Application of artificial neural networks to the prediction of tunnel boring machine penetration rate

Rate of penetration of a Tunnel Boring Machine(TBM)in a rock environment is generally a key parameter for the successful accomplishment of a tunneling project.This paper presents the results of a study into the application of an Artificial Neural Network(ANN)technique for modeling the penetration rate of tunnel boring machines.A database,including actual,measured TBM penetration rates,uniaxial compressive strengths of the rock,the distance between planes of weakness in the rock mass and rock quality designation was established.Data collected from three different TBM projects(the Queens Water Tunnel,USA,the Karaj-Tehran water transfer tunnel,Iran,and the Gilgel Gibe Ⅱ hydroelectric project,Ethiopia).A five-layer ANN was found to be optimum,with an architecture of three neurons in the input layer,9,7 and 3 neurons in the first,second and third hidden layers,respectively,and one neuron in the output layer.The correlation coefficient determined for penetration rate predicted by the ANN was 0.94.     

Optimal extraction sequence modeling for open pit mining operation considering the dynamic cutoff grade

The cutoff grade problem is an important research challenge and vital optimization task in the yearly operational planning of open pit mines due to its combinatorial nature. Because of it's influenced by the economic parameters, the capacities of stages in the mining operation, mining sequence, and grade distribution of the deposit. Essentially, it asserts that the dynamic cutoff grade at any given period is a function of the ore availability and the needs of the mill at that period. Consequently, cutoff grades strategy and extraction sequence should be considered, simultaneously. Due to its goal, various attempts have been made to develop a computerized procedure for the extraction sequence of open pit mine. None of the resulting approaches appear to enjoy wide acceptance because of it's the numerous associated variables. A new model is proposed to overcome this shortcoming. This model solves the problem in the three steps: 1) the actual economic loss associated with each type of processing for each block, 2) the probabilities distribution and average grade for each type processing is computed from independent realization, and 3) each block with its expected economic loss is developed as a binary integer programming model. Using this model, the optimum extraction sequences in each period are identified based on the optimum processing decisions. A case study is presented to illustrate the applicability of the model developed. Results showed that the extraction sequences obtained using the suggested model will be realistic and practical. This model allows for the solution of very large problem in reasonable time with very high solution quality in terms of optimal net present value.     

Biogenic amines in Iranian white brine cheese: modelling and optimisation of processing factors

The simultaneous effects of processing factors such as ripening time (25–75 days), ripening temperature (4–14 °C) and brine concentration (10–13%) on biogenic amines content, proteolysis and sensory score of Iranian white brine cheese were studied, in 12 cheeses. Response surface methodology (RSM) was used to minimise biogenic amines content. At low level of ripening time, biogenic amines content decreased with increasing levels of brine concentration but at high level of ripening time, brine concentration had inverse effect. Ripening time showed quadratic effect on biogenic amines content. Based on biogenic amines content and sensory score, the optimum conditions were 13% brine and ripening at 9–14 °C for 43–65 days.     

Isolation of lactic acid bacteria from Lighvan cheese, a semihard cheese made from raw sheep milk in Iran

The lactic acid bacteria contributing to Lighvan cheese ripening during the different stages of production were investigated. Isolated strains from different culture media were identified phenotypically to species and subspecies level. In total, 413 strains were isolated from raw milk, 1-day-old cheese and fully ripened cheese. The most abundant species belonged to Enterococcus faecium (87 isolates), Lactococcus lactis ssp. lactis (68 isolates), Enterococcus faecalis (55 isolates) and Lactobacillus plantarum (48 isolates). E. faecium, Lc. lactis and Lb. plantarum were the predominantly isolated strains from ripened cheese. Therefore, they may contribute considerably to the aroma and flavour development of Lighvan cheese.     

Proteolysis in ultra-filtered and conventional Iranian white cheese during ripening

Conventional and ultra-filtered (UF) Iranian white cheeses were made with almost identical gross chemical composition and the extent and characteristics of proteolysis were studied during ripening. UF cheeses exhibited a lower rate of development of pH 4.6-soluble nitrogen than conventional cheeses. The rates of degradation of α_s1-casein and particularly β-casein were lower in UF cheeses than in conventional cheeses. Plasmin activity was lower in UF cheeses than that in conventional cheese, whereas coagulant activity was higher in the former. Noticeable qualitative and quantitative differences were observed in reverse-phase high performance liquid chromatography (RP-HPLC) peptide profiles between UF and conventional white cheeses and chemometric analysis of peak height data distributed the cheeses into two separate groups. The levels of free amino acids in UF cheeses were lower than in conventional cheeses. Lower peptide degradation and production of amino acids suggested slower ripening, which may have been associated with the weak aroma development characteristic of UF cheeses.     

Optimum blending of iron ore from Choghart stockpiles by using goal programming

Iron and steel industry is an important sector of Iran’s economy. Choghart iron ore mine is an important iron ore producer of Iran steel industry. Phosphorous contained in the iron ore concentrates of Choghart mine has a detrimental effect on the steel making process, whereby this causes cracks to form in the refractory lining of blast furnaces. In the past, about 1.43 Mt of low-grade and 4.53 Mt of high-phosphorous materials had been transported to low grade and high phosphorous stockpiles, respectively, for future beneficiation. As a result of the progressive depletion of high-grade ore and establishment of beneficiation plant in Choghart, exploitation of these two stockpiles in this mine became an important issue. In this work, a linear goal programming (GP) model was developed in order to determine the optimum iron ore blend in terms of quality from low grade and high phosphorous stockpiles of Choghart mine. The model was solved by the SOLVER V.9 program. Results show that feeding with acceptable quality (w(Fe)≥50% and w(P)≤1.2%, mass fraction) materials can be blended from stockpiles that satisfy the needs of the Choghart processing line.