PROJECT PROFILE

BlueScope is Australia’s largest steel producer.  Its Port Kembla plant is a vertically integrated manufacturing facility incorporating iron making and downstream finished products including household brands such as COLORBOND® steel, strip steel coils and flat steel plate products.

​

BlueScope's Port Kembla plant manufactures high quality plate steel tested to Australian and International Standards.  Standards require samples to be taken at regular intervals to build up build up a statistical profile of the manufacturing performance of plant as well as the particular properties of the steel batches produced.  The quality process therefore requires cutting samples from selected plates based on plate attributes such as steel grade, thickness and batch size.  

​

Plates must be taken off the processing lines for the samples to be cut.  BlueScope wanted to investigate the possibility of installing a robot to automatically cut the required samples whilst plates remained on the processing line.  For this to feasible the length of time required for the robot to cut a sample must not introduce a bottleneck in the process.  This would be fairly straight forward if time the needed to cut a sample was consistent for all plate sizes and grades.  But alas - things are rarely that simple.    Cutting time has an exponential relationship to plate gauge.      As the plant processes a wide range of plates in varying grades and thickness, sampling cutting time vary widely making it near impossible - (using simple methods) to determine whether or not the robot would improve the situation or adversely impact production.

​

The solution?

  

BlueScope had already had a digital twin of the plate manufacturing process.   This model was used as a basis for testing the efficacy of the proposed sample robot.  A database was developed using AnyLogic's internal SQL tables that included the testing requirements defined in the Australian Standard for all grade and gauge combinations.  When a plate arrives at the sampling station, the model would queries the database to ascertain what - if any - tests are required.  An algorithm then calculates the length of time required to cut the sample using the plate's gauge and grade parameters along with the required sample type as variables in the equation.    

 

The source data for this model is same as the digital twin - real time production data covering both WIP inventory and planned rolling schedules.  The model produces detail log files enabling technicians to verify if all the tests taken are consistent with Standard.  The log also lists along with each sample the length of time the robot needed to cut the sample.  The result is a dynamic model that accurately mimics how the proposed robot would impact the manufacturing process using real time production data.  Engineers and production personnel can now easily and accurately assess the merits of the proposed robot station. 

​

​

In this case, simulation modelling was used to tame uncertainty, and shed on a complex, dynamic problem.  It enabled BlueScope to make the right decision.   Are you dealing with a complex dynamic problem?    Could simulation modelling help you to make right decision?  Give us call - we'd be happy to help. 

​