You can try the following design approach if it works for your application.
A = 0;
B = 16.6;
C = 475;
D = 0;
sys = ss(A, B, C, D)
The design task is to analytically compute the control gain K, or tune it
K = 10/16.6
until the desired pole is obtained
eig(A - B*K) % eigenvalue of compensated system
If you understand the theory behind the design, then this place() function is not needed:
place(A, B, -10) % should return the same value as K
sysD = ss(A-B*K, B, C, D); % dummy
scale = 1/dcgain(sysD) % scale for input signal
Td = 550; % Desired Temperature (input signal)
input = Td*scale % scaled input signal
Compensated System:
sysC = ss(A-B*K, B*(input), C, D);
step(sysC, 1)
S = stepinfo(sysC)