Biomass Recognition And Check Model On Coal And Biomass Co-firing Process

Biomass co-firing, a potential technology, has been paid high attention and utilized in the world. There are 211 biomass co-fring power plants installed in 17 countries. Pure biomass burning has made a great development in china by 2006. However, biomass co-firing project is limited. The amount of consumped biomass in co-firing boiler is not measured by some proper methods, which leads to the subsidy not paid. Therefore a new technology is needed to measure the quantity of biomass in real time.A new technology is developed by image recognition theory. It can be used to account the weight of biomass with the help of weight system. In this thesis an auto image acquisition system was built up and the software panel can fulfill many founctions such as image frame capture, display, storage, query etc. The acquired image was pre-processed with different algorithms for example smooth, filtering, histogram equalization etc. 20 statisical values (gray, HSV system values, area, perimeter, counter number, texture parameters) can be used to discriminated various feed stock except Hue. Because the distribution of Hue value is haphazard and it has no clear regular pattern.A specific experiment was carried out in the auto image recognition system. First of all, 108 images of 9 feedstock were captured and then 9×12×20 characteristic parameters were calculated with principal component analysis algorithm to build a model library for feedstock recognition. After sample training, 18 images of 2 group feedstock were selected randomly, and projected the eigenvectors of these images to the principal space. The results show that: principal components Z1, Z2, Z3 represent gray, morphology, HSV system respectively. Six biomass and three coals were recognized by subspace of Z1 and Z2 easily.Another method was studied. The sulfur content of biomass and coal has a big difference, so the amount of consumped biomass can be calculated by the exhausted sulfur dioxide. The author built a pipe still—FTIR system and studied SO₂ release in cornstalk,bark and Shenhua coal co-firing process under different co-firing rates. It is indicated that SO₂ yield decrease with co-firing rate increase and the formula y=A₂+(A₁-A₂)/(1+exp(x-x₀)/x₁) represents the relationship between SO₂ yield and co-firing rate. According to this formula and emission data of SO₂ an experience model can be set up, which could compute the co-firing rate and check the result of the first method.