Dr R.P. Jenner, J. Krabicka, University of Greenwich 1 Engineering Mathematics 2 Electronic and Computer Engineering Challenge Introduction The main aims of the computer modelling challenge are twofold; firstly, by completing the challenge you will increase your understanding of frequency analysis and transfer functions in context of an engineering application. Secondly, you will gain experience of using an industry leading mathematical simulation package such as Matlab®. All electronic/computer/business engineering challenge groups have essentially the same challenge to complete although there will be some differences to provide originality in the challenge solutions. The key to successful completion of the challenge will be planning and teamwork. As a group, you need to invest time to properly understand what you have been asked to do and put a plan in place to achieve it. Group members need to agree to this plan and stick to it! When devising your challenge plan, you need to account for…  Research  Development of systems and algorithms  Implementation of algorithms in Matlab® (and other computing languages)  Testing of results  Analysis and evaluation of results  Optimisation of algorithms based on results  Reporting The challenge! Materials in powder form are often transported in industry by blowing them through pipelines (pneumatic conveying), but it is important that they travel at the right speed. Too fast will cause unnecessary wear on the pipeline, and too slow will cause the powder to drop out of suspension and sit at the bottom of the pipeline. It is, however, difficult to know how fast the material is flowing through the pipeline. Simply measuring air speed is inaccurate because there is often some ‘slip’ between the air stream and the powder particles, causing the powder to travel more slowly than the air. A widely used method to determine the powder speed uses electrostatic sensors. When powders are pneumatically conveyed through a pipeline, they naturally build up static charge. As they travel past the metallic electrode of an electrostatic sensor, a small fluctuating voltage is produced. With suitable electronics, this signal can be captured and then digitised for computer processing. Using upstream andDr R.P. Jenner, J. Krabicka, University of Greenwich 2 downstream sensors, similar (but not identical) signals are produced and the time delay between these signals can be used to infer flow velocity (see Figure 1). Figure 1: Electrostatic velocity measurement Your challenge is to use suitable filtering of given upstream/downstream sensor signals with FIR filters and the digital processing technique of cross correlation in order to infer the powder speed. You must try to develop an optimal system that uses the least amount of processing and resources whilst still producing good results. Consider that the processing system will ultimately be implemented on a either on a processor running C code or on FPGA hardware, so few (if any) Matlab® specific functions should be used in the final system (although built-in Matlab® functions may be very useful for system development and testing). This process can be broken down into four main sections… 1. Investigate the design and implementation of FIR filters 2. Investigate the implementation of cross correlation processing 3. Apply the FIR filters and cross correlation algorithm to the given signals, analyse the results of using various processing parameters. 4. Optimise the system for the best balance of accuracy, speed and resource utilisation. Remember! To be successful, you must work as a team and you must derive a detailed plan of work; make full use of your laboratory time and make sure you meet (as a group) with your tutors regularly. Meet with your tutors during scheduled laboratories and via the online appointment system; bookable appointments will be available every Monday throughout term 2. Submission Deadlines Group challenge specification report, upload to Moodle by midnight Sunday 5th February 2017 Individual report, upload to Moodle by midnight Sunday 9th April 2017 .