All devices and equipment must pass the oldest EMC standard which is the flicker-test. This standard defines how an equipment is allowed to draw power from the network without generating voltage fluctuations which would disturb lightning.
In most modern devices flicker is controlled by software (e.g. switching on and off of a heater), also hardware can be affected, e.g. if performance specification and flicker standard can only be reached through optimization of the form of the switching transient.
Late discovery: In most cases the non-compliance is discovered only when hardware and software prototypes are available, in the worst case only at the compliance measurements.
Complexity: The flicker-measurement is highly non-linear and rather difficult to understand.
One can trust ones luck and wait until hard- and software are ready.
A certain overview is achievable by using the approximation formula and curves from the standard.
A much more exact prediction is available through the simulation of the flicker measurement, i.e. of a flickermeter.
With this approach, the input data can be measured or simulated results from control system simulation and software development or from measurements on prototypes and rapid prototypes on one hand.
On the other hand generic generic waveforms (rectangle, trapezoidal,...) can be simulated.
The result is a first pass at the real flicker test.
Solcept has developed a flicker-measurement simulator and provides it here as open source under the Boost license.
The simulator runs on MathWorks MATLAB or the open source tool Octave.
It resulted in the same values as the measurement with real software and electronics for a 2.5 kW heater control system!
In addition on this site more information about flicker for developers plus three application examples: