WO2019050726A1

WO2019050726A1 - Method and system to estimate boiler tube failures

Info

Publication number: WO2019050726A1
Application number: PCT/US2018/048445
Authority: WO
Inventors: Sivaramanivas Ramaswamy; Debasish Mishra; Vasant Kumar JAIN; Ashwin Raman; Sreenivasa Rao GUBBA
Original assignee: General Electric Company
Priority date: 2017-09-08
Filing date: 2018-08-29
Publication date: 2019-03-14

Abstract

According to some embodiments, a system (600) and method (400) are provided to estimate boiler tube (110) failures. The method comprises receiving fuel properties (S210) associated with a type of fuel being used to fire a boiler. The boiler comprises a plurality of tubes (130). Material properties (S220) associated with the plurality of tubes and operational data (S220a) associated with the boiler are also received (S220). A first model (408) based on the received fuel properties, the received material properties and operational data is determined. Tuning parameters (410) associated with the first model based on first validation data may be modified using Bayesian learning to determine a second model (412). Based on second operational data input into the second model, a likelihood of failure associated with the plurality of tubes is automatically determined and an alert associated with the boiler may be transmitted.

Description

METHOD AND SYSTEM TO ESTIMATE BOILER TUBE FAILURES

BACKGROUND

[0001] Boilers such as, but not limited to, coal-fired boilers, gas-fired boilers and oil-fired boilers use a fuel source to heat water. Boilers may suffer from problems such as, but not limited to, corrosion, erosion of metal, fatigue, weld failures, deposits of material, and creep. Early detection and diagnosis of problems associated with a boiler may help minimize downtime of the boiler as well as prevent secondary damage. For example, when a boiler under pressure fails, the boiler may release water causing damage to its surroundings such as other boiler components of the building where the boiler is housed. This may create longer downtimes. It would therefore be desirable to provide a system to determine problems or potential problems associated with a boiler as early as possible to provide time for a repair crew to address the determined or potential problems associated with the boiler prior to a boiler causing damage to its surrounding.

SUMMARY

[0002] Some embodiments described herein relate to a system and method to aid in estimating boiler tube failures. The method comprises receiving fuel properties associated with a type of fuel being used to fire a boiler. First model based on the received fuel properties, received material properties, and operational data may be determined with its tuning parameters. A second model may comprise the first model with the modified tuning parameters based on Bayesian learning. Based on additional inputs either from operational data and /or validation data -into the second model, a likelihood of failure associated with the boiler tubes may be automatically determined and an alert associated with the boiler may be transmitted. [0003] A technical advantage of some embodiments disclosed herein are improved systems and methods for the early alerting of potential problems associated with a boi ler prior to secondary damage being done to the boiler and the boiler's surroundings.

BRIEF DESCRIPTION OF THE DRAWINGS

[0004] FIG. 1 is an overview of a boiler monitoring system in accordance with some embodiments.

[0005] FIG. 2 illustrates a method in accordance with some embodiments.

[0006] FIG. 3 illustrates boiler failure modes according to some embodiments.

[0007] FIG. 4 illustrates a process according to some embodiments.

[0008] FIG. 5 illustrates a typical result according to some embodiments.

[0009] FIG. 6 illustrates a system according to some embodiments.

[0010] FIG. 7 illustrates a portion of a database according to some embodiments.

DETAILED DESCRIPTION

[0011] In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of embodiments. However, it will be understood by those of ordinary skill in the art that the embodiments may^¬ be practiced without these specific details. In other instances, well-known methods, procedures, components and circuits have not been described in detail so as not to obscure the embodiments.

[0012] One or more specific embodiments of the present invention will be described below. In an effort to provide a concise description of these embodiments, all features of an actual implementation may not be described in the specification. It should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another. Moreover, it should be appreciated that such a development effort might be complex and time consuming, but would nevertheless be a routine undertaking of design, fabrication, and manufacture for those of ordinary skill having the benefit of this disclosure.

[0013] The present embodiments described herein relate to a novel system and method to estimate failures associated a boiler and, more specifically, the tubing associated with a boiler. Detection of failures in boilers can be challenging. Erosion of metal that makes up boiler tubes is often a cause for forced outages. For example, erosion is prevalent in boilers that burn high ash content coals and it may be difficult to determine when a boiler and/or its tubes may be about to fail (e.g., an imminent failure). An amount of erosion may be based on flow properties of flue gas (e.g., gas including ash particles) which may vary based on current operational load conditions as well as variations in coal blends. These operational conditions can change on a daily basis and this creates uncertainty in the prediction of erosion rates. In addition, boiler tubes may be configured in banks of tubes where each bank comprises a plurality of rows. Configurations, such as these, may create a challenge for the inspection of boiler tubes as well as being able to physically monitor w ll loss and implement any repairs or replacement of tubes. The embodiments described herein relate to an intelligent method to predict erosion rates and/or wall loss associated with boiler tubes as well as a location within a boiler. When an imminent failure is determined, a user may be alerted and, as such, boiler operation and maintenance may be improved by shorter downtimes and less damage to boiler components or the boiler's surroundings (e.g., water or structural damage to walls, ceilings, floors).

[0014] Now referring to FIG. 1, an embodiment of a system 100 is illustrated. As illustrated in FIG. 1, the system 100 comprises a boiler 1 10 and a monitoring platform 120. The boiler 110 may comprise any fuel-fired boiler such as, but not limited to, a coal-fired boiler, an oil-fired boiler or a gas-fired boiler. The boiler 110 may provide heat based on circulating water heated externally by a fire through a plurality of tubes 130. Fuel (e.g., coal, oil or gas) may be burned inside a furnace which creates hot gas to heat water in the plurality of tubes. In some boilers, water may surround a heat source (e.g., a fire) and gases from combustion may pass through tubes within the water space.

[0015] In a case of a coal-fired boiler, burning of coal may produce coal ash. Coal ash may include a number of by-products produced from the burning coal, such as fly ash (e.g., a powdery material composed mostly of silica), bottom ash ( e.g., an angular ash particle that is too large to be carried up into the smoke stacks and may form in the bottom of the coal furnace), boiler slag (e.g., a molten bottom ash) and flue gas material (e.g., a sludge consisting of calcium sulfite or calcium sulfate or a dry powered material that is a mixture of sulfites and sulfates). Coal ash, collectively, may have a detrimental effect on a boiler's tubing by wearing away material from the tubing causing the tubing to become thinner and making the tubing more susceptible to rupture.

[0016] The monitoring platform 120 may receive data from a plurality of sensors associated with the boiler (e.g., time-series) or may receive periodic data that is input into the monitoring platform. For example, in one embodiment, the monitoring platform 120 may receive data from ultrasonic sensors that monitor wall thickness of one or more tubes associated with the boiler. In some embodiments, the monitoring platform 120 may receive multiple streams of data (e.g., one stream for each tube) and the monitoring platform 120 may compare the data to an existing model to indicate the likelihood of one or more of the tubes failing as well as a location of a potential failure. The process of estimating a likelihood of one or more of the tubes failing will now be described in more detail with respect to FIG. 2.

[0017] FIG. 2 illustrates a method 200 that might be performed by the monitoring platform 120 of the system 100 described with respect to FIG. 1. The flow chart described herein does not imply a fixed order to the steps, and embodiments of the present invention may be practiced in any order that is practicable. Note that any of the methods described herein may be performed by hardware, software, or any combination of these approaches. For example, a non-transitory computer- readable storage medium may store thereon instructions that when executed by a machine result in performance according to any of the embodiments described herein.

[0018] At S210, fuel properties associated with a type of fuel being used to fire a boiler may be received. Since fuel may vary depending on the load, the received fuel properties may be based on a chemical makeup of each individual load of fuel. For example, each load of oil may have a different level of sulfur and each load of coal may produce different amounts of ash (e.g., ash percentage) and different types of ash. The chemical makeup may be determined by a laboratory and entered into the monitoring platform or may be provided with each load of fuel and then entered into the monitoring platform. Regarding the use of coal, the received fuel properties may further comprise a velocity of ash particles within the plurality of tubes and an angle of input associated with the ash particles. The velocity of ash particles and the angle of input may be measured by one or more sensors or may be based on a table of known angles and velocities which is stored in a repository of the monitoring platform.

[0019] As described above, a boiler may comprise a plurality of tubes for distributing heat and at S220, material properties associated with the plurality of tubes may be received. The plurality of tubes may be comprised of one of a variety of metals. For example, the plurality of tubes may comprise steel, copper, stainless steel, galvanized steel or any other metal substance capable of transferring heat to or from a liquid. Tubes may be classified by their outside diameter measurement and thickness. For example, a tube could have an outer diameter of 10 mm and a thickness of 2-mm. The received material properties may comprise a spacing of the tubes as well as a type of material (e.g., copper, steel) and a thickness of the tubes (e.g., 2 mm). As liquids and/or heat pass in or around the tubes, the tubes may begin to erode. For example, the outside of the tubes may erode due to the abrasiveness of coal ash against the tubes or the inside of the tubes may erode due to acids or due to friction caused by material (e.g., water, ash, etc.) moving through the tubes. Furthermore, the material properties may include historical data associated with a type of material being used for the boiler tubes. In some embodiments, the historical data may be related to oxidation of the type of metal, a hardness and or yield strength of the type of metal based on temperature, data associated with reactions to other substances (e.g., acids) or potential to hydrogen (pH) levels.

[0020] At S220a, operational data of the boiler may be received. Depending upon the required plant load factor, the coal flow rate, the air-to-fuel, ratio and related combustion parameters may be set. This may dictate a number and a configuration of mills that are operational.

[0021] At S230, a first model that is based on the received fuel properties received material properties and operational parameters, may be determined. In some embodiments, the first model may be based on a Bayesian model. According to some embodiments, a Bayesian model comprises a model that draws its inferences from a posterior distribution. For example, the Bayesian model may utilize a distribution of historical data as well as a likelihood of an event occurring which are related by Bayes' theorem. The first model may be

K xⁿp_m p™<ySinfif

based on the formula

wiiere £ is erosion rate, xⁿ represents the erosion index of ash, wherein x represents the weight percentage of silica in ash. p_m is density of the tube material, p_P is density of the impinging ash particle, V is the velocity of impact, SinP is sine of the angle of impact, o_y is the yield strength of the tube material. The factor 'n' the exponent of x, that is determined from lab experiments, represents the contribution of ash-content.

Factor K may represent the tuning parameter to be estimated, taking into account the effect of size and shape variations of the impacting particles. All the listed parameters may assume a distribution in real-life conditions, instead of single value. [0022] In another embodiment, the first model may be based on the following formula:

Where, OD' represents the outer diameter of the eroded tube, OD and ID are the outer and inner diameter of a new tube, respectively. FACTi may comprise a constant derived from OD and ID, FACT2 may comprise a constant that describes the cumulative erosion, and K may represent a tuning parameter to be estimated based on an effect of size and shape variations of impacting particles. For a given set of input parameters and their distributions, and assuming a prior distribution for tuning parameter K, the posterior distribution of may be determined from Bayes' Theorem.

[0023] Given a set of input parameters and their distributions, as described by variables of FACTi, FACT2, and OD¹ above, except variable K, and assuming a prior distribution for K, the posterior distribution of K may be determined from Bayes' Theorem. Given a set of inputs parameters are represented by vector 'D', and their distributions, erosion prediction may be formulated as a parameter estimation problem for parameter K in this instance.

[0024] Taking the likelihood function P(D|K, σ²) as a normal distribution, it may represent the probabilistic statement on measured values D given the predicted values determined from, for example, equation 1, and parameter K. The variance σ² represents the variance arising from the uncertainties of modeling and measurement uncertainties.

[0025] Using Bayes' Theorem, the posterior distribution of the parameters for a given set of measured D, can be constructed as, [0026] In this case i = 1. Ρπ (K), represents the joint prior distribution of and σ². The variance term may be ignored in the above expression for brevity.

[0027] For each parameter to be estimated, the marginal posterior distribution can be obtained by integrating above equation with respect to the rest parameters and σ² over the domain of interest. Monte-Carlo Markov Chain approach is used to solve for the posterior distributions, with multi-dimensional parameter sets as above. In an embodiment, the prior parameter distribution ¾¾, taken to be normal distribution.

[0028] In some cases, the first model may not be an accurate representation of the boiler since every boiler and every load of fuel is slightly different. Therefore, the first model may be adjusted through the use of tuning parameters which may be adjusted based on validation data. Thus, a first set of validation data associated with the boiler may be received. This first validation data may be compared to the first model output and the first model may be adjusted by changing the tuning parameters to better reflect the validation data. Thus, at S240, tuning parameters associated with the first model may be modified based on the first validation data to determine a second model. As such, using the first validation data associated with the boiler, the first model may be adjusted to more accurately represent that particular boiler using a particular load of fuel. This new model may be saved as the second model.

[0029] As more operational data is input into the second model, the second model may learn and self-adjust. In some embodiments, operational data input into the second model may indicate a likelihood of failure associated with the plurality of tubes and this likelihood may automatically be determined at S250.

[0030] At S260, an alert associated with the boiler may be transmitted. The alert may be transmitted to a repair crew and the alert may comprise an indication that a boiler has a high likelihood of failure and that tubing should be inspected and/or replaced prior to any secondary damage occurring. While the method of FIG. 2 has been described with respect to the erosion of material associated with boiler tubes, the aforementioned method may be applied to other failure modes as well. In some embodiments, the alert may be associated with a dashboard (e.g., a graphical user interface) that allows a user to visualize real-time tube life. For example, the dashboard may comprise a graphical user interface that illustrates a gauge to display millimeters of material removed per day. Furthermore, the dashboard may illustrate a gauge to display a remining useful life for a tube. Operational data may be based on inspection data that indicates the particular tubes being inspected. Therefore, the dashboard may also indicate a location of the tube or tubes that are in need of inspection and/or replacement.

[0031] Now referring to FIG. 3, failure modes 300 associated with boilers tubes are illustrated. For example, the failure modes 300 may comprise corrosion 302, erosion 304, corrosion-fatigue 306, creep 308, over-heating 310, deposits 312 and welding defects 314. Corrosion 302 may comprise a natural destruction of metal by chemical and/or electrochemical reaction with its environment. Erosion 304 may comprise a surface processes that removes material through the flow of a fluid, a gas (e.g., flue gas) or solvent. Corrosion-fatigue 306 may comprise fatigue in a corrosive environment which is a degradation of a material based on both corrosion and repeated or fluctuating stresses. Creep 308 may comprise a time-dependent deformation at elevated temperature and constant stress. Overheating 310 may comprise a boiler's temperature rising higher than a determined temperature setting for a prolonged period of time causing an increase in pressure. Deposits 312 may comprise contaminants that can form in a boiler include calcium, magnesium, iron, copper, aluminum, silica, silt and oil. Deposits 312 can create scale that is crystallized directly onto tube surfaces or become sludge deposits that are transported to a surface of the tube and may be baked in place if exposed to high heat. Deposits 312 can cause material damage by insulating a heat transfer path from a boiler flame as well as restrict boiler water circulation. Welding defects 314 may relate to the use of dissimilar metals causing an entire tube to fail across a circumference of the tube. Each of the aforementioned failures modes 300 can be modeled based on physical properties that are unique to the particular failure mode 300.

[0032] Now referring FIG. 4, an embodiment of a process flow 400 is illustrated. The process flow 400 may illustrate a process to estimate failure associated with boiler tubes based on a particular failure mode. Furthermore, in some embodiments, the process flow 400 may illustrate a method to predict failure in boiler tubes using physics based kernels and machine learning. In the illustrated embodiment, machine learning may estimate location specific failures in boilers based on adjusting tuning parameters. These tuning parameters may factor in variations in fuel properties (e.g., coal, oil or gas), a boiler load factor, size and shape of ash particles and other characteristics that are unique to an individual boiler design. The process flow 400 may further be based on Bayesian learning techniques, and self-tuning using operational data and by employing prior distributions of failure attributes (e.g., historical data associated with previous failures of the individual boiler). In one embodiment, the process flow 400 may provide information for inspection planning and planned outages as opposed to the present practice of inspecting a boiler and the boiler's tubes after a forced outage has already occurred.

[0033] At 402, information associated with a physics based kernel may be received. The information associated with the physics based kernel may comprise information related to a particular failure mode (e.g., creep, corrosion, overheating, etc.) and each failure mode may comprise properties that are unique to the particular failure mode. In a case of erosion in a coal-fired boiler, coal properties, a blend ratio, a velocity of ash particles within a plurality of tubes and an angle of input associated with the ash particles may be received. The information may be received at a monitoring platform where it may be used to generate a model of the boiler. Velocity profiles of ash may be generated from a table of location specific velocity distributions associated with different boiler operating conditions. The velocity profile table may be stored in a database or repository associated with the monitoring platform such as monitoring platform 120.

[0034] At 404 operational data is received. Operational data may comprise information associated with the particular boiler that is being modeled. For example, operational data may comprise information about the boiler tubes that are specific to the boiler being modeled such as a type of material, an outside diameter measurement and thickness of one or more boiler tubes. Baseline wall thickness of tubes may be generated from ultrasonic thickness measurements or may be based on nominal wall thickness data. Furthermore, the operational data may also include operation data of the boiler such as an average operating temperature, boiler load, an amount of up time for the boiler, etc.

[0035] At 406, prior distributions may be received. Prior distributions may comprise historical data associated with the particular boiler. For example, the prior distributions may comprise a length of time that a boiler operated before experiencing a failure, historical operating temperatures, historical plant loads, environmental conditions, materials used for construction, historical tube thickness data and any other known historical data regarding features or operation of the particular boiler. In some embodiments, the prior distributions may comprise data based upon on lab and or computational fluid dynamics (CFD) calibration tests for velocity and erosion rate. In some embodiments, reduced order models generated from CFD may provide a table for operating conditions (PLF/ burner turn down ratio) and velocity' as well as angle distributions for required locations.

[0036] At 408, a Bayesian based model may be produced. The Bayesian based model may estimate tuning parameter that accounts for variation in the physics based kernel (e.g., coal properties and blend ratio, velocity, plant load factor, etc.) as well as estimate location specific failure (e.g., erosion and wall loss). The Bayesian based model may estimate tuning parameter based on a Bayesian learning algorithm utilizing the information from the physics based kernel, the operational data and the prior distributions as the input. Using statistical sampling techniques of data (e.g., physics based kernel, the operational data and the prior distributions) with Bayesian probabilistic methods to determine a model may narrow a gap between observed failure rates and predicted erosion rates (e.g., observed erosion rates v predicted erosion rates). Step 408 may comprise a predictor of the posteriori probabilities of tuning parameters from operational data as will be described with respect to 410.

[0037] At 410, tuning parameters may be adjusted. The tuning parameters associated with the first model may be adjusted based on received validation data. The received validation data associated with the boiler may be used to adjust the first model to more accurately represent a real-life operation of the boiler. The tuning may be based on comparing error rates 414 (e.g., a percentage error) that are calculated between the current tuning parameters and validation data. Once the parameters of the Bayesian model 408 are tuned, this new model may serve as a learning model at 412.

[0038] Referring back to FIG. 4, the learning model 412 may continually or periodically receive validation data 418 associated with a boiler that is being modeled. The validation data may comprise a set of historical thickness data obtained from offline measurements or from set of online sensors. The learning model 412 may predict 416 inspection locations based on estimated erosion FIG. 5. Prediction 416 may be based on comparing error rates 414 (e.g., a percentage error) that is calculated between tuning parameters and validation data.

[0039] Note the embodiments described herein may be implemented using any number of different hardware configurations. For example, FIG. 6 illustrates a monitoring platform 600 that may be, for example, associated with the system 100 of FIG. 1. The monitoring platform 600 may provide a technical and commercial advantage by creating a prediction tool, that causes minimal changes to existing plant digital control systems and that is field deployable on existing coal fired power plants. Furthermore, the monitoring platform 600 may provide advisory alerts on erosion rates and location specific information that can be used to plan outages instead of reacting to a failure in a boiler that is currently in operation.

[0040] The monitoring platform 600 may comprise a processor 610 ("processor"), such as one or more commercially available Central Processing Units (CPUs) in the form of one-chip microprocessors, coupled to a communication device 620 configured to communicate via a communication network (not shown in FIG. 6). The communication device 620 may be used to communicate, for example, with one or more users. The monitoring platform 600 further includes an input device 640 (e.g., a mouse and/or keyboard to enter information about the boiler and or the boiler tubing) and an output device 650 (e.g., to output and display the data and/or alerts).

[0041] The processor 610 also communicates with a memory/storage device 630 that stores data 616. The storage device 630 may comprise any appropriate information storage device, including combinations of magnetic storage devices (e.g., a hard disk drive), optical storage devices, mobile telephones, and/or semiconductor memory devices. The storage device 630 may store a program 612 and/or processing logic 616 for controlling the processor 610. The processor 610 performs instructions of the programs 612, 616, and thereby operates in accordance with any of the embodiments described herein. For example, the processor 610 may receive data such as, but not limited to, sensor data or data from a repository and may create a model based on the data and/or may also detect a likelihood of a boiler tube failing via the instructions of the programs 612 and processing logic 616.

[0042] The programs 612, 616 may be stored in a compiled, compressed, uncompiled and/or encrypted format or a combination. The programs 612, 616 may furthermore include other program elements, such as an operating system, a database management system, and/or device drivers used by the processor 610 to interface with peripheral devices.

[0043] FIG. 7 is a tabular view of a portion of a database 700 in accordance with some embodiments of the present invention. The table includes entries associated with boilers such as fuel related properties, material related data and/or operational parameters. The table also defines fields 702, 704, 706, 708, 710, 712, 714 and 716 for each of the entries. The fields, for example, may specify: a boiler ID 702, a first fuel related property 704, a second fuel related property 706, a first material related property 707, a second material related property 710, an nth material related property 712, a first operational parameter714, an nth operational parameter 716. The information in the database 700 may be periodically created and updated based on information collection during operation of a boiler and as they are received from one or more sensors.

[0044] The boiler ID 702 might be a unique alphanumeric code identifying a specific boiler of a plurality of boilers and the properties 704/706/708/710/712/714/716 might provide information related to features associated with a specific material or fuel/load of fuel or operational parameters.

[0045] As used herein, information may be "received" by or "transmitted" to, for example", (i) the monitoring platform 600 from another device; or (ii) a software application or module within the monitoring platform 600 from another software application, module, or any other source.

[0046] As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method or computer program product. Accordingly, aspects of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a "circuit," "module" or "system." Furthermore, aspects of the present invention may take the form of a computer program product embodied in one or more computer readable medium(s) having computer readable program code embodied thereon.

[0047] The process flow and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). it should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart il lustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

[0048] It should be noted that any of the methods described herein can include an additional step of providing a system comprising distinct software modules embodied on a computer readable storage medium; the modules can include, for example, any or all of the elements depicted in the block diagrams and/or described herein; by way of example and not limitation, a geometrical compensation module. The method steps can then be carried out using the distinct software modules and/or sub-modules of the system, as described above, executing on one or more hardware processors. Further, a computer program product can include a computer-readable storage medium with code adapted to be implemented to carry out one or more method steps described herein, including the provision of the system with the distinct software modules.

[0049] This written description uses examples to disclose the invention, including the preferred embodiments, and also to enable any person skilled in the art to practice the invention, including making and using any devices or systems and performing any incorporated methods. The patentable scope of the invention is defined by the claims, and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal languages of the claims. Aspects from the various embodiments described, as well as other known equivalents for each such aspects, can be mixed and matched by one of ordinary skill in the art to construct additional embodiments and techniques in accordance with principles of this application.

[0050] Those in the art will appreciate that various adaptations and modifications of the above-described embodiments can be configured without departing from the scope and spirit of the claims. Therefore, it is to be understood that the claims may be practiced other than as specifically described herein.

Claims

WE CLAIM :

1. A system for classifying an anomaly, the system comprising:

a processor (610); and

a non-transitory computer-readable medium comprising instructions that when executed by the processor perform a method to automatically detect anomalies, the method comprising:

receiving fuel properties (S210) associated with a type of fuel being used to fire a boiler (1 10), the boiler comprising a plurality of tubes (130).

receiv ing material properties (S220) associated with the plurality of tubes;

receiving operational data (S220a) associated with the boiler; determining, via the processor (610), a first model (408) based on the received fuel properties, received material properties and operational data.

modifying tuning parameters (410) associated with the first model based on validation data to determine a second model;

based on operational data input into the second model (412), automatically determining a likelihood of failure associated with the plurality of tubes; and

transmitting an alert associated with the boiler.

2. The system of claim 1 , wherein the type of fuel comprises coal and the fuel properties comprise (i) a material analysis of the coal to determine an amount of ash that will be produced, (ii) a velocity of the ash particles within the plurality of tubes and (iii) an angle of input associated with the ash particles.

3. The system of claim 1, wherein the material properties comprise (i) a thickness of the plurality tubes, (ii) a material that makes up the plurality of tubes, and (iii) historical data (406) associated with the material.

4. The system of claim 3, wherein the historical data relates to erosion of the material over time.

5. The system of claim 1, wherein the first model is constructed in accordance

Kx"p_m p° (VSin f

F =

with the expression - wherein £ is an erosion rate, xⁿ represents an erosion index of ash, x represents a weight percentage of silica in ash, p_m is a density of the tube material, p_P is a density' of an impinging ash particle, V is the velocity of impact, Sin is sine of an angle of impact, σ>- is a yield strength of the tube material, n represents a contribution of ash-content and K represents tuning parameter to be estimated based on an effect of size and shape variations of impacting particles wherein a posterior distribution of K may be determined from Bayes' Theorem.

6. The system of claim 1, wherein the first model is constructed in accordance with the expression:

®B* = ^FACT^ -h K.FA€T₂ + |^f¾)

wherein OD' represents an outer diameter of an eroded tube, OD represents an outer diameter of a new tube, ID represents an inner diameter of a new tube. FACTi is a constant derived from OD and I D, FACT2 is a constant that describes the cumulative erosion, and K represents a tuning parameter to be estimated based on an effect of size and shape variations of impacting particles wherem a posterior distribution of may be determined from Bayes' Theorem.

7. The system of claim 1 , wherein the first model is a regression model and the type of fuel comprises oil and the fuel properties comprise (i) a material analysis of the oil to determine an amount of sulfur.

8. The system of claim 1, wherein the alert is transmitted to a graphical user interface that illustrates a gauge to display millimetres of material removed per day and a gauge to display a remaining useful life of one of the plurality of boiler tubes.

9. A method of estimating boiler tube failures, the method comprising:

receiving fuel properties (S210) associated with a type of fuel being used to fire a boiler (110), the boiler comprising a plurality of tubes (130).

receiving material properties (S220) associated with the plurality of tubes; receiving operational data (S220a) associated with the boiler;

determining, via a processor (610), a first model (408) based on the received fuel properties and the received material properties.

modifying tuning parameters (410) associated with the first model based on first validation data using Bayesian learning to determine a second model: based on second operational data input into the second model (412), automatically determining a likelihood of failure associated with the plurality of tubes; and

transmitting an alert associated with the boiler.

10. The method of claim 9, wherein the type of fuel comprises coal and the fuel properties comprise (i) a material analysis of the coal to determine an amount of ash that will be produced, (ii ) a velocity of the ash particles within the plurality of tubes and (iii) an angle of input associated with the ash particles, wherein the material properties comprise (i) a thickness of the plurality' tubes, (ii) a material that makes up the plurality of tubes, and (iii) historical data associated with the material, and wherein the historical data relates to erosion of the material over time.