Open Collections

Expert system control of a flotation circuit

University of British Columbia

Expert systems technology is a discipline of artificial intelligence that has recently emerged from the research environment and is currently making significant inroads into business and industry. The proponents of expert systems make many attractive claims. Two of the claims are that expert systems can capture the knowledge of the experts and can be programmed by non-programmers. To date, most uses of this technology in the process industry are in off-line applications, that is, applications that are not directly tied to operating environments. Moreover, those that are used in on-line environments are used as advisors and only suggest changes, human operators are required to close-the-loop to the process. This technology should be applicable to operating environments and should the claims of its proponents be valid, then it should also be better than existing tools currently in use. In an operating environment it could use the knowledge of experienced operators as an intelligent controller and apply it directly to a process without the intervention of human operators. In this thesis the prospect of using an expert system as an intelligent controller is investigated. The thesis offers background to expert systems, how expert systems are related to artificial intelligence and what the generic components of an expert system are. As a test of the technology an expert system was developed as an intelligent controller in a mineral processing application. The prototype expert system was developed as a supervisory controller in the copper flotation circuit in the concentrator at Brenda Mines Ltd.. The expert system operated on-line and controlled the process in real-time. It read sensor data and using the operating experience of Brenda's flotation operators, manipulated regulatory controller setpoints as deemed necessary. The expert system was able to manipulate directly reagent flowrates and process air flowrates. Also, it suggested changes to other process variables not directly under its control. The manipulated variables were collector and frother reagent flowrates, the air flowrate to a bank of scavenger flotation cells and the air flowrate to four flotation columns. The operators were allowed to, and sometimes required to, intervene in instances of large process upsets. Results of a month long trial period in the flotation circuit indicated the success of the application. The expert system was indeed able to maintain metallurgical performance at a level approximating that of the operating experts, however, no direct comparison between the performance of the circuit under expert system control, versus performance under operator control, was possible. A rudimentary comparison was made between the circuit's performance under expert system control and its historic performance. The results were favourable. A less tangible measure, though as important, were the attitudes of the operators toward the system. Almost unanimously, the operators felt the system eased the task of circuit operation and was responsible for better circuit performance.

Text

2010-10-27

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/29600

Mining Engineering

University of British Columbia

Graduate