Abstract:As an important geometric error index,roundness error directly affects assembly accuracy and service life of mechanical parts.On-line measurement for intelligent manufacturing processes puts forward higher requirements for the speed and accuracy of roundness evaluation methods.Aiming at the online roundness error evaluation,a minimum area evaluation method based on least squares support vector machine (LSSVM) was proposed in combination with the grinding active measuring instrument.For LSSVM,the second norm of error and equality constraints were used to replace the error and inequality constraints in traditional support vector machines,the quadratic programming problem was transformed into solving linear equations,reducing the computational complexity and effectively improving the solution speed.By comparing the roundness error evaluation results of the four algorithms of simplex algorithm,genetic algorithm,support vector machine and LSSVM,the accuracy and feasibility of the evaluation method of minimum area roundness error based on LSSVM are verified,its high efficiency in processing huge extracted data is found,therefore it can be used to realize the on-line evaluation of the roundness error of the grinding process active measuring instrument in the production process,and the processing efficiency is improved.