Auto-testing of code

Abstract: The topic of this thesis is software in systems of automated machines that produce products in a company. Testing of these systems is crucial for the machines to work as intended. Usually testing is done manually, both while the system is developed and while the production is running. This is called traditional testing. Testing is done to improve the machines program code. The machine needs to do something wrong to fail the test. Certain faults lead to the change of certain parts of the code. Test-cases have to be developed to address different parts of the code. Nowadays testing is usually done completely manually. This thesis will explore the possibilities and the limitations with partly automatic testing, as well as develop a theoretical machine to test. The thesis was done at a machine company in Helsingborg. The approach is to use an automatic regression testing of the code. Firstly an analysis was done which went through the existing options and refined the approach. One approach was selected and used for the thesis. The main task of the thesis is to facilitate the way of testing the machines for companies, or at least help companies to improve machines, with earlier commissioning and enable distance maintenance. The analysis went through different software aspects and old reports written about automatic testing before implementing a testing strategy to a control system. The control system needed to be developed and a theoretical machine were needed to be tested. The next part was to simulate the machine and lastly illustrate the scenario with a pictorial view. By developing the machine code it was possible to create automatic tests on it, called unit testing. This was combined with a simulation of the machine. This was successfully shown on a human machine interface. By doing this it is possible to save time and money but it will depend on the programmer creating the tests. The simulation was done through software that coded the controller, called TIA portal. In the future a better and more suitable environment could be used for simulation purposes. Unit testing and simulation proved to work well for automatic testing. Remote maintenance and testing requires simulating the entire machine with all of its complex parts. All physical hardware can be simulated. Simulating and testing on a machine was implemented. The combination is considered to be time and money saving and that there are options for making the tests partly automatic and machines with higher quality.