Research And Experiment On Forming Properties Of Pellet Feed Basing On Model Theory

With the development of the feed industry,higher requirements have been put forward for the quality consistency of the pellet feed products in domestic feed enterprises in the sight of enlargement of single plant scale,the popularization of pellet feed production and the intelligentization of production quality control.However,in the actual production process,instability of the product quality exists under the same formula and production process parameters due to the difference in raw material characteristics.Additionally,the adjustment of the processing parameters for different feed formula on the same production line will also cause instability in quality.In the view of the above problems,experimental research on the quality characteristics of the main raw materials of feed and the mixture were conducted,basing on that the forming constitutive model is built in the perspective of model theory and method,in order to study the rheological properties of the material in the forming process,also the numerical simulation and optimization of the forming process and the key parts of the pelleting equipment are investigated.On account of fruit above,the quality prediction model basing on artificial intelligence model theory and control method of pellet feed products with NSGA-II algorithm focus on process parameters are established.The main results are as follows:The laboratory forming test platform was set up to study the forming characteristics of main raw materials of feed and mixture,and the effect of fat addition and processing parameters on the strength of pellet feed under the condition of scale production was studied.The results revealed that the index of pellet density,forming ratio and pellet hardness were most optimum for corn,soybean meal,and wheat respectively,while the DDGS forming characteristics were the poorest.Among all the factors,the moisture content of the material had a superior influence on each index.The addition of fat had an extremely significant influence on the pellet density and forming ratio of the finisher pigs feed(P<0.001),but insignificant on the pellet hardness(P>0.05).According to mixture experiment design,the optimum ratio of corn/soybean meal/wheat bran and cassava/soybean meal/wheat bran which aim at the pellet quality were 41.5%/25%/33.5%and 38.9%/20.5%/40.6%,respectively.The shear failure strength of the pellet feed fluctuated during the pellet process.Meanwhile,the strength index of followed factor s and its levels are significantly higher than others(P<0.001):the amount of fat addition of 1%,the length-diameter ratio of die of 12.7:1,the pellet temperature of 92?,and the condition temperature of 82.2?.On the other hand,there is a very significant and positive correlation relationship between the shear strength and the pellet durability and pellet productivity(P<0.001).The constitutive model of pellet feed based on phenomenological analysis method was constructed,and the rheological properties of the feed material in the forming process were studied,which included elastoplastic,viscosity and friction characteristics.The results show that the model determination coefficient R2 is above 0.99 and the average relative error between the model value and the test value is 3.378%,combine with ?2 examination indicated that the model has excellent performance.For the sensitivity analysis of the constitutive model,results suggested that the viscosity coefficient and strain hardening exponent have the highest impact on the model,followed by the plastic modulus and the elastic modulus.As far as it is concerned,the moisture content has a significant effect on the elastic modulus and the integrated plasticity coefficient,same as the particle size on the viscosity coefficient(P<0.05).A simulation model for pellet feed in the circumstance of laboratory and a dynamic simulation model for ring pelleting process were built based on the constitutive model,experimental data,and numerical simulation method.The results illustrated that the stress and strain index of the material inside the pellet is stratified or lumpy,and the stress,strain in the surface and interface of die hole are concentrated,which means the failure phenomenon is easy to appear in those positions.In virtue of the response surface test design method,the optimization structural parameters of the straight hole and straight decompression hole considering the total stress and strain are obtained.Moreover,the results state clearly that the force,velocity,and overlap of particles in the pelleting process exhibited a regular change,and the forming effect is better when the gap between die and roller,the opening rate of die and the inlet angle of the die hole are 0.1 mm,44.1%and 30 degrees respectively,but the increase of the extrusion effect or opening area deteriorate the fatigue life and safety factor of the ring die in a certain way.The pellet feed quality prediction model of MIV-PSO-BPNN(mean impact value-particle swarm optimization-back propagation neural network)on the basis of the artificial neural network as well as the multi-objective optimization method of processing parameters are established.The MIV-PSO-BPNN model was combined with the MIV independent variable selection,the optimization method of the PSO algorithm and neural network prediction regard the formula and process parameters.The results indicated that the determination coefficient R2 is more than 0.94,and the average prediction error of each output index is 0.442,2.185%,0.5481,as well as the optimum amplitude of error compared to the multiple linear regression model and the basic BPNN model is 84.99%and 56.95%respectively.While the stability increased by 91.46%relate to the basic BPNN model.By the aidance of MIV-PSO-BPNN model as the approximate model of processing parameters and pellet feed quality,a multi-objective optimization method based on the NSGA-? algorithm is constructed.The optimization results of the processing parameters include the length-diameter ratio of die 7:1,the diameter of the die hole 3 mm,the pore size of the shredder 2 mm and 1.5 mm,and the condition temperature 76.2?.The optimum range of productivity and PDI were 22.04%and 6.04% respectively.