How to convert from plot3 to surf?
Show older comments
Hello!
I want to convert my plot3 graph to surf graph but it says data is too much. Can anyone help me to convert it a surf graph?
Thanks for any response.
Accepted Answer
More Answers (1)
format short g;
format compact;
fontSize = 18;
counter= 0;
counter1=0;
for ky=0:1:1000
for i=0:1:89
angle=deg2rad(i);
c=ky*cos(angle);
k1ky=0.85*ky;
p1= 105*sin(angle)+105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
p2= -105*sin(angle)+105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
p3= 105*sin(angle)-105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
p4= -105*sin(angle)-105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
es1 = (0.003*p1)/c;
es2 = (0.003*p2)/c;
es3 = (0.003*p3)/c;
es4 = (0.003*p4)/c;
ss1 = max(min(es1*200000,420),-420);
ss2 = max(min(es2*200000,420),-420);
ss3 = max(min(es3*200000,420),-420);
ss4 = max(min(es4*200000,420),-420);
fs1 = ss1*100*pi*10^-3;
fs2 = ss2*100*pi*10^-3;
fs3 = ss3*100*pi*10^-3;
fs4 = ss4*100*pi*10^-3;
fs1y = fs1*0.105;
fs1x = fs1*-0.105;
fs2y = fs2*0.105;
fs2x = fs2*0.105;
fs3y = fs3*-0.105;
fs3x = fs3*-0.105;
fs4y = fs4*-0.105;
fs4x = fs4*0.105;
Tfs = fs1+fs2+fs3+fs4;
Msy = fs1x+fs2x+fs3x+fs4x;
Msx = fs1y+fs2y+fs3y+fs4y;
if k1ky <= 300
if 300*tan(angle)< k1ky
Area= (((k1ky-(300*tan(angle)))+k1ky)/2)*300;
Centroidx= (300*tan(angle)*150*(-50))/Area;
Centroidy= (((150-((k1ky-300*tan(angle))/2)) *300* (k1ky-300*tan(angle)))+ (300*tan(angle)*150* (150-((k1ky-300*tan(angle))+(300*tan(angle)/3)))))/Area;
else
Area= ((k1ky*cot(angle))*k1ky)/2;
Centroidx= (((k1ky*cot(angle))/3)-150);
Centroidy= 150-(k1ky/3);
end
else
if 300*tan(angle)<= k1ky
if k1ky-300*tan(angle) >= 300
Area = 90000;
Centroidx = 0;
Centroidy = 0;
else
Area= 90000-(((300-(k1ky-(300*tan(angle))))*((300-(k1ky-(300*tan(angle))))*cot(angle)))/2);
Centroidx = -((150-(((300-(k1ky-(300*tan(angle))))*cot(angle))/3))*(90000-Area))/Area;
Centroidy = -((((300-(k1ky-(300*tan(angle))))/3)-150)*(90000-Area))/Area;
end
else
Area= (k1ky*cot(angle))*300-((300*cot(angle))*150);
Centroidx = ((((k1ky*cot(angle))*300)*(((k1ky*cot(angle))/2)-150)) - (((300*cot(angle))*150)*((k1ky*cot(angle))-((300*cot(angle))/3)-150)))/Area;
Centroidy = -(((300*cot(angle))*150)*-50)/Area;
end
end
fc = 0.85*20*Area*10^-3;
N = max(0,fc+Tfs);
Mx = Msx + fc*Centroidy*10^-3;
My = Msy + fc*Centroidx*10^-3;
counter = counter + 1;
xAll(counter) = Mx;
yAll(counter) = My;
zAll(counter) = N;
end
end
figure
plot3(xAll,-yAll,zAll,'m.');
grid on
hold on
plot3(-yAll,xAll,zAll,'m.');
set(gca, 'CameraPosition', [607.33 607.6 2775.3]);
xlabel('Mx(kNm)')
ylabel('My(kNm)')
zlabel('N(kN)')
min(zAll), max(zAll)
figure
N = 50;
C1a = xAll.';
C2a = -yAll.';
C3a = zAll.';
mask = isfinite(C1a) & isfinite(C2a) & isfinite(C3a);
C1a = C1a(mask); C2a = C2a(mask); C3a = C3a(mask);
xmin = min(C1a); xmax = max(C1a);
xvec = linspace(xmin, xmax, N);
ymin = min(C2a); ymax = max(C2a);
yvec = linspace(ymin, ymax, N);
[X, Y] = meshgrid(xvec, yvec);
F = scatteredInterpolant(C1a, C2a, C3a, 'natural');
Z = F(X, Y);
surf(X, Y, Z, 'edgecolor', 'none')
xlabel('Mx(kNm)')
ylabel('My(kNm)')
zlabel('N(kN)')
hold on
min(C3a), max(C3a)
min(Z(:)), max(Z(:))
C1b = -yAll.';
C2b = xAll.';
C3b = zAll.';
mask = isfinite(C1b) & isfinite(C2b) & isfinite(C3b);
C1b = C1b(mask); C2b = C2b(mask); C3b = C3b(mask);
xmin = min(C1b); xmax = max(C1b);
xvec = linspace(xmin, xmax, N);
ymin = min(C2b); ymax = max(C2b);
yvec = linspace(ymin, ymax, N);
[X, Y] = meshgrid(xvec, yvec);
F = scatteredInterpolant(C1b, C2b, C3b, 'natural');
surf(X, Y, Z, 'edgecolor', 'none')
hold off
xlim auto; ylim auto; zlim auto
set(gca, 'CameraPosition', [607.33 607.6 2775.3]);
min(C3b), max(C3b)
min(Z(:)), max(Z(:))
figure
subplot(3,1,1); histogram(C1a)
subplot(3,1,2); histogram(C2a)
subplot(3,1,3); histogram(C3a)
Your coordinates are very biased, and when sorted into order, is not very continuous. That leads to sharp changes between values, and that leads to really big spurious peaks when you interpolate in 3D.
3 Comments
Ege Çetiner
on 18 Apr 2021
Walter Roberson
on 19 Apr 2021
10000 is the histogram count. Your coordinates are not equally distributed, they are mostly very small and very closely packed, but also not very continuous.
DELTA = 0.002;
format short g;
format compact;
fontSize = 18;
counter= 0;
counter1=0;
for ky=0:1:1000
for i=0:1:89
angle=deg2rad(i);
c=ky*cos(angle);
k1ky=0.85*ky;
p1= 105*sin(angle)+105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
p2= -105*sin(angle)+105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
p3= 105*sin(angle)-105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
p4= -105*sin(angle)-105*cos(angle)-150*sin(angle)+(ky-150)*cos(angle);
es1 = (0.003*p1)/c;
es2 = (0.003*p2)/c;
es3 = (0.003*p3)/c;
es4 = (0.003*p4)/c;
ss1 = max(min(es1*200000,420),-420);
ss2 = max(min(es2*200000,420),-420);
ss3 = max(min(es3*200000,420),-420);
ss4 = max(min(es4*200000,420),-420);
fs1 = ss1*100*pi*10^-3;
fs2 = ss2*100*pi*10^-3;
fs3 = ss3*100*pi*10^-3;
fs4 = ss4*100*pi*10^-3;
fs1y = fs1*0.105;
fs1x = fs1*-0.105;
fs2y = fs2*0.105;
fs2x = fs2*0.105;
fs3y = fs3*-0.105;
fs3x = fs3*-0.105;
fs4y = fs4*-0.105;
fs4x = fs4*0.105;
Tfs = fs1+fs2+fs3+fs4;
Msy = fs1x+fs2x+fs3x+fs4x;
Msx = fs1y+fs2y+fs3y+fs4y;
if k1ky <= 300
if 300*tan(angle)< k1ky
Area= (((k1ky-(300*tan(angle)))+k1ky)/2)*300;
Centroidx= (300*tan(angle)*150*(-50))/Area;
Centroidy= (((150-((k1ky-300*tan(angle))/2)) *300* (k1ky-300*tan(angle)))+ (300*tan(angle)*150* (150-((k1ky-300*tan(angle))+(300*tan(angle)/3)))))/Area;
else
Area= ((k1ky*cot(angle))*k1ky)/2;
Centroidx= (((k1ky*cot(angle))/3)-150);
Centroidy= 150-(k1ky/3);
end
else
if 300*tan(angle)<= k1ky
if k1ky-300*tan(angle) >= 300
Area = 90000;
Centroidx = 0;
Centroidy = 0;
else
Area= 90000-(((300-(k1ky-(300*tan(angle))))*((300-(k1ky-(300*tan(angle))))*cot(angle)))/2);
Centroidx = -((150-(((300-(k1ky-(300*tan(angle))))*cot(angle))/3))*(90000-Area))/Area;
Centroidy = -((((300-(k1ky-(300*tan(angle))))/3)-150)*(90000-Area))/Area;
end
else
Area= (k1ky*cot(angle))*300-((300*cot(angle))*150);
Centroidx = ((((k1ky*cot(angle))*300)*(((k1ky*cot(angle))/2)-150)) - (((300*cot(angle))*150)*((k1ky*cot(angle))-((300*cot(angle))/3)-150)))/Area;
Centroidy = -(((300*cot(angle))*150)*-50)/Area;
end
end
fc = 0.85*20*Area*10^-3;
N = max(0,fc+Tfs);
Mx = Msx + fc*Centroidy*10^-3;
My = Msy + fc*Centroidx*10^-3;
counter = counter + 1;
xAll(counter) = Mx;
yAll(counter) = My;
zAll(counter) = N;
end
end
figure
plot3(xAll,-yAll,zAll,'m.');
grid on
hold on
plot3(-yAll,xAll,zAll,'m.');
set(gca, 'CameraPosition', [607.33 607.6 2775.3]);
xlabel('Mx(kNm)')
ylabel('My(kNm)')
zlabel('N(kN)')
min(zAll), max(zAll)
figure
N = 50;
C1a = xAll.';
C2a = -yAll.';
C3a = zAll.';
mask = isfinite(C1a) & isfinite(C2a) & isfinite(C3a) & abs(C1a) <= DELTA & abs(C2a) <= DELTA;
C1a = C1a(mask); C2a = C2a(mask); C3a = C3a(mask);
xmin = min(C1a); xmax = max(C1a);
xvec = linspace(xmin, xmax, N);
ymin = min(C2a); ymax = max(C2a);
yvec = linspace(ymin, ymax, N);
[X, Y] = meshgrid(xvec, yvec);
F = scatteredInterpolant(C1a, C2a, C3a, 'natural');
Z = F(X, Y);
surf(X, Y, Z, 'edgecolor', 'none')
xlabel('Mx(kNm)')
ylabel('My(kNm)')
zlabel('N(kN)')
hold on
min(C3a), max(C3a)
min(Z(:)), max(Z(:))
C1b = -yAll.';
C2b = xAll.';
C3b = zAll.';
mask = isfinite(C1b) & isfinite(C2b) & isfinite(C3b) & abs(C1b) <= DELTA & abs(C2b) <= DELTA;
C1b = C1b(mask); C2b = C2b(mask); C3b = C3b(mask);
xmin = min(C1b); xmax = max(C1b);
xvec = linspace(xmin, xmax, N);
ymin = min(C2b); ymax = max(C2b);
yvec = linspace(ymin, ymax, N);
[X, Y] = meshgrid(xvec, yvec);
F = scatteredInterpolant(C1b, C2b, C3b, 'natural');
surf(X, Y, Z, 'edgecolor', 'none')
hold off
xlim auto; ylim auto; zlim auto
set(gca, 'CameraPosition', [607.33 607.6 2775.3]);
min(C3b), max(C3b)
min(Z(:)), max(Z(:))
figure
subplot(3,1,1); histogram(C1a)
subplot(3,1,2); histogram(C2a)
subplot(3,1,3); histogram(C3a)
figure
scatter3(C1a(:), C2a(:), C3a(:))
So what I have done here at the bottom is to restrict the output to places where the first coordinate (which is either xAll or -yall) has within +/- 0.002, and the second coordinate (which is either -yall or xall) is as well, and look at the output then. Notice from the bottom plot that even within +/- 0.0005 that you have zAll values from 0 to 2047. If you look at the histogram directly above the bottom plot, you will see that in that narrow range, some of the zAll values are "close" to 0, but others are up around 2000.
So you have rapid changes in Z values in a small area.
[C1as,idx] = sort(C1a);
C2as = C2a(idx);
C3as = C3a(idx);
plot3(C1as, C2as, C3as)
You have a number of different places pretty much at (0,0) with different Z values.
Any attempt to interpolate a smooth surface is going to have problems with changes that rapid, unless it does something like an envelope calculation (a binned maximimum intensity projection might help, for example)
Categories
Find more on Surface and Mesh Plots in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!








