Modèle mathématique1

Réponses à toutes vos questions après le Bac (Fac, Prépa, etc.)
arthurzz
Messages: 8
Enregistré le: 03 Aoû 2017, 01:50

modèle mathématique1

par arthurzz » 05 Aoû 2017, 03:08

bonsoir à tous!
voici le problème formulé bien qu'il soit en anglais.
Applying non-explos ive expans ion material (NEEM) is one of the safest methods for rock fracturing, espec ially in a hard rock quarry. An experimental study was conducted to determine the pressure exerted by NEEM in a hole, which is necessary for analysing rock fracturing, particularly in a numerical modelling. Stress d istribution was developed in a th ick-wall cylinder in order to measure the tangential strain on the external boundary. In the experiments, nine pipes of different materials and varying d iameters were selected and
then loaded internally with a general type of NEEM. Through electrical strain gauges, the tangential strains were recorded on the external boundary of the pipes.
The results obtained from these tests and existing data from previous researchers were utilized for the statistical analysis. A new experimentally-derived formula is proposed for determining the pressure exerted by NEEM by multiple parameter regression with a high correlation coeffic ient. The proposed model has a nonlinear form with three independent parameters i.e. the hole d iameter, the time, and the modulus of elasticity of the material.

Experimental model
The physical modelling proved that the hole pressure depends on the hole diameter, the time, and the elastic modulus .
In other words, the hole pressure function is given by:
P=f(t,d,E)
where P is the pressure exerted by the NEEM, t is the time, d is the hole diameter, and E is the elastic modulus .
All of theacquired data were converted into logarithmic form, and then a linear multiple parameter regression was utilized as:
y=a_{0} + a_{1}x_{1}+ a_{2}x_{2} + a_{3}x_{3}
The hole pressure function can therefore be written as:
P=f(t,d,E)=a_{0} + a_{1}t + a_{2}d+ a_{3}E [*]
Equation [*] is another form of the following relationship:
logP=a_{0 } + a_{1}logt + a_{2}logd + a_{3}logE
where a0 is a constant value and could be replaced by log k, which is 10^a_{0}, (a_{0} = log k) . As such, Equation [19] can be rewritten as:
logP= logk + logt^{a_{1}} + logd^{a_{2}} + logE^{a_{3}}
logP= log(k.t^{a_{1}}.d^{a_{2}}.E^{a_{3}})
P= k.t^{a_{1}}.d^{a_{2}}.E^{a_{3}}
Analysing the matrix of coefficients yielded a correlation
of 0 .935 . For determining the NEEM pressure, the proposed
experimental model is as follow:
P= 0,566t^{0,933}.d^{0,407}.E^{0,493}
where P is the pressure accruing from the NEEM in
megapascals, t is the time in hours, d is the hole diameter in
metres, and E is Young’s modulus in gigapascals .

j'espère avoir bien formulé le problème. je l'ai pris tel dans le fichier. désolé si les fonction ne sont pas bien écrites mais j'ai essayé et j'ai pas pu. là il y'a des indices _{ } et des exposants symbolisés par ^{ }. merci pour votre compréhension



pascal16
Membre Légendaire
Messages: 6663
Enregistré le: 01 Mar 2017, 12:58
Localisation: Angoulème : Ville de la BD et du FFA. gare TGV

Re: modèle mathématique1

par pascal16 » 05 Aoû 2017, 09:11

Vu que tu écriture respecte le format tex, tu peux sélectionner une ligne, appuyer sur 'tex' et une fois posté, les formules s'afficherons normalement.

finalement, c'est quoi la question ?
Ils proposent de faire des régressions affines, mais donnent le résultat, alors je pige pas.

arthurzz
Messages: 8
Enregistré le: 03 Aoû 2017, 01:50

Re: modèle mathématique1

par arthurzz » 05 Aoû 2017, 10:45

ce que j'aimerais savoir Pascal c'est pourquoi ils prennent des solutions de la forme de Y tel que je l'ai écrit et puis l'introduction des logarithme

pascal16
Membre Légendaire
Messages: 6663
Enregistré le: 01 Mar 2017, 12:58
Localisation: Angoulème : Ville de la BD et du FFA. gare TGV

Re: modèle mathématique1

par pascal16 » 05 Aoû 2017, 19:35

y est approximé par un polynôme de degré 3.
P est elle des paramètres en puissance, passer en log linéarise le problème.

exemple bidon
si l=a^x, a inconnu à déterminer.
ln(l)=ln(a^x)
ln(l)=xln(a), fonction affine, il est facile de trouver a avec plusieurs expériences

arthurzz
Messages: 8
Enregistré le: 03 Aoû 2017, 01:50

Re: modèle mathématique1

par arthurzz » 06 Aoû 2017, 18:38

explique moi un peu bien cette discrétisation s'il te plait

pascal16
Membre Légendaire
Messages: 6663
Enregistré le: 01 Mar 2017, 12:58
Localisation: Angoulème : Ville de la BD et du FFA. gare TGV

Re: modèle mathématique1

par pascal16 » 06 Aoû 2017, 19:58

exemple où l'on part d'une forme et où on choisi le type de regression :
https://support.office.com/fr-fr/article/Choisir-la-meilleure-courbe-de-tendance-pour-vos-donn%C3%A9es-1bb3c9e7-0280-45b5-9ab0-d0c93161daa8

En physique, on a des modèles théoriques qui nous donnent le type de régression à appliquer. et quand on sait pas, on fait un polynomiale (c'est à dire un développement limité)

arthurzz
Messages: 8
Enregistré le: 03 Aoû 2017, 01:50

Re: modèle mathématique1

par arthurzz » 08 Aoû 2017, 14:27

je voudrais que tu me demontres bien là où ils ont introduit les logarithme, je ne comprend pas bien ça

 

Retourner vers ✯✎ Supérieur

Qui est en ligne

Utilisateurs parcourant ce forum : Aucun utilisateur enregistré et 24 invités

Tu pars déja ?



Fais toi aider gratuitement sur Maths-forum !

Créé un compte en 1 minute et pose ta question dans le forum ;-)
Inscription gratuite

Identification

Pas encore inscrit ?

Ou identifiez-vous :

Inscription gratuite