the appropriate number of quality reading points required to adequately capture the dynamic behavior of the parameters in the startup period 316, and then randomly identify a specific time within the startup period 316 (starting at a dwell time) for the identified number of quality reading points. This random selection process may be repeated for the steady-state period 318 and the shutdown period 320.

[0048] like Figure 4 In the example shown, although the time intervals between adjacent quality reading points 402 vary according to the degree of dynamic changes associated with the quality parameters 302, 304, 306, the intervals between the quality reading points 402 are less than the residence time 324. As a result, different historical sampling batches generated for different quality reading points 402 contain overlapping portions of the historical process data. For example, as described above, the first portion of the historical process data 404 corresponds to the value of the process parameter from time 0 to 10 minutes. In the example shown, the second portion of the historical process data 406 is shown, which corresponds to the ninth quality reading point 402 (Q9) set at 18 minutes into the process. When the quality reading point 402 defines the end time of the associated historical sampling batch, the start time is determined by backtracking in time according to the residence time 324 (10 minutes in this example). Therefore, the second portion of the historical process data 406 in the example shown begins at 8 minutes into the process, resulting in a two-minute overlap with the first portion of the historical process data 404. The third portion of the historical process data 408 is shown extending from 15 minutes into the process to 25 minutes, resulting in a three minute overlap with the second portion of the historical process data 406. The fourth portion of the historical process data 410 is shown extending from 30 minutes into the process to 40 minutes, so as not to overlap with the second portion of the historical process data 406. Figure 4The first three portions of the historical process data are shown to overlap. However, additional portions of the historical process data associated with other quality reading points 402 acquired during the corresponding time period will cause each portion of the historical process data to overlap with at least one other portion of the data. Although the portions of the historical process data overlap, the time shift or time interval between adjacent portions of the data results in different values ​​of the quality parameters and process parameters at the corresponding quality reading points 402, as well as differences in the process parameters at the beginning of each portion of the data (which serves as the initial condition for each associated historical sampling batch).

[0049] In some examples, the time interval of the quality reading points 402 may be equal to or greater than the dwell time when there is substantially no dynamic change in the quality parameter and / or process parameter during the time period corresponding to the dwell time 324 (e.g., during the steady-state period 318). In this case, the portions of the historical process data of different historical sampling batches of the corresponding historical sampling batches do not overlap, and as described above, the duration of the historical sampling batch generated based on the portions of the historical process data is equal to the dwell time. Although the above examples are described using historical sampling batches assumed to have a duration corresponding to the dwell time, in other examples, the historical sampling batches may be defined as having a duration greater than the dwell time (e.g., twice the dwell time, three times the dwell time, etc.).

[0050] Despite the above Figure 4 The discussion is limited to selecting quality read points 402 during the startup period 316 and the steady state period 318, but a similar approach can be used to identify the quality read points 402 during the startup period 316 and the steady state period 318. Figure 3 In other words, the timing and spacing of the additional quality reading points 402 within the closed period 320 may be selected based on the degree of dynamic change of the quality parameters 302, 304, 306 and / or the process parameters 310, 312, 314.

[0051] The exemplary historical sampling batch generator 206 can generate any suitable number of batches associated with each of the startup period 316, the steady-state period 318, and the shutdown period 320. In some examples, the number of historical sampling batches associated with the steady-state period 318 will be greater than the number of historical sampling batches associated with the startup and shutdown periods 316, 320. For example, in some examples, approximately 50% of all historical sampling batches generated by the historical sampling batch generator 206 are associated with the steady-state period 318, and 25% of all historical sampling batches are associated with each of the startup and shutdown periods 316, 320. In some examples, the historical sampling batches associated with the startup period 316 can be obtained from historical process data corresponding to a single startup event. In other examples, the historical sampling batches associated with the startup period 316 can be obtained from multiple different instances of continuous process system startup. Similarly, the historical sampling batches associated with the steady-state and shutdown periods 318, 320 can all be obtained from a single operation of the continuous process or from multiple different operations separated by multiple shutdowns and subsequent startups. In some examples, the historical sampling batches generated by the historical sampling batch generator 206 are stored in the example database 214 for subsequent use.

[0052] exist Figure 2 In an example of the present invention, the predictive analysis engine 105 includes a batch model generator 208 to generate an analysis model that can be used to analyze real-time sampling batches associated with a continuous control system process operating in real time. In some examples, the analysis model is developed using the historical sampling batches generated by the exemplary historical sampling batch generator 206. More specifically, in some examples, the batch model generator 208 generates the analysis model using multivariate data analysis of the historical sampling batch pool. The pool can include any suitable number of historical sampling batches to provide statistically reliable results (e.g., at least 25 batches). In some examples, the pool can be divided into two batch sets; the first is used for model training and the second is used for cross-validation. In some examples, all historical sampling batches generated by the historical sampling batch generator 206 are included in the pool. In other examples, a subset of all historical sampling batches is selected to be included in the pool. In some examples, the historical sampling batches are selected and / or generated so that the pool includes more historical sampling batches associated with the steady-state period 318 than the historical sampling batches associated with the startup period 316 or the shutdown period 320. In some examples, the number of historical sampling batches associated with the startup period 316 and the shutdown period 320 is approximately equal, while the number of historical sampling batches associated with the steady-state period 318 is approximately twice the number of historical sampling batches associated with the startup period 316 and the shutdown period 320 .

[0053] In some examples, the batch model generator 208 corresponds to (or at least operates similar to) existing analytical software, firmware, and / or hardware for generating models of traditional batch processes. In other words, in some examples, the batch model generator 208 generates analytical models based on historical sampling batches using PCA and / or PLS techniques. As described above, PCA and PLS cannot be directly applied to continuous processes because continuous processes involve dynamic behaviors that change over time and cannot be explained by PCA and PLS. However, the examples described herein overcome this difficulty by generating multiple fixed-length snapshots of discrete portions of the continuous process corresponding to the above-mentioned historical sampling batches. In addition, the dynamic behavior of the parameters associated with the continuous process is considered based on different historical sampling batches, and these different historical sampling batches have different start and end times that are spaced apart in time at a granularity suitable for capturing the changes in the parameters of the continuous process over time. In other words, instead of analyzing multiple batches with the same start and the same end to generate analytical models (as done in traditional batch analysis), the analysis in the disclosed examples is associated with multiple sampling batches that are slightly time-shifted relative to each other during the relevant period of the continuous process of interest. In some examples, analytical models generated by batch model generator 208 are stored in example database 214 for subsequent use.

[0054] Figure 2 The exemplary predictive analysis engine 105 of the embodiment includes an exemplary virtual batch unit controller 210 for generating a virtual batch unit that serves as a basis for the virtual batch unit controller 210 to implement a sample batch corresponding to the real-time operation of the continuous process in real time. In some examples, the virtual batch unit controller 210 corresponds to existing batch control software, firmware and / or hardware (e.g., DeltaV TM ) (or at least operates similarly thereto). According to the ISA-88 standard for conventional batch process control, a batch is defined in the context of a recipe that includes at least one unit procedure that defines a program control associated with the operation of equipment in a physical process control unit. Consistent with the standard, the virtual batch unit controller 210 of the illustrated example defines a virtual batch unit to replace the physical unit defined for a standard batch process. As used herein, a virtual batch unit refers to a data structure that defines or includes all relevant parameters of a continuous control system process (e.g., process parameters and / or inputs and quality parameters and / or outputs) to enable sampling batches to be implemented in a manner consistent with ISA-88 so that batch-like analytical techniques can be applied to the sampling batches. Figure 5A schematic diagram of an exemplary virtual batch unit 500 is shown in FIG. Similar to a standard batch unit and associated unit procedures, the exemplary virtual batch unit 500 includes input parameters 502 that are monitored substantially in real time and output parameters 504 that are used to measure and / or indicate the quality of the output of the process. In some examples, the input parameters 502 correspond to process parameters associated with a continuous process control system, such as in conjunction with Figure 3 and / or Figure 4 The process parameters 310, 312, 314 are shown and described. In addition, in some examples, the output parameter 504 corresponds to a quality parameter associated with a continuous process control system, such as in conjunction with Figure 3 and / or Figure 4 Quality parameters 302, 304, 306 are shown and described.

[0055] Unlike a standard batch unit and associated unit procedures, the exemplary virtual batch unit 500 includes a separate set of inputs referred to herein as initial conditions 506. Conventional batch processes do not have to specifically define initial conditions for the same set of process parameters because the start of a batch process is substantially the same each time with respect to these parameters. In contrast, as discussed further below, different sampling batches implemented on the virtual batch process unit 500 start at different points in time within a single continuous batch process. As a result, the values ​​of the process parameters at the start of any particular sampling batch are not necessarily the same as the values ​​of the process parameters at the start of a different sampling batch. Thus, the exemplary virtual batch unit 500 includes initial conditions 506, which are defined as the process parameters at the start time of the current sampling batch (e.g., Figure 3 and / or Figure 4 310, 312, 314) of the process parameters 310, 312, 314). In other words, the virtual batch unit 500 includes a repeated set of process parameters as input. However, while the input parameters 502 are substantially updated in real time as the associated continuous process proceeds, the initial conditions 506 are fixed values ​​defined when a sampling batch is first started. Although fixed throughout the implementation of a single sampling batch, the initial conditions 506 can be reset with new values ​​each time a new sampling batch is started.

[0056] In some examples, the virtual batch unit controller 210 implements a sampling batch on the virtual batch unit 500 substantially in real time in parallel with the operation of the continuous process. The sampling batch corresponds to the value of the process parameter currently present in the continuous process. The purpose of implementing the sampling batch using the virtual batch unit by the virtual batch unit controller 210 is to enable the use of batch-like analysis techniques for the continuous process. In other words, the implementation of the sampling batch on the virtual batch unit does not directly affect or control the operation of the continuous process, but runs in parallel with the control and operation of the continuous process to provide predictive analysis. As a result, the recipe used in conjunction with the virtual batch unit 500 is referred to herein as a simulated recipe because the recipe does not actually control the operation of the system. However, in some examples, the results of the predictive analysis information generated by analyzing the sampling batch running on the virtual batch unit 500 can be used to adjust or adapt the process controlled by the standard continuous process control technology.

[0057] Each individual sampling batch implemented by the virtual batch unit controller 210 is defined to have the same duration as the historical sampling batches used to generate the analytical model used to generate the predictive analysis. In addition, the real-time sampling batches implemented by the virtual batch unit controller 210 are implemented back-to-back in a continuous manner. In other words, the end of one sampling batch corresponds to the beginning of the next subsequent sampling batch, wherein a new sampling batch is started each time the current sampling batch ends for the entire duration of the associated continuous process to be analyzed. In other words, the continuous process is viewed as the activity of many consecutive batches running on the virtual batch unit 500.

[0058] For example, Figure 6 600 is a graph illustrating process parameters 310, 312, 314 of a continuous process operating in real time. In the example shown, the first sampling batch begins at the same time as the continuous process begins (time 0) and extends corresponding to the above in conjunction with Figure 3 and Figure 46. As described above, a historical sampling batch may be defined as having a length corresponding to the residence time of a continuous process (e.g., ten minutes in the illustrated example). As described above, the initial conditions 506 defined for the virtual batch unit 500 when implementing this first sampling batch (during the first time period 602) correspond to the initial values ​​of the process parameters 310, 314, 312 (e.g., at time 0). At the ten minute mark, the first sampling batch ends, and a second sampling batch is started on the virtual batch unit 500 during the second time period 604. At the transition between the first time period 602 and the second time period 604 (associated with the transition from the first sampling batch to the second sampling batch), the initial conditions 506 of the virtual batch unit 500 are reset to the values ​​of the process parameters at the beginning of the second time period 604 (e.g., at 10 minutes into the process). The process of starting subsequent sampling batches at subsequent time periods 606, 608 corresponds to successive ten minute time increments, and resetting the initial conditions of the virtual batch unit 500 for each new sampling batch effectively enables the continuous process to be viewed as a series or campaign of a single batch process.

[0059] The examples discussed herein focus primarily on scenarios where a sampling batch is defined as having a duration corresponding to the dwell time of a continuous process (e.g., 10 minutes in the illustrated example). However, as described above, in some examples, the duration of a sampling batch may be longer than the dwell time. More specifically, in some examples, the duration of a single sampling batch may be a multiple of the dwell time (e.g., two dwell times, three dwell times, four dwell times, etc.). In some such examples, a sampling batch associated with a single program unit (e.g., virtual batch unit 500) may be divided into multiple stages having a duration of a single dwell time. In other words, rather than defining a sampling batch as a single process unit, a sampling batch may be defined as a single process unit. Figure 6 602 and the second time period 604 in the example are considered to correspond to separate sampling batches, but in some examples, the first time period 602 and the second time period 604 may correspond to separate phases associated with a single sampling batch extending the first extended time period 610. In this example, the second extended time period 612 defines the duration of the second sampling batch, wherein the third time period 606 and the fourth time period 608 correspond to separate phases within the second sampling batch. In some examples, the duration of the sampling batch and / or the phases within the sampling batch may not correspond to multiples of the residence time, but may be any suitable length, depending on the degree of dynamic change of the process and / or quality parameters and / or the level of accuracy of the predictive analysis to be generated.

[0060] Figure 2The exemplary predictive analysis engine 105 includes an exemplary sampling batch analyzer 212 for analyzing a sampling batch currently implemented on the virtual batch unit 500 by the virtual batch unit controller 210 using the analysis model generated by the batch model generator 208. In some examples, the sampling batch analyzer 212 corresponds to (or at least operates similarly to) existing batch analysis software, firmware, and / or hardware for analyzing a standard batch based on a corresponding analysis model for a batch process. In examples where the sampling batch analyzer 212 corresponds to existing batch analysis software, firmware, and / or hardware, it may not be necessary to disable the dynamic time warping functionality implemented for the standard batch analysis, as described above, because the sampling batch being analyzed and all historical sampling batches used to generate the analysis model are defined to have the same duration.

[0061] In some examples, the results output by the sampling batch analyzer 212 provide fault detection and / or quality prediction information (collectively referred to as predictive analysis information) associated with the current sampling batch within the duration of the sampling batch. When a new sampling process is started, the sampling batch analyzer 212 analyzes the new sampling batch to generate predictive analysis information for the new sampling batch. Therefore, the future time period (e.g., prediction time domain) for which quality prediction is provided corresponds to the duration of the sampling batch. Therefore, when a new sampling batch is started with a dwell time of 10 minutes in the above example, the predictive analysis information provides a prediction of the quality of the output of the continuous process for the next 10 minutes. As time advances in the current sampling batch, the distance to the future represented by the predictive analysis information will decrease until the time reaches the end of the current sampling batch. Thereafter, a new sampling batch is started, and a new prediction for the next 10 minutes can be generated.

[0062] In some examples, the predictive analysis information provided by the example sampling batch analyzer 212 is stored in a database 214. Additionally or alternatively, the example communication interface 202 can send the predictive analysis information to a continuous historical database 218. Furthermore, in some examples, the user interface 216 generates and / or presents a graphical representation of the predictive analysis information via an associated display screen. In some examples, the graphical representation of the predictive analysis information can be similar to the graphical representation provided for a standard batch process analysis. However, in some examples, the predictive analysis information for consecutive sampling batches can be appended to each other to provide a predicted duration line over the duration of the entire monitored continuous control system process.

[0063] For example, Figure 7 An exemplary quality prediction interface 700 is illustrated that provides predictions for a particular quality parameter (eg, Figure 3706 of the quality parameters 302, 304, 306 of the process). In some examples, the start and end of a separate sampling batch can be identified by a sampling batch line 710. In this example, the first time period 702 corresponds to the initial startup of the process. As a result, there is a significant change in the prediction of the quality parameter indicated by the middle solid line 712, and there is a relatively low confidence in the prediction as indicated by the relatively wide confidence range indicated by the upper dashed line 714 and the lower dashed line 716. However, as time passes and the process approaches a steady state, the prediction for the quality parameter also converges to a substantially steady state, which is within a shaded area 718, which defines the boundaries of the quality parameter that meets the specifications of the output of the process. In some examples, a current time indicator 720 (e.g., a straight line) indicates the current time. In this example, the current time is 1:47 pm, which is two minutes into the fourth time period 708 corresponding to the fourth sampling batch. As shown in the illustrated example, the prediction for the quality parameter (e.g., solid line 712) extends beyond the current time 720 to 1:55 pm, corresponding to the end time of the current sampling batch. Thus, the operator has a prediction of the quality of the output of the continuous process extending from the current time 720 to eight minutes in the future. In some examples, once the current time reaches the end of the fourth time period 708, the graph in the quality prediction interface 700 can move to represent the prediction associated with the new sampling batch.

[0064] return Figure 2 , the exemplary user interface 216 also enables a user to configure and / or adjust the operation of the components of the predictive analysis engine 105. For example, in some examples, a user may use the user interface 216 to adjust or specify the dwell time and / or specify whether the length of a sampling batch corresponds to the dwell time or corresponds to a duration greater than the dwell time. In some examples, a user may specify via the user interface 216 the number of historical sampling batches to be generated by the historical sampling batch generator 206 and / or the time intervals of the quality reading points 402 used to define the time locations of different historical sampling batches. In some examples, a user may specify via the user interface 216 the number of historical sampling batches used by the batch model generator 208 to generate the analytical model. In some examples, a user may specify via the user interface 216 the proportion of historical sampling batches acquired from each of the startup period 316, the steady-state period 318, and the shutdown period 320. In some examples, a user may specify via the user interface 216 a continuous process parameter associated with the continuous process, which will be used as an input parameter and an output parameter of the virtual batch unit 500.

[0065] Despite Figure 2 The implementation is illustrated in Figure 1The predictive analysis engine 105 is an exemplary embodiment of the present invention, but may be combined, divided, rearranged, omitted, eliminated and / or implemented in any other manner. Figure 2 In addition, the exemplary communication interface 202, the exemplary residence time analyzer 204, the exemplary historical sampling batch generator 206, the exemplary batch model generator 208, the exemplary virtual batch unit controller 210, the exemplary sampling batch analyzer 212, the exemplary database 214, the exemplary user interface 216 and / or more generally, Figure 1 The exemplary predictive analysis engine 105 can be implemented by hardware, software, firmware, and / or any combination of hardware, software, and / or firmware. Thus, for example, any of the exemplary communication interface 202, the exemplary residence time analyzer 204, the exemplary historical sampling batch generator 206, the exemplary batch model generator 208, the exemplary virtual batch unit controller 210, the exemplary sampling batch analyzer 212, the exemplary database 214, the exemplary user interface 216, and / or more generally, the exemplary predictive analysis engine 105 can be implemented by one or more analog or digital circuits, logic circuits, programmable processors, programmable controllers, graphics processing units (GPUs), digital signal processors (DSPs), application specific integrated circuits (ASICs), programmable logic devices (PLDs), and / or field programmable logic devices (FPLDs). Implementation. When any apparatus or system claim of this patent is read to cover pure software and / or firmware implementations, at least one of the exemplary communication interface 202, exemplary residence time analyzer 204, exemplary historical sampling batch generator 206, exemplary batch model generator 208, exemplary virtual batch unit controller 210, exemplary sampling batch analyzer 212, exemplary database 214, and / or exemplary user interface 216 is expressly defined herein to include a non-transitory computer-readable storage device or storage disk (such as a memory, digital versatile disk (DVD), compact disk (CD), Blu-ray disk, etc.) that includes software and / or firmware. Further, Figure 1 An exemplary predictive analysis engine 105 may include Figure 2 In addition to or in place of the components, processes and / or equipment shown in Figure 2 The present invention may include one or more of the elements, processes and / or devices shown in the embodiment of the present invention, and / or may include more than one of any or all of the elements, processes and devices shown. As used herein, the phrase "in communication", including variations thereof, covers direct communication and / or indirect communication through one or more intermediate components, and does not require direct physical (e.g., wired) communication and / or constant communication, but may additionally include selective communication at regular intervals, scheduled intervals, non-periodic intervals and / or one-time events.

[0066] exist Figure 8-Figure 10 The diagram shows a method for implementing Figure 1 and / or Figure 2 105 of the predictive analysis engine 105. The machine-readable instructions may be one or more executable programs or (one or more) portions of executable programs, such as the following in combination with the state machine implemented by the hardware logic, machine-readable instructions, hardware-implemented state machine and / or any combination thereof. Fig.11 The program may be executed by a computer processor such as the processor 1112 shown in the exemplary processor platform 1100 discussed above. The program may be embodied in software stored in a non-transitory computer-readable storage medium (e.g., a CD-ROM, floppy disk, hard drive, DVD, Blu-ray disk, or memory associated with the processor 1112), but the entire program and / or portions thereof may alternatively be executed by a device other than the processor 1112 and / or embodied in firmware or dedicated hardware. In addition, although reference is made to Figure 8-Figure 10 , but many other methods of implementing the exemplary predictive analysis engine 105 may be used instead. For example, the order of execution of the blocks may be changed, and / or some of the blocks described may be changed, eliminated, or combined. Additionally or alternatively, any or all of the blocks may be implemented by one or more hardware circuits (e.g., discrete and / or integrated analog and / or digital circuits, FPGAs, ASICs, comparators, operational amplifiers (op-amps), logic circuits, etc.) that are configured to perform the corresponding operations without executing software or firmware.

[0067] The machine-readable instructions described herein may be stored in one or more of a compressed format, an encrypted format, a segmented format, a packaged format, and the like. The machine-readable instructions described herein may be stored as data (e.g., portions of instructions, code, representations of code, and the like) that may be used to create, manufacture, and / or generate machine-executable instructions. For example, the machine-readable instructions may be segmented and stored on one or more storage devices and / or computing devices (e.g., servers). The machine-readable instructions may require one or more of installation, modification, adaptation, updating, combination, supplementation, configuration, decryption, decompression, unpacking, distribution, redistribution, and the like so that they may be directly read and / or executed by a computing device and / or other machine. For example, the machine-readable instructions may be stored in a plurality of portions that are each compressed, encrypted, and stored on separate computing devices, wherein the portions, when decrypted, decompressed, and combined, form an executable instruction set that implements a program such as described herein. In another example, the machine-readable instructions may be stored in a state in which they can be read by a computer, but require the addition of a library (e.g., a dynamic link library (DLL)), a software development kit (SDK), an application programming interface (API), etc., in order to execute the instructions on a particular computing device or other device. In another example, the machine-readable instructions and / or corresponding program may need to be configured (e.g., stored settings, data inputs, recorded network addresses, etc.) before the machine-readable instructions and / or corresponding program can be executed in whole or in part. Therefore, the disclosed machine-readable instructions and / or corresponding program are intended to cover such machine-readable instructions and / or programs regardless of the specific format or state of the machine-readable instructions and / or programs when stored or in dormancy or in transit.

[0068] As described above, the present invention may be implemented using executable instructions (e.g., computer and / or machine readable instructions) stored on a non-transitory computer and / or machine readable medium (e.g., a hard drive, flash memory, read-only memory, compact disk, digital versatile disk, cache, random access memory, and / or any other storage device or storage disk). Figure 8-Figure 10 The illustrated process of the present invention is a non-transitory computer and / or machine readable medium in which information is stored for any duration (e.g., for an extended period of time, permanently, temporarily, for temporary buffering and / or for information caching). As used herein, the term non-transitory computer readable medium is expressly defined to include any type of computer readable storage device and / or storage disk, and to exclude propagating signals and to exclude transmission media.

[0069] "Include" and "comprising" (and all forms and tenses thereof) are used herein as open-ended terms. Thus, whenever a claim employs any form of "include" or "comprising" (e.g., includes, comprises, contains, includes, has, etc.) as a preamble or in any form of a claim recitation, it is understood that additional elements, items, etc. may be present without falling outside the scope of the corresponding claim or recitation. As used herein, when the phrase "at least" is used as a transitional term, such as in the preamble of a claim, it is open-ended in the same manner as the terms "include" and "comprising" are open-ended. When used, for example, in a form such as A, B, and / or C, the term "and / or" refers to any combination or subset of A, B, and C, such as (1) only A, (2) only B, (3) only C, (4) A and B, (5) A and C, (6) B and C, and (7) A and B and C. As used herein in the context of describing structures, components, items, objects, and / or things, the phrase "at least one of A and B" is intended to refer to implementations that include any of the following: (1) at least one A, (2) at least one B, and (3) at least one A and at least one B. Similarly, as used herein in the context of describing structures, components, items, objects, and / or things, the phrase "at least one of A or B" is intended to refer to implementations that include any of the following: (1) at least one A, (2) at least one B, and (3) at least one A and at least one B. As used herein in the context of describing the execution or performance of processes, instructions, actions, activities, and / or steps, the phrase "at least one of A and B" is intended to refer to implementations that include any of the following: (1) at least one A, (2) at least one B, and (3) at least one A and at least one B. Similarly, as used herein in the context of describing the operation or execution of a process, instruction, action, activity, and / or step, the phrase "at least one of A or B" is intended to refer to an implementation that includes any of the following: (1) at least one A, (2) at least one B, and (3) at least one A and at least one B.

[0070] Figure 8 The exemplary process of begins at block 802, where the exemplary predictive analysis engine 105 generates an analysis model to analyze a sample batch associated with a continuous control system process. Fig. 9 and Fig.10 Further details are provided regarding the implementation of block 802. At block 804, the example virtual batch unit controller 210 generates a batch unit based on parameters associated with the continuous control system process (e.g., Figure 3 and / or Figure 4 The quality parameters 302, 304, 306 and process parameters 310, 312, 314) of the batch are virtual batch units (e.g., Figure 5The example virtual batch unit controller 210 may define input parameters and output parameters for the sample batch. At block 806, the example virtual batch unit controller 210 determines a start time for the sample batch. The start time for the sample batch may correspond to a start time for the continuous control system process or an end time for a previous sample batch. The end time for the current sample batch is defined based on a fixed duration set for the sample batch. In some examples, the historical sample batch generator 206 and / or the virtual batch unit controller 210 determines the duration of the sample batch based on a residence time for the continuous control system process. At block 808, the example virtual batch unit controller 210 sets initial conditions for the virtual batch unit based on parameters associated with the continuous control system process at the start time of the sample batch.

[0071] At block 810, the example virtual batch unit controller 210 implements a sampling batch on the virtual batch unit 500 in parallel with the continuous control system process. At block 812, the example sampling batch analyzer 212 analyzes the sampling batch based on the analytical model. At block 814, the example sampling batch analyzer 212 generates predictive analysis information for the sampling batch. The predictive analysis information may indicate fault detection and / or quality prediction for the continuous control system process. At block 816, the example user interface 216 presents a graphical representation of the predictive analysis information.

[0072] At block 818, the example virtual batch unit controller 210 determines whether to implement another sampling batch. If so, control returns to block 806. Otherwise, control advances to block 820, where the predictive analysis engine 105 determines whether to update the analysis model. If so, control returns to block 802. Otherwise, Figure 8 The exemplary process ends.

[0073] Fig. 9 Yes means Figure 8 Flowchart of an exemplary implementation of block 802 of . Fig. 9The exemplary process begins at box 902, where the communication interface 202 accesses historical process data of the continuous control system process. At box 904, the exemplary residence time analyzer 204 determines the residence time of the continuous control system process. At box 906, the exemplary historical sampling batch generator 206 and / or the exemplary virtual batch unit controller 210 determines the length of the sampling batch based on the residence time. At box 908, the exemplary historical sampling batch generator 206 generates a historical sampling batch associated with a startup period (e.g., startup period 316) of the continuous control system process. At box 910, the exemplary historical sampling batch generator 206 generates a historical sampling batch associated with a steady-state period (e.g., steady-state period 318) of the continuous control system process. At box 912, the exemplary historical sampling batch generator 206 generates a historical sampling batch associated with a shutdown period (e.g., shutdown period 320) of the continuous control system process. The following is combined with Fig.10 More details are provided regarding the implementation of blocks 908, 910, and 912. At block 914, the example batch model generator 208 selects a pool of historical sampled batches for model generation. At block 916, the example batch model generator 208 generates an analytical model based on the pool of historical sampled batches. Thereafter, Fig. 9 The process ends and returns to complete Figure 8 process.

[0074] Fig.10 Yes means Fig. 9 A flowchart of an exemplary implementation of any one of blocks 908, 910, and 912 of FIG. Fig.10 The example process begins at block 1002, where the example historical sampling batch generator 206 determines quality reading points (e.g., Figure 4 In this context, the relevant time period corresponds to the time interval between the quality reading points 402 of the implementation. Fig. 9 The start-up period 316 of the frame 908, in the implementation Fig. 9 The steady state period 318 of block 910 and the Fig. 9 At block 1004, the example historical sampling batch generator 206 identifies a quality reading point 402 within a relevant period of the historical process data. At block 1006, the example communication interface 202 obtains a historical sampling batch (at block 1006) corresponding to the length of the sampling batch and ending at the quality reading point. Fig. 9At block 1008, the example historical sampling batch generator 206 sets initial conditions for the historical sampling batch. At block 1010, the example database 214 stores the historical sampling batch. At block 1012, the example historical sampling batch generator 206 determines whether to generate another historical sampling batch associated with the relevant time period. If so, control returns to block 1002. Otherwise, Fig.10 The exemplary process ends and returns to complete Fig. 9 process.

[0075] Fig.11 is a block diagram of an exemplary processor platform 1100 configured to perform Figure 8-Figure 10 Instructions to achieve Figure 1 and / or Figure 2 The processor platform 1100 may be, for example, a server, a personal computer, a workstation, a self-learning machine (e.g., a neural network), a mobile device (e.g., a cellular phone, a smart phone, a tablet computer (such as an iPad), or a computer system (e.g. ... computer system)). TM ), a personal digital assistant (PDA), an Internet appliance, or any other type of computing device.

[0076] The processor platform 1100 of the illustrated example includes a processor 1112. The processor 1112 of the illustrated example is hardware. For example, the processor 1112 can be implemented by one or more integrated circuits, logic circuits, microprocessors, GPUs, DSPs, or controllers from any desired series or manufacturer. The hardware processor can be a semiconductor-based (e.g., silicon-based) device. In this example, the processor implements the exemplary residence time analyzer 204, the exemplary historical sampling batch generator 206, the exemplary batch model generator 208, the exemplary virtual batch unit controller 210, and the exemplary sampling batch analyzer 212.

[0077] The processor 1112 of the illustrated example includes a local memory 1113 (e.g., cache). The processor 1112 of the illustrated example communicates with a main memory including a volatile memory 1114 and a non-volatile memory 1116 via a bus 1118. The volatile memory 1114 may be comprised of a synchronous dynamic random access memory (SDRAM), a dynamic random access memory (DRAM), Dynamic Random Access Memory The main memory 1114 and 1116 may be implemented by a memory controller.

[0078] The processor platform 1100 of the illustrated example also includes an interface circuit 1120. The interface circuit 1120 may be implemented by any type of interface standard (such as an Ethernet interface, a universal serial bus (USB), a Bluetooth In this example, the interface circuit 1120 implements the exemplary communication interface 202 and the exemplary user interface 216.

[0079] In the example shown, one or more input devices 1122 are connected to the interface circuit 1120. The input device(s) 1122 allow a user to enter data and / or commands into the processor 1112. The input devices may be implemented by, for example, an audio sensor, a microphone, a camera (still or video), a keyboard, buttons, a mouse, a touch screen, a trackpad, a trackball, isopoint, and / or a voice recognition system.

[0080] One or more output devices 1124 are also connected to the interface circuit 1120 of the illustrated example. The output device 1124 can be implemented, for example, by a display device (e.g., a light emitting diode (LED), an organic light emitting diode (OLED), a liquid crystal display (LCD), a cathode ray tube display (CRT), a plane switching (IPS) display, a touch screen, etc.), a tactile output device, a printer, and / or a speaker. Therefore, the interface circuit 1120 of the illustrated example typically includes a graphics driver card, a graphics driver chip, and / or a graphics driver processor.

[0081] The interface circuitry 1120 of the illustrated example also includes communication devices, such as transmitters, receivers, transceivers, modems, residential gateways, wireless access points, and / or network interfaces to facilitate exchanging data with external machines (e.g., any kind of computing device) via the network 1126. Communications may be via, for example, an Ethernet connection, a digital subscriber line (DSL) connection, a telephone line connection, a coaxial cable system, a satellite system, a line-of-sight wireless system, a cellular telephone system, and the like.

[0082] The processor platform 1100 of the illustrated example also includes one or more mass storage devices 1128 for storing software and / or data. Examples of such mass storage devices 1128 include floppy disk drives, hard disk drives, compact disk drives, Blu-ray disk drives, redundant array of independent disks (RAID) systems, and digital versatile disk (DVD) drives. In this example, the mass storage device includes the exemplary database 214.

[0083] Figure 8-Figure 10The machine-executable instructions 1132 may be stored in the mass storage device 1128, the volatile memory 1114, the non-volatile memory 1116, and / or on a removable, non-transitory computer-readable storage medium such as a CD or DVD.

[0084] From the foregoing it will be appreciated that exemplary methods, apparatus, and articles of manufacture have been disclosed that enable the generation of predictive analytics for continuous process control systems based on analytical techniques traditionally used for batch process control systems. The disclosed methods, apparatus, and articles of manufacture improve the operation of a continuous control system process by continuously providing an operator with future predictions of the output of the process. More specifically, such information may enable an operator to more quickly recognize and respond to potential deviations of a continuous process so that appropriate remedial action can be taken sooner rather than later when the deviation becomes more severe.

[0085] Although certain exemplary methods, apparatus, and articles have been disclosed herein, the scope of coverage of this patent is not limited thereto. On the contrary, this patent covers all methods, apparatus, and articles that fully fall within the scope of the claims of this patent.

Claims

1. A device comprising: A historical sampling batch generator for generating a historical sampling batch pool based on historical process data of a continuous control system process; a virtual batch unit controller for implementing a sampling batch on a virtual batch unit, the virtual batch unit running in parallel with a continuous control system process, the sampling batch corresponding to a discrete time period of the continuous control system process, the virtual batch unit including input and output parameters corresponding to parameters associated with the continuous control system process; as well as a sampling batch analyzer for generating predictive analytical information indicative of a predicted quality of an output of the continuous control system process at the end of the discrete time period based on an analysis of the sampling batch relative to an analytical model; Wherein, the analysis model is generated based on the historical sampling batch pool.

2. The apparatus according to claim 1, further comprising: A residence time analyzer is used to approximate the residence time of the continuous control system process, and the length of the discrete time period is based on the residence time.

3. The device according to claim 2, wherein: The length of the discrete time period is equal to an approximate value of the dwell time.

4. The device according to claim 2, wherein: The length of the discrete time period is greater than an approximate value of the dwell time.

5. The device according to claim 4, wherein: The length of the discrete time period is a multiple of the approximate value of the dwell time, ranging from two to four times the approximate value of the dwell time.

6. The device according to claim 1, wherein: The virtual batch unit controller specifies a value of an initial condition of the virtual batch unit at the beginning of the discrete time period to correspond to a value of the parameter associated with the continuous control system process, the initial condition being separate from the input and output parameters.

7. The device according to claim 1, wherein: The discrete time period is a first discrete time period and the sampling batch is a first sampling batch, the virtual batch unit controller implements a second sampling batch on the virtual batch unit, the second sampling batch corresponds to a second discrete time period of the continuous control system process, a start of the second discrete time period corresponds to an end of the first discrete time period, and the predictive analysis information indicates a predicted quality of an output of the continuous control system process at an end of the second discrete time period based on an analysis of the second sampling batch relative to the analysis model.

8. The apparatus according to claim 7, further comprising: A user interface is provided for presenting a quality prediction interface, wherein the quality prediction interface graphically represents the predictive analysis information along a timeline corresponding to the first discrete time period and the second discrete time period.

9. The device according to claim 7, wherein: The first discrete time period has the same duration as the second discrete time period.

10. The device according to claim 1, wherein: The historical sampling batch pool includes: a first historical sampling batch set from historical process data associated with a startup period of the continuous control system process; a second historical sampling batch set from historical process data associated with a steady-state period of the continuous control system process; and a third historical sampling batch set from historical process data associated with a shutdown period of the continuous control system process; and A batch model generator is used to generate the analysis model based on the first historical sampling batch set, the second historical sampling batch set and the third historical sampling batch set.

11. The device according to claim 10, wherein: A first time interval between start times of consecutive historical sampling batches in the first historical sampling batch set is smaller than a second time interval between start times of consecutive historical sampling batches in the second historical sampling batch set.

12. The device according to claim 11, wherein The first time interval is less than a time length of the historical sampling batch, so that different historical sampling batches in the first historical sampling batch set include overlapping portions of the historical process data.

13. The device according to claim 10, wherein: The time lengths of some of the historical sampling batches correspond to the lengths of the discrete time periods.

14. The device according to claim 10, wherein: The number of historical sampling batches in the second historical sampling batch set is greater than the number of historical sampling batches in the first historical sampling batch set or the third historical sampling batch set.

15. A non-transitory computer readable medium comprising instructions that, when executed, cause a machine to at least: Generate a historical sampling batch pool based on historical process data of a continuous control process system; implementing a sampling batch on a virtual batch unit that runs in parallel with a continuous control system process, the sampling batch corresponding to a discrete time period of the continuous control system process, the virtual batch unit including input and output parameters corresponding to parameters associated with the continuous control system process; and generating predictive analytical information indicative of a predicted quality of an output of the continuous control system process at the end of the discrete time period based on an analysis of the sample batch relative to an analytical model; in, The analysis model is generated based on the historical sampling batch pool.

16. The non-transitory computer readable medium of claim 15, wherein: The instructions further cause the machine to approximate a dwell time of the continuous control system process, the length of the discrete time period being based on the dwell time.

17. The non-transitory computer readable medium of claim 15, wherein: The instructions further cause the machine to assign a value of an initial condition of the virtual batch unit to a value corresponding to the parameter associated with the continuous control system process at the beginning of the discrete time period, the initial condition being separate from the input and output parameters.

18. The non-transitory computer readable medium of claim 15, wherein: The discrete time period is a first discrete time period and the sampled batch is a first sampled batch, the instructions further causing the machine to implement a second sampled batch on the virtual batch unit, the second sampled batch corresponding to a second discrete time period of the continuous control system process, the start of the second discrete time period corresponding to an end of the first discrete time period, the predictive analysis information indicating a predicted quality of an output of the continuous control system process at an end of the second discrete time period based on an analysis of the second sampled batch relative to the analysis model.

19. The non-transitory computer readable medium of claim 18, wherein: The instructions also cause the machine to present, via a display, a quality prediction interface that graphically represents the predictive analysis information along a timeline corresponding to the first discrete time period and the discrete time period.

20. The non-transitory computer readable medium of claim 18, wherein: The first discrete time period has the same duration as the second discrete time period.

21. The non-transitory computer readable medium of claim 15, wherein: The instructions for generating a pool of historical sample batches further cause the machine to: generating a first historical sampling batch set based on historical process data associated with a startup period of the continuous control system process; generating a second historical sampling batch set based on historical process data associated with a steady-state period of the continuous control system process; generating a third historical sampling batch set based on historical process data associated with a shutdown period of the continuous control system process; as well as The analysis model is generated based on the first historical sampling batch set, the second historical sampling batch set, and the third historical sampling batch set.

22. The non-transitory computer readable medium of claim 21, wherein: A first time interval between start times of consecutive historical sampling batches in the first historical sampling batch set is smaller than a second time interval between start times of consecutive historical sampling batches in the second historical sampling batch set.

23. The non-transitory computer readable medium of claim 22, wherein: The first time interval is smaller than the time length of the historical sampling batch, so that different historical sampling batches in the first historical sampling batch set include overlapping parts of the historical process data.

24. The non-transitory computer readable medium of claim 21, wherein: The time lengths of some of the historical sampling batches correspond to the lengths of the discrete time periods.

25. The non-transitory computer readable medium of claim 21, wherein: The number of historical sampling batches in the second historical sampling batch set is greater than the number of historical sampling batches in the first historical sampling batch set or the third historical sampling batch set.

26. A method comprising: generating a historical sampling batch pool based on historical process data of the continuous control system process; implementing a sampling batch on a virtual batch unit that runs in parallel with a continuous control system process, the sampling batch corresponding to a discrete time period of the continuous control system process, the virtual batch unit including input and output parameters corresponding to parameters associated with the continuous control system process; and generating predictive analytical information indicative of a predicted quality of an output of the continuous control system process at the end of the discrete time period based on an analysis of the sample batch relative to an analytical model; Wherein, the analysis model is generated based on the historical sampling batch pool.

27. The method according to claim 26, further comprising: A residence time of the continuous control system process is approximated, and the length of the discrete time period is based on the residence time.

28. The method of claim 26, further comprising: At the beginning of the discrete time period, an initial condition of the virtual batch unit is assigned a value corresponding to the value of the parameter associated with the continuous control system process, the initial condition being separate from the input and output parameters.

29. The method according to claim 26, wherein: The discrete time period is a first discrete time period and the sampling batch is a first sampling batch, the method further comprising implementing a second sampling batch on the virtual batch unit, the second sampling batch corresponding to a second discrete time period of the continuous control system process, a start of the second discrete time period corresponding to an end of the first discrete time period, the predictive analysis information indicating a predicted quality of an output of the continuous control system process at an end of the second discrete time period based on an analysis of the second sampling batch relative to the analysis model.

30. The method of claim 29, wherein: Also included is presenting, via a display, a quality prediction interface that graphically represents the predictive analysis information along a timeline corresponding to the first discrete time period and the second discrete time period.

31. The method of claim 29, wherein: The first discrete time period has the same duration as the second discrete time period.

32. The method of claim 26, wherein generating a historical sampling batch pool further comprises: generating a first historical sampling batch set based on historical process data associated with a startup period of the continuous control system process; generating a second historical sampling batch set based on historical process data associated with a steady-state period of the continuous control system process; generating a third historical sampling batch set based on historical process data associated with a shutdown period of the continuous control system process; as well as The analysis model is generated based on the first historical sampling batch set, the second historical sampling batch set, and the third historical sampling batch set.

33. The method of claim 32, wherein: A first time interval between start times of consecutive historical sampling batches in the first historical sampling batch set is smaller than a second time interval between start times of consecutive historical sampling batches in the second historical sampling batch set.

34. The method of claim 33, wherein: The first time interval is smaller than the time length of the historical sampling batch, so that different historical sampling batches in the first historical sampling batch set include overlapping parts of the historical process data.

35. The method of claim 32, wherein: The time lengths of some of the historical sampling batches correspond to the lengths of the discrete time periods.

36. The method of claim 32, wherein: The number of historical sampling batches in the second historical sampling batch set is greater than the number of historical sampling batches in the first historical sampling batch set or the third historical sampling batch set.

Images