FIELD OF THE INVENTION:-
The present invention relates to an automated system to estimate percentage of unburned carbon in ash produced in tangential fossil fuel fired boilers based on properties of coal being fired and operating conditions.
BACKGROUND OF THE INVENTION:-
Power plant boilers are known for their nonlinear behavior due the nature of varying operating conditions involving high-pressure and high temperatures. Fuel i.e. coal burnt along with air is an important contributor to generate effective heat in the boiler system. Coal has an important role in the design and the performance of the boiler. Subsequently, operating conditions also influence the performance of the boiler. Unburned Carbon is a direct indication of the performance of the combustion process in the boiler. Percentage of Unburned carbon is one form of the energy losses that occur in power plant whose measurement is carried-out by analyzing the amount of unburned carbon in various ash types namely bottom ash, duct ash, air-heater ash and fly ash. Among these four kinds, the amount of unburned carbons in bottom ash and fly ash are more significant as compared to the unburned carbons in other two types of ash collections. In the existing art, there are no tools for predicting unburned carbon in ash. To know the unburned carbon in ash, it has to be analyzed in the laboratory after combustion and assess the unburned carbon, which is a time taking process and cannot be used for any online parameter correction to improve the efficiency of the power plant. Further, while operating the power plant, it becomes a difficult task for an operator to keep all the conditions in mind and relate how the particular change in any selected parameter would influence the efficiency of overall operation.

The existence of more unburned carbon in ash not only increases loss of energy but also makes Fly Ash unusable in Cement and brick productions and thereby unsalable for use as building material; as a consequence more unburned carbon in ash, loading on the bottom ash and fly ash disposal systems increases and stack opacity may also increase. These reasons are driving factors for investigation of methods and apparatus for timely estimation of unburned carbon (UBC) which helps an operator in making corrective actions and improve the performance of the combustion in the boiler. There is a need for a tool that enables operators to optimise operational parameters, improve combustion and reduce unburned carbon.
Further, one of the important factors to be considered while designing power-plant is how much unburned carbon would be resulting in the combustion process of the coal and its range of coal properties for desired operational conditions of the boiler. As a known fact the coal properties wide spread in the range even within the same coal mine. This leads to a generic design of the boiler and provide scope for adjustment of certain operational conditions through which the combustion performance can be improved.
Several methods and systems are in use or proposed to use for estimating unburned carbon in ash produced after combustion process. These approaches include the following methods.
Using mathematical model for combustion process and predict unburned carbon or Physical microscopic examination of the unburned carbon in sampled ash loading or Loss On Ignition (LOI) method (Usual method that is in Lab /offline method of burning the ash and weighing the leftover)

Abstract of JP2011013197A (title: Method and System for evaluating coal ash properties) describes a patent wherein method for evaluating coal ash properties includes the steps of determining influence items that specify the density of coal ash, Blaine value, flow value ratio, and activity index of coal ash for expressing the density, the Blaine value, the flow value ratio and the activity index based on the particle size, the ash composition, the concentration of unburned combustible contents, and the ash content of coal ash; and evaluating the properties of coal ash to be targeted. A tentative estimate equation containing a predetermined coefficient via the particle size, the ash composition, the concentration of unburned combustible contents, and the ash content of coal ash from the known actual measurement data, determining each of the coefficients, forming the tentatively estimated equation for estimating the density, the Blaine value, the flow value ratio, and the activity index.
Abstract of JP2949096 (title: Method for measuring quantity of unburnt carbon of fly ash and apparatus therefore describes a patent wherein method and apparatus for measuring quality of unburned carbon of fly ash. The apparatus measures the quantity of reflected light of fly ash is constituted of a sampling device set in a transfer path, a probe set at a rear end part of the sampling device for measuring the quantity of reflected light, a light source for use in obtaining the quantity of reflected light and an amplifier for amplifying an output signal from the probe. The light source uses a blue light-emitting body of a peak wavelength of 450 nm which little influences the quantity of totally reflected light of other components. The quantity of unburned carbon of fly ash is obtained from the measured quantity of reflected light of the fly ash on the basis of a preliminarily formed working curve showing the quantity of reflected light

and quantity of unburned carbon correspondingly to each other. The quantity of reflected light of the fly ash is measured online. The quantity of unburned carbon of the fly ash is obtained correctly from the measurement result.
Abstract of US5988079A (title: Unburned Carbon and other combustibles monitor) describes a patent wherein method to monitor Unburned Carbon and other combustibles of a fossil-fuelled boiler, such as a pulverized coal fired boiler and a control system based on the infrared imaging camera. It describes on how the combustibles are estimated using the correlations between hot particles moving along entrained hot gases exiting the boiler.
Abstract of EP0507060B1 (title: Method and apparatus for determining the amount of unburned in-ash component in waste gases of a powdered coal combustion system) and US5231939A (title: Apparatus for estimating an unburned carbon component amount in Ash in a coal-fired furnace) describe an invention wherein an in-ash unburned component estimating device for a coal-fired furnace which monitors the density of in-ash unburned component contained in burning waste gases to operate the furnace efficiently. A furnace temperature, a load band in the furnace, a furnace contamination coefficient, a ratio of two-stage combustion air supplied to the furnace, and a coal mixture ratio are taken in as fuzzy quantities to infer fuel ratio data and correction data used to correct predetermined reference values of reference in-furnace temperature distribution, reference in-furnace air ratio distribution and reference powdered coal grain diameter distribution.
Abstract of US6490909B1 (title: Method and apparatus for calculating carbon content of fly ash) describes a method and apparatus for

calculating the volume fraction of carbon in the fly ash using a carbon in fly ash sensor that has a resonant cavity. The method first determines the volume fraction of ash in the fly ash. The method then determines the real and imaginary components of the dielectric constant of a mixture of pure carbon and pure fly ash, and the transmission factor of a signal from the oscillator in the sensor transmitted through the cavity due to absorption by material in the cavity. The method then determines the volume fraction of carbon in the fly ash by using the volume fraction of ash, the real and imaginary components of the pure mixture dielectric constant, the absorption transmission factor and the length of the cavity, the speed of light and the frequency of the oscillator.
Whereas in the present invention it does not need to develop a mathematical model for combustion, or does not need to perform any microscopic ash loading examination, or does not need to perform any burning of ash to find unburned carbon percentage in the ash. Further, it does not need to estimate a tentative equation using coal particle size, ash contents etc like done in [JP2011013197A]. or does not need to measure the quantity of reflected light from ash sampling device like in [JP2949096], or does not need to use infrared imaging camera to estimate unburned carbon like done in [US5988079A], or does not need to know quantities like furnace temperature, a load band in the furnace, a furnace contamination coefficient etc needed in [EP0507060B1, US5231939A], or does not need to any additional sensors or arrangement like the one described in [US6490909B1] which works based on microwave absorption property of carbon to estimate the unburned carbon volume fraction in ash, where it is claimed to be taking 12 minutes to complete the estimation of carbon presence on the fly ash loadings. These above methods, approaches and inventions can only

examine as they occur and estimate the unburned carbon and would not
be helpful when operator wants to know what would be unburned carbon when certain conditions are changed or to be changed. Further, these above methods, approaches and inventions would not help on estimating how the input conditions of the power plant would influence the percentage unburned carbon in the ash loadings.
The invention addresses the need for an apparatus which would support the operator in making decisions while setting certain operating points during the general operations of the power plant. In case operator is not satisfied with the combustion performance for the given conditions, he can still change to other optimised operating conditions using this invention without incurring any resource loss including time.
OBJECT OF THE INVENTION;-
It is therefore an object of the invention is to provide a system developed based on artificial neural networks which learns from the historical datasets collected from the power plants and laboratory analysis, that can be deployed in predicting the percentage of unburned carbon in fly ash and bottom ash produced by tangential fossil fuel fired boiler and support the plant operator in making certain decisions while operating the power plant.
Another object of the invention is to propose system to visualize the sensitivity effect of each and every influencing factors of unburned carbon independently both on fly ash and bottom ash.
Another object of the invention is that to predict and visualize the sensitivity effect of combination of two parameters simultaneously.

Yet another object of the invention is to provide a system that does not require extra instrumentation.
A still further object of invention to estimate percentage of unburned coal at various operating conditions.
SUMMARY OF THE INVENTION:-
In the invention, a method, using Artificial Neural Network is an automated system approach which learns from the historical datasets collected from the power plants and laboratory analysis, are described for predicting percentage of unburned carbon (UBC) in fly ash and bottom ash produced by tangential fossil fuel fired boilers. The apparatus in the invention has the capability of predicting the %UBC in both bottom ash and fly ash for set of given initial conditions in almost real-time. This capability of the invention helps the operator to check the possible outcome for various operating conditions of the boiler before actual setting for the boiler operations. Further, this invention helps to assess and to visualize the sensitivity effect of each influencing variable of the process and estimate its contribution in controlling unburned carbon and provide support to the plant designer and operator in making decisions efficiently. This invention helps the power plant operator in estimating unburned carbon in ash with interactive computing and visualizing the contribution of each one of the influencing parameters on unburned carbon; and thereby supporting in making appropriate decision for reducing the unburned carbon in the fly ash and bottom ash simultaneously using fuel analysis and operational parameters as input. The artifact that is developed as a part of invention is based on one of the Machine Learning Techniques known as Artificial Neural Networks.

DETAILED DESCRIPTION OF ACCOMPANYING DRAWINGS:-
1. Figure 1 shows Basic Architecture of the apparatus in accordance to the invention.
2. Figure 2 shows Schematic of Boiler with depiction of ash locations in accordance to the invention.
3. Figure 3 shows Graphical User Interface of the apparatus in
accordance to the invention.
4. Figure 4 shows Comparison of Predictions by the apparatus and
Laboratory analysis for %UBC in Bottom ash in accordance to the
Invention.
5. Figure 5 shows Comparison of Predictions by the apparatus and
Laboratory analysis for %UBC in Fly ash for various test
conditions in accordance to the invention.
6. Figure 6 shows Sample of Sensitivity curves generated by the
apparatus with average burner tilt as sensitivity parameters for
%UBC in Bottom ash with one set of initial conditions given in Table 2 in accordance to the invention.
7. Figure 7 shows Sample of Sensitivity curves generated by the
apparatus with average burner tilt as sensitivity parameters for
%UBC in Fly ash with one set of initial conditions given in Table 2
in accordance to the invention.

8. Figure 8 shows Sample of Sensitivity curves generated by the apparatus with average burner tilt and %Ash in coal as two sensitivity parameters for %UBC in Bottom ash with one set of initial conditions given in Table 3 in accordance to the invention.
9. Figure 9 shows Sample of Sensitivity curves generated by the apparatus with average burner tilt and %Ash in coal as two sensitivity parameters for %UBC in Fly ash with one set of initial conditions given in Table 3 in accordance to the invention.

10. Figure 10 shows Table with list of input variables used for building Artificial networks model for predicting both %UBC in bottom ash and %UBC in fly ash in accordance to the invention.
11. Figure 11 shows One set of initial conditions for generating sensitivity curves for bottom ash and fly ash with respect to average burner tilt in accordance to the invention.
12. Figure 12 shows One set of initial conditions for generating sensitivity curves for bottom ash and fly ash with respect both average burner tilt and %Ash in coal in accordance to the invention.
DETAILED DESCRIPTION OF A PREFERED EMOBODIMENT OF THE INVENTION
Figure 1 shows the architecture of the embodiment of the invention

disclosed herein. Invention is a combination of computing hardware and programming code implemented on the computing hardware and which works on the hardware to interact with the operator who is using the same.
Figure 2 is a simple schematic view of Thermal power plant with control room and sensor signal connection for information collection on various parameters. The power plant is monitored and controlled from a control room (1) from where all the states of the power plant i.e. pressure, temperature, flow rates etc. at various locations consisting hundreds of parameters are measured and stored into a database (2) for various purposes including modeling, monitoring, diagnostics and for devising a better control strategy to encounter odds. Visualizing device (3) displays all the collected information in database (2) for the ease of operators. In this invention out of many hundreds of the collected parameters, proximate analysis of the coal that is being fed into the furnace is performed in Coal Analysis lab (4), primary air flow rate (5), secondary air flow rate (6), coal flow rate (7) measured at each coal feeder (8) which feeds coal to corresponding mills making mill combination (9), that intern feeds coal and air mix into the boiler furnace area through the entrance of Windbox (10), burner tilt angle (11), percentage of excess air measured at economiser outlet (12) before the air-preheater (13). All variable measured and derived from the above mentioned list are considered as input variables listed in Table 1. Similarly the targeted parameters i.e. unburned carbon in fly ash and bottom ash are analysed based on the samples collected from the hopper of Electrostatic Precipitator [ESP] (14), the fly ash (15) and from the bottom of the furnace, the bottom ash (16). Figure 2 also depicts ID fan (17) which drives the flue gas out through chimney (18). Further, Figure 2 shows the display screen (19) of the

invention in the visualizing device (3) in the control room(l). A screenshot of the display screen (19) in the control room gives access to the graphical user interface of the framework as shown in detail in Figure 3 which takes all the variables listed in Table 1 as inputs and predicts the percentage unburned carbon (%UBC) in fly ash and bottom ash as outputs which the operator can utilize to first verify the outcome of the operating settings that the operator would like to set on the actual boiler. Upon the satisfactory results, those input conditions can be set on actual boiler. Figure 4 shows the performance of the apparatus in predicting the %UBC in bottom ash compared with laboratory analysis estimated. Similarly, Figure 5 shows the performance of the apparatus in predicting the %UBC in fly ash compared with laboratory analysis estimated. The capability of the framework in predicting %UBC in fly ash and bottom ash is leveraged to quantify and understand the influence of each one of the input variable on the both outputs. Further, the framework supports sensitivity analysis for quantification of the influence of each one of the variable on the outputs with given initial conditions.
Sensitivity analysis performed using the framework is useful for the operator who is controlling the power plant in making appropriate control actions as the sensitivity analysis is useful for verifying a prior on how the current initial conditions are driving the power plant performance. Figure 6 and Figure 7 show sensitivity analysis depicting the influence of single variable, in this example, Average Burner tilt angle is chosen with set of initial conditions listed in Table 2. Figure 6 shows the sensitivity curve representing the relationship between Average burner tilt and %UBC in Bottom ash and Figure 7 shows sensitivity

curve representing the relationship between Average burner tilt and %UBC in Fly ash. Further, this frameworks supports the sensitivity analysis with any two input variables on the outputs i.e. %UBC in bottom ash and fly ash. Figure 8 and 9 show sensitivity curves for %UBC in bottom ash and fly ash respectively with the initial conditions listed in Table 3.

We claim:
1. An automated system configured to calculate percentage of unburned
carbon separately both in fly ash and bottom ash produces in power
plant boiler, wherein fly ash (15) is collected from an electrostatic
precipitator (ESP) (14) and bottom ash (16) is collected at the bottom of
the boiler furnace;
wherein percentage of unburned carbon in the fly ash and the bottom ash are estimated using on the Artificial neural networks artifact for given conditions including fuel properties and operational conditions; wherein influence of the variable or plurality of variables on the variation of % unburned carbon can be examined selectively within the range between minimum and maximum values;
wherein the percentage of unburned carbon is computed and stored for various analysis.
2. The automated system as claimed in claim 1, wherein multiple combination of input values enable the operator to determine the percentage of unburned carbon in ash and select optimum parameters to minimize unburned carbon.
3. The automated system as claimed in claim 1, wherein the computed values are visualized in a x-y plot to form sensitivity curves.
4. The automated system as claimed in claim 1, substantially as herein described and illustrated.

5. The automated system as claimed in claim 1, wherein the system has the capability to predict the % UBC in both bottom ash and fly ash for set of given initial conditions in real- time.