FeatureRequest: Consideration of disturbances in lateral direction

[XaverR]

Hi Andres,

would it be possible for you to integrate disturbances in lateral direction? I'm thinking of simulating the effect of cross slopes and/or cross winds. For this I need at least a new input for the lateral forces on the axles/tires. I currently solved it like this:

% Characteristic curve
`FyF = nF * self.tire.Characteristic(ALPHAF, FzF/nF, muy) + FyFDisturb;
FyR = nR * self.tire.Characteristic(ALPHAR, FzR/nR, muy) + FyRDisturb;

FyFDisturb and FyRDisturb are new inputs to the model, which currently is VehicleSimpleNonLinear.

What do you think of it?

Best regards,
Xaver

[andresmendes]

Dear Xavier,

I believe that the lateral disturbances should be considered directly in the equations of motion and not in the characteristic curve of the tires (Barnard, Richard Harry. Road vehicle aerodynamic design: An introduction. 2001.). If you don't have access to this book take a look at this doctoral thesis, section 2.2.

Considering the wind flow parallel to the ground, we can define an absolute wind direction BETA and a center of pressure CP located somewhere over the longitudinal axis of the vehicle (for simplicity). See this figure. This way, the wind generates forces in the X and Y directions (inertial reference frame) and yaw moment.

Check the Fwind terms in equations (4), (5) and (6) of the VehicleSimpleNonlinear.m:

            BETA = pi/2; % Lateral wind

            Fwind = 1000;

            THETA = pi/2  - (BETA - PSI);

            lCP = 0; % Center of gravity coincident with center of pressure

            dx(1,1) = v * cos(ALPHAT + PSI); % X
            dx(2,1) = v * sin(ALPHAT + PSI); % Y
            dx(3,1) = dPSI; % dPSI
            dx(4,1) = (FxF * cos(ALPHAT - DELTA) + FxR * cos(ALPHAT) + FyF * sin(ALPHAT - DELTA) + FyR * sin(ALPHAT) - Fwind*cos(BETA) )/(m);
            dx(5,1) = ( - FxF * sin(ALPHAT - DELTA) - FxR * sin(ALPHAT) + FyF * cos(ALPHAT - DELTA) + FyR * cos(ALPHAT) - m * v * dPSI - Fwind*sin(BETA)) / (m * v);
            dx(6,1) = (FxF * a * sin(DELTA) + FyF * a * cos(DELTA) - FyR * b - lCP*Fwind*cos(THETA)) / I;

Specifically, in your case, the initial cross wind generates a force in the lateral direction of the vehicle (beta = 90 deg).

Try different values of BETA, Fwind and lCP with the vehicle moving in a straight line without input.

Let me know if it worked.

OBS: To properly implement cross slopes effects, the model should consider lateral load transfer. This could be another issue.