Systems and methods for detecting buildup on a surface of a heater

Abstract

A method for monitoring a surface characteristic of a heater provided in a fluid flow conduit of a fluid flow system includes obtaining sensor data from one or more sensors of the fluid flow system, and generating simulation data based on one or more inputs, a mathematical model of the fluid flow system, and the sensor data. The method further includes training a machine learning model based on the sensor data and determining whether the machine learning model satisfies one or more model conditions based on the simulation data. The method also includes predicting, in response to the machine learning model satisfying the one or more model conditions, the surface characteristic of the heater at a plurality of locations of the heater based on the machine learning model, wherein the plurality of locations includes an undetectable location of the heater.

Description

Attorney Docket No.: 0100TS-000017-WO-POASYSTEMS AND METHODS FOR DETECTING BUILDUP ON A SURFACE OF A HEATERCROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims priority to U.S. provisional application number 63 / 683,956 filed on August 16, 2024. The disclosure of the above application is incorporated herein by reference.FIELD

[0002] The present disclosure relates to systems and methods for operating a thermal system having a process control system and a heater system, and more particularly to systems and methods for detecting coking or fouling within fluid heat exchangers.BACKGROUND

[0003] The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.

[0004] A thermal system for an industrial process typically includes a heater system and a process control system to monitor and control the operations of the heater system. In some heater systems, tubular heaters are used in heat exchangers. In operation, the heat capacity rate of the heat exchanger depends on the heat generation capability of the tubular heater. However, the tubular heaters may have performance issues, such as coking or fouling due to overheating, which eventually leads to failure. In portions of the tubular heater that do not include sensors, it is difficult to determine if and when the coking or fouling may occur.

[0005] These issues related to coking or fouling within fluid heat exchangers are addressed by the present disclosure.SUMMARY

[0006] This section provides a general summary of the disclosure and is not a comprehensive disclosure of its full scope or all of its features.

[0007] The present disclosure provides systems and methods for monitoring surface characteristics of a heater provided in a fluid flow conduit of a fluid flow system. The method includes obtaining sensor data from one or more sensors of the fluid flow system, and generating simulation data based on one or more inputs, aAttorney Docket No.: 0100TS-000017-WO-POA mathematical model of the fluid flow system, and the sensor data. The mathematical model of the fluid flow system associates the sensor data with a known position of each of the one or more sensors, and the simulation data includes at least one data value associated with an undetectable position or running condition of the fluid flow system. The method further includes training a machine learning model based on the sensor data and determining whether the machine learning model satisfies one or more model conditions based on the simulation data. The method also includes predicting, in response to the machine learning model satisfying the one or more model conditions, the surface characteristic of the heater at a plurality of locations of the heater based on the machine learning model, wherein the plurality of locations includes an undetectable location of the heater.

[0008] In variations of the method of the above paragraph, which may be implemented individually or in any combination: the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, a conduit layout or dimension, a gas component, or a combination thereof; the simulation data indicates a simulated condition at any approximate location of the heater (using a three-dimensional simulation in some forms), including at least a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, a simulated running condition of the heater, or a combination thereof, among others; the simulated electrical characteristic of the heater includes a simulated voltage of the heater, a simulated current of the heater, or a combination thereof; the sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof; the electrical characteristic of the heater includes a voltage of the heater, a current of the heater, a power of the heater, a leakage current of the heater, a resistance of the heater, or a combination thereof; the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof; the machine learning model satisfies the one or more model conditions in response to the machine learning model beingAttorney Docket No.: 0100TS-000017-WO-POA associated with an accuracy indicator that is greater than a threshold accuracy indicator; the heater is one of a heat exchanger, a fired heater, a layered heater, a tubular heater, a cartridge heater, a polymer heater, a flexible heater, a line heater, and a cable heater; the surface characteristic indicates an amount of material buildup on the heater surface; the mathematical model is a physics-based model of the environment, a thermodynamic model of the environment, or a combination thereof; the method further includes validating the simulated data based on a comparison between the simulated data and the sensor data; and / or the method also includes selectively adjusting the mathematical model based on a difference between the simulation data and the sensor data.

[0009] The present disclosure provides a system for monitoring a surface characteristic of a heater surface of a heater of an environment, where the environment further includes a fluid flow conduit. The system includes one or more processors and one or more non-transitory computer-readable mediums comprising instructions that are executable by the one or more processors. The instructions include generating simulation data based on one or more inputs and a mathematical model of the environment and obtaining sensor data from one or more sensors of the environment. The instructions further include training a machine learning model based on the sensor data and determining whether the machine learning model satisfies one or more model conditions based on the simulation data. The instructions also include predicting, in response to the machine learning model satisfying the one or more model conditions, the surface characteristic of the heater surface based on the machine learning model, wherein the surface characteristic indicates an amount of material buildup on the heater surface.

[0010] In variations of the system of the above paragraph, which may be implemented individually or in any combination: the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, or a combination thereof; the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, or a combination thereof; the sensor data indicates an electricalAttorney Docket No.: 0100TS-000017-WO-POA characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof; the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof; the machine learning model satisfies the one or more model conditions in response to the machine learning model being associated with an accuracy indicator that is greater than a threshold accuracy indicator; and / or the mathematical model is a physics-based model of the environment, a thermodynamic model of the environment, or a combination thereof.

[0011] The present disclosure provides a method for monitoring a surface characteristic of a heater provided in a fluid flow conduit of a fluid flow system. The method includes obtaining a first set of sensor data from one or more sensors of the fluid flow system, where the first set of sensor data corresponds to a detectable location of the fluid flow system; predicting, using a machine learning model, a surface characteristic of the heater at an undetectable location of the heater based on the first set of sensor data; and selectively performing a corrective action based on the surface characteristic of the heater.

[0012] In variations of the method of the above paragraph, which may be implemented individually or in any combination: where the first set of sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof; the electrical characteristic of the heater includes a voltage of the heater, a current of the heater, a power of the heater, a leakage current of the heater, a resistance of the heater, or a combination thereof; the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof; the surface characteristic indicates an amount of material buildup on a heater surface; the heater is one of a heat exchanger, a fired heater, a layered heater, a tubular heater, a cartridge heater, a polymer heater, a flexible heater, a line heater, and a cable heater; the machine learning model is trained based on a second set of sensor data from the one or more sensors of the fluid flow system; the machine learning model is validated based on simulation data that corresponds to an undetectable position of the fluid flow system or a running condition of the fluid flow system; the simulation data is based on one or more inputs, a mathematical model of the fluid flow system, and the second setAttorney Docket No.: 0100TS-000017-WO-POA of sensor data, where the mathematical model of the fluid flow system associates the second set of sensor data with a known position of each of the one or more sensors; the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, or a combination thereof; and / or the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, or a combination thereof.

[0013] The present disclosure provides a system for monitoring a surface characteristic of a heater provided in a fluid flow conduit of a fluid flow system. The system includes one or more processors and one or more non-transitory computer-readable mediums comprising instructions that are executable by the one or more processors. The instructions include obtaining a first set of sensor data from one or more sensors, where the first set of sensor data corresponds to a detectable location of the fluid flow system; predicting, using a machine learning model, a surface characteristic of the heater at an undetectable location of the heater based on the first set of sensor data; and selectively performing a corrective action based on the surface characteristic of the heater.

[0014] In variations of the system of the above paragraph, which may be implemented individually or in any combination: where the first set of sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof; the electrical characteristic of the heater includes a voltage of the heater, a current of the heater, a power of the heater, a leakage current of the heater, a resistance of the heater, or a combination thereof; the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof; the surface characteristic indicates an amount of material buildup on a heater surface; the heater is one of a heat exchanger, a fired heater, a layered heater, a tubular heater, a cartridge heater, a polymer heater, a flexible heater, a line heater, and a cable heater; the machine learning model is trained based on a second set of sensor data from the one or more sensors of the fluid flow system; the machine learning model is validatedAttorney Docket No.: 0100TS-000017-WO-POA based on simulation data that corresponds to an undetectable position of the fluid flow system or a running condition of the fluid flow system; the simulation data is based on one or more inputs, a mathematical model of the fluid flow system, and the second set of sensor data, where the mathematical model of the fluid flow system associates the second set of sensor data with a known position of each of the one or more sensors; the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, or a combination thereof; and / or the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, or a combination thereof.

[0015] Further areas of applicability will become apparent from the description provided herein. It should be understood that the description and specific examples are intended for purposes of illustration only and are not intended to limit the scope of the present disclosure.DRAWINGS

[0016] In order that the disclosure may be well understood, there will now be described various forms thereof, given by way of example, reference being made to the accompanying drawings, in which:

[0017] FIG. 1 is a block diagram of a thermal system having a process control system, a heater system, and a monitoring system according to the present disclosure;

[0018] FIG. 2 is a side view of a heater capable of being monitored according to the present disclosure;

[0019] FIG. 3 is a partial cross-sectional view of the heater of FIG. 2;

[0020] FIG. 4 is a block diagram of a heater control module according to the present disclosure;

[0021] FIG. 5 is a flowchart for monitoring a surface characteristic of a heater according to the present disclosure; andAttorney Docket No.: 0100TS-000017-WO-POA

[0022] FIG. 6 is another flowchart for monitoring a surface characteristic of a heater according to the present disclosure.

[0023] The drawings described herein are for illustration purposes only and are not intended to limit the scope of the present disclosure in any way.DETAILED DESCRIPTION

[0024] The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses. It should be understood that throughout the drawings, corresponding reference numerals indicate like or corresponding parts and features.

[0025] Referring to FIG. 1 , a thermal system 10 including a process control system 100, a monitoring system 200, and a fluid flow system 300 having one or more heater(s) 302 therein is shown. In one form, the process control system 100 is configured to control the operation of the heater system 300 and more particularly, the heater 302. As described further herein, the monitoring system 200 of the present disclosure is configured to use sensor data from one or more sensors 304 to predict the surface characteristics (e.g., buildup on the surface) of the heater 302 at different locations, including at one or more undetectable locations of the heater 302 (e.g., locations not monitored by sensors). More particularly, as described in more detail herein, the monitoring system 200 generates simulation data based at least in part on a model (e.g., a mathematical model, a machine learning model, or a combination thereof) of the fluid flow system 300 associated with the heater 302.

[0026] The thermal system 10 may be part of various types of industrial processes for controlling a thermal characteristic of a load being heated. In one example, the thermal system 10 may be used in a fluid heat exchanger to heat fluid flowing through one or more fluid flow conduits 306, such as in the fluid flow system 300. In one form, the one or more heaters 302 may include multiple flexible heaters that wrap around the fluid flow conduits 306 to heat the fluid therein. In yet another example, the thermal system 10 may employ cartridge heaters to directly heat fluid (e.g., gas and / or liquid) flowing through conduits 306 or provided within a container. Other heaters and heater configurations are contemplated by the present disclosure as described in more detail herein.

[0027] In some examples, the process control system 100 may be configured to control a thermal profile of the heater(s) 302, which may vary based onAttorney Docket No.: 0100TS-000017-WO-POA external thermal controls and internal thermal controls. For example, internal thermal controls may include, but are not limited to, power provided to the heater 302 (e.g., a tubular or cartridge heater), an operational mode of the thermal system 10 (e.g., a manual mode to control power to the heater 302 based on inputs from a user, a coldstart mode to gradually increase temperature of the heater, a steady-state mode to maintain the heater at a temperature setpoint, among other defined operation modes for controlling the heater 302), and / or operational state of different zones of the heater, among other parameters controllable by the thermal system 10. Examples of external thermal controls include, but are not limited to, controls for heater applications, particularly heat exchanger applications as set forth in greater detail below.

[0028] Various process control systems 100 may be used to control the heater 302 using sensor data from the sensors 304. An example is disclosed in U.S. Patent No. 11 ,550,346, filed February 10, 2020, and titled “SMART HEATER SYSTEM,” which is commonly owned with the present application and the contents of which are incorporated herein by reference in its entirety. In the control system, a heater control unit controls the heater elements differently, via switches, based on the heater information and the measured temperature from temperature sensor(s).

[0029] While specific applications of the thermal system 10 are provided herein, the present disclosure may be applicable to other industrial processes having a thermal system to heat a load. These applications may include, by way of example, injection molding processes, heat exchangers, exhaust gas aftertreatment systems, and energy processes, among others. Furthermore, the heater 302 should not be limited to the examples provided herein and may include other types of heater constructions such as by way of example, a heat exchanger, a layered heater, a fired heater, a cartridge heater, a tubular heater, a polymer heater, a flexible heater, a cable heater, and a line heater, among others. In each application, the external controls and internal controls may be identified and utilized by the monitoring system 200, as described herein. In one form, for example, a cable heater is used as a clog detection sensor to apply pulses (e.g., from setting temperature to setting temperature (30C ~50C)), or a line heater (silicone rubber heater) is used to heat the pipe to maintain a setting temperature.

[0030] In some examples, the sensor(s) 304 are one or more discrete sensors for acquiring sensor data for the heater 302. For example, a temperature sensor may be provided at the heater 302 and is communicably coupled to the processAttorney Docket No.: 0100TS-000017-WO-POA control system 100 to provide the temperature measurement as a process variable. In one form, at least one of a pressure sensor, a flow rate meter, a voltage meter, and / or a current meter, among others, is provided at the heater 302 to obtain a corresponding performance characteristic.

[0031] As used herein, the term “sensor data” or “meter data” refers to data acquired by the one or more sensors 304 (or meters) and used by the process control system 100 for determining an output control and / or by the monitoring system 200 to predict one or more surface characteristics of the heater 302. For example, the sensor data may include any data that indicates an electrical characteristic (e.g., voltage, current, power, leakage current, resistance, etc.) of the heater 302, an inlet temperature of the fluid flow conduit 306 of the fluid flow system 300, an outlet temperature of the fluid flow conduit 306, an inlet pressure of the fluid flow conduit 306, an outlet pressure of the fluid flow conduit 306, a fluid flow rate, or a combination thereof, among others.

[0032] The process control system 100 is configured to control the heater 302 based on one or more control processes and that can use predicted surface characteristics (e.g., predicted buildup) of the heater 302 as determined by the monitoring system 200 as part of the one or more control processes. In one form, the process control system 100 includes a control loop module 102 and a heater control module 104. The control loop module 102 is configured to perform the control process(es) to determine the output control based on one or more inputs, such as a setpoint (temperature, power, resistance, among other measurable parameters to control the heater 302). That is, one or more inputs may include a power setpoint of the heater 302, a setpoint inlet temperature of the fluid flow conduit 306, a setpoint outlet temperature of the fluid flow conduit 306, a setpoint inlet pressure of the fluid flow conduit 306, a setpoint outlet pressure of the fluid flow conduit 306, a setpoint flow rate, a pipe / conduit layout and dimensions, a gas component, etc.

[0033] In some examples, a buildup detection system includes the sensors 304 and a buildup detection module 106 (see also FIG. 2) for detecting buildups on the surface of the heater 302 based on data from the sensors 304. The sensors 304 may be provided at a plurality of predetermined locations at or adjacent to the heater 302 to measure an operating characteristic at the plurality of predetermined locations. In one form, the plurality of predetermined locations may be multiple locations on or adjacent to an outer surface of the heater 302 along aAttorney Docket No.: 0100TS-000017-WO-POA longitudinal direction of the heater 302 and / or multiple locations on or adjacent to the outer surface of the heater 302 along a circumferential direction of the heater 302. Alternatively, the plurality of predetermined locations may be multiple locations inside the heater 302 along a longitudinal direction and / or a circumferential direction of the heater 302. Alternatively, when the heater 302 includes a plurality of heating zones, the plurality of predetermined locations may be multiple locations corresponding to the plurality of heating zones or a plurality groups of heating zones, whether the sensors 304 are disposed inside or outside the heater 302.

[0034] If the sensors 304 are disposed outside the heater 302, the operating characteristic to be measured by the sensors 304 may include, but not limited to, a temperature of the heater 302, a thermal response time of the heater 302, a fluid pressure, and / or an actual fluid flow rate of the fluid flowing proximate the heater 302. When the sensors 304 are disposed inside the heater 302, the operating characteristic to be measured by the sensors 304 may be a temperature or a thermal response time of the heater 302. The sensors 304 send data indicative of the operating characteristic at the predetermined locations to the buildup detection module 106 for analysis.

[0035] The buildup detection module 106 in some examples is configured to receive the data from the sensors 304 and determine whether a buildup exists at the predetermined locations based on the data from the sensors 304. When a buildup exists at a particular location, the temperature and thermal response of the heater 302 at the particular location may be lower than designed because a portion of the heat from the heater 302 is absorbed by the buildups and less heat is transferred to the fluid flowing over the buildups. Moreover, the buildups cause obstruction to the fluid, resulting in a pressure drop in the fluid flow and a reduced flow rate. Therefore, the buildup detection module 106 may determine whether a buildup exists on the surface of the heater 302 by comparing the values of the operating characteristic obtained by the sensors 304 at the plurality of predetermined locations. The buildup detection module 106 is configured to determine that a potential buildup is about to be formed at one or more of the predetermined locations when the values of the operating characteristic obtained at the one or more of the predetermined locations deviate from the values from the other locations. The buildup detection module 106 is configured to determine that the buildup exists at one or more of the predetermined locations when the values of the operating characteristic exceed a threshold. The threshold may beAttorney Docket No.: 0100TS-000017-WO-POA determined based on experimental and / or statistical data, or by any method known in the art without departing from the scope of the present disclosure.

[0036] In operation, based on the setpoint and the process variable, the control loop module 102 is configured to calculate an operational power level to be applied to the heater 302, such that the power applied to the heater 302 is approximately equal to a power setpoint. In some examples, the control loop module 102 selectively monitors and controls the heater 302 using the sensor data and, optionally, predicted surface characteristics, as described in more detail herein. It should be appreciated that the predicted surface characteristics can be used in different control and operation processes for the heater 302 or other components of the thermal system 10, as well as other processes.

[0037] Thus, in one form, the control loop module 102 determines the output control, or more particularly, controls power to the heater 302 based on, for example, the setpoint input and information related monitored conditions of the heater 302, such as monitored surface characteristics of the heater 302 at detectable locations (where sensors 304 are present) and predicted surface characteristics of the heater 302 at undetectable locations (where sensors 304 are not present, such as due to space limitations, visual limitations, operational conditions, environmental conditions, etc.). In one example, the surface characteristics are indicative of an amount of material buildup on one or more surfaces of the heater 302, such as coking or fouling of the surfaces.

[0038] FIGS. 2 and 3 illustrate an example of the heater 302 configured as a heat exchanger having a plurality of tubular heaters 310. In one form, the heater 302 is a direct electric heat exchanger, which includes an outer tube 312 surrounding the plurality of tubular heaters 310. The outer tube 312 includes an inlet 314 and an outlet 316. The fluid to be heated flows in and out of the outer tube 312 through the inlet 314 and the outlet 316. The tubular heaters 310 extend from the inlet 314 to the outlet 316 and in some examples have bends disposed proximate the outlet 316. As the fluid enters the inlet 314, the fluid is gradually heated by the tubular heaters 310 until the fluid leaves the outer tube 312 through the outlet 316. The fluid proximate the inlet 314 is cooler than the fluid proximate the outlet 316. It should be appreciated that the configuration of the heater 302 illustrated in FIGS. 2 and 3 is merely an example of one form of a heating arrangement. The systems and methods described hereinAttorney Docket No.: 0100TS-000017-WO-POA can be implemented with different types and configurations of heaters and heater elements.

[0039] Accordingly, a variety of different forms of heaters, sensors, control systems, and related devices and methods can be implemented for use in the fluid flow system 300. Many of the different forms can be combined with each other and may also include additional features specific to the data, equations, and configurations as described herein.

[0040] Referring to FIG. 4 in conjunction with FIG. 1 , the heater control module 104 may be configured to control operation of the heater 302, such as to modulate power to the heater 302. In some forms, the heater control module 104 includes a control device 400 and may be configured to control one or more heater elements or members of the heater 302 that takes into consideration predicted coking and fouling locations or buildup situations of the heater 302 or other components of the thermal system 10. The one or more heaters 302 are disposed in the fluid flow pathway, such as in the fluid flow conduits 306.

[0041] The buildup detection module 106 and a buildup prediction module 108 operate in combination with the heater control module 104 is some implementations (e.g. integrated into the process control system 100) for controlling operation of the heater 302 for improving operation of the heater 302 based on the predicted buildup information (e.g., predicted surface characteristics of undetectable locations of the heater 302). In one form, process control system 100 includes the heater control module 104, the buildup detection module 106, and the buildup prediction module 108. The heater control module 104 is configured to control the temperature of the heater 302 according to a desired temperature profile and based one or more control parameters, such as electrical characteristics of the heating apparatus, temperature setpoint, voltage, and power output to the heating apparatus.

[0042] The buildup detection module 106 is in communication with the sensors 304 and is configured to receive data indicative of one or more operating characteristic from the sensors 304 disposed at different locations along or across the heater 302. The buildup detection module 106 is further configured to compare the values of the one or more operating characteristic from among the plurality of locations and determine whether the value of the one or more operating characteristics for a particular location deviate from the values for the other locations. When the deviated value for the particular location exceeds or is below a threshold (depending on whatAttorney Docket No.: 0100TS-000017-WO-POA the operating characteristic is), the buildup detection module 106 then determines that a buildup exists on the heater 302 at that particular location. One specific example is that when elements within the same zone of control (all either on or off at the same time) diverge in temperature (higher) over time in one or more elements the divergence in temperature can indicate build up in that area.

[0043] The buildup detection module 106 in some examples include a buildup model configured to compare each of the temperature deviations associated with the plurality of heating zones. In one form, the buildup model may be configured to perform a zone-to-zone check in which the difference in temperature between adjacent heating zones is monitored. In this form, if the difference in temperature exceeds a material buildup threshold, the buildup model determines that a potential material buildup exists and determines that a corrective action is necessary based on an existence of the potential material buildup. The buildup detection module 106 then sends data relating to the buildup for the particular location to the heater control module 104.

[0044] The heater control module 104 in some examples controls the plurality of heating zones in response to the buildup information to slow the formation of buildups. For example, the heater control module 104 may power down a particular heater unit / heating zone that starts to experience a buildup and power up the other heater units to compensate for the reduced heat output from the particular heater unit in order to maintain a desired heat distribution along the length of the heater 302 while slowing the formation of the buildups.

[0045] The buildup detection module 106 may also send the buildup information to the buildup prediction module 108 to predict the level of buildup. The buildup prediction module 108 may predict the level of buildup based on the data from the sensors 304 and the operating parameters, such as heater duty cycle, from the heater control module 104. In some examples, as described in more detail herein, a determination is made as to whether the machine learning model satisfies one or more model conditions based on the simulation data, and in response to the machine learning model satisfying the one or more model conditions, predicting the surface characteristic of the heater 302 at a plurality of locations of the heater 302 based on the machine learning model (e.g., a plurality of locations that includes an undetectable location of the heater 302).Attorney Docket No.: 0100TS-000017-WO-POA

[0046] Thus, in some examples, the data points along with the operating parameters can be used to predict buildup levels at different undetectable locations of the heater 302. For example, knowing the coking and fouling or buildup from process metrics, one or more implementations described herein use a hybrid method to predict the unknown location buildup level based on known location buildup level. This “picture” of the buildup level at undetectable locations provides an indication of predictive maintenance. The buildup prediction module 108 may pre-determine a "threshold level" of a buildup. The buildup prediction module 108 may also pre-store data relating to different levels of buildups and a correlation between the different levels of buildups and the operating characteristics. The buildup prediction module 108 may determine the level of buildup based on the measured value of the operating characteristic and the pre-stored correlation between the different levels of buildups and the operating characteristics, and then predict the time when the buildup would reach the "threshold level" that would require preventive maintenance using the hybrid method. The buildup prediction module 108 can use the current or most recent data obtained by the sensors 304 during operation of the heater 302 to thereby predict further level of buildups, particularly in undetectable locations.

[0047] In operation, the heater control module 104 includes a control device 400 that in one form is configured to receive at least one input relating to the conditions of the heater 302, or other components of the thermal system 10, or the temperature along one or more fluid flow pathways and to modulate power to the heater 302 accordingly. The control device 400 uses the input(s) to control operation of the heater 302 (e.g., adjust or modify the power output of the heater 302). The control device 400 may include a power switch 402, a controller 404, and a lookup table 406. The controller 404 may be any type of controller, such as (proportional-integral-derivative) PID controller, a predictive feedback controller, a model-based controller, or any controller that can control heater power output.

[0048] Using the input(s), such as the temperature readings or other sensor data, the controller 404 may be used to control the heater power based on sensed conditions and / or desired or required operating characteristics. In one form, the controller 404 is configured to access the lookup table 406 and use a machine learning model 408 as described in more detail herein to control the heater 302, based at least in part on one or more surface characteristics of the heater 302. It should be noted that the input to the heater control module 104 is constantly monitored in someAttorney Docket No.: 0100TS-000017-WO-POA forms to provide dynamic control of the heater 302. For example, a power switch 402 of the control device 400 may be used to control the power to the heater 302. The power switch 402 may cause pulsed current / power to be supplied to the heater 302. Power is pulsed to control current supply to the heater 302. By using the pulsed power along with power measurements, combined with predicted surface characteristic information, improved control of the heater 302 can be performed.

[0049] As used herein, the term “model” should be construed to mean an equation or set of equations, a tabulation of values representing the value of a parameter at various operating conditions, an algorithm, a computer program or a set of computer instructions, a signal conditioning device or any other device that modifies the controlled variable (e.g., power to the heater) based on predicted / projected / future conditions, wherein the prediction / projection is based on a combination of a priori and in-situ measurements.

[0050] Referring again to FIG. 1 , the monitoring system 200 is configured to monitor the sensor data and use a simulation data generator 210 to generate simulation data based on the sensor data in combination with one or more inputs (e.g., control inputs) and a mathematical model 208. In one form, a relationship module 204 uses the mathematical model 208, which may be a mathematical model of the fluid flow system 300, to associate the sensor data with one or more known positions of the sensors 304 and generate simulation data. A training module 206 trains the machine learning model 408 based on the sensor data, which is then used to predict surface characteristic(s) of the heater 302 at a plurality of locations of the heater 302 based on the machine learning model 408 in response to the machine learning model 408 satisfying one or more model conditions based on the simulation data. As such, surface characteristics of the heater 302 at one or more undetectable locations are predicted using a coking and fouling hybrid detection method in some examples.

[0051] As described further herein, the process control system 100 receives sensor data from the heater 302 and predicted surface characteristic data from the monitoring system 200 and controls the power of the heater 302. As an example, the predicted surface characteristic data may indicate a predicted amount of material buildup on the surface of the heater at an undetectable location, which may then be used to control operation (e.g., adjust settings) of the heater 302 by the heater control module 104, perform maintenance of the heater 302, perform other operations,Attorney Docket No.: 0100TS-000017-WO-POA etc. In one form, the control loop module 102 may change the output control to the heater control module 104 to adjust or turn-off power to the heater 302 and may issue an alert (e.g., a message relating to the predicted surface characteristic).

[0052] Thus, based on the output control from the control loop module 102, the heater control module 104 controls power to the heater 302 using a hybrid method as described herein. In one form, the heater control module 104 is electrically coupled to a power source (not shown) and may include a power regulator circuit (not shown) to adjust the power from the power source to a selected power level and apply the adjusted power to the heater 302.

[0053] While the process control system 100 is shown and described utilizing particular controls, the process control system 100 may include other control systems. In addition, the process control system 100 may be configured to include a separate module for analyzing or performing different operations as described herein. Furthermore, while specific modules are provided as forming the process control system 100, the process control system 100 may include other modules for controlling operations of the heater 302, such as a diagnostic module for detecting abnormal operation of the heater 302.

[0054] In one form, the simulation data generator 210 is configured to generate simulation data based on or more of the inputs 202, the mathematical model 208 of the fluid flow system 300 (e.g., a physics-based model of the environment, a thermodynamic model of the environment, or a combination thereof), and the sensor data. The mathematical model 208 in one example associates the sensor data with one or more known positions of each of the one or more sensors 304 (e.g., based on known installed locations of the sensor(s) 304 within in the fluid flow system 300 and relative to the heater 302). The simulation data includes data associated with an undetectable position of the fluid flow system 300. For example, the simulation data includes one of an undetectable data value associated with buildup at the undetectable location.

[0055] In some forms, one or more mathematical models 208 are utilized. For example, a final math model in one form is an equation including setting temperature, pressure, flow rate, gas, process, etc. The values for each variable are the inputs, for example the input(s) 202, such that equation determines a coking or fouling level. As such, with the sensor 304 able to detect an accurate coking or fouling level, in one form, deep learning is used to predict a future coking or fouling conditionAttorney Docket No.: 0100TS-000017-WO-POA(particularly at an undetectable location), including a threshold level when maintenance is needed (e.g., three weeks later the system will need to be taken down, since the predicted coking or fouling level will be at a level requiring maintenance based on the deep learning model prediction).

[0056] In one form, lab testing and / or simulation facilitates generates an accurate mathematical model 208. As such, once defined, real-time data inputs (e.g., real-time sensor data) are analyzed by the trained machine learning model(s) 408 to predict the surface characteristics of the heater 302, such as surface characteristics of undetectable locations of the heater.

[0057] Thus, in operation, based on the known locations of the sensor data, simulation data is used to acquire data from other locations (where sensors 304 are not available), such as temperature / flow rate / pressure data, to evaluate / predict the overall coking / fouling system buildup conditions using machine learning and a deep learning method in one form. For example, simulation data is considered testing data to predict the overall coking and fouling for a large size thermal system (wherein training data is used from known sensors 304). The simulation can be used to acquire all dimensional process metrics data based on an input set of control parameters (e.g., power supply, setting temperature, setting flow rate, setting pressure, etc.) and with metrics from known locations sensor signals processed (e.g., used in the training process). With one or more herein described examples, simulation data can thereby be separated from real testing data as two parts, wherein real sensor data is used as training data for the machine learning / deep learning model (e.g., the machine learning model(s) 408); while simulation data is used as testing data for the machine learning / deep learning model (e.g., the machine learning model(s) 408). As such, all locations of coking and fouling or buildup situations of the thermal system 10 (e.g., in the heater 302) can be predicted as described herein, including for undetectable positions that do not have sensor(s) 304 for monitoring. In one form, a hybrid method for labeling data to determine coking and fouling or buildup is thereby provided.

[0058] It should be noted that in some examples, the simulation results are validated, for example, when the results are close to the real signal or deviate from the real signal (e.g., a significant deviation). Additionally, the simulation results from different situations, such as high setting / low setting, different hydrocarbon fluid / gas, different production processes, etc. are also similarly validated in some examples. In one or more examples, the simulated data is validated based on a comparisonAttorney Docket No.: 0100TS-000017-WO-POA between the simulated data and the sensor data and the mathematical model is adjusted based on the difference between the simulation data and the sensor data.

[0059] It should also be noted that in some example, other methods (to confirm the known locations buildup level from known sensor signals) are integrated or combined with one or more herein described methods. For example, the coking and fouling or buildup are then known from the process metrics. And with one or more examples of the hybrid method described herein, the unknown locations buildup level is predicted based on the known location buildup level.

[0060] To perform the functions described herein, the process control system 100 and / or the monitoring system 200 may include one or more databases (such as having the look-up table 406) for storing data, such as historical data, validation data, measured data, simulation data, sensor data, etc. With the data, the relationship module 204 is configured to determine a relationship or more particularly, the mathematical model(s) 208 between the sensor data and the known positions of the sensor(s) 304. That is, using suitable data modeling techniques, the relationship module 204 identifies trends / relationship between the data in some examples, which is used as an input to the training module 206 to generate the machine learning model(s) 408. As an example, the relationship module 204 and / or the training module 206 is configured to perform various supervised machine learning routines, such as a linear, nonlinear, or logistic regression routine, a random forest routine, a support vector routine, among others, to generate one or more models (as the model(s) 208, 408) during a data simulation data generation routine and / or a training routine, respectively. In some examples, deep learning models, neural networks, artificial neural network (ANN) regression, convolutional neural network (CNN) regression, etc. can be used to perform various learning routines. It should be noted that one or more routines can be performed online or offline, and may be performed in real-time, and can be performed locally (e.g., at the thermal system 10) or at a remote location (e.g., cloud processing). It should be understood that the relationship module 204 and / or the training module 206 may perform unsupervised or semi-supervised learning routines to generate the model(s) 208, 408 (e.g., a clustering routine).

[0061] Using the machine learning model(s) 408 to determine if one or more model conditions are satisfied (e.g., the machine learning model(s) 408 satisfies the one or more model conditions based on the simulation data and in response to the machine learning model(s) 408 being associated with an accuracy indicator that isAttorney Docket No.: 0100TS-000017-WO-POA greater than a threshold indicator), surface characteristics of the heater 302 at a plurality of locations are predicted (e.g., one or more undetectable or non-monitored locations of the heater 302) based on the machine learning model(s) 408. As such, real-time coking / fouling buildup can be predicted using a combination of machine learning and deep learning as described herein. While the modules 204, 206 are provided as part of the monitoring system 200, once defined, the modules 204, 206 may be employed with the process control system 100 to improve control of the operation of the heater 302.

[0062] It should be understood that to perform the functionality described herein, the modules 204, 206 may iteratively perform a machine learning training routine to associate the outputs of the model(s) 208, 408 with defined locations of the heater 302. As an example, the machine learning training routine may be a supervised learning routine which does not include tagging or labeling data with conditions of the heater 302 when using the hybrid method of various examples as described herein.

[0063] While the process control system 100, the monitoring system 200, and the heater system 300 are shown as separate systems, it should be understood that at least one of the process control system 100, the monitoring system 200, and the heater system 300 may be implemented as a single system. For example, the monitoring system 200 may be provided as part of the process control system 100. In one form, the monitoring system 200 may be disposed at a different location from that of the process control system 100 and may communicate with the process control system 100 via a wireless communication network, one or more edge computing devices disposed within the thermal system 10, or a combination thereof.

[0064] With reference to FIG. 5, a flowchart illustrating an example routine 500 (e.g., a hybrid method) for monitoring a surface characteristic of a heater (e.g., a surface characteristic indicating an amount of a material buildup on a surface of the heater 302) provided in a fluid flow conduit (e.g., the fluid flow conduit 306) of a fluid flow system (e.g., the fluid flow system 300) is shown. At 502 sensor data is obtained, for example from one or more of the sensors 304 of the fluid flow system 300 as described in more detail herein. At 504, simulation data is generated based on one or more inputs, a mathematical model of the fluid flow system 300, and the sensor data, such as by the simulation data generator 210 as described in more detail herein. For example, the simulation data indicates a simulated electrical characteristic of the heater 302 (e.g., a simulated voltage of the heater, a simulated current of theAttorney Docket No.: 0100TS-000017-WO-POA heater, or a combination thereof), a simulated inlet temperature of the fluid flow conduit 306 (e.g., one or more fluid flow conduits), a simulated outlet temperature of the fluid flow conduit 306, a simulated inlet pressure of the fluid flow conduit 306, a simulated outlet pressure of the fluid flow conduit 306, a simulated fluid flow rate, or a combination thereof. It should be appreciated that simulated results (e.g., simulated data) in various examples is acquired at locations other than the inlet and outlet. For example, in some implementations, a three-dimensional (3D) simulation is performed that allows for a determination of simulation data at any approximate location of the heater, such as simulated temperature, flow rate, etc. at the approximate location of the heater. In some examples, simulation operations to locate one or more unrun conditions as described herein provide missing data for one or more running conditions that cannot otherwise be duplicated. The simulation data also provides information relating to coking or fouling conditions in some examples as described in more detail herein.

[0065] In one form, the mathematical model of the fluid flow system 300 associates the sensor data with a known position of each of the one or more sensors 304 as described herein, and wherein the simulation data includes at least one data value (e.g., undetectable value, non-determined running condition value, nondetermined idle state value, etc.) associated with an undetectable position of the fluid flow system 300.

[0066] At 506, a machine learning model (e.g., the machine learning model 408, such as a linear regression model, a nonlinear regression model, or a combination thereof) is trained based on the sensor data. For example, the training module 206 performs model training processes as described in more detail herein. At 508, a determination is made whether the machine learning model 408 satisfies one or more model conditions based on the simulation data, such as being associated with an accuracy indicator that is greater than a threshold accuracy indicator as described herein. For example, an accuracy value (e.g., likelihood that the result is accurate) is used to determine whether the one or more model conditions are satisfied.

[0067] At 510, in response to the machine learning model 408 satisfying the one or more model conditions, the surface characteristic of the heater 302 at a plurality of locations of the heater 302 are predicted based on the machine learning model 408 as described herein. In one form, the plurality of locations includes an undetectable location of the heater 302.Attorney Docket No.: 0100TS-000017-WO-POA

[0068] With reference to FIG. 6, a flowchart illustrating an example routine 600 (e.g., a hybrid method) for monitoring a surface characteristic of a heater (e.g., a surface characteristic indicating an amount of a material buildup on a surface of the heater 302) provided in a fluid flow conduit (e.g., the fluid flow conduit 306) of a fluid flow system (e.g., the fluid flow system 300) is shown. At 602, a first set of sensor data (e.g., a set of sensor data that is different from the sensor data used to train the machine learning model) is obtained from one or more of the sensors 304 of the fluid flow system 300 as described in more detail herein. At 604, the machine learning model (i.e., a trained machine learning model as described herein) predicts a surface characteristic of the heater at an undetectable location of the heater based on the first set of sensor data. At 606, the thermal system 10 selectively performs a corrective action based on the surface characteristic of the heater 302. As an example, when the machine learning model detects a threshold level of buildup, the process control system 100 may broadcast a notification to a user interface (Ul) indicating the presence, location, and / or amount of buildup. As another example, when the machine learning model detects a threshold level of buildup, the process control system 100 (e.g., the heater control module 104) may adjust the operation of the heater 302 to accommodate for the material buildup. As yet another example, when the machine learning model detects a threshold level of buildup, the process control system 100 may automatically initiate a maintenance routine, which may include temporarily ceasing operation of at least a portion of the thermal system 10 and providing instructions (e.g., on a Ul) for remediating the material buildup.

[0069] Unless otherwise expressly indicated herein, all numerical values indicating mechanical / thermal properties, compositional percentages, dimensions and / or tolerances, or other characteristics are to be understood as modified by the word “about” or "approximately" in describing the scope of the present disclosure. This modification is desired for various reasons including industrial practice, material, manufacturing, and assembly tolerances, and testing capability.

[0070] As used herein, the phrase at least one of A, B, and C should be construed to mean a logical (A OR B OR C), using a non-exclusive logical OR, and should not be construed to mean “at least one of A, at least one of B, and at least one of C.”

[0071] In this application, the term “controller” and / or “module” may refer to, be part of, or include: an Application Specific Integrated Circuit (ASIC); a digital,Attorney Docket No.: 0100TS-000017-WO-POA analog, or mixed analog / digital discrete circuit; a digital, analog, or mixed analog / digital integrated circuit; a combinational logic circuit; a field programmable gate array (FPGA); a processor circuit (shared, dedicated, or group) that executes code; a memory circuit (shared, dedicated, or group) that stores code executed by the processor circuit; other suitable hardware components (e.g., op amp circuit integrator as part of the heat flux data module) that provide the described functionality; or a combination of some or all of the above, such as in a system-on-chip.

[0072] The term memory is a subset of the term computer-readable medium. The term computer-readable medium, as used herein, does not encompass transitory electrical or electromagnetic signals propagating through a medium (such as on a carrier wave); the term computer-readable medium may therefore be considered tangible and non-transitory. Non-limiting examples of a non-transitory, tangible computer-readable medium are nonvolatile memory circuits (such as a flash memory circuit, an erasable programmable read-only memory circuit, or a mask readonly circuit), volatile memory circuits (such as a static random access memory circuit or a dynamic random access memory circuit), magnetic storage media (such as an analog or digital magnetic tape or a hard disk drive), and optical storage media (such as a CD, a DVD, or a Blu-ray Disc).

[0073] The apparatuses and methods described in this application may be partially or fully implemented by a special purpose computer created by configuring a general-purpose computer to execute one or more particular functions embodied in computer programs. The functional blocks, flowchart components, and other elements described above serve as software specifications, which can be translated into the computer programs by the routine work of a skilled technician or programmer.

[0074] The description of the disclosure is merely exemplary in nature and, thus, variations that do not depart from the substance of the disclosure are intended to be within the scope of the disclosure. Such variations are not to be regarded as a departure from the spirit and scope of the disclosure.

Claims

Attorney Docket No.: 0100TS-000017-WO-POACLAIMSWhat is claimed is:1 . A method for monitoring a surface characteristic of a heater provided in a fluid flow conduit of a fluid flow system, the method comprising: obtaining sensor data from one or more sensors of the fluid flow system; generating simulation data based on one or more inputs, a mathematical model of the fluid flow system, and the sensor data, wherein the mathematical model of the fluid flow system associates the sensor data with a known position of each of the one or more sensors, and wherein the simulation data includes at least one data value associated with an undetectable position or running condition of the fluid flow system; training a machine learning model based on the sensor data; determining whether the machine learning model satisfies one or more model conditions based on the simulation data; and predicting, in response to the machine learning model satisfying the one or more model conditions, the surface characteristic of the heater at a plurality of locations of the heater based on the machine learning model, wherein the plurality of locations includes an undetectable location of the heater.

2. The method of Claim 1 , wherein the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, a conduit layout or dimension, a gas component, or a combination thereof.

3. The method of Claim 1 , wherein the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, a simulated running condition of the heater, or a combination thereof.Attorney Docket No.: 0100TS-000017-WO-POA4. The method of Claim 3, wherein the simulated electrical characteristic of the heater includes a simulated voltage of the heater, a simulated current of the heater, or a combination thereof.

5. The method of Claim 1 , wherein the sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof.

6. The method of Claim 5, wherein the electrical characteristic of the heater includes a voltage of the heater, a current of the heater, a power of the heater, a leakage current of the heater, a resistance of the heater, or a combination thereof.

7. The method of Claim 1 , wherein the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof.

8. The method of Claim 1 , wherein the machine learning model satisfies the one or more model conditions in response to the machine learning model being associated with an accuracy indicator that is greater than a threshold accuracy indicator.

9. The method of Claim 1 , wherein the heater is one of a heat exchanger, a fired heater, a layered heater, a tubular heater, a cartridge heater, a polymer heater, a flexible heater, a line heater, and a cable heater.

10. The method of Claim 1 , wherein the surface characteristic indicates an amount of material buildup on a heater surface.

11. The method of Claim 1 , wherein the mathematical model is a physicsbased model of an environment, a thermodynamic model of the environment, or a combination thereof.

12. The method of Claim 1 further comprising validating simulated data based on a comparison between the simulated data and the sensor data.Attorney Docket No.: 0100TS-000017-WO-POA13. The method of Claim 12 further comprising selectively adjusting the mathematical model based on a difference between the simulation data and the sensor data.

14. A system for monitoring a surface characteristic of a heater surface of a heater of an environment, the environment further comprising a fluid flow conduit, the system comprising: one or more processors; and one or more non-transitory computer-readable mediums comprising instructions that are executable by the one or more processors, wherein the instructions comprise: generating simulation data based on one or more inputs and a mathematical model of the environment; obtaining sensor data from one or more sensors of the environment; training a machine learning model based on the sensor data; determining whether the machine learning model satisfies one or more model conditions based on the simulation data; and predicting, in response to the machine learning model satisfying the one or more model conditions, the surface characteristic of the heater surface based on the machine learning model, wherein the surface characteristic indicates an amount of material buildup on the heater surface.

15. The system of Claim 14, wherein the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, or a combination thereof.

16. The system of Claim 14, wherein the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulatedAttorney Docket No.: 0100TS-000017-WO-POA inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, or a combination thereof.

17. The system of Claim 14, wherein the sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof.

18. The system of Claim 14, wherein the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof.

19. The system of Claim 14, wherein the machine learning model satisfies the one or more model conditions in response to the machine learning model being associated with an accuracy indicator that is greater than a threshold accuracy indicator.

20. The system of Claim 14, wherein the mathematical model is a physicsbased model of the environment, a thermodynamic model of the environment, or a combination thereof.21 . A method for monitoring a surface characteristic of a heater provided in a fluid flow conduit of a fluid flow system, the method comprising: obtaining a first set of sensor data from one or more sensors of the fluid flow system, wherein the first set of sensor data corresponds to a detectable location of the fluid flow system; predicting, using a machine learning model, a surface characteristic of the heater at an undetectable location of the heater based on the first set of sensor data; and selectively performing a corrective action based on the surface characteristic of the heater.

22. The method of Claim 21 , wherein the first set of sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flowAttorney Docket No.: 0100TS-000017-WO-POA conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof.

23. The method of Claim 22, wherein the electrical characteristic of the heater includes a voltage of the heater, a current of the heater, a power of the heater, a leakage current of the heater, a resistance of the heater, or a combination thereof.

24. The method of Claim 21 , wherein the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof.

25. The method of Claim 21 , wherein the surface characteristic indicates an amount of material buildup on a heater surface.

26. The method of Claim 21 , wherein the heater is one of a heat exchanger, a fired heater, a layered heater, a tubular heater, a cartridge heater, a polymer heater, a flexible heater, a line heater, and a cable heater.

27. The method of Claim 21 , wherein the machine learning model is trained based on a second set of sensor data from the one or more sensors of the fluid flow system.

28. The method of Claim 27, wherein the machine learning model is validated based on simulation data that corresponds to an undetectable position of the fluid flow system or a running condition of the fluid flow system.

29. The method of Claim 28, wherein the simulation data is based on one or more inputs, a mathematical model of the fluid flow system, and the second set of sensor data, wherein the mathematical model of the fluid flow system associates the second set of sensor data with a known position of each of the one or more sensors.

30. The method of Claim 29, wherein the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flowAttorney Docket No.: 0100TS-000017-WO-POA conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, or a combination thereof.

31. The method of Claim 28, wherein the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, or a combination thereof.

32. A system for monitoring a surface characteristic of a heater provided in a fluid flow conduit of a fluid flow system, the system comprising: one or more processors; and one or more non-transitory computer-readable mediums comprising instructions that are executable by the one or more processors, wherein the instructions comprise: obtaining a first set of sensor data from one or more sensors, wherein the first set of sensor data corresponds to a detectable location of the fluid flow system; predicting, using a machine learning model, a surface characteristic of the heater at an undetectable location of the heater based on the first set of sensor data; and selectively performing a corrective action based on the surface characteristic of the heater.

33. The system of Claim 32, wherein the first set of sensor data indicates an electrical characteristic of the heater, an inlet temperature of the fluid flow conduit, an outlet temperature of the fluid flow conduit, an inlet pressure of the fluid flow conduit, an outlet pressure of the fluid flow conduit, a fluid flow rate, or a combination thereof.

34. The system of Claim 33, wherein the electrical characteristic of the heater includes a voltage of the heater, a current of the heater, a power of the heater, a leakage current of the heater, a resistance of the heater, or a combination thereof.Attorney Docket No.: 0100TS-000017-WO-POA35. The system of Claim 32, wherein the machine learning model is a linear regression model, a nonlinear regression model, or a combination thereof.

36. The system of Claim 32, wherein the surface characteristic indicates an amount of material buildup on a heater surface.

37. The system of Claim 32, wherein the heater is one of a heat exchanger, a fired heater, a layered heater, a tubular heater, a cartridge heater, a polymer heater, a flexible heater, a line heater, and a cable heater.

38. The system of Claim 32, wherein the machine learning model is trained based on a second set of sensor data from the one or more sensors of the fluid flow system.

39. The system of Claim 38, wherein the machine learning model is validated based on simulation data that corresponds to an undetectable position of the fluid flow system or a running condition of the fluid flow system.

40. The system of Claim 39, wherein the simulation data is based on one or more inputs, a mathematical model of the fluid flow system, and the second set of sensor data, wherein the mathematical model of the fluid flow system associates the second set of sensor data with a known position of each of the one or more sensors.41 . The system of Claim 40, wherein the one or more inputs include a power setpoint of the heater, a setpoint inlet temperature of the fluid flow conduit, a setpoint outlet temperature of the fluid flow conduit, a setpoint inlet pressure of the fluid flow conduit, a setpoint outlet pressure of the fluid flow conduit, a setpoint fluid flow rate, or a combination thereof.

42. The system of Claim 39, wherein the simulation data indicates a simulated electrical characteristic of the heater, a simulated inlet temperature of the fluid flow conduit, a simulated outlet temperature of the fluid flow conduit, a simulated inlet pressure of the fluid flow conduit, a simulated outlet pressure of the fluid flow conduit, a simulated fluid flow rate, or a combination thereof.

Images