Simulation - Algorithm to introduce noise?

Have a new project and I'd like to get my simulation logic a bit more realistic, as such I'd like to add noise to the simulated analog feedback the process requires (temperatures, pressures). It Would be nice to configure things like amplitude, rate, offset... so I could vary the noise across all the different analog feedback devices I want to try this on.

Anyone have a good place to start? I can already generate nice simulated analog signals (sine waves/ramps), but I'd like to dirty this up. Probably the opposite of a low-pass filter?

Reason for this, the process we are upgrading has a very long cycle time. I've got very basic captures of the data and as I build the SCADA and trending getting a sense of what real data might look like on-screen and in the reports will be helpful. I'd like to leave this simulation run for 24+ hours and look at the data and not just see a straight line.

Plus, I always like to add a littl' something to my development simulation as new projects come at me.

Suggestions? Note, I'm not trying to model this process, just some dirty analog feedback.

One idea that comes to mind...
A poor mans modified pseudorandom noise generator.
Create math code that will continuously output 2 lots of random numbers at a input time value (tick). I faintly remember that there may be some code for this floating around on this forum somewhere. These numbers will be your x and y values.
Use those numbers to create a pulse train. The amplitude varies with the y value, and the pulse width will vary with the x value. x is scaleable to your chosen time resolution.
Put the randomised pulse train through a simple first order filter and add that to your analog signal.
Rough, but will give you some very random noise

I always like using the least significant bits of the clock as my noise generator. PLCs often reply back down to the basement these days

i guess use Triangle wave formula

Look up Gaussian noise generator. I wanted to simulate sample jitter once. I used a Weibull. There are different noise distributions for different types of noise.

mk42 said:

PLCs often reply back down to the basement these days

Click to expand...

That was supposed to be "nanoseconds" not "basement". Way too late to edit, but that was a heck of an autocorrect.

Thanks for the suggestions!