Gedvilas R. Determination of fatigue curves parameters and lifetime at low cycle straining for stainless steels: Master thesis of mechanical engineer / research advisor associate professor dr. R. Šniuolis; Šiauliai University, Technological Faculty, Mechanical Engineering Department. – Šiauliai, 2006. - 58 p. Cycle fracture parameters of structural materials at low cycle straining are understood as lifetime or low cycle fatigue curves that are constructed in coordinates or according to the number of loading cycles N. The lifetime of material at low cycle straining in elastic plastic area is expressed by Coffin’s equation. Parameter of Coffin’s equation C markedly differs from the experimental, therefore high errors occur in evaluating lifetime of the material. It is noticed in many works that parameter proposed by Coffin is close to experimental. Parameter m=0,5, proposed by Coffin, is determined according the low cycle fatigue curves in coordinates for structural materials. Parameters of low cycle fatigue curves mp, Cp for stainless steels according to Coffin were specified and parameters me, Ce according to elastic strain at low cycle straining at room and elevated temperature were determined. Analytic dependences of low cycle fatigue curve parameters on modified plasticity (su/sy)y were proposed, lifetime for stainless steels at room and elevated temperature was calculated in this work. Statistical analysis of results of low cycle fatigue fracture parameters and... [to full text]

Accidental errors are inevitable in the detenninatiorr of the fatigue curves parameters and mechanical characteristics of welded joint materials. That's why the investigation results are accidental and are analysed by means of statistical method. The use of statistical methods includes analysis of literature about test methods and treatment of experimental data. The purpose of this work is to propose consequent methods of statistical treatment of investigation results by the means of the main computation methods.

Konstruojant modelį, susiduriama su optimalios architektūros parinkimo problema. Optimalus modelis reiškia optimalų netiesinės regresijos komponenčių skaičių ir svorių reikšmes su kuriomis modelis „geriausiai“ prognozuoja tirpalo koncentracijas. Optimaliam komponenčių skaičiui nustatyti pasitelkiame pažingsninės regresijos metodą. Skaičiavimams buvo pasirinkta netiesinės regresijos sigmoidinė aktyvacijos funkcija, šis pasirinkimas buvo pagrįstas P.L. Bartlet [1] darbo rezultatais. Modelyje taip pat taikomas nepilnai apibūdintų duomenų analizės metodas, pagrįstas Volker Tresp [3] darbo rezultatais. Šis metodas yra plačiai taikomas šiuolaikinėje informacinėse technologijose bei matematinių modelių konstravime, kai turima pilna informacija apie modelių įėjimus tačiau išėjimo duomenys yra žinomi ne visais atvejais. Mūsų atveju, yra žinomi tikslūs amperometro parodymai – srovės stipris tam tikru laiko momentu (realybėje šiuos duomenis nesunku pamatuoti), tačiau duomenys apie tikslias tirpalų koncentracijas yra žinomi ne visi. Panaudojome pagrindinių komponenčių analizės metodą (PCA) ir sumažinome turimų duomenų rinkinius, taip išsaugodami didžiąją dalį informacijos apie tuos duomenis. Tam tikri modelio parametrai : , buvo parenkami bandymų keliu, stebint modelio vidurkių kvadratų sumą(MSE). Gauti rezultatai rodo, kad sukurtas regresijos modelis pakankamai gerai klasifikuoja duomenis, tačiau tiksliems koncentracijų prognozavimams reikalingi papildomi tyrimai. / The purpose of this work is to create non liner regression model with optimal number of coefficients and optimal values of these coefficients, and predict liquor concentration having values of amperimetric data. The main task of the work: • To use Semi-supervised learning algorithm while creating the model (Kai Yu, Volker Tresp [4]) • To apply P.L. Bartlet [1] idea, that in non-linear regression model the value of the weights is more important than number of these weights. • To create mathematical model that could calculate optimal non-liner regression’ weights using stochastic search, and could determine the concentration of liquor according biosensors response curve While creating the model, we met model optimization problem. The optimal model means the optimal number of model’ weights and the values of weights with whom the model best predicts the concentration of the liquor. We used step wise regression method to determine the optimal number of weights. For the calculations we choose non-liner regressions sigmoid function according P.L. Bartlet [1] work results. We also implemented semi supervised data analysis method that was chosen based on Volker Tresp [3] work results. These methods are widely applied in the mathematical modeling and information technologies. What is more, we also applied partial component analysis and reduced the data sets of response curve, without loosing significant information about these data. The parameters , of the model we chosen in... [to full text]