Description:
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a prediction method for the evaluation of the coke quality, more particularly to a coke quality prediction obtained through wet and dry quenching methods based on coal blend parameters.

BACKGROUND OF THE INVENTION
The uncertainty in coke quality while preparing coal blend, and to prepare an optimum blend in terms of coal cost and coke quality has been challenging due to the complex and non-linear relationships between coal blend properties and coke quality parameters. Manual methods for predicting coke quality are labor-intensive, time-consuming, and prone to errors. Furthermore, these methods may not fully utilize the available data or account for all relevant factors affecting coke quality.
The variability in coke quality within Coke Ovens and Blast Furnaces, which can lead to inefficiencies and inconsistencies in the steelmaking process. Coke quality is crucial for the performance of these processes, as it directly affects the efficiency of the furnace and the quality of the produced steel. Variations in coke quality can result from changes in coal blend composition and properties, which are influenced by factors such as coal mining conditions and blending practices.
Prior art patent to refer:
CN114841460A discloses a coke quality prediction method and system based on machine learning, and the method comprises the steps: carrying out the coke quality index feature analysis of basic coking sample data in a basic coking sample data set through combining a coke oven coke quality prediction model, outputting a coke quality index feature cluster, and carrying out the prediction of the coke quality through combining with the coke quality index feature cluster.
CN117350568A discloses coke quality index prediction method comprises the following steps: collecting various coal blending indexes, process parameters and corresponding coke quality indexes of each production in actual industrial production as original industrial data; performing feature normalization processing on each feature index, and dividing a processed data sample set into a training set and a test set; performing regression analysis on each coke quality index by adopting regression analysis to obtain a regression equation about each characteristic index and the coke quality index; and screening and sorting the characteristic variables by adopting a recursive characteristic elimination method so as to determine a characteristic variable combination which has the most obvious influence on the coke quality index, and establishing a coke quality index prediction model.
CN111950854B discloses a coke quality index prediction method based on a multilayer neural networks, and belongs to the technical field of industrial information. Adopting industrial actual production data, firstly cleaning the data, adopting gradient reinforced tree to make correlation analysis on the factors influencing coke quality index, selecting the parameters of ash content, sulfur content and M10、M40And the most relevant variables such as CRI and CSR, further constructing a training sample, establishing a multilayer neural network prediction model to predict coke quality indexes, and optimizing the variables in the model by adopting an intelligent optimization algorithm to give a final coke quality index prediction result.
US7803627B2 discloses process for evaluating the coke and/or bitumen yield and quality of a plurality of refinery feedstocks, by (i) providing a plurality of refinery feedstocks, (ii) placing a sample of each of the plurality of refinery feedstocks on a heating device, (iii) heating each sample under vacuum to a temperature in the range 300° C. to 420° C. using the respective heating device while measuring the weight loss of the sample, and then (iv) (a) measuring the bitumen quality of the vacuum residues formed, and/or (b)(i) heating the vacuum residues to a temperature in the range 450° C. to 600° C. using the heating device, while measuring the weight loss of the sample, and then (ii) measuring the coke quality of the products formed.
Hence there is a need for a method to be developed to predict the coke qualities like CSR, CRI, M10, and M40 based on coal blend composition. The present invention provides a system and method to address the existing problem of uncertainty in coke quality while preparing coal blend and to prepare an optimum blend in terms of coal cost and coke quality.

OBJECT OF THE INVENTION
It is an object of the present invention to overcome the shortcomings of the prior art.
It is an object of the present invention to provide a method for the prediction of a coke qualities like CSR, CRI, M10 and M40 based on coal blend composition.
Yet another object of the invention to mitigate the effect of changes in battery oven conditions on coke quality.
Yet another object of the invention to provide an algorithm to predict the coke qualities with an accuracy of above 95% and hit score of above 99%.
Yet another object of the invention is to help in preparing optimized coal blend with cost and quality optimization.

SUMMARY OF THE INVENTION
The following disclosure presents a simplified summary of the invention in order to provide a basic understanding of some aspects of the invention. This summary is not an extensive overview of the present invention. It is not intended to identify the key/critical elements of the invention or to delineate the scope of the invention. Its sole purpose is to present some concept of the invention in a simplified form as a prelude to a more detailed description of the invention presented later.
According to one implementation, a method for prediction of coke quality, the method comprises the steps of:
collecting, coal blend data and corresponding coke analysis data;
cleaning, to remove all erroneous data from the said collecting;
predicting, the coke qualities based on coal blend chemical parameters; and
optimizing, to prepare the optimum blend in terms of cost and quality.

According to one implementation, a system for prediction of coke quality, comprising:
a collection module, configured to collect coal blend data and corresponding coke analysis data;
a cleaning module, configured to remove all erroneous data; and
a prediction module, configured to predict the coke qualities based on coal blend chemical parameters.

Other aspects, advantages, and salient features of the invention will become apparent to those skilled in the art from the following detailed description, which, taken in conjunction with the annexed drawings, discloses exemplary embodiments of the invention.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
The above and other aspects, features and advantages of the embodiments of the present disclosure will be more apparent in the following description taken in conjunction with the accompanying drawings, in which:
Figure 1 depicts the method validation report of coke sorting plants -1 data of Bhilai Steel Plantof a method for prediction of coke quality in accordance with an embodiment of the present invention.
Figure 2 depicts method validation report of coke sorting plants -2 data of Bhilai Steel Plant of a method for prediction of coke quality in accordance with an embodiment of the present invention.
Figure 3 depicts method validation report of coke sorting plants -3 data of Bhilai Steel Plant of a method for prediction of coke quality in accordance with an embodiment of the present invention.
Figure 4 depicts method validation report of coke sorting plants -4 data of Bhilai Steel Planta method for prediction of coke quality in accordance with an embodiment of the present invention.
Figure 5 shows data collection of a method for prediction of coke quality in accordance with an embodiment of the present invention.
Figure 6 depicts shows data optimization of a method for prediction of coke quality in accordance with an embodiment of the present invention.
Figure 7 depicts shows data prediction of a method for prediction of coke quality in accordance with an embodiment of the present invention.

Persons skilled in the art will appreciate that elements in the figures are illustrated for simplicity and clarity and may not have been drawn to scale. For example, the dimensions of some of the elements in the figure may be exaggerated relative to other elements to help to improve understanding of various exemplary embodiments of the present disclosure.

DETAILED DESCRIPTION OF THE PRESENT INVENTION
The following description with reference to the accompanying drawings is provided to assist in a comprehensive understanding of various embodiments of the present disclosure as defined by the claims and their equivalents. It includes various specific details to assist in that understanding, but these are to be regarded as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the various embodiments described herein can be made without departing from the scope and spirit of the present disclosure. In addition, descriptions of well-known functions and constructions may be omitted for clarity and conciseness.
All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which various embodiments belong. Further, the meaning of terms or words used in the specification and the claims should not be limited to the literal or commonly employed sense but should be construed in accordance with the spirit of the disclosure to most properly describe the present disclosure.
The terminology used herein is for the purpose of describing particular various embodiments only and is not intended to be limiting of various embodiments. As used herein, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising" used herein specify the presence of stated features, integers, steps, operations, members, components, and/or groups thereof, but do not preclude the presence or addition of one or more other features, integers, steps, operations, members, components, and/or groups thereof.
The present disclosure will now be described more fully with reference to the accompanying drawings, in which various embodiments of the present disclosure are shown.
The present invention provides a system that collects and analyzes data on coal blend properties and corresponding coke quality parameters. By using advanced data analytics and artificial intelligence techniques, the system can predict coke quality with a high degree of accuracy. This enables steelmakers to better control and optimize their processes, leading to improved efficiency, reduced costs, and higher-quality steel production.
According to one implementation, a method for prediction of coke quality, the method comprises the steps:
collecting, coal blend data and corresponding coke analysis data;
cleaning, to remove all erroneous data from the said collecting;
predicting, the coke qualities based on coal blend chemical parameters; and
optimizing, to prepare the optimum blend in terms of cost and quality.

According to one implementation, a system for prediction of coke quality, comprising:
a collection module, configured to collect coal blend data and corresponding coke analysis data;
a cleaning module, configured to remove all erroneous data; and
a prediction module, configured to predict the coke qualities based on coal blend chemical parameters.
The present invention plays a pivotal role in reducing process variance and enhancing process control within Coke Ovens and Blast Furnaces by predicting in advance coke qualities like CSR, CRI, M10 and M40 based on chemical properties of coal blend parameters like PR (Plastic Range), MMR (Min Max Reflectance), BAR, and Fluidity with an accuracy of over 95% and hit score of above 99%.
Firstly, historical coal blend parameters and corresponding coke analysis data of only last three years is collected. Data collection is dynamic in nature and runs with minimal manual intervention. The data collection module of the model collects data automatically online in real-time from different sources like LIMS(Laboratory Information Management System), Web based GUI and Process SCADA systems, processes it, automatically removes any erroneous data and generates alarm in case of erroneous data. Figure 5 shows data collection module collects data online. Whenever a new data enters into the system, the data collection and cleaning module processes the data in conjunction with historical dataset and generates an alarm if data integrity factor (DIF) is lower than the accepted limits set by the system. The erroneous data is automatically removed from the system and is accepted only after correction through manual intervention. Coal blend compositin data from Laboratory Infromation Management System(LIMS). On data collection, data cleaning is performed to remove all erroneous data.
Figure 6 shows coal blend optimization by processing all possible blend compositions in the chosen optimization range, this module gives the cheapest possible blend composition optimized in terms of cost and quality. The output of coke quality prediction module is used by the blend optimization module to prepare the optimum blend in terms of cost and quality.

An algorithm as shown in Fig. 7 using data analytics and artificial intelligence principles is developed to correctly predict the coke qualities based on coal blend chemical parameters. Various development tools like R-programming, Python, PL/SQL etc. were used for developing the algorithm and Post-model development. The innovative algorithm has auto-correction and self-improvement features. It keeps on improving itself till a 100% hit score is achieved.
In the process of predicting the coke qualities associated with a newly input coal blend composition, a stepwise recursive data analysis is employed by the innovative algorithm to evaluate each historical blend. This analysis assigns an impact score to each historical blend based on their chemical parameters and temporal influence (to take care of oven health). Subsequently, these historical blends are prioritized in descending order of impact score, and the top n blends are selected for further analysis. Among these chosen blends, a subset of m blends with minimal standard deviation in the three highest impact parameters (chemical parameters based on correlation factor adjusted for age of blend) is identified. These m historical blends are used to forecast the coke qualities of the newly input coal blend, incorporating an adjustment factor (AF). The value of m and n are decided during model training phase. The algorithm iteratively adjusts the values of AF to enhance the hit score (explained in point-2) calculated for last 10 blends gradually. The algorithm also predicts the accuracy of its prediction based on the impact score of m chosen blends called Model Confidence Factor (MCF). Notably, all stages of this algorithm are executed automatically, without requiring any manual intervention. The algorithm has been tested for different coal-coking regimes like-
a. Tall battery with dry quenching process
b. Tall battery with wet quenching process
c. Small battery with wet quenching process

In all the above scenarios the algorithm achieved a hit score of above 99%. Subsequently, approximately 100 blends have been prepared with the assistance of this model, demonstrating its practical utility at BSP.
Hit Score is to ensure a meaningful and dependable assessment of method accuracy. This involved calculating the average variation in LAB test results for CSR, CRI, M10, and M40 values, pertaining to each blend over the last three years. Only instances where the error in method prediction was less than this calculated average lab result deviation was considered for a successful hit score. The specific method accuracy required to achieve a successful hit score varied for different shops, with individual criteria established accordingly. Hit score is calculated as follows:
Hit Score (%) = (No. of blends with successful prediction hits/Total number of blends) *100

Table 1 shows lab result variation and accuracy required for successful prediction hit is calculated online in real-time as shown in below table:

Shop Name CSR CRI M10
(Lab Result deviation) M40
(Lab Result deviation)
Lab Result deviation Successful Hit score accuracy Lab Result deviation 100% Hit score Lab Result deviation 100% Hit score accuracy Lab Result deviation 100% Hit score accuracy
Coke sorting plant1 1.3% 98.7 % 3.7% 96.3 % 7.9% 92.1 % 2.3% 97.7 %
Coke sorting plant2 1.6% 98.4 % 6.0% 94 % 10.9% 89.1 % 2.9% 97.1 %
Coke sorting plant3 1.5% 98.5 % 3.8% 96.2 % 9.1% 90.9 % 2.5% 97.5 %
Coke sorting plant4 2.4% 97.6 % 6.8% 93.2 % 17.8% 82.2 % 5.6% 94.4 %

Figure 1 to Figure 4 shows the method validation report of four coke sorting plants data of Bhilai Steel Plant of a method for prediction of coke quality. On the secondary Y-Axis (right side), method prediction accuracy has been plotted.

Outputs and Results:
The accuracy of the present invention is validated during the next 6 months using actual field data. On the validation, of the output of the method with average accuracy of more than 95% and hit-score of over 99%, a web based application is developed for user data entry and method visualization using the latest web development tools. The web based application is hosted on SAIL enterprise-wide network and is accessible by all the sister units of SAIL throughout the country. The user from any sister unit can access the model through user id and password validation and use the model output based on their own dataset.

Novel and inventive features:
The present invention provides a method for prediction of coke quality
1. To predict coke qualities like CSR, CRI, M10 and M40 with an accuracy of over 95% and hit score of above 99%.
2. To help in preparing optimized coal blend with cost and quality optimization.
3. Equally effective for coke obtained through different coking processes like dry and wet quenching methods and for different types of coke oven batteries like 7mtr.tall batteries, 4.5 mtr. tall batteries etc.

The following advantages can be list out from the present invention:
Technical / technological:
a) Users can check the predicted coke qualities in advance before making the coal blend and make blend accordingly which leads to:
i) Better and consistent coke quality.
ii) Less wastage of coal which leads to cost saving and lesser greenhouse gas emission.
b) Any constraint in coal blend supply can be handled better.
c) Advance information about the supplied coke quality leads to better process management at Blast-furnaces.
d) The capability of easily being tuned for any disparate regime of coke making and come out with reasonably accurate coke quality predictions.

Economic:
1. Maintaining steady quality of coke leading to steady regime of operation in downstream of blend preparations, viz. Coke Ovens and Blast Furnaces.
2. It helps overcome supply and stock related constraints for various quality of coal as in-feed material.
3. Foresightedness of output of Coke Ovens, leads to reduced wastage of in-feed coal – this leads not just to cost savings but also lesser greenhouse gases emission.
4. The cost and quality optimization feature of the present invention also leads to significant cost saving without compromising in coke quality.

Application:
The present invention can be easily implemented in any coal-coking industry. This invention has the potential to change the way coal blends are prepared for coking process in steel industry and increase the profitability by the following application:
1. Process Optimization: By accurately predicting coke quality based on coal blend properties, the invention enables steelmakers to optimize their processes. They can adjust coal blend compositions and blending practices to produce coke with desired qualities, such as CSR, CRI, M10, and M40. This optimization leads to improved efficiency and cost savings.
2. Quality Control: The invention helps in maintaining consistent coke quality, which is essential for ensuring the quality of the produced steel. By predicting coke quality with high accuracy, steelmakers can implement better quality control measures and minimize variations in the steelmaking process.
3. Resource Management: The invention allows for better management of resources, such as coal blends. By predicting coke quality, steelmakers can make informed decisions about the selection and use of coal blends, optimizing the use of resources and reducing waste.
4. Risk Reduction: The invention reduces the risk of producing low-quality coke, which can lead to operational issues and product defects. By predicting coke quality accurately, steelmakers can mitigate risks associated with inconsistent coke quality and ensure smoother operations.
, Claims:
1. A method for prediction of coke quality, the method comprises the steps of:
collecting, coal blend data and corresponding coke analysis data;
cleaning, to remove all erroneous data from the said collecting;
predicting, the coke qualities based on coal blend chemical parameters; and
optimizing, to prepare the optimum blend in terms of cost and quality.

2. The method as claimed in claim 1, wherein said collecting comprises data of only the last three years.

3. The method as claimed in claim 1, wherein said predicting includes an algorithm based on data analytics and artificial intelligence principles comprises the steps of:
autocorrecting and self-improving; and
achieving a hit score, wherein the hit score is calculated as
Hit Score (%) = (No. of blends with successful prediction hits/Total number of blends) *100

4. The method as claimed in claim 1, wherein the said collecting dynamically collect data in real time from different sources like LIMS (Laboratory Information Management System), web-based GUI (Graphical user interface), and Process SCADA systems.

5. The method as claimed in claim 1, wherein said cleaning further comprises to generate an alarm.

6. A system for prediction of coke quality, comprising:
a collection module, configured to collect coal blend data and corresponding coke analysis data;
a cleaning module, configured to remove all erroneous data; and
a prediction module, configured to predict the coke qualities based on coal blend chemical parameters.

7. The system of claim 6, wherein the prediction module predicts coke qualities including CSR (Coke Strength after Reaction), CRI (Coke Reactivity Index), M10, and M40 based on coal blend chemical parameters including PR (Plastic Range), MMR (Min Max Reflectance), BAR (Banded Area Ratio), and Fluidity with an accuracy of over 95%.

8. The system of claim 6 or 7, wherein the prediction module includes an algorithm based on data analytics and artificial intelligence principles comprises the steps of:
autocorrecting and self-improving; and
achieving a hit score, wherein the hit score is calculated as
Hit Score (%) = (No. of blends with successful prediction hits/Total number of blends) *100.

9. The system of claim 6 or 7, wherein the collection module dynamically collects data in real time from different sources like LIMS (Laboratory Information Management System), web-based GUI (Graphical user interface), and Process SCADA systems.

10. The system of claim 6 or 7, wherein the cleaning module is further configured to generate an alarm.

11. The system of any of claim 6 to 7, wherein the collection module includes historical coal blend parameters and corresponding coke analysis data of only from the last three years.