Improvement And Implementation Of The Coke Oven Straight-line Temperature Inspection Robot

Coke is the main raw material for steel production.Coal is obtained by high temperature dry distillation in a coke oven.During the coking process,the heating temperature of the coal directly affects the quality of the output coke.In order to control the combustion chamber heating in real time and to improve the uniformity of coke heating,the temperature distribution of the coke oven’s inline riser channel needs to be checked in real time to track the temperature changes in the coke oven.The high temperature at the top of the coke oven,high dust levels and the presence of sulphides in the air led to accelerated deterioration of the electrical components of the inspection equipment.After one year of field operation,the previous generation of inspection and temperature measurement robots,experienced problems such as slow charging and frequent pushrod failures.Due to the inadequacy of the robot’s equipment condition detection methods and the lack of a sound fault diagnosis strategy,the cause of the faults that occur cannot be accurately determined.For these reasons,this project redesigned the hardware circuit,dividing the system into a power supply board,a control board,and a pusher drive board according to function.The supply voltage and current detection means of the travel mechanism,charging and discharging circuit,pusher mechanism and fan were refined,the current detection accuracy of the charging and discharging circuit was improved,and the power supply status detection of the pyrometer,card reader and alarm lamp was added.On this basis,the key threads and control logic of the software system have been optimised based on the RT-Thread operating system to improve the accuracy of battery power estimation,and a more complete system self-diagnostic mechanism has been designed to determine the causes of faults such as charging current,travel slippage and derailment,pusher action,power supply to peripheral equipment,etc.to facilitate rapid troubleshooting by field personnel and ensure the smooth progress of automatic inspection and temperature measurement.In order to further improve the accuracy of furnace temperature prediction for more accurate control of the standing flame channel temperature,this project uses the historical temperature of the standing flame channel,coke strength,coal fineness,coal moisture,blast furnace gas flow and pressure as inputs to the Long Short Term Memory(LSTM)neural network to predict the temperature of all standing flame channels for the next measurement.After sixty days of trial operation,the test results show that the improved inspection robot can accurately distinguish faults affecting the stable operation of the system such as charging,travel,pusher action,fan blocking and temperature collection,achieve rapid fault location and alarm,automatically deal with travel slippage and abnormal reset faults,effectively extending the fault-free operation cycle of the system and improving operational reliability and stability.The introduction of the control volume LSTM standing fire channel temperature prediction model has improved the accuracy of furnace temperature prediction,laying the foundation for further improvement of coke quality,energy saving and consumption reduction...