A machine learning-based warm cloud flare catalyst formulation optimization method and system

Abstract

The application discloses a kind of based on machine learning's warm cloud flame agent catalyst formula optimization method and system, belong to formula data processing technical field.Acquire the combustion position and particle size spectrum when basic formula propellant and local formula adjustment propellant continuous combustion, establish corresponding relationship between basic formula combustion state, component mass ratio change and particle size spectrum change, respectively form benchmark particle size prediction model and particle size spectrum difference prediction model, and when local formula influence range overlaps, utilize combination adjustment propellant to establish correction model;In the allowable range of component mass ratio, candidate formula is screened, discrete combustion position corresponding formula is converted into the axial formula section that can be prepared, the measured particle size spectrum of verification propellant is updated to corresponding model, and the optimization formula that satisfies target particle size spectrum determination condition is obtained.The application can improve the consistency of formula prediction result and propellant continuous combustion measured particle size spectrum.