Votre navigateur n'accepte pas le Javascript.La navigation sur ce site risque de ne pas fonctionner correctement.

248. Trouvez une équation cartésienne d’hyperbole connaissant ses asymptotes et un de ses points

Dans un repère orthonormé $(O, \vv{i}, \vv{j})$ considérez la droite $\mathscr{D}_1$ d’équation $y =x+2$ et la droite $\mathscr{D}_2$ d’équation $y=-2x+4.$

Soit $A(4,3)$ un point du plan n’appartenant ni à la droite $\mathscr{D}_1$, ni à la droite $\mathscr{D}_2.$

Vous obtenez le schéma suivant :

06/04/2022 - Construction dune hyperbole connaissant ses asymptotes et un de ses points

Déterminez une équation de la réunion des deux droites $\mathscr{D}_1$ et $\mathscr{D}_2$

Pour tout point $M$ de coordonnées $(x,y)$ dans le repère $(O, \vv{i}, \vv{j}):$

\begin{align*}
M\in \mathscr{D}_1\cup \mathscr{D}_2 &\Longleftrightarrow (x-y+2)(2x+y-4)=0\\
&\Longleftrightarrow  2x^2+xy-4x-2xy-y^2+4y+4x+2y-8=0\\
&\Longleftrightarrow  2x^2-xy-y^2+6y-8=0.
\end{align*}

Déterminez une équation du second degré satisfaite par les coordonnées du point $A$

Quand vous choisissez $A$, vous avez $x=4$ et $y=3.$

Comme $A$ n’appartient ni à la droite $\mathscr{D}_1$, ni à la droite $\mathscr{D}_2$ en remplaçant dans l’équation obtenue précédemment, vous ne trouverez pas $0.$ Calculez précisément ce que vous obtenez :

\begin{align*}
  2x^2-xy-y^2+6y-8 &= 2\times 16-12-9 + 18-8\\
&= 32-21+10\\
&=21.
\end{align*}

Proposez une équation du second degré

Soit $\mathscr{H}$ l’ensemble des points $M$ du plan de coordonnées $(x,y)$ dans le repère $(O, \vv{i}, \vv{j})$ satisfaisant l’équation :

\boxed{2x^2-xy-y^2+6y-29=0.}

Les calculs précédents montrent que $A\in\mathscr{H}.$

Vous allez prouver que $\mathscr{H}$ est une hyperbole qui admet les droites $\mathscr{D}_1$ et $\mathscr{D}_2$ pour asymptotes.

Changez d’origine

Cherchez le point d’intersection des droites $\mathscr{D}_1$ et des droites $\mathscr{D}_2.$ Cela conduit à la résolution du système :

\begin{align*}
x-y &=-2\\
2x+y&=4.
\end{align*}

La somme des deux lignes fournit $3x = 2$ donc $x = \frac{2}{3}.$

Puis $y=x+2$ donc $y=\frac{2}{3}+2$ d’où $y=\frac{8}{3}.$

Ainsi le point $\Omega\left(\frac{2}{3}, \frac{8}{3}\right)$ est bien le point d’intersection des droites $\mathscr{D}_1$ et $\mathscr{D}_2.$

Pour tout point $M$ du plan vous notez $(x,y)$ ses coordonnées dans le repère $(O, \vv{i}, \vv{j})$ et $(x’,y’)$ ses coordonnées dans le repère $(\Omega, \vv{i}, \vv{j}).$

Les relations suivantes traduisent ce changement :

\begin{align*}
x' &= x-\frac{2}{3}\\
y' &= y-\frac{8}{3}.
\end{align*}

De là vous déduisez une équation de la réunion $\mathscr{D}_1\cup \mathscr{D}_2:$

\begin{align*}
M\in \mathscr{D}_1\cup \mathscr{D}_2 &\Longleftrightarrow  2\left(x'+\frac{2}{3}\right)^2-\left(x'+\frac{2}{3}\right)\left(y'+\frac{8}{3}\right)-\left(y'+\frac{8}{3}\right)^2+6\left(y'+\frac{8}{3}\right)-8=0\\
&\Longleftrightarrow  2\left(x'^2+\frac{4}{9}+\frac{4}{3}x'\right)-\left(x'y'+\frac{8}{3}x'+\frac{2}{3}y'+\frac{16}{9}\right)\\
&\qquad -\left(y'^2+\frac{64}{9}+\frac{16}{3}y'\right)+6y'+16-8=0\\
&\Longleftrightarrow 2x'^2+\frac{8}{9}+\frac{8}{3}x'-x'y'-\frac{8}{3}x'-\frac{2}{3}y'-\frac{16}{9}\\
&\qquad -y'^2-\frac{64}{9}-\frac{16}{3}y'+6y'+8=0\\
&\Longleftrightarrow  18x'^2+8-9x'y'-6y'-16-9y'^2-64-48y'+54y'+72=0\\
&\Longleftrightarrow  18x'^2-9x'y'-9y'^2=0\\
&\Longleftrightarrow  2x'^2-x'y'-y'^2=0.
\end{align*}

Note. Vous constatez que vous obtenez une équation qui commence par les mêmes termes que l’équation initiale, sans les termes de degré $1.$

Comme l’équation de l’ensemble $\mathscr{H}$ diffère d’un terme constant par rapport à l’équation proposée de $\mathscr{D}_1\cup \mathscr{D}_2$ vous obtenez :

\begin{align*}
M\in \mathscr{H} &\Longleftrightarrow  2\left(x'+\frac{2}{3}\right)^2-\left(x'+\frac{2}{3}\right)\left(y'+\frac{8}{3}\right)-\left(y'+\frac{8}{3}\right)^2+6\left(y'+\frac{8}{3}\right)-29=0\\
&\Longleftrightarrow  2x'^2-x'y'-y'^2=21.
\end{align*}

Effectuez une rotation des axes du repère

Soit $\theta$ un nombre réel que vous choisirez plus tard. Faites subir à la base $(\vv{i}, \vv{j})$ une rotation de l’angle $\theta$ ce qui fournit une nouvelle base $(\vv{u},\vv{v}).$

Pour tout point $M$ du plan, notez $(x’,y’)$ ses coordonnées dans le repère $(\Omega, \vv{i}, \vv{j})$. Notez $(X,Y)$ ses coordonnées dans le repère $(\Omega, \vv{u}, \vv{v}).$

Le lien entre $(x’,y’)$ et $(X,Y)$ est le suivant :

\begin{align*}
x'&=X\cos\theta-Y\sin\theta \\
y'&=X\sin\theta+Y\cos\theta.
\end{align*}

Alors, il vient :

\begin{align*}
M\in \mathscr{D}_1\cup \mathscr{D}_2 &\Longleftrightarrow  2x'^2-x'y'-y'^2=0 \\
&\Longleftrightarrow  2(X\cos\theta-Y\sin\theta)^2 - (X\cos\theta-Y\sin\theta)(X\sin\theta+Y\cos\theta)\\
&\qquad -(X\sin\theta+Y\cos\theta)^2=0\\
&\Longleftrightarrow 2(X^2\cos^2 \theta+Y^2\sin^2\theta - 2XY\sin\theta\cos\theta)\\
&\qquad -(X^2\sin\theta\cos\theta + XY(\cos^2\theta-\sin^2\theta) - Y^2\sin\theta\cos\theta)\\
&\qquad -(X^2\sin^2\theta+Y^2\cos^2\theta+2XY\sin\theta\cos\theta)=0\\
&\Longleftrightarrow (2\cos^2\theta - \sin\theta\cos\theta - \sin^2\theta)X^2\\
&\qquad +(-6\sin\theta\cos\theta +\sin^2\theta - \cos^2\theta )XY\\
&\qquad +(2\sin^2\theta+\sin\theta\cos\theta-\cos^2\theta) Y^2=0\\
&\Longleftrightarrow (2\cos^2\theta - \sin\theta\cos\theta - \sin^2\theta)X^2\\
&\qquad +(-3\sin2\theta - \cos 2\theta )XY\\
&\qquad +(2\sin^2\theta+\sin\theta\cos\theta-\cos^2\theta) Y^2=0.
\end{align*}

Vous allez obtenir de même que :

\begin{align*}
M\in \mathscr{H}
&\Longleftrightarrow (2\cos^2\theta - \sin\theta\cos\theta - \sin^2\theta)X^2\\
&\qquad +(-3\sin2\theta - \cos 2\theta )XY\\
&\qquad +(2\sin^2\theta+\sin\theta\cos\theta-\cos^2\theta) Y^2=21 \\
&\Longleftrightarrow (4\cos^2\theta - 2\sin\theta\cos\theta - 2\sin^2\theta)X^2\\
&\qquad +(-6\sin2\theta - 2\cos 2\theta )XY\\
&\qquad +(4\sin^2\theta+2\sin\theta\cos\theta-2\cos^2\theta) Y^2=42 \\
&\Longleftrightarrow \left(4\frac{1+\cos 2\theta}{2} - \sin2\theta - 2\frac{1-\cos 2\theta}{2}\right)X^2\\
&\qquad +(-6\sin2\theta - 2\cos 2\theta )XY\\
&\qquad +\left(4\frac{1-\cos 2\theta}{2}+\sin2\theta-2\frac{1+\cos 2\theta}{2}\right) Y^2=42 \\
&\Longleftrightarrow \left(2+2\cos 2\theta - \sin2\theta - 1+\cos 2\theta\right)X^2\\
&\qquad +(-6\sin2\theta - 2\cos 2\theta )XY\\
&\qquad +\left(2-2\cos 2\theta+\sin2\theta-1-\cos 2\theta\right) Y^2=42 \\
&\Longleftrightarrow \left(3\cos 2\theta - \sin2\theta + 1\right)X^2+(-6\sin2\theta - 2\cos 2\theta )XY\\
&\qquad +\left(-3\cos 2\theta+\sin2\theta+1\right) Y^2=42.
\end{align*}

Vous obtenez donc :

\begin{align*}
M\in \mathscr{D}_1\cup \mathscr{D}_2 &\Longleftrightarrow \left(3\cos 2\theta - \sin2\theta + 1\right)X^2+(-6\sin2\theta - 2\cos 2\theta )XY\\
&\qquad +\left(-3\cos 2\theta+\sin2\theta+1\right) Y^2=0.
\end{align*}

Note. Excepté le membre de droite, vous constatez que l’ensemble $\mathscr{H}$ et la réunion $\mathscr{D}_1\cup \mathscr{D}_2$ admettent des termes strictement identiques.

Choisissez convenablement l’angle de rotation

Le point clé est de choisir le réel $\theta$ afin d’annuler le terme croisé. Vous cherchez donc $\theta\in\R$ tel que :

\begin{align*}
-6\sin2\theta - 2\cos 2\theta&=0\\
-3\sin2\theta - \cos 2\theta&=0\\
3\sin2\theta + \cos 2\theta&=0\\
\frac{3}{\sqrt{10}}\sin2\theta +\frac{1}{\sqrt{10}} \cos 2\theta&=0.
\end{align*}

Comme le point de coordonnées $\left(\frac{3}{\sqrt{10}}, -\frac{1}{\sqrt{10}}\right)$ appartient au cercle trigonométrique, il existe un réel $\varphi \in]-\pi, \pi]$ tel que :

\left\{
\begin{align*}
\cos\varphi &= \frac{3}{\sqrt{10}}\\
\sin\varphi &= -\frac{1}{\sqrt{10}}.
\end{align*}\right.

Note. Après conversion du réel $\varphi$ en degrés, vous obtenez $-18,43°$ environ.

En choisissant $\theta = \frac{\varphi}{2}$ (ce qui correspond à environ $-9,22°$) vous obtenez :

\begin{align*}
\cos 2\theta &= \cos \varphi = \frac{3}{\sqrt{10}}\\
\sin 2\theta &= \sin \varphi = -\frac{1}{\sqrt{10}}.
\end{align*}

Du coup :

\begin{align*}
\frac{3}{\sqrt{10}}\sin2\theta +\frac{1}{\sqrt{10}} \cos 2\theta&= \cos 2\theta\sin2\theta - \sin2\theta\cos2\theta\\
&=0.
\end{align*}

Le terme croisé est bien annulé.

Vous calculez maintenant le coefficient de $X^2:$

\begin{align*}
3\cos 2\theta - \sin2\theta + 1 &= \frac{9}{\sqrt{10}}+\frac{1}{\sqrt{10}}+1\\
&=  \frac{10}{\sqrt{10}}+1\\
&= \sqrt{10}+1.
\end{align*}

Vous calculez maintenant le coefficient de $Y^2:$

\begin{align*}
-3\cos 2\theta+\sin2\theta+1 &= \frac{-9}{\sqrt{10}}-\frac{1}{\sqrt{10}}+1\\
&=  \frac{-10}{\sqrt{10}}+1\\
&= 1-\sqrt{10}.
\end{align*}

Dans le repère $(\Omega, \vv{u}, \vv{v})$ vous obtenez :

\begin{align*}
M\in \mathscr{H}
&\Longleftrightarrow (1+\sqrt{10})X^2 +(1-\sqrt{10}) Y^2=42\\
M\in \mathscr{D}_1\cup \mathscr{D}_2 &\Longleftrightarrow (1+\sqrt{10})X^2 +(1-\sqrt{10}) Y^2=0.
\end{align*}

Vous posez $a = \frac{\sqrt{42}}{\sqrt{1+\sqrt{10}}}$ et $b = \frac{\sqrt{42}}{\sqrt{\sqrt{10}-1}.}$

Vous obtenez finalement :

\boxed{\begin{align*}
M\in \mathscr{H}
&\Longleftrightarrow \frac{X^2}{a^2} - \frac{Y^2}{b^2}=1\\
M\in \mathscr{D}_1\cup \mathscr{D}_2 &\Longleftrightarrow \frac{X^2}{a^2}-\frac{Y^2}{b^2}=0.
\end{align*}}

Ainsi, l’ensemble $\mathscr{H}$ est bien une hyperbole, dont les asymptotes sont les droites $\mathscr{D}_1$ et $\mathscr{D}_2.$

Tracez l’hyperbole $\mathscr{H}$ et ses asymptotes

Numériquement, $a\approx 3,18$ et $b\approx 4,41.$

L’égalité $c^2=a^2+b^2$ permet de placer les foyers $F$ et $F’$ sur l’axe transversal qui passe par le point $\Omega$ étant donné que $c = \Omega F = \Omega F’$ et que $\Omega$ est le milieu du segment $[FF’].$

08/04/2022 - Trace dune hyperbole passant par un point a et connaissant ses asymptotes

247. Etudiez la nature d’une conique à centre (2/3)

Soient deux réels $a$ et $b$ fixés et un réel $h$ tel que $ab-h^2\neq 0.$

Dans un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez la conique $\mathscr{C}$ formée par l’ensemble des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2=1.

Dans cette chronique vous supposez que l’équation caractéristique $\boxed{\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2}$ d’inconnue $r\in\C^{*}$ admet deux solutions réelles opposées et deux solutions imaginaires pures opposées.

Vous allez démontrer que $\mathscr{C}$ est une hyperbole, en déterminant ses sommets, son axe transversal (celui qui contient ses sommets) et son axe conjugué qui lui est perpendiculaire.

Le cas où $h$ est nul

Le traitement de ce paragraphe vous est laissé.

Traitez le cas où $h$ est non nul

Soit $\theta$ un nombre réel, il sera choisi plus tard. Considérez la base $(\vv{u}, \vv{v})$ obtenue en faisant subir à la base $(\vv{i}, \vv{j})$ une rotation d’angle $\theta.$

Vous notez $(x,y)$ les coordonnées d’un point $M$ du plan dans le repère $(O,\vv{i},\vv{j})$ et $(X,Y)$ ses coordonnées dans le repère $(O,\vv{u},\vv{v}).$

Dès lors :

\begin{align*}
x&=X\cos\theta-Y\sin\theta \\
y&=X\sin\theta+Y\cos\theta.
\end{align*}

Vous obtenez :

\begin{align*}
ax^2+2hxy+by^2 = 1 &\Longleftrightarrow a(X\cos\theta-Y\sin\theta)^2\\
&\qquad + 2h(X\cos\theta - Y\sin\theta)(X\sin\theta+Y\cos\theta)\\
&\qquad +b(X\sin\theta + Y\cos\theta)^2=1\\
&\Longleftrightarrow a(X^2\cos^2\theta+Y^2\sin^2\theta-2XY\sin\theta\cos\theta)^2\\
&\qquad +2h(X^2\sin\theta\cos\theta +XY(\cos^2\theta - \sin^2\theta)-Y^2\sin\theta\cos\theta)\\
&\qquad +b(X^2\sin^2\theta+Y^2\cos^2\theta+2XY\sin\theta\cos\theta) = 1\\
&\Longleftrightarrow (a\cos^2\theta +2h\sin\theta\cos\theta + b\sin^2\theta)X^2\\
&\qquad +2((b-a)\sin\theta \cos\theta + h(\cos^2\theta-\sin^2\theta))XY\\
&\qquad +(a\sin^2\theta -2h\sin\theta\cos\theta + b\cos^2\theta)Y^2 = 1.
\end{align*}

Explicitez les nombres obtenus avec l’angle $2\theta$

Posez :

A = a\cos^2\theta +2h\sin\theta\cos\theta + b\sin^2\theta.

Vous utilisez les formules de l’angle double :

\begin{align*}
\cos^2\theta &= \frac{1+\cos 2\theta}{2}\\
2\sin\theta \cos\theta &= \sin 2\theta\\
\sin^2\theta &=\frac{1-\cos 2\theta}{2}.
\end{align*}

Par suite :

\begin{align*}
A &= a\frac{1+\cos 2\theta}{2} +h\sin 2\theta + b\frac{1-\cos 2\theta}{2}\\
&=\frac{a+b+(a-b)\cos 2\theta + 2h\sin 2\theta}{2}.
\end{align*}

Posez :

H =(b-a)\sin\theta \cos\theta + h(\cos^2\theta-\sin^2\theta).

Alors :

\begin{align*}
H &= \frac{1}{2}(b-a)\sin 2\theta + h\cos 2\theta\\
&=\frac{(b-a)\sin2\theta + 2h\cos2\theta}{2}.
\end{align*}

Posez :

B = a\sin^2\theta -2h\sin\theta\cos\theta + b\cos^2\theta.

Alors :

\begin{align*}
B &=a\frac{1-\cos2\theta}{2}-h\sin2\theta+b\frac{1+\cos2\theta}{2}\\
&=\frac{a+b+(b-a)\cos 2\theta-2h\sin2\theta}{2}.
\end{align*}

Choisissez $\theta$ pour annuler le nombre $H$

Vous posez $\boxed{\delta = \sqrt{(b-a)^2+4h^2}.}$ Remarquez que $\delta > 0$ vu que $h\neq 0.$

Le point de coordonnées $\left(\frac{b-a}{\delta}, \frac{-2h}{\delta}\right)$ est situé sur le cercle trigonométrique puisque :

\begin{align*}
\left(\frac{b-a}{\delta}\right)^2 + \left(\frac{-2h}{\delta}\right)^2 &= \frac{(b-a)^2+4h^2}{\delta^2}\\
&=1.
\end{align*}

Donc il existe un réel $\varphi$ tel que :

\begin{align*}
\sin \varphi &= \frac{-2h}{\delta}\\
\cos \varphi &= \frac{b-a}{\delta}. 
\end{align*}

Pour annuler le nombre $H$ il suffit de choisir $\theta$ pour avoir :

(b-a)\sin2\theta + 2h\cos2\theta = 0.

En divisant par $\delta$ cela s’écrit :

\cos\varphi\sin2\theta - \sin\varphi \cos2\theta = 0.

En vertu de la formule de soustraction du sinus, vous obtenez :

\sin(2\theta - \varphi) = 0.

En définitive, vous choisissez $\boxed{\theta = \frac{\varphi}{2}}$ où $\varphi$ est un réel appartenant à la réunion d’intervalles $]-\pi, 0[\cup ]0, \pi[$ satisfaisant les conditions:

\boxed{\begin{align*}
\sin \varphi &= \frac{-2h}{\delta}\\
\cos \varphi &= \frac{b-a}{\delta}. 
\end{align*}}

En effet, $\varphi$ peut être choisi dans l’intervalle $[-\pi,\pi]$ mais il convient d’exclure les valeurs $\pi$, $-\pi$ et $0$ étant donné que $h\neq 0$ implique $\sin\varphi \neq 0.$

Ainsi le réel $\theta$ appartient à la réunion d’intervalles $]-\pi/2, 0[\cup ]0, \pi/2[.$ Cette précision sera importante dans la recherche des axes de la conique $\mathscr{C}.$ En effet, $\sin \theta$ et $\cos \theta$ sont non nuls.

Explicitez les nombres $A$ et $B$ en fonction des données de départ

Vous avez :

\begin{align*}
A &= \frac{a+b+(a-b)\cos 2\theta + 2h\sin 2\theta}{2}\\
&=\frac{a+b+(a-b)\cos \varphi + 2h\sin \varphi}{2}\\
&=\frac{a+b+(a-b)\displaystyle \frac{b-a}{\delta} + 2h\frac{-2h}{\delta}}{2}\\
&=\frac{\delta(a+b)-(a-b)^2-4h^2 }{2\delta}\\
&=\frac{\delta(a+b)-\delta^2 }{2\delta}\\
&=\frac{a+b- \delta}{2}.
\end{align*}

D’autre part :

\begin{align*}
B &= \frac{a+b+(b-a)\cos 2\theta-2h\sin2\theta}{2} \\
&=  \frac{a+b+(b-a)\cos \varphi-2h\sin \varphi}{2} \\
&=  \frac{a+b+(b-a)\frac{b-a}{\delta}-2h\frac{-2h}{\delta}}{2} \\
&=  \frac{\delta(a+b)+(b-a)^2+4h^2}{2\delta} \\
&=  \frac{\delta(a+b)+\delta^2}{2\delta} \\
&=  \frac{a+b+\delta}{2}.
\end{align*}

Remarque. Le produit $AB$ fournit :

\begin{align*}
AB &=   \frac{a+b+\delta}{2} \times \frac{a+b+\delta}{2}\\
&= \frac{(a+b)^2-\delta^2}{4}\\
&=\frac{a^2+b^2+2ab-(a-b)^2-4h^2}{4}\\
&=\frac{a^2+b^2+2ab-a^2-b^2+2ab-4h^2}{4}\\
&=\frac{4ab-4h^2}{4}\\
&=ab-h^2.
\end{align*}

Comme $ab-h^2\neq 0$ vous déduisez $A\neq 0$ et $B\neq 0.$

Revenez à l’équation caractéristique

Pour tout nombre complexe $r$ non nul :

\begin{align*}
\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2 &\Longleftrightarrow (r^2a-1)(br^2-1)=h^2r^4\\
&\Longleftrightarrow abr^4-(a+b)r^2+1=h^2r^4\\
&\Longleftrightarrow (ab-h^2)r^4-(a+b)r^2+1=0.
\end{align*}

Le trinôme $(ab-h^2)X^2-(a+b)X+1$ a un discriminant égal à :

(a+b)^2-4(ab-h^2).

En développant, vous trouvez :

\begin{align*}
(a+b)^2-4(ab-h^2) &= a^2+b^2+2ab-4ab+4h^2\\
&= a^2+b^2-2ab+4h^2\\
&= (a-b)^2+4h^2\\
&= \delta^2.
\end{align*}

Le trinôme $(ab-h^2)X^2-(a+b)X+1$ admet deux racines réelles distinctes $x_1$ et $x_2$ vu que son discriminant est strictement positif ce qui donne :

\begin{align*}
x_1 &= \frac{a+b-\delta}{2(ab-h^2)} = \frac{A}{ab-h^2} = \frac{A}{AB} = \frac{1}{B}\\
x_2 &= \frac{a+b+\delta}{2(ab-h^2)} = \frac{B}{ab-h^2} = \frac{B}{AB} = \frac{1}{A}.
\end{align*}

Ainsi les nombres $x_1$ et $x_2$ sont non nuls.

S’ils étaient tous les deux négatifs, vous auriez $-x_1>0$ et $-x_2>0$ l’équation caractéristique admettrait quatre racines complexes imaginaires pures et donc aucune racine réelle : $i\sqrt{-x_1}$, $-i\sqrt{-x_1}$, $i\sqrt{-x_2}$ et $-i\sqrt{-x_2}$ contradiction.

S’ils étaient tous les deux positifs, vous l’équation caractéristique admettrait quatre racines réelles deux à deux distinctes : $\sqrt{x_1}$, $-\sqrt{x_1}$, $\sqrt{x_2}$ et $-\sqrt{x_2}$ contradiction.

Donc les nombres $x_1$ et $x_2$ ne peuvent avoir le même signe. Leur produit étant égal à $\frac{1}{ab-h^2}$ vous déduisez que le nombre $ab-h^2$ est strictement négatif.

Explicitez les signes des nombres $A$ et $B$

Les calculs suivants montrent que $\delta^2$ est strictement supérieur à $\vert a+b\vert ^2:$

\begin{align*}
\delta^2 &\geq (a-b)^2+4h^2\\
&\geq a^2+b^2-2ab+4h^2\\
&\geq a^2+b^2+2ab-4ab+4h^2\\
&\geq  a^2+b^2+2ab + 4 (h^2-ab)\\
&>  a^2+b^2+2ab\\
&\geq (a+b)^2\\
&\geq \vert a+b \vert^2.
\end{align*}

En prenant la racine carrée qui est une fonction strictement croissante vous obtenez :

\delta > \vert a+b \vert \geq a+b.

Donc $a+b-\delta < 0$ et par suite $A<0$ et $x_2<0.$

Comme $AB < 0$ il vient $B > 0$ et $x_1>0.$

Déterminez la nature de cette conique

\begin{align*}
ax^2+2hxy+by^2 = 1 
&\Longleftrightarrow AX^2+2HXY+BY^2=1\\
&\Longleftrightarrow AX^2+BY^2=1.
\end{align*}

Comme $B$ est strictement positif et $A$ est strictement négatif, vous obtenez :

\begin{align*}
ax^2+2hxy+by^2 = 1 
&\Longleftrightarrow (\sqrt{B})^2Y^2 - (\sqrt{-A})^2X^2=1\\
&\Longleftrightarrow \frac{Y^2}{(1/\sqrt{B})^2} - \frac{X^2}{(1/\sqrt{-A})^2}=1.
\end{align*}

Posez :

\begin{align*}
k = \frac{1}{\sqrt{B}} = \sqrt{x_1}\\
\ell = \frac{1}{\sqrt{-A}} = \sqrt{-x_2}.
\end{align*}

Il s’agit d’une hyperbole. Son axe transversal a pour équation $X=0$ dans le repère $(O, \vv{u}, \vv{v}).$ Son axe conjugué a pour équation $Y=0$ dans ce même repère.

Cette hyperbole a pour équation réduite, toujours dans ce même repère :

\begin{align*}
\frac{Y^2}{k^2} - \frac{X^2}{\ell^2}=1.
\end{align*}

Les deux sommets de cette hyperbole ont pour coordonnées $(0,k)$ et $(0,-k)$ dans le repère $(O, \vv{u}, \vv{v}).$

Trouvez une équation de l’axe transversal

Les nombres $k = \sqrt{x_1}$ et $-k$ sont les deux racines réelles opposées de l’équation caractéristique.

Par rotation d’angle $-\theta$ la relation :

\begin{align*}
x&=X\cos\theta-Y\sin\theta \\
y&=X\sin\theta+Y\cos\theta.
\end{align*}

fournit :

\begin{align*}
X&=x\cos\theta+y\sin\theta \\
Y&=-x\sin\theta+y\cos\theta.
\end{align*}

L’axe transversal a pour équation $X=0.$ Comme $\cos \theta \neq 0$ vous obtenez :

\begin{align*}
X = 0 &\Longleftrightarrow x\cos^2\theta + y\sin\theta\cos\theta = 0\\
&\Longleftrightarrow 2x\cos^2\theta + 2y\sin\theta\cos\theta = 0\\
&\Longleftrightarrow x(1+\cos 2\theta) + y\sin2\theta = 0\\
&\Longleftrightarrow x\left(1+\frac{b-a}{\delta}\right) + y\frac{-2h}{\delta} = 0\\
&\Longleftrightarrow x\left(\delta+b-a\right) -2h y = 0\\
&\Longleftrightarrow x\left(a-b-\delta\right) +2h y = 0\\
&\Longleftrightarrow x\left(2a-a-b-\delta\right) +2h y = 0\\
&\Longleftrightarrow \left(a-\frac{a+b+\delta}{2}\right)x +h y = 0\\
&\Longleftrightarrow \left(a-B\right)x +h y = 0\\
&\Longleftrightarrow \left(a-\frac{1}{x_1}\right)x +h y = 0\\
&\Longleftrightarrow \left(a-\frac{1}{k^2}\right)x +h y = 0.
\end{align*}

En définitive, si $k$ est l’unique solution positive de l’équation caractéristique, alors l’axe transversal de l’hyperbole admet pour équation :

\boxed{\left(a-\frac{1}{k^2}\right)x +h y = 0.}

Trouvez une équation de l’axe conjugué

Les nombres $i\ell = i\sqrt{-x_2}$ et $-i\ell$ sont les deux racines imaginaires pures opposées de l’équation caractéristique.

L’axe conjugué a pour équation $Y=0.$ Comme $\sin \theta \neq 0$ vous obtenez :

\begin{align*}
Y = 0 &\Longleftrightarrow -x\sin\theta+y\cos\theta = 0\\
&\Longleftrightarrow -x\sin^2\theta +y\sin\theta\cos\theta = 0\\
&\Longleftrightarrow -2x\sin^2\theta +2y\sin\theta\cos\theta = 0\\
&\Longleftrightarrow -x(1-\cos2\theta) +y\sin2\theta = 0\\
&\Longleftrightarrow -x\left(1-\frac{b-a}{\delta}\right) +y\frac{-2h}{\delta} = 0\\
&\Longleftrightarrow x\left(1-\frac{b-a}{\delta}\right) +y\frac{2h}{\delta} = 0\\
&\Longleftrightarrow x\left(\delta-b+a\right) +2hy = 0\\
&\Longleftrightarrow x\left(\delta-b-a+2a\right) +2hy = 0\\
&\Longleftrightarrow x\left(\frac{\delta-b-a}{2}+a\right) +hy = 0\\
&\Longleftrightarrow \left(a-\frac{a+b-\delta}{2}\right)x +hy = 0\\
&\Longleftrightarrow \left(a-A\right)x +hy = 0\\
&\Longleftrightarrow \left(a-\frac{1}{x_2}\right)x +hy = 0\\
&\Longleftrightarrow \left(a-\frac{1}{(i\ell)^2}\right)x +hy = 0\\
\end{align*}

En définitive, si $i\ell$ est l’unique solution imaginaire pure de partie imaginaire positive de l’équation caractéristique, alors l’axe conjugué, qui n’a aucun point d’intersection avec l’hyperbole, admet pour équation :

\boxed{\left(a-\frac{1}{(i\ell)^2}\right)x +h y = 0.}

246. Etudiez la nature d’une conique à centre (1/3)

Soient deux réels $a$ et $b$ fixés et un réel $h$ tel que $ab-h^2\neq 0.$

Dans un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez la conique $\mathscr{C}$ formée par l’ensemble des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2=1.

Dans cette chronique vous supposez que l’équation caractéristique $\boxed{\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2}$ d’inconnue $r\in\R$ n’a pas de solution réelle.

Traitez le cas où $h$ est nul

Si $h$ est nul, alors l’équation caractéristique devient :

\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = 0.

Du coup, il est impossible d’avoir $a>0$ (sinon vous choisissez $r = \frac{1}{\sqrt{a}}$ qui est solution.) De même il est impossible d’avoir $b>0$.

Donc $a$ et $b$ sont négatifs.

D’autre part, $ab \neq h^2$ s’écrit $ab \neq 0$ donc $a<0$ et $b<0.$

Sous cette hypothèse, la conique admet pour équation :

ax^2+by^2=1.

Le membre de gauche est négatif et celui de droite est strictement positif donc la conique est vide.

Traitez le cas où $h$ est non nul

Pour tout réel $r$ non nul, l’équation caractéristique n’est pas vérifiée donc :

\begin{align*}
\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right)&\neq h^2 \\
(ar^2-1)(br^2-1) &\neq h^2r^4\\
abr^4-(a+b)r^2+1 &\neq h^2r^4\\
(ab-h^2)r^4-(a+b)r^2+1 &\neq 0.
\end{align*}

Cela suggère d’étudier le polynôme $(ab-h^2)X^2-(a+b)X+1$ qui est bien de degré $2$ puisque $ab-h^2$ est non nul.

Le discriminant de ce trinôme vaut :

\begin{align*}
(a+b)^2-4(ab-h)^2 &= a^2+b^2+2ab-4ab+4h^2\\
&=(a-b)^2+4h^2. 
\end{align*}

Comme $h$ est non nul, $h^2$ est strictement positif et par somme le discriminant est strictement positif.

Le trinôme $(ab-h^2)X^2-(a+b)X+1$ admet ainsi deux racines réelles distinctes $x_1$ et $x_2.$

Utilisant les relations entre les coefficients et les racines, vous avez :

\begin{align*}
x_1+x_2 &= \frac{a+b}{ab-h^2}\\
x_1 x_2 &= \frac{1}{ab-h^2}.
\end{align*}

Si $x_1$ est strictement positif, vous posez $r = \sqrt{x_1}$ et déduisez que $r^2 = x_1$ et $(ab-h^2)r^4-(a+b)r^2+1 = 0.$ En divisant par $r^4$ qui est non nul vous déduisez $\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2$ donc l’équation caractéristique n’est pas sans solution, ce qui est absurde.

Le même raisonnement tient pour $x_2.$

Vous déduisez que $x_1$ et $x_2$ sont négatifs.

Leur produit est égal à $\frac{1}{ab-h^2}$ qui est non nul. Donc $x_1<0$, $x_2<0$ et $ab-h^2 > 0.$

Comme $ab > h^2 > 0$ les réels $a$ et $b$ ont le même signe.

Comme $x_1<0$ et $x_2<0$, vous avez par somme $x_1+x_2< 0.$ Comme $x_1+x_2 = \frac{a+b}{ab-h^2}$ vous déduisez $a+b < 0$ et du coup $a<0$ et $b<0.$

Quels que soient les réels $x$ et $y:$

\begin{align*}
ax^2+2hxy+by^2=1 &\Longleftrightarrow a^2x^2+2ahxy+aby^2=a \\
 &\Longleftrightarrow  (ax+hy)^2+(ab-h^2)y^2=a.
\end{align*}

Le membre de gauche est positif comme somme de deux nombres positifs et celui de droite est strictement négatif donc la conique est vide.

Concluez

Vous déduisez de cette étude que lorsque l’équation caractéristique $\boxed{\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2}$ d’inconnue $r\in\R$ n’a pas de solution réelle, alors la conique à centre $\mathscr{C}$ est vide.

245. Equation caractéristique d’une conique à centre

Soient deux réels $a$ et $b$ fixés et ainsi qu’un réel $h$ non nul. Dans un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez la conique $\mathscr{C}$ formée par l’ensemble des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2=1.

Vous constatez déjà que pour tout point $M(x,y)$ appartenant à $\mathscr{C}$, le point $M'(-x,-y)$ appartient aussi à $\mathscr{C}.$ L’origine du repère est centre de symétrie de $\mathscr{C}.$

Découvrez l’origine de l’équation caractéristique

Dans le cas où la conique $\mathscr{C}$ est une ellipse, vous trouvez ses deux sommets du grand axe comme étant les points d’intersection de la conique $\mathscr{C}$ et d’un cercle de centre $O$ et de rayon $r$ strictement positif bien choisi, pour que ce cercle soit tangent à la conique au niveau des sommets recherchés. Vous changez le rayon et trouverez les deux sommets du petit axe.

Soit $r$ un réel strictement positif. Notez $C$ le cercle de centre $O$ et de rayon $r$ qui a pour équation $\frac{x^2+y^2}{r^2}=1.$

Les points d’intersection de $\mathscr{C}$ et de $C$ vérifient l’équation :

\begin{array}{l}
\displaystyle ax^2+2hxy+by^2=\frac{x^2+y^2}{r^2}\\
\displaystyle \left(a-\frac{1}{r^2}\right)x^2+2hxy+\left(b-\frac{1}{r^2}\right)y^2 = 0.
\end{array}

Notez alors $\mathscr{C’}$ la conique dégénérée représentée par l’équation :

\left(a-\frac{1}{r^2}\right)x^2+2hxy+\left(b-\frac{1}{r^2}\right)y^2 = 0.

Ecartez la possibilité du déterminant strictement positif

Supposez un instant que $r$ soit choisi pour obtenir $\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) – h^2 > 0.$

Alors une rotation de centre $O$ du repère amène à l’équation $AX^2+BY^2=0$ avec $A$ et $B$ de même signe et non nuls. Cela conduit à avoir $X=Y=0$, et donc $x=y=0$, cas à rejeter, puisque l’intersection de la conique $\mathscr{C}’$ et de la conique $\mathscr{C}$ serait vide et ne permettrait pas d’arriver sur les sommets de l’ellipse.

Ecartez la possibilité du déterminant strictement négatif

Supposez un instant que $r$ soit choisi pour obtenir $\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) – h^2 < 0.$

Alors une rotation de centre $O$ du repère amène à l’équation $AX^2+BY^2=0$ avec $A$ et $B$ de signes contraires et non nuls. Cela conduit à avoir la réunion de deux droites sécantes à l’origine qui formeront 4 points d’intersection avec la conique $\mathscr{C}$ ce que vous souhaitez éviter.

Concluez

De ce qui précède, vous choisissez $r>0$ pour annuler le déterminant $2\times 2$ ce qui fournit :

\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2.

Vous vous attendez, dans le cas de l’ellipse, à ce que cette conique soit celle du grand axe ou celle du petit axe lorsque $r$ vérifie cette équation. Cela n’est à ce stade qu’une conjecture.

Cela semble suffisamment important pour suggérer la définition suivante :

L’équation $\boxed{\left(a-\frac{1}{r^2}\right)\left(b-\frac{1}{r^2}\right) = h^2}$ d’inconnue $r\in\R$ est appelée équation caractéristique de la conique $\mathscr{C}.$

Prolongement

Vous souhaitez savoir ce qu’il advient lorsque l’équation caractéristique susmentionnée n’admet aucune solution réelle ? Allez jeter un coup d’oeil dans l'article 246.

244. Construisez le foyer d’une parabole

Soit $p$ un nombre réel strictement positif.

Vous vous intéressez à la représentation graphique de la courbe d’équation $\boxed{y = \frac{x^2}{4p}}$ dans un repère orthonormé.

Il s’agit d’une parabole admettant l’origine du repère pour sommet.

Déterminez l’équation réduite d’une tangente

Soit $a$ un nombre réel non nul, de sorte que le point de coordonnées $A\left(a,\frac{a^2}{4p}\right)$ soit situé sur la parabole et soit distinct du sommet.

La parabole précitée est la représentation graphique de la fonction $f$ définie par $\forall x\in\R, f(x) = \frac{x^2}{4p}.$

Cette fonction est dérivable sur $\R$ et pour tout réel $x$, $f'(x) = \frac{x}{2p}.$

Le coefficient directeur de la tangente à la parabole en $A$ est $f'(a)=\frac{a}{2p}$ et l’équation réduite de celle-ci est :

y=\frac{a}{2p}(x-a)+\frac{a^2}{4p}.

En développant vous obtenez :

y=\frac{2ax-a^2}{4p}.

Notez $B$ le point d’intersection de cette tangente avec l’axe des ordonnées. Vous avez $B\left(0,\frac{-a^2}{4p}\right).$

En appelant $H$ le projeté orthogonal du point $A$ sur l’axe des ordonnées, il apparaît que le sommet de la parabole est le milieu de $[BH].$

Déterminez l’équation réduite de la médiatrice du segment $[AB]$

Pour tout point $M$ du plan de coordonnées $(x,y)$ vous avez la série d’équivalences :

\begin{align*}
MA = MB & \Longleftrightarrow MA^2=MB^2\\
& \Longleftrightarrow \left(x-a\right)^2+\left(y-\frac{a^2}{4p}\right)^2=x^2+\left(y+\frac{a^2}{4p}\right)^2\\
& \Longleftrightarrow x^2-2ax+a^2+y^2+\frac{a^4}{16p^2}-\frac{a^2y}{2p} = x^2+y^2+\frac{a^4}{16p^2}+\frac{a^2y}{2p}\\
& \Longleftrightarrow  a^2-2ax= \frac{a^2}{p}y\\
& \Longleftrightarrow  p-\frac{2p}{a}x=y.
\end{align*}

Vous constatez que la médiatrice du segment $[AB]$ passe par un point fixe $F$ de coordonnées $\boxed{(0,p)}$ qui s’appelle le foyer de la parabole.

01/04/2022 - Trace de la parabole dequation yx^24p avec p0.5
Tracé de la parabole d’équation $y=\frac{x^2}{4p}$ avec $p=1/2.$

Note. Partant d’un point $A$ distinct du sommet de la parabole, vous construisez le point $H,$ projeté orthogonal du point $A$ sur l’axe de symétrie de la parabole, puis vous construisez le point $B$ symétrique de $H$ rapport au sommet de la parabole. Le foyer $F$ apparaît comme étant le point de l’axe de symétrie de la parabole équidistant des points $A$ et $B.$ Le triangle $FAB$ est alors isocèle en $F.$

Prolongement

A partir du sommet d’une parabole et de son tracé, sauriez-vous construire géométriquement l’axe de symétrie de celle-ci ?

243. Déterminez une équation de l’union des deux axes d’une ellipse ou d’une hyperbole

Etant donnés six réels $a, b, c, f, g$ et $h$ et un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez la conique $\mathscr{C}$ formée par l’ensemble des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2+2gx+2fy+c=0.

Vous supposez que :

\left\{\begin{array}{ll}
\begin{vmatrix}a & h  \\
h & b \\
\end{vmatrix} \neq 0\\
\\
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} \neq 0.
\end{array}\right.

Dans ces conditions, l’ensemble $\mathscr{C}$ est soit une ellipse (si $ab-h^2>0$), soit une hyperbole (si $ab-h^2<0$), soit l’ensemble vide $\varnothing.$

En effet, comme $ab-h^2\neq 0$ le système ci-dessous admet une solution unique $(u,v)$ telle que :

\left\{\begin{align*}
au+hv+g&=0\\
hu+bv+f&=0.
\end{align*}\right.

Effectuez un changement d’origine du repère

Posez $X = x-u$ et $Y=y-v$ et considérez le point $\Omega$ de coordonnées $(u,v)$ dans le repère $(O,\vv{i},\vv{j}).$

\begin{align*}
ax^2+2hxy+by^2+2gx+2fy+c=0 &\Longleftrightarrow a(X+u)^2+2h(X+u)(Y+v)+b(Y+v)^2\\&\qquad+2g(X+u)+2f(Y+v)+c=0 \\
&\Longleftrightarrow a(X^2+u^2+2uX)\\&\qquad+2h(XY+vX+uY+uv)
\\&\qquad +b(Y^2+v^2+2vY)
\\&\qquad +2g(X+u)+2f(Y+v)+c=0\\
&\Longleftrightarrow aX^2+2hXY+bY^2
\\&\qquad +(2au+2hv+2g)X+(2hu+2bv+2f)Y\\
&\qquad +au^2+2huv +bv^2 + 2gu +2fv+c=0\\
&\Longleftrightarrow aX^2+2hXY+bY^2 \\
&\qquad +au^2+2huv +bv^2 + 2gu +2fv+c = 0\\
&\Longleftrightarrow aX^2+2hXY+bY^2 \\
&\qquad +au^2+huv +huv +bv^2 +gu + gu +fv +fv+c = 0\\
&\Longleftrightarrow aX^2+2hXY+bY^2 \\
&\qquad +u(au+hv+g)+v(hu+bv+f)   + gu  +fv+c = 0\\
&\Longleftrightarrow aX^2+2hXY+bY^2  + gu  +fv+c = 0.
\end{align*}

De part les propriétés du déterminant :

\begin{align*}
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix}  &= \begin{vmatrix}a & h & au+hv+g \\
h & b & hu+gv+f \\
g& f & gu+fv+c
\end{vmatrix} \\
 &= \begin{vmatrix}a & h & 0 \\
h & b & 0 \\
g& f & gu+fv+c
\end{vmatrix} \\
&=  (gu+fv+c) \begin{vmatrix}a & h \\
h & b 
\end{vmatrix} .
\end{align*}

Comme $\begin{vmatrix}a & h & g \\h & b & f \\ g& f & c \end{vmatrix}\neq 0$ vous déduisez que $gu+fv+c \neq 0.$

Posez alors :

\begin{align*}
A &= \frac{a}{gu+fv+c}\\
H &= \frac{h}{gu+fv+c}\\
B &= \frac{b}{gu+fv+c}.
\end{align*}

L’ensemble $\mathscr{C}$ est donc décrit par l’équation $\boxed{AX^2+2HXY+BY^2 =1.}$

Par souci de lisibilité vous allez noter encore par $x$ et $y$ les coordonnées du point $M$ mais cette fois-ci dans le repère $(\Omega,\vv{i},\vv{j}).$

D’après ce qui a été établi, vous avez montré que, pour tout point $M$ de coordonnées $(x,y)$ dans le repère $(\Omega,\vv{i},\vv{j}):$

M\in\mathscr{C} \Longleftrightarrow Ax^2+2Hxy+By^2=1.

Effectuez une rotation du repère $(\Omega,\vv{i},\vv{j})$

Soit $\theta$ un nombre réel, il sera choisi plus tard. Considérez la base $(\vv{u}, \vv{v})$ obtenue en faisant subir à la base $(\vv{i}, \vv{j})$ une rotation d’angle $\theta.$

Vous notez $(x,y)$ les coordonnées d’un point $M$ du plan dans le repère $(\Omega,\vv{i},\vv{j})$ et $(X,Y)$ ses coordonnées dans le repère $(\Omega,\vv{u},\vv{v}).$

Dès lors :

\begin{align*}
x&=X\cos\theta-Y\sin\theta \\
y&=X\sin\theta+Y\cos\theta.
\end{align*}

Note. La rotation d’angle $-\theta$ permet d’inverser les relations ci-dessus, donc :

\begin{align*}
X&=x\cos\theta+y\sin\theta \\
Y&=-x\sin\theta+y\cos\theta.
\end{align*}

Vous obtenez :

\begin{align*}
Ax^2+2Hxy+By^2 = 1 &\Longleftrightarrow A(X\cos\theta-Y\sin\theta)^2\\
&\qquad + 2H(X\cos\theta - Y\sin\theta)(X\sin\theta+Y\cos\theta)\\
&\qquad +B(X\sin\theta + Y\cos\theta)^2=1\\
&\Longleftrightarrow A(X^2\cos^2\theta+Y^2\sin^2\theta-2XY\sin\theta\cos\theta)^2\\
&\qquad +2H(X^2\sin\theta\cos\theta +XY(\cos^2\theta - \sin^2\theta)-Y^2\sin\theta\cos\theta)\\
&\qquad +B(X^2\sin^2\theta+Y^2\cos^2\theta+2XY\sin\theta\cos\theta) = 1\\
&\Longleftrightarrow (A\cos^2\theta +2H\sin\theta\cos\theta + B\sin^2\theta)X^2\\
&\qquad +2((B-A)\sin\theta \cos\theta + H(\cos^2\theta-\sin^2\theta))XY\\
&\qquad +(A\sin^2\theta -2H\sin\theta\cos\theta + B\cos^2\theta)Y^2 = 1.
\end{align*}

L’annulation du terme croisé $XY$ est possible.

En effet, si $a=b$ alors $A=B$ et vous choisissez $\theta = \pi/4.$

Si $a\neq b$ vous avez aussi $B-A\neq 0$ et vous posez $\theta = \frac{1}{2}\mathrm{Arctan}\frac{-2H}{B-A}.$

Alors $2\theta = \mathrm{Arctan} \frac{-2H}{B-A}$ et $2\theta \in ]-\pi/2, \pi/2[$ du coup $\cos 2\theta \neq 0$ et $\tan 2\theta = \frac{-2H}{B-A}$ ce qui se traduit par $\frac{\sin 2\theta}{\cos 2\theta} = \frac{-2H}{B-A}$ et donc:

\begin{align*}
(B-A)\sin2 \theta  + 2H \cos 2\theta&= 0\\
2(B-A)\sin \theta \cos\theta + 2H (\cos^2\theta - \sin^2\theta)&= 0\\
(B-A)\sin \theta \cos\theta + H (\cos^2\theta - \sin^2\theta)&= 0.
\end{align*}

Dans le repère $(\Omega,\vv{u},\vv{v})$ l’ensemble $\mathscr{C}$ est décrit par l’équation suivante:

 (A\cos^2\theta +2H\sin\theta\cos\theta + B\sin^2\theta)X^2+(A\sin^2\theta -2H\sin\theta\cos\theta + B\cos^2\theta)Y^2 = 1.

Il a été démontré qu’une rotation du repère ne modifie pas la valeur du déterminant d’ordre $2.$ Vous êtes invité à vous référer au contenu écrit dans l'article 241.

Ainsi :

\begin{align*}
 (A\cos^2\theta +2H\sin\theta\cos\theta + B\sin^2\theta)(A\sin^2\theta -2H\sin\theta\cos\theta + B\cos^2\theta) - 0^2 &=AB-H^2\\
&=\frac{ab-h^2}{(gu+fv+c)^2}.
\end{align*}

Comme $ab-h^2$ est non nul, les deux nombres $A\cos^2\theta +2H\sin\theta\cos\theta + B\sin^2\theta$ et $A\sin^2\theta -2H\sin\theta\cos\theta + B\cos^2\theta$ sont non nuls.

Posez alors $k = \sqrt{\vert A\cos^2\theta +2H\sin\theta\cos\theta + B\sin^2\theta \vert}$ et $\ell = \sqrt{\vert A\sin^2\theta -2H\sin\theta\cos\theta + B\cos^2\theta \vert}$ pour conclure que $\mathscr{C}$ admet pour équation :

\pm \frac{X^2}{k^2} \pm \frac{Y^2}{\ell^2} = 1.

Selon les signes, vous obtenez soit l’ensemble vide $\varnothing$ soit une ellipse, soit une hyperbole.

Déterminez une équation de la réunion des deux axes

Supposez que $\mathscr{C}$ soit non vide.

Si $h$ est nul, l’axe des abscisses et l’axe des ordonnées sont déjà les deux axes de $\mathscr{C}.$

Supposez maintenant que $h$ soit non nul.

Un point $M$ de coordonnées $(X,Y)$ dans le repère $(\Omega,\vv{u},\vv{v})$ appartient à l’un des deux axes de $\mathscr{C}$, si et seulement si, $X=0$ ou $Y = 0.$ Notez $(x,y)$ les coordonnées de $M$ dans le repère $(\Omega,\vv{i},\vv{j}).$

Cette condition fournit les équivalences :

\begin{align*}
X=0\text{ ou }Y=0 &\Longleftrightarrow XY=0\\ 
 &\Longleftrightarrow (x\cos\theta+y\sin\theta)(-x\sin\theta+y\cos\theta) = 0\\
&\Longleftrightarrow (y^2-x^2)\sin\theta\cos\theta+xy(\cos^2\theta-\sin^2\theta) = 0\\
&\Longleftrightarrow (y^2-x^2)\sin\theta\cos\theta+xy \cos 2\theta= 0\\
&\Longleftrightarrow  (y^2-x^2)\sin2 \theta+2xy \cos 2\theta= 0.
\end{align*}

Dans le cas où $a=b$, vous avez $\theta = \pi/4$ donc $2\theta = \pi/2$ donc :

\begin{align*}
X=0\text{ ou }Y=0  &\Longleftrightarrow  y^2-x^2= 0.
\end{align*}

Dans le cas où $a\neq b$, il a été vu que $\cos 2\theta \neq 0$ et que $\tan 2\theta = \frac{-2H}{B-A} = \frac{-2h}{b-a}.$

Ainsi :

\begin{align*}
X=0\text{ ou }Y=0 &\Longleftrightarrow  (y^2-x^2)\tan2 \theta+2xy = 0\\
 &\Longleftrightarrow  (y^2-x^2)\frac{-2h}{b-a}+2xy = 0\\
&\Longleftrightarrow  -2h(y^2-x^2)+2(b-a)xy = 0\\
&\Longleftrightarrow  hx^2+(b-a)xy -hy^2= 0.
\end{align*}

Remarquez que lorsque $a=b$ et après division par $h$ vous retrouvez l’équation $x^2-y^2=0.$

Concluez

Ellipse ou hyperbole centrées à l’origine

Soient trois nombres réels $a, b$ et $h$ tels que :

\begin{vmatrix}a & h  \\
h & b \\
\end{vmatrix} \neq 0.

Alors la conique $\mathscr{C}$ formée par l’ensemble des points $M$ tel que :

ax^2+2hxy+by^2 = 1

est soit une hyperbole, soit une ellipse, soit l’ensemble vide $\varnothing.$

Si la conique $\mathscr{C}$ est non vide et et si $h$ est non nul la réunion des deux axes de $\mathscr{C}$ a pour équation :

\boxed{ hx^2+(b-a)xy -hy^2= 0. }

Si la conique $\mathscr{C}$ est non vide et et si $h$ est nul la réunion des deux axes de $\mathscr{C}$ a pour équation $xy=0.$

Cas général

Soient six réels $a, b, c, f, g$ et $h$ tels que :

\left\{\begin{array}{ll}
\begin{vmatrix}a & h  \\
h & b \\
\end{vmatrix} \neq 0\\
\\
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} \neq 0.
\end{array}\right.

La conique $\mathscr{C}$ formée par l’ensemble des points $M$ de coordonnées $(x,y)$ tels que :

ax^2+2hxy+by^2+2gx+2fy+c=0.

est soit une hyperbole, soit une ellipse, soit l’ensemble vide $\varnothing.$

Dans le cas où la conique $\mathscr{C}$ est non vide, elle admet un centre $\Omega$ de coordonnées $(u,v)$ qui vérifient le système :

\left\{\begin{align*}
au+hv+g&=0\\
hu+bv+f&=0.
\end{align*}\right.

Note. Il est rappelé que le point $\Omega$ n’appartient pas à la conique $\mathscr{C}$ à cause de la non-nullité du déterminant $3\times 3$ précité.

En passant du repère $(\Omega, \vv{i}, \vv{j})$ au repère $(O,\vv{i}, \vv{j})$ vous déduisez dans ces conditions les résultats suivants.

Si $h$ est non nul, la réunion des deux axes de la conique $\mathscr{C}$ est :

\boxed{ h(x-u)^2+(b-a)(x-u)(y-v) -h(y-v)^2= 0. }

Si $h$ est nul, la réunion des deux axes de la conique $\mathscr{C}$ est : $(x-u)(y-v)=0.$

Visualisez

31/03/2022 - Trace de lellipse dequation 2x^2xyy^21 et de ses deux axes 9db3c84903f000e47855f23e60fef753054273a1
Tracé de l’ellipse d’équation $2x^2+xy+y^2=1$ et de ses deux axes d’équation $x^2-y^2-2xy=0.$
31/03/2022 - Trace de lhyperbole dequation 2x^24xyy^21 et de ses deux axes e03c2c520af3d3b70027d0a29a5b64e870cb78ab
Tracé de l’hyperbole d’équation $2x^2+4xy+y^2=1$ et de ses deux axes d’équation $2x^2-2y^2-xy=0.$

242. Lorsqu’une équation générale du second degré à deux variables admet deux déterminants nuls

Etant donnés six réels $a, b, c, f, g$ et $h$ et un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez l’ensemble $\mathscr{C}$ des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2+2gx+2fy+c=0.

Vous établissez la réciproque partielle des résultats obtenus dans l'article 241 et dans l'article 240.

Note. Cette réciproque est effectivement partielle dans la mesure où il sera possible d’obtenir l’ensemble vide ou le plan tout entier.

Vous supposez dans cette chronique que :

\left\{\begin{array}{ll}
\begin{vmatrix}a & h  \\
h & b \\
\end{vmatrix} = ab-h^2 = 0\\
\\
\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} = 0.
\end{array}\right.

Développez le déterminant $\Delta$

Comme $ab-h^2=0$ le calcul de $\Delta$ sera facilité en le développant par rapport à la troisième ligne :

\begin{align*}
\Delta &= g
\begin{vmatrix}  h & g \\
 b & f 
\end{vmatrix}-f\begin{vmatrix}  a & g \\
 h & f 
\end{vmatrix} + c \begin{vmatrix}  a & h \\
 h & b 
\end{vmatrix}\\
&= g (hf-gb)-f(af-gh) + c (ab-h^2) \\
&=g (hf-gb)-f(af-gh) +c\times 0\\
&=2fgh-bg^2-af^2.
\end{align*}

Supposez $a$ non nul

Comme $\Delta = 0$ vous déduisez après multiplication par $a:$

\begin{align*}
0 &= -a\times \Delta \\
&=a^2f^2 +abg^2 - 2afgh\\
&=a^2f^2 +h^2g^2 - 2afgh\\
&=(af-gh)^2.
\end{align*}

Vous déduisez de ce résultat que $af-gh = 0.$

\begin{align*}
M\in\mathscr{C}&\Longleftrightarrow ax^2+2hxy+by^2+2gx+2fy+c=0  \\
&\Longleftrightarrow a^2x^2+2ahxy+aby^2+2agx+2afy+ac=0  \\
&\Longleftrightarrow (ax+hy)^2 +(ab-h^2)y^2+2agx+2afy+ac=0  \\
&\Longleftrightarrow (ax+hy)^2 +2gax+2afy+ac=0  \\
&\Longleftrightarrow (ax+hy)^2 +2g(ax+hy)+(2af-2gh)y+ac=0  \\
&\Longleftrightarrow (ax+hy)^2 +2g(ax+hy)+2(af-gh)y+ac=0  \\
&\Longleftrightarrow (ax+hy)^2 +2g(ax+hy)+ac=0  \\
&\Longleftrightarrow (ax+hy)^2 +2g(ax+hy) + g^2 +ac - g^2=0  \\
&\Longleftrightarrow (ax+hy+g)^2 +ac - g^2=0.
\end{align*}

Trois cas se présentent.

Si $ac-g^2 > 0$ alors :

\forall (x,y)\in\R^2, (ax+hy+g)^2 +ac - g^2>0.

Par suite $\mathscr{C} = \varnothing.$

Si $ac-g^2 = 0$ alors :

\begin{align*}
M\in\mathscr{C}
&\Longleftrightarrow (ax+hy+g)^2 =0  \\
&\Longleftrightarrow ax+hy+g =0.
\end{align*}

Comme $a\neq 0$ vous avez $(a,h)\neq (0,0)$ donc $\mathscr{C}$ est une droite.

Si $ac-g^2 <0$ alors le nombre $\delta = \sqrt{g^2-ac}$ est bien défini et est non nul.

\begin{align*}
M\in\mathscr{C}
&\Longleftrightarrow (ax+hy+g)^2 -\delta^2=0  \\
&\Longleftrightarrow (ax+hy+g+\delta)(ax+hy+g-\delta) = 0\\
&\Longleftrightarrow ax+hy+g+\delta = 0\text{ ou } ax+hy+g-\delta = 0.
\end{align*}

Comme $(a,h)\neq (0,0)$ les équations $ax+hy+g+\delta = 0$ et $ax+hy+g-\delta = 0$ sont bien des équations cartésiennes de droites.

Notez $\mathscr{D}_1$ la droite d’équation $ax+hy+g+\delta = 0.$

Notez $\mathscr{D}_2$ la droite d’équation $ax+hy+g-\delta = 0.$

Elles admettent toutes les deux le même vecteur directeur non nul $\vv{d}\begin{pmatrix}-h \\ a \end{pmatrix}$ donc elles sont parallèles.

De plus elles n’ont aucun point en commun. En effet si vous notiez $(p,q)$ les coordonnées d’un tel point, vous auriez :

\begin{align*}
ap+hq+g+\delta = 0\\
ap+hq+g-\delta = 0
\end{align*}

Et par soustraction vous déduisez $2\delta = 0$ et $\delta = 0$ contradiction avec $\delta > 0.$

Donc les droites $\mathscr{D}_1$ et $\mathscr{D}_2$ sont strictement parallèles et $\mathscr{C}$ est leur réunion.

Supposez $a$ nul

La condition $ab-h^2 = 0$ force la nullité de $h.$

La nullité de $\Delta$ fournit :

\begin{align*}
0 &= \Delta\\
 &=2fgh-bg^2-af^2\\
&=-bg^2.
\end{align*}

Si $b$ est nul

Vous avez :

\begin{align*}
M\in\mathscr{C}&\Longleftrightarrow 2gx+2fy+c=0.
\end{align*}

Donc $\mathscr{C}$ est soit une droite, soit l’ensemble vide, soit le plan tout entier.

Si $b$ est non nul

Comme $bg^2 = 0$ le nombre $g$ est nul. Du coup :

\begin{align*}
M\in\mathscr{C}&\Longleftrightarrow by^2+2fy+c=0.
\end{align*}

Pour étudier le trinôme en $y$, obtenu il y a trois possibilités, qui sont récapitulées en détail par souci de complétude.

Si $f^2-bc$ est strictement négatif

Vous écrivez que :

\begin{align*}
\forall y\in\R, b(by^2+2fy+c) &= b^2y^2+2bfy+bc\\
&= (by+f)^2+bc-f^2
\end{align*}

Comme $bc-f^2$ est strictement positif, la somme $(by+f)^2+bc-f^2$ est supérieure à $bc-f^2$ donc elle est strictement positive.

Dès lors :

\forall y\in\R, by^2+2fy+c\neq 0.

Vous déduisez que $\mathscr{C}$ est vide.

Si $f^2-bc$ est nul

Vous avez :

\begin{align*}
M\in\mathscr{C}&\Longleftrightarrow b^2y^2+2bfy+bc=0\\
&\Longleftrightarrow (by+f)^2+bc-f ^2=0\\
&\Longleftrightarrow (by+f)^2=0\\
&\Longleftrightarrow by+f=0\\
&\Longleftrightarrow y =-\frac{f}{b}.
\end{align*}

Vous déduisez que $\mathscr{C}$ est une droite horizontale.

Si $f^2-bc$ est strictement positif

Le réel $\delta = \sqrt{f^2-bc}$ est bien défini et il est strictement positif.

D’autre part :

\begin{align*}
M\in\mathscr{C}&\Longleftrightarrow b^2y^2+2bfy+bc=0\\
&\Longleftrightarrow (by+f)^2+bc-f ^2=0\\
&\Longleftrightarrow (by+f)^2 - \delta^2=0\\
&\Longleftrightarrow (by+f+\delta)(by+f-\delta)=0\\
&\Longleftrightarrow by+f+\delta = 0 \text{ ou } by+f-\delta = 0\\
&\Longleftrightarrow y=\frac{-f-\delta}{b} \text{ ou } y = \frac{-f+\delta}{b}.
\end{align*}

Vous déduisez que $\mathscr{C}$ est la réunion de deux droites horizontales qui sont strictement parallèles.

Concluez

Supposez :

\left\{\begin{array}{ll}
ab-h^2 = 0\\
\\
\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} = 0.
\end{array}\right.

Il y a quatre possibilités en tout pour l’ensemble $\mathscr{C}$ qui peut être :

  • la réunion de deux droites strictement parallèles,
  • une droite,
  • l’ensemble vide $\varnothing$,
  • le plan tout entier.

241. Lorsqu’une équation générale du second degré à deux variables admet deux droites strictement parallèles pour représentation graphique

Le résultat obtenu dans l'article 240 subsiste avec une démonstration similaire.

Dans cette chronique vous allez obtenir une autre démonstration en utilisant un point $\Omega$ bien choisi géométriquement parlant.

Etant donnés six réels $a, b, c, f, g$ et $h$ et un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez l’ensemble $\mathscr{C}$ des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2+2gx+2fy+c=0.

Vous supposez que l’ensemble $\mathscr{C}$ est la réunion de deux droites strictement parallèles notées $\mathscr{D}_1$ et $\mathscr{D}_2.$

Effectuez un changement d’origine du repère

Soit $A$ un point quelconque de la droite $\mathscr{D}_1$ et soit $B$ un point quelconque de la droite $\mathscr{D}_2.$

Vous appelez $\Omega$ le milieu du segment $[AB].$ Notez $(u,v)$ ses coordonnées dans le repère $(O,\vv{i},\vv{j}).$

Pour tout point $M$ du plan, vous notez $(x_M,y_M)$ ses coordonnées dans le repère $(O,\vv{i},\vv{j})$ et $(X_M,Y_M)$ ses coordonnées dans le repère $(\Omega,\vv{i},\vv{j})$ et vous avez les relations :

\begin{align*}
X_M = x_M - u\\
Y_M = y_M - v.
\end{align*}

Alors la série d’équivalences suivantes fournit :

\begin{align*}
M\in\mathscr{C} &\Longleftrightarrow ax_M^2+2hx_My_M+by_M^2+2gx_M+2fy_M+c=0\\
&\Longleftrightarrow a(X_M+u)^2+2h(X_M+u)(Y_M+v)+b(Y_M+v)^2+2g(X_M+u)+2f(Y_M+v)+c=0\\
&\Longleftrightarrow a(X_M^2+u^2+2uX_M)+2h(X_MY_M+vX_M+uY_M+uv)+b(Y_M^2+v^2+2vY_M)\\&\qquad+2g(X_M+u)+2f(Y_M+v)+c=0\\
&\Longleftrightarrow aX_M^2+2hX_MY_M+bY_M^2\\&\qquad+2(au+hv+g)X_M+2(hu+bv+f)Y_M\\&\qquad+(au^2+2huv+bv^2+2gu+2fv+c)=0.
\end{align*}

Notez $(\alpha,\beta)$ les coordonnées du point $A$ dans le repère $(\Omega,\vv{i},\vv{j}).$ Comme $A\in\mathscr{C}$ vous avez :

\begin{align*}
 & a\alpha^2+2h\alpha\beta+b\beta^2\\ &\qquad+2(au+hv+g)\alpha+2(hu+bv+f)\beta\\ &\qquad+(au^2+2huv+bv^2+2gu+2fv+c)=0.
\end{align*}

Le point $B$ symétrique du point $A$ par la symétrie de centre $\Omega$ a pour coordonnées $(-\alpha,-\beta)$ dans le repère $(\Omega,\vv{i},\vv{j})$ et par suite :

\begin{align*}
 & a\alpha^2+2h\alpha\beta+b\beta^2\\ &\qquad-2(au+hv+g)\alpha-2(hu+bv+f)\beta\\ &\qquad+(au^2+2huv+bv^2+2gu+2fv+c)=0.
\end{align*}

Par soustraction, vous déduisez :

\begin{align*}
 4(au+hv+g)\alpha+4(hu+bv+f)\beta=0 \\
 (au+hv+g)\alpha+(hu+bv+f)\beta=0.
\end{align*}

Soit maintenant un point $A’$ distinct du point $A$ et appartenant à la droite $\mathscr{D}_1.$ Notez $(\alpha’,\beta’)$ ses coordonnées dans le repère $(\Omega,\vv{i},\vv{j}).$ Vous avez encore :

\begin{align*}
 (au+hv+g)\alpha'+(hu+bv+f)\beta'=0.
\end{align*}

Or, les vecteurs $\vv{\Omega A}$ et $\vv{\Omega A’}$ ne peuvent être colinéaires.

Si c’était le cas, les points $A$, $A’$ et $\Omega$ seraient alignés et par suite $\Omega \in (AA’)$ donc $\Omega \in\mathscr{D}_1.$ Les points $\Omega$ et $A$ appartenant à $\mathscr{D}_1$ le point $B$, symétrique de $A$ par la symétrie de centre $\Omega$ appartiendrait encore à $\mathscr{D}_1$ et donc $B \in \mathscr{D}_1\cap \mathscr{D}_2$, contredisant le fait que les droites $\mathscr{D}_1$ et $\mathscr{D}_2$ sont strictement parallèles.

Vous déduisez que :

\begin{vmatrix}\alpha & \beta \\ \alpha' & \beta'\end{vmatrix}\neq 0.

Par suite le système suivant est vérifié :

\begin{align*}
(au+hv+g)\alpha+(hu+bv+f)\beta=0\\
 (au+hv+g)\alpha'+(hu+bv+f)\beta'=0.
\end{align*}

Comme il a un déterminant non nul, vous obtenez la double égalité :

\left\{
\begin{align*}
au+hv+g=0\\
 hu+bv+f=0.
\end{align*}\right.

Comme $\Omega\notin \mathscr{C}$, il ne vérifie pas l’équation de $\mathscr{C}$ donc en posant $C = au^2+2huv+bv^2+2gu+2fv+c$ qui est non nul, vous déduisez que, pour tout point $M$ du plan :

\begin{align*}
M\in\mathscr{C} 
&\Longleftrightarrow aX_M^2+2hX_MY_M+bY_M^2+C = 0.
\end{align*}

Effectuez une rotation du repère

Afin d’analyser la situation, il est commode de rendre les deux droites $\mathscr{D}_1$ et $\mathscr{D}_2$ horizontales par une rotation du repère. Soit donc $\vv{u}$ un vecteur directeur unitaire commun aux droites $\mathscr{D}_1$ et $\mathscr{D}_2.$

Soit $\vv{v}$ le vecteur unitaire obtenu en faisant subit au vecteur $\vv{u}$ une rotation d’angle $\pi/2.$

Grâce au schéma que vous pouvez retrouver dans l'article 236, vous utilisez les notations déjà vues, complétées par un nouveau repère.

Pour tout point $M$ du plan vous notez $(x_M,y_M)$ ses coordonnées dans le repère $(O,\vv{i},\vv{j})$, $(X_M,Y_M)$ ses coordonnées dans le repère $(\Omega,\vv{i},\vv{j})$ et enfin $(X’_M,Y’_M)$ ses coordonnées dans le repère $(\Omega,\vv{u},\vv{v}).$

Il est rappelé que :

\begin{align*}
X_M&=X'_M\cos \theta - Y'_M\sin \theta \\
Y_M&=X'_M\sin \theta+Y'_M\cos \theta.
\end{align*}

Du coup :

\begin{align*}
M\in\mathscr{C} 
&\Longleftrightarrow aX_M^2+2hX_MY_M+bY_M^2+C = 0 \\
&\Longleftrightarrow a(X'_M\cos \theta - Y'_M\sin \theta)^2\\
&\qquad+2h(X'_M\cos \theta - Y'_M\sin \theta)(X'_M\sin \theta+Y'_M\cos \theta)\\
&\qquad + b(X'_M\sin \theta+Y'_M\cos \theta)^2+C=0\\
&\Longleftrightarrow  a({X'}_M^2 \cos^2\theta+{Y'}_M^2\sin^2\theta-2X'_MY'_M\sin\theta\cos\theta)\\
&\qquad +2h({X'}_M^2\sin\theta\cos\theta + X'_MY'_M(\cos^2 \theta - \sin^2\theta)-{Y'}_M^2\sin\theta\cos\theta)\\
&\qquad +b({X'_M}^2\sin^2\theta + {Y'}_M\cos^2\theta+2X'_MY'_M\sin\theta\cos\theta)+C=0\\
&\Longleftrightarrow (a\cos^2\theta +2h\sin\theta\cos\theta+b\sin^2\theta){X'_M}^2\\
&\qquad +(2(b-a)\sin\theta\cos\theta+2h(\cos^2\theta-\sin^2\theta))X'_MY'_M\\
&\qquad (a\sin^2\theta-2h\sin\theta \cos\theta+b\cos^2\theta){Y'_M}^2+C=0.
\end{align*}

Pour le choix de $\theta$ obtenu, vous avez :

Pour plus de lisibilité, vous allez poser :

\begin{align*}
R&= a\cos^2\theta +2h\sin\theta\cos\theta+b\sin^2\theta \\
T&=a\sin^2\theta-2h\sin\theta \cos\theta+b\cos^2\theta\\
S&=(b-a)\sin\theta\cos\theta+h(\cos^2\theta-\sin^2\theta).
\end{align*}

Ainsi, vous avez obtenu :

\begin{align*}
M\in\mathscr{C} 
&\Longleftrightarrow R{X'}_M^2+2SX'_MY'_M+T{Y'}_M^2+C=0.
\end{align*}

Grâce à un calcul similaire mené dans l'article 240 vous avez la conservation d’un déterminant de taille $2\times 2$ soit :

ab-h^2 = RT-S^2.

Par souci de complétude, ce résultat va être prouvé en utilisant la propriété multiplicative du déterminant.

Prouvez l’égalité $ab-h^2 = RT-S^2$

En effet, partez de la matrice suivante :

\begin{pmatrix}
a & h\\
h& b
\end{pmatrix}.

Son déterminant étant égal à $ab-h^2$, vous multipliez cette matrice à droite par une matrice de rotation, qui a un déterminant égal à $1:$

\begin{pmatrix}
a & h\\
h& b
\end{pmatrix}
\begin{pmatrix}
\cos\theta & -\sin\theta\\
\sin\theta & \cos\theta
\end{pmatrix} = 
\begin{pmatrix}
a\cos \theta + h\sin\theta & -a\sin\theta + h\cos\theta\\
h\cos\theta + b\sin \theta & -h\sin\theta+b\cos\theta
\end{pmatrix}.

Multipliez à gauche cette matrice par la matrice de rotation d’angle $-\theta:$

\begin{align*}
&\begin{pmatrix}
\cos\theta & \sin\theta\\
-\sin\theta & \cos\theta
\end{pmatrix}\begin{pmatrix}
a\cos \theta + h\sin\theta & -a\sin\theta + h\cos\theta\\
h\cos\theta + b\sin \theta & -h\sin\theta+b\cos\theta
\end{pmatrix} \\
&\qquad= \begin{pmatrix}
a\cos^2\theta+2h\sin\theta\cos\theta +b\sin^2\theta & (b-a)\sin\theta\cos\theta + h(\cos^2\theta-\sin^2\theta)\\
 (b-a)\sin\theta\cos\theta + h(\cos^2\theta-\sin^2\theta) & a\sin^2\theta-2h\sin\theta\cos\theta+b\cos^2\theta
\end{pmatrix}\\
&\qquad= \begin{pmatrix}
R & S \\S & T
\end{pmatrix}.
\end{align*}

Du coup :

\begin{align*}
RT-S^2 &= \begin{vmatrix}
R & S \\S & T
\end{vmatrix}\\
&=\begin{vmatrix}
\cos\theta & \sin\theta\\
-\sin\theta & \cos\theta
\end{vmatrix}\begin{vmatrix}
a\cos \theta + h\sin\theta & -a\sin\theta + h\cos\theta\\
h\cos\theta + b\sin \theta & -h\sin\theta+b\cos\theta
\end{vmatrix}\\
&=\begin{vmatrix}
a\cos \theta + h\sin\theta & -a\sin\theta + h\cos\theta\\
h\cos\theta + b\sin \theta & -h\sin\theta+b\cos\theta
\end{vmatrix}\\
&=\begin{vmatrix}
a & h\\
h& b
\end{vmatrix}
\begin{vmatrix}
\cos\theta & -\sin\theta\\
\sin\theta & \cos\theta
\end{vmatrix} \\
&=\begin{vmatrix}
a & h\\
h& b
\end{vmatrix}\\
&=ab-h^2.
\end{align*}

Prouvez que $ab-h^2$ est nul

Dans le repère $(\Omega,\vv{u},\vv{v})$ les droites $\mathscr{D}_1$ et $\mathscr{D}_2$ sont horizontales et ne passent pas par l’origine du repère.

De part la position du point $\Omega$, il existe un nombre réel $k$ non nul tel que :

\begin{align*}
\mathscr{D}_1 &= \{M, Y'_M = k\} \\
\mathscr{D}_2 &= \{M, Y'_M = -k\}.
\end{align*}

Pour tout réel $x$, le point de coordonnées $(x,k)$ dans le repère $(\Omega,\vv{u},\vv{v})$ appartient à $\mathscr{D}_1$ donc à $\mathscr{C}$ donc :

\forall x\in\R, Rx^2+2Skx+Tk^2+C=0.

Le polynôme $RX^2+2SkX+Tk^2+C$ admet une infinité de racines donc ses coefficients sont tous nuls, donc $R = 2Sk = 0.$

Or $k$ est non nul, donc $S=0.$

Il en résulte que $RT-S^2 = 0$ et par suite $\frac{ab-h^2 = 0.}$

Démontrez la nullité du déterminant d’ordre $3$

Posez :

\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix}.

Vous allez montrer que $\Delta$ est nul grâce aux propriétés de $u$ et de $v.$

Un déterminant n’est pas modifié lorsque vous remplacez une colonne par elle-même augmentée d’une combinaison linéaire des autres. Cela conduit à :

\begin{align*}
\Delta &=
\begin{vmatrix}a & h & au+hv+g \\
h & b & hu+gv+f \\
g& f & gu+fv+c
\end{vmatrix}\\
&=\begin{vmatrix}a & h & 0 \\
h & b & 0 \\
g& f & gu+fv+c
\end{vmatrix}\\
&=(gu+fv+c)\begin{vmatrix}a & h \\
h & b \\
\end{vmatrix}\\
&=(gu+fv+c)(ab-h^2) \\
&=(gu+fv+c)\times 0\\
&= 0.
\end{align*}

Concluez

Si l’ensemble des points $M$ du plan de coordonnées $(x,y)$ satisfaisant l’équation :

ax^2+2hxy+by^2+2gx+2fy+c=0

est la réunion de deux droites strictement parallèles, alors nécessairement :

\boxed{\begin{array}{ll}
ab-h^2 = 0\\
\\
\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} = 0.
\end{array}}

Remarque. Si l’ensemble précédent est composé d’une seule droite, alors la même conclusion subsiste, d’après le contenu écrit dans l'article 240.

240. Lorsqu’une équation générale du second degré à deux variables admet une droite pour représentation graphique

Etant donnés six réels $a, b, c, f, g$ et $h$ et un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez l’ensemble $\mathscr{C}$ des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2+2gx+2fy+c=0.

Dans cette chronique vous supposez que l’ensemble $\mathscr{C}$ est une droite notée $\mathscr{D}.$

Supposez que cette droite est verticale

Il existe alors un nombre réel $k$ fixé tel que $\mathscr{D} = \{M(x,y), x=k\}.$

Soit maintenant $y$ un réel fixé. Le point de coordonnées $(k,y)$ appartient à $\mathscr{D}$ donc vous déduisez :

ak^2+2hky+by^2+2gk+2fy+c=0.

Vu comme un polynôme en $y$, vous obtenez :

by^2 +(2hk+2f)y+ak^2+2gk+c=0.

Ainsi le polynôme $bX^2+(2hk+2f)X+(ak^2+2gk+c)$ admet une infinité de racines, ce qui prouve que ses coefficients sont tous nuls. Du coup :

\begin{array}{lll}
b=0 \\
hk+f = 0\\
ak^2+2gk+c=0.
\end{array}

Compte tenu des conditions précédentes, la droite $\mathscr{D}$ est l’ensemble des points $M(x,y)$ tels que :

ax^2+2hxy+2gx-2hky-ak^2-2gk=0.

Or, quels que soient les réels $x$ et $y:$

\begin{align*}
ax^2+2hxy+2gx-2hky-ak^2-2gk=0 &\Longleftrightarrow a(x^2-k^2)+2h(x-k)y+2g(x-k)=0 \\
&\Longleftrightarrow a(x-k)(x+k)+2h(x-k)y+2g(x-k)=0 \\
&\Longleftrightarrow (x-k)(a(x+k)+2hy+2g)=0 \\
&\Longleftrightarrow (x-k)(ax+2hy+ak+2g)=0.
\end{align*}

Soit maintenant $x$ un réel différent de $k.$ Alors :

\forall y\in\R, M(x,y)\notin \mathscr{D}.

Donc :

\forall y\in\R, 2hy+ak+2g+ax\neq 0.

Si $h$ était non nul, alors :

\forall y\in\R, y \neq \frac{-ak-2g-ax}{2h}.

Ceci est absurde.

Vous obtenez : $h=0.$

Comme $b=h=0$ vous déduisez $\boxed{ab-h^2 = 0.}$

Il reste à s’occuper maintenant du dicriminant $\Delta.$

Dans la suite du contenu rédigé dans l'article 239 vous le définissez de la façon suivante :

\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix}.

Or, $kh+f = 0$ donc $f=0$ et finalement :

\Delta =
\begin{vmatrix}
a & 0 & g \\
0 & 0 & 0 \\
g& 0 & c
\end{vmatrix}.

Vous déduisez $\boxed{\Delta = 0.}$

Supposez que cette droite n’est pas verticale

Notez $\vv{v}$ un vecteur unitaire directeur de la droite $\mathscr{D}.$ Notez $\vv{u}$ le vecteur obtenu par rotation du vecteur $\vv{v}$ d’angle $-\pi/2.$

Notez $\theta$ une mesure de l’angle $(\vv{i},\vv{u}).$

Pour tout point $M$ du plan, vous notez $(x_M,y_M)$ ses coordonnées dans le repère $(O,\vv{i},\vv{j})$ et $(X_M,Y_M)$ ses coordonnées dans le repère $(O,\vv{u},\vv{v}).$

Vous avez les relations en vertu du contenu écrit dans l'article 236 :

\left\{\begin{align*}
x_M &= X_M\cos\theta-Y_M\sin\theta\\
y_M &= X_M\sin\theta+Y_M\cos\theta.
\end{align*}\right.

La série d’équivalences suivante va fournir successivement, pour tout point $M$ du plan :

\begin{align*}
M\in\mathscr{D} &\Longleftrightarrow ax_M^2+2hx_My_M+by_M^2+2gx_M+2fy_M+c=0\\
&\Longleftrightarrow a(X_M^2\cos^2\theta-2X_MY_M\sin\theta\cos\theta+Y_M^2\sin^2\theta)\\
&\qquad+2h(X_M^2\sin\theta\cos\theta + X_MY_M(\cos^2\theta - \sin^2\theta)-Y_M^2\sin\theta\cos\theta)\\
&\qquad+b(X_M^2\sin^2\theta + 2X_MY_M\sin\theta\cos\theta+Y_M^2\cos^2\theta)\\
&\qquad+2g(X_M\cos\theta-Y_M\sin\theta)+2f(X_M\sin\theta+Y_M\cos\theta)+c=0\\
&\Longleftrightarrow (a\cos^2\theta+2h\sin\theta\cos\theta+b\sin^2\theta)X_M^2\\
&\qquad +2((b-a)\sin\theta\cos\theta+h(\cos^2\theta-\sin^2\theta))X_MY_M\\
&\qquad +(a\sin^2\theta-2h\sin\theta\cos\theta+b\cos^2\theta)Y_M^2\\
&\qquad +2(g\cos\theta+f\sin\theta)X_M+2(-g\sin\theta+f\cos\theta)Y_M+c=0.
\end{align*}

Dans le repère $(O,\vv{u},\vv{v})$ la droite $\mathscr{D}$ est verticale.

D’après la section précédente :

\begin{align*}
 &(a\cos^2\theta+2h\sin\theta\cos\theta+b\sin^2\theta)(a\sin^2\theta-2h\sin\theta\cos\theta+b\cos^2\theta)&\\
&\qquad -((b-a)\sin\theta\cos\theta+h(\cos^2\theta-\sin^2\theta))^2 = 0\\
& a^2\sin^2\theta\cos^2\theta  -2ah\sin\theta \cos^3\theta+ab\cos^4\theta\\
&\qquad +2ah\sin^3\theta\cos \theta-4h^2\sin^2\theta\cos^2\theta+2bh\sin\theta\cos^3\theta\\
&\qquad +ab\sin^4\theta-2bh\sin^3\theta\cos\theta+b^2\sin^2\theta\cos^2\theta\\
&\qquad -((b-a)^2\sin^2\theta\cos^2\theta+h^2(\cos^4\theta+\sin^4\theta-2\sin^2\theta\cos^2\theta)+2(b-a)h(\sin\theta\cos^3\theta-\sin^3\theta\cos\theta)) = 0\\
&(a^2-4h^2+b^2-(b-a)^2+2h^2)\sin^2\theta\cos^2\theta + (-2ah+2bh-2(b-a)h)\sin\theta\cos^3\theta\\
&\qquad +(ab-h^2)\cos^4\theta + (2ah-2bh+2(b-a)h)\sin^3\theta\cos\theta + (ab-h^2)\sin^4\theta = 0\\
&(-2h^2+2ab)\sin^2\theta\cos^2\theta + (ab-h^2)(\cos^4\theta+\sin^4\theta) = 0\\
&2(ab-h^2)\sin^2\theta\cos^2\theta + (ab-h^2)((\cos^2\theta+\sin^2\theta)^2-2\sin^2\theta\cos^2\theta) = 0\\
&2(ab-h^2)\sin^2\theta\cos^2\theta + (ab-h^2)(1-2\sin^2\theta\cos^2\theta) = 0\\
&ab-h^2=0.
\end{align*}

Note. Il est possible de justifier ce résultat en utilisant les propriétés d’un déterminant $2\times 2.$

Pour démontrer la nullité de $\Delta$, vous allez toujours utiliser la section précédente :

\begin{align*}
0 &= \begin{vmatrix}
a\cos^2\theta+2h\sin\theta\cos\theta+b\sin^2\theta & (b-a)\sin\theta\cos\theta+h(\cos^2\theta-\sin^2\theta) & g\cos\theta+f\sin\theta \\
(b-a)\sin\theta\cos\theta+h(\cos^2\theta-\sin^2\theta) & a\sin^2\theta-2h\sin\theta\cos\theta+b\cos^2\theta & -g\sin\theta+f\cos\theta \\
g\cos\theta+f\sin\theta& -g\sin\theta+f\cos\theta & c
\end{vmatrix}
\end{align*}

Au lieu de développer l’ensemble vous allez effectuer deux produits matriciels.

Le premier fournit :

\begin{pmatrix}
a & h & g \\
h& b& f \\
g & f & c
\end{pmatrix}
\begin{pmatrix}
\cos\theta & -\sin\theta & 0 \\
\sin\theta& \cos\theta& 0 \\
0 & 0 & 1
\end{pmatrix}
= \begin{pmatrix}
a\cos\theta +h \sin\theta& -a\sin\theta+h\cos\theta & g \\
h\cos\theta+b\sin\theta& -h\sin\theta+b\cos\theta& f \\
g\cos\theta+f\sin\theta & -g\sin\theta+f\cos\theta & c
\end{pmatrix}.

Puis :

\begin{align*}
&\begin{pmatrix}
\cos\theta & \sin\theta & 0 \\
-\sin\theta& \cos\theta& 0 \\
0 & 0 & 1
\end{pmatrix}
\begin{pmatrix}
a\cos\theta +h \sin\theta& -a\sin\theta+h\cos\theta & g \\
h\cos\theta+b\sin\theta& -h\sin\theta+b\cos\theta& f \\
g\cos\theta+f\sin\theta & -g\sin\theta+f\cos\theta & c
\end{pmatrix}\\
&\qquad=\begin{pmatrix}
a\cos^2\theta+2ah\sin\theta \cos\theta + b\sin^2\theta & (b-a)\sin\theta\cos\theta+h\cos^2\theta-h\sin^2\theta & g\cos\theta+f\sin\theta \\
(b-a)\sin\theta\cos\theta-h\sin^2\theta+h\cos^2\theta & a\sin^2\theta-2h\sin\theta\cos\theta+b\cos^2\theta & -g\sin\theta+f\cos\theta\\
g\cos\theta+f\sin\theta & -g\sin\theta+f\cos\theta & c
\end{pmatrix}.
\end{align*}

De ce qui précède :

\begin{align*}
0 &=\begin{vmatrix}
\cos\theta & \sin\theta & 0 \\
-\sin\theta& \cos\theta& 0 \\
0 & 0 & 1
\end{vmatrix}
\begin{vmatrix}
a\cos\theta +h \sin\theta& -a\sin\theta+h\cos\theta & g \\
h\cos\theta+b\sin\theta& -h\sin\theta+b\cos\theta& f \\
g\cos\theta+f\sin\theta & -g\sin\theta+f\cos\theta & c
\end{vmatrix} \\
&=
\begin{vmatrix}
a\cos\theta +h \sin\theta& -a\sin\theta+h\cos\theta & g \\
h\cos\theta+b\sin\theta& -h\sin\theta+b\cos\theta& f \\
g\cos\theta+f\sin\theta & -g\sin\theta+f\cos\theta & c
\end{vmatrix}\\
&=
\begin{vmatrix}
a & h & g \\
h& b& f \\
g & f & c
\end{vmatrix}
\begin{vmatrix}
\cos\theta & -\sin\theta & 0 \\
\sin\theta& \cos\theta& 0 \\
0 & 0 & 1
\end{vmatrix}\\
&=\begin{vmatrix}
a & h & g \\
h& b& f \\
g & f & c
\end{vmatrix}.
\end{align*}

Concluez

Si l’ensemble des points $M$ du plan de coordonnées $(x,y)$ satisfaisant l’équation :

ax^2+2hxy+by^2+2gx+2fy+c=0

est une droite, alors nécessairement :

\boxed{\begin{array}{ll}
ab-h^2 = 0\\
\\
\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} = 0.
\end{array}}

Remarque. Si l’ensemble précédent est composé de deux droites strictement parallèles, alors la même conclusion subsiste, d’après le contenu écrit dans l'article 241.

239. Caractérisez une conique dégénérée dans le cas où elle est l’union de deux droites sécantes (2/2)

Dans cette chronique vous établissez la réciproque du résultat établi dans l'article 238.

Etant donnés six réels $a, b, c, f, g$ et $h$ et un plan muni d’un repère $(O,\vv{i},\vv{j})$, vous étudiez l’ensemble $\mathscr{C}$ des points $M$ de coordonnées $(x,y)$ dans ce repère qui vérifient l’équation suivante :

ax^2+2hxy+by^2+2gx+2fy+c=0.

Vous considérez le discriminant noté $\Delta$ en posant :

\Delta =
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix}.

Vous supposez de plus :

\boxed{\begin{array}{ll}
ab-h^2 < 0\\
\Delta = 0.
\end{array}}

Vous allez démontrer que la conique $\mathscr{C}$ est l’union de deux droites sécantes. C’est un cas où elle est qualifiée de dégénérée.

Construisez un point appartenant à la conique $\mathscr{C}$

Comme $ab-h^2$ est non nul, le système suivant d’inconnues $(u,v)\in\R^2$ admet une solution :

\left\{\begin{align*}
au+hv+g&=0\\
hu+bv+f&=0.
\end{align*}\right.

Dès lors, la nullité du discriminant fournit :

\begin{align*}
0 &=
\begin{vmatrix}a & h & g \\
h & b & f \\
g& f & c
\end{vmatrix} \\
0 &=
\begin{vmatrix}a & h & au+hv+g \\
h & b & hu+bv+f \\
g& f & gu+fv+c
\end{vmatrix}\\
0 &=
\begin{vmatrix}a & h & 0 \\
h & b & 0 \\
g& f & gu+fv+c
\end{vmatrix} \\
0 &= \begin{vmatrix}a & h \\
h & b 
\end{vmatrix}  \times (gu+fv+c) \\
0 &= (ab-h^2)(gu+fv+c) \\
0 &= gu+fv+c.
\end{align*}

Vous allez maintenant prouver que le point de coordonnées $\Omega(u,v)$ dans le repère $(O,\vv{i},\vv{j})$ appartient à la conique $\mathscr{C}:$

\begin{align*}
au^2+2huv+bv^2+2gu+2fv+c &= au^2+2huv+bv^2+gu+fv+(gu+fv+c) \\
 &= au^2+2huv+bv^2+gu+fv \\
&= u(au+hv+g) +huv+bv^2+fv \\
&=huv+bv^2+fv \\
&=v(hu+bv+f) \\
&=0.
\end{align*}

Déterminez les équations des deux droites recherchées

Pour tout point $M$ de coordonnées $(x_M,y_M)$ dans le repère $(O,\vv{i},\vv{j})$, vous notez $(X_M,Y_M)$ ses coordonnées dans le repère $(\Omega,\vv{i},\vv{j})$ de sorte que :

\begin{align*}
X_M = x_M - u\\
Y_M = y_M - v.
\end{align*}

Alors la série d’équivalences suivantes fournit :

\begin{align*}
M\in\mathscr{C} &\Longleftrightarrow ax_M^2+2hx_My_M+by_M^2+2gx_M+2fy_M+c=0\\
&\Longleftrightarrow a(X_M+u)^2+2h(X_M+u)(Y_M+v)+b(Y_M+v)^2+2g(X_M+u)+2f(Y_M+v)+c=0\\
&\Longleftrightarrow a(X_M^2+u^2+2uX_M)+2h(X_MY_M+vX_M+uY_M+uv)+b(Y_M^2+v^2+2vY_M)\\&\qquad+2g(X_M+u)+2f(Y_M+v)+c=0\\
&\Longleftrightarrow aX_M^2+2hX_MY_M+bY_M^2\\&\qquad+2(au+hv+g)X_M+2(hu+bv+f)Y_M\\&\qquad+(au^2+2huv+bv^2+2gu+2fv+c)=0\\
&\Longleftrightarrow aX_M^2+2hX_MY_M+bY_M^2 = 0.
\end{align*}

Si $a$ est nul

Vous obtenez :

\begin{align*}
M\in\mathscr{C} &\Longleftrightarrow 2hX_MY_M+bY_M^2 = 0 \\
 &\Longleftrightarrow Y_M(2hX_M+bY_M) = 0\\
&\Longleftrightarrow (Y_M = 0) \text{ ou } (2hX_M+bY_M = 0).
\end{align*}

Notez que $ab-h^2< 0$ s’écrit $-h^2 < 0$ ce qui force la non nullité de $h.$ Donc le vecteur $\vv{d}_1\begin{pmatrix}-b \\ 2h\end{pmatrix}$ est non nul. Ainsi $2hX_M+bY_M = 0$ est bien l’équation d’une droite $\mathscr{D}_1$ qui admet $\vv{d}_1$ pour vecteur directeur.

La droite $\mathscr{D}_2$ d’équation $Y_M = 0$, qui désigne la droite horizontale passant par $\Omega$ admet pour vecteur directeur $\vv{d}_2 \begin{pmatrix}1 \\ 0\end{pmatrix}.$

Le déterminant $\mathrm{det}({\vv{d}_1, \vv{d}_2})$ est non nul :

\begin{align*}
\mathrm{det}({\vv{d}_1, \vv{d}_2}) &= \begin{vmatrix} -b & 1\\ 2h  & 0\end{vmatrix}\\
&=-2h.
\end{align*}

Les droites $\mathscr{D}_1$ et $\mathscr{D}_2$ sont bien deux droites sécantes, le point $\Omega$ est leur point d’intersection et vous avez :

\begin{align*}
M\in\mathscr{C} &\Longleftrightarrow M\in\mathscr{D}_1 \cup \mathscr{D}_2.
\end{align*}

Si $a$ est non nul

Comme $ab-h^2<0$ vous pouvez considérer $\delta = \sqrt{h^2-ab}$ qui est bien défini et est non nul. Alors :

\begin{align*}
M\in\mathscr{C} &\Longleftrightarrow aX_M^2+2hX_MY_M+bY_M^2 = 0\\
& \Longleftrightarrow a^2X_M^2+2ahX_MY_M+abY_M^2 = 0 \\
& \Longleftrightarrow (aX_M+hY_M)^2+(ab-h^2)Y_M^2=0 \\
& \Longleftrightarrow (aX_M+hY_M)^2-\delta^2Y_M^2=0 \\
& \Longleftrightarrow (aX_M+hY_M)^2-(\delta Y_M)^2=0 \\
& \Longleftrightarrow (aX_M+hY_M + \delta Y_M)(aX_M+hY_M - \delta Y_M)=0 \\
& \Longleftrightarrow (aX_M+(h+\delta)Y_M)(aX_M+(h- \delta) Y_M)=0\\
& \Longleftrightarrow (aX_M+(h+\delta)Y_M=0)\text{ ou }(aX_M+(h- \delta) Y_M=0).
\end{align*}

Comme $a$ est non nul, aucun des vecteurs parmi $\vv{d}_1\begin{pmatrix}-h-\delta \\ a\end{pmatrix}$ et $\vv{d}_2\begin{pmatrix}-h+\delta \\ a\end{pmatrix}$ n’est nul.

L’ensemble des points $M$ tels que $aX_M+(h+\delta)Y_M=0$ est donc une droite $\mathscr{D}_1$ qui passe par le point $\Omega$ et qui est dirigée par le vecteur $\vv{d}_1.$

L’ensemble des points $M$ tels que $aX_M+(h-\delta)Y_M=0$ est une droite $\mathscr{D}_2$ qui passe par le point $\Omega$ et qui est dirigée par le vecteur $\vv{d}_2.$

Le déterminant $\mathrm{det}({\vv{d}_1, \vv{d}_2})$ est non nul :

\begin{align*}
\mathrm{det}({\vv{d}_1, \vv{d}_2}) &= \begin{vmatrix} -h-\delta & -h+\delta\\ a  & a\end{vmatrix}\\
&= a \begin{vmatrix} -h-\delta & -h+\delta\\ 1  & 1\end{vmatrix}\\
&= a  \begin{vmatrix} -2\delta & -h+\delta\\ 0  & 1\end{vmatrix}\\
&= -2a \delta.
\end{align*}

Ainsi les droites $\mathscr{D}_1$ et $\mathscr{D}_2$ sont bien deux droites sécantes, encore au point $\Omega.$

Compte tenu de la série d’équivalences ci-dessus, vous avez encore :

\begin{align*}
M\in\mathscr{C} &\Longleftrightarrow M\in\mathscr{D}_1 \cup \mathscr{D}_2.
\end{align*}

Concluez

Vous avez démontré qu’il existe deux droites $\mathscr{D}_1$ et $\mathscr{D}_2$ sécantes en un point $\Omega(u,v)$ satisfaisant les conditions suivantes :

\boxed{\begin{array}{lll}
au+hv+g=0\\
hu+bv+f=0\\
\mathscr{C} = \mathscr{D}_1 \cup \mathscr{D}_2.
\end{array}}