Title of Invention

A METHOD FOR HEAT INPUT CONTROL FOR A COKE OVEN BATTERY BY MEANS OF CONTROL OF PAUSE PERIOD BETWEEN REVERSALS

Abstract

The present invention relates to a method for heat control for a coke oven battery by means of control of pause period between reversals, said method comprises the steps of measurement of heating parameters, such as, regenerator temperature, gas flow, gas calorific value of the battery and converting the above parameters as suitable input to the fuzzy model by application of heating principles and using fuzzy logic base model for quantitative control of heat input into the coke oven battery by application of appropriate pause period characterised in that the said method maintains the set target battery temperature by application of an appropriate pause period taking into account the thermal behaviour of the battery, such as, past temperature variation, present temperature variation and total heat input supplied and the said model is versatile enough to take new inputs such as moisture, coal quality etc.

Full Text

FIELD OF INVENTION: The prtstnt invention generally relate* to a method for operating heating system of a coke oven battery by controlling pause time between reversals. More specifically, the present invention relates to the use of fuzzy logic to control the heat input into the coke oven battery for converting coal into coke. BACKGROUND OF THE INVENTION! Coke making is a complex heating process. The process of converting coal to coke involves soaking of coal at high temperature (around 1200 deg Celsius) for a certain period of time called the coking time and then quenching it to room temperature. In a coke plant, a coke oven battery may consist of 70—72 ovens. Each oven is lined with refractory material and is heated upto 1200 degree Celsius. The battery is heated from below by supplying fuel, which in this case can be coke oven gas or rich gas from blast furnace. Each oven has 28 flues, and at any point-in—time, one flue acts as combustion flue where fuel is burnt and the flue adjacent to it acts as an exhaust flue from where the burnt gases are sucked. Thus, each oven chamber has 14 combustion flues and 14—exhaust flues. Coal mass is charged into ovens and heated for approximately 20 hours. During this coking time, after each 30 minutes, a phenomenon called reversal takes place, when the roles of combustion and exhaust flues are reversed and combustion flue becomes exhaust flue and exhaust flue becomes combustion flue. There is a time at the start of reversal when no fuel to the battery is supplied. This period is called pause—time. Charged ovens are pushed out of an oven chamber when their coking time is complete. The pushed ovens are taken in quenching loco for quenching with water to bring it to ambient temperature. A major challenge in coke making process, is controlling the quality of coke produced. There are several factors, which influence quality of coke. The coke end temperature (CET), measured by means of a radiation pyrometer as quenching loco passes under the pyrometer, is a true indicator of quality of combustion taken place inside an oven chamber and consequently quality of coke. However, coke end temperature is not directly controllable. Most of the available models for heating control of coke ovens are proportional + integral + differential (PID) based. Battery temperature is measured by means of thermocouples embedded along the length of the battery in regenerators, which is a source and sink of energy to the battery. Proportional + integral + BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS; Fig. 1 shows a pictorial representation of a coke oven battery Fig. 2 shows the control room of a coke plant Fig. 3 shows the architecture of a fuzzy logic control application Fig. 4 shows the parts of the fuzzy model. Fig. 5 shows a view of interface application Fig. 6 shows a variation of actual applied pausetime and regenerator temperature error with respect to time Fig. 7 shows a variation of total heat input and regenerator temperature error with respect to time. DETAILED DESCRIPTION OF THE INVENTION* Fig. 2 shows the centrol control room of coke plant where the higher-level automation system forms an integral part of operation along with basic automation systems. The central server communicates with various machines and programmable logic controller for events and data. The following hardware environment is necessary for level — 2 and heat input model. Fuzzy server is a Compaq Proliant H1530, dual CPU Pentium III Xenon 1.0 8Hz, 1 6B RAM, 10/100 Hops network card. Excel for developing the interface application between Level - 2 and fuzzy model Cubicalc — the fuzzy logic software tool. The software environment for Level - 2 consist of the following : VAX VMS — operating system for the server Windows NT for user terminals Oracle for data storage Process data acquisition : Through PLNM,CM50S, process parameters like gas flow, calorific value, regenerator temperature etc. are sent to server. Pushing charging data acquisition s This data is acquired over RF network,using appropriae Motorola supplied hardware installed on each machine and gateway software on the central Alpha server. Hardware unit at machine level is termed as MO6CAD and at the ground server level is called HCPT. Coke end temperature : This data is acquired over RS485 serial link radiation pyrometer, which measures the temperature and communicates it to the central server. Fuzzy model is developed on top of the level—2 automation. The system relies on the data and events acquired by the level-2 automation server. Fig.3 shows the architecture of the fuzzy logic based model. As shown in the figure, the system consists of three main parts, viz. fuzzy model, data gatherer and interface application. Fuzzy model is the heart of the model and is developed using cubical software.Fuzzy model has been developed with the following variables as inputs : Current temperature error : This value is defined as the difference between current regenerator temperature and target regenerator temperature. The main aim of fuzzy model is to have the actual regenerator temperature as close to the target as possible. The exact value of deviation, hence, is an important parameter that would determine the pause that is to be applied between reversals. If this value is highly positive, it means that pausetime has to be more and conversely, a high negative value of this variable indicates that pausetime has to be as less as possible. J Past temperature error : This value is defined as the difference between current regenerator temperature and target regenerator temperature during the previous reversal. Coking process has a lot of inertia when it comes to reacting to a certain process parameter change. So, even if the pausetime is changed for the current reversal the effect of it may/may not be seen before the next reversal. The fuzzy model takes into account this characteristic feature of the process and hence considers some amount of history in its inputs for calculation of pausetime between reversals, i.e. the period at the start of a 'reversal' when no fuel is supplied to the battery. Actual heat input into battery during the current reversals Heat input into the battery which happens to be a very dynamic quantity, does determine the amount of pausetime that should be applied. If during a reversal the average heat input is more, then pausetime should also be adjusted to take care of the more heat input into the battery. Output from the model is pausetime. This is the fuzzy calculated pausetime that is to be applied in the next reversal. Fuzzy model is developed using the fuzzy tool — cubical. Fig. 4 shows the various parts that have been built into the fuzzy model. The processing function in cubical reads data from interface application and stores it in local variables. Design of the processing function involves defining adjectives for the various inputs/output variables,defining the fuzzy rule base of the model. Adjective of a variable would imply defining the following for various variables t Number of fuzzy subsets and range of each of the subsets. Depending on the sensitivity of the output variable to a certain input, and the inputs possible range of values, the number of fuzzy sets are defined for the input variable. Same criteria are used to define the adjectives of output variable also. Each subsets name and membership function. Subsets are named depending on the range as "HiHi', 'MedHi', 'High', 'NZ', 'Low', 'HedLo', 'LoLo' etc. For example,a temperature error, value of —0.5 to 0.5 would be considered 'NZ' subset value, whereas value of 0—2.5 would be considered high (High), a value of 2.5—4 would be considered medium high (MedHi) etc. A similar nomenclature would be used when the values are on the negative side. All membership functions defined are triangular for both inputs and output variables. Definition of fuzzy rule base t Development of fuzzy models relies very much on the knowledge base available with the experts in the area for which the model is being developed, or wealth of data containing the required information. The current model has been developed with knowledge of experts in coke making domain. Also when the level—2 systems -10- are already in place during the time of model development there is quite an amount of data available for knowledge deciphering. The rule base hence, is an excellent representation of both the expert's knowledge and information hidden in data. Fuzzy rules define for various combinations of the input variable's fuzzy subsets, the output variable's fuzzy subset. Execution of the model involves the following : Fuzzification : The inputs passed to processing by preprocessing function are fuzzified. Fuzzification in short calculates the belongingness of the input's current value to each of the fuzzy subset defined for the same input variable. Inference engine : Fuzzy arithmetic using weighted average works on fuzzified inputs and rule base to calculate the belongingness of the output to each of its subset. Defuzzification s This function converts the calculated belongingness of the output to its subsets, into a crisp value of output. This crisp value would represent the pause time that is calculated for te current input values. Post processing : This function in cubical, writes the output value onto interface application. Data gatherer and interface application Modules are developed as peripherals to fuzzy model, mainly to handle the input and output data from the model. Data gatherer, as the name suggests,collects the data required by model from level—2 data store. Data gatherer also registers the pause time calculated by fuzzy model in level-2 database. The data collected by the data gatherer is communicated to fuzzy model through an interface application. Interface application is an Excel application which connects to data gatherer and fuzzy model through DDE technology. The present invention has been successfully validated and implemented. Though the present model of the invention considers only three inputs, more process inputs like coal fineness, moisture content etc., can be considered. Since fuzzy logic has been used for the model, more inputs can easily be considered, by framing rules for each of the inputs. Fig. 6 shows the variation of actual applied pausetime with respect to time. The graph clearly shows how regenerator temperature error determines the pause time that is to be given in controlling the regenerator error to a minimum value. A dip in the regenerator temperature error, has also dipped the pausetime to be applied. Amd.page.... WE CLAIM 1. A method for heat control for a coke oven battery by means of control of pause period between reversals, said method comprises the steps of: - measurement of heating parameters, such as, regenerator temperature, gas flow, gas calorific value of the battery; - converting the above parameters as suitable input to the fuzzy model by application of heating principles; - using fuzzy logic base model for quantitative control of heat input into the coke oven battery by application of appropriate pause period; Characterised in that the said method maintains the set target battery temperature by application of an appropriate pause period taking into account the thermal behaviour of the battery, such as, past temperature variation, present temperature variation and total heat input supplied and the said model is versatile enough to take new inputs such as moisture, coal quality etc. 2. The method as claimed in claim 1, wherein the step of deriving inferences by the fuzzy logic is based on deviations of battery temperature from set points. 3. The method as claimed in claim 2, wherein the battery temperature derivation during prior reversals are also considered for deriving fuzzy logic inferences. Amd.page... -15- 4. The method as claimed in the preceding claims, wherein some amount of history of the inputs to the fuzzy logic is taken into account for controlling the pause time. 5. A method for operating heating systems of a coke oven battery by controlling pause time between reversals, substantially as herein described and illustrated. The present invention relates to a method for heat control for a coke oven battery by means of control of pause period between reversals, said method comprises the steps of measurement of heating parameters, such as, regenerator temperature, gas flow, gas calorific value of the battery and converting the above parameters as suitable input to the fuzzy model by application of heating principles and using fuzzy logic base model for quantitative control of heat input into the coke oven battery by application of appropriate pause period characterised in that the said method maintains the set target battery temperature by application of an appropriate pause period taking into account the thermal behaviour of the battery, such as, past temperature variation, present temperature variation and total heat input supplied and the said model is versatile enough to take new inputs such as moisture, coal quality etc.

Documents:

490-KOL-2004-(05-12-2011)-FORM-27.pdf

490-KOL-2004-(22-08-2012)-FORM-27.pdf

490-kol-2004-granted-abstract.pdf

490-kol-2004-granted-claims.pdf

490-kol-2004-granted-correspondence.pdf

490-kol-2004-granted-description (complete).pdf

490-kol-2004-granted-drawings.pdf

490-kol-2004-granted-examination report.pdf

490-kol-2004-granted-form 1.pdf

490-kol-2004-granted-form 13.pdf

490-kol-2004-granted-form 18.pdf

490-kol-2004-granted-form 2.pdf

490-kol-2004-granted-form 3.pdf

490-kol-2004-granted-form 5.pdf

490-kol-2004-granted-gpa.pdf

490-kol-2004-granted-reply to examination report.pdf

490-kol-2004-granted-specification.pdf