To balance power generation and consumption, traditional power plants such as thermal and hydro have governors that regulate the steam/water that flows into the turbine.
The energy that renewables harness is intermittent, so how is generation matched with consumption for them?
The other answers are correct, but one crucial thing is not mentioned.
If you increase or decrease the amount of active1 power you feed into the grid then this must be compensated for somehow. As others have stated, electricity must be used instantaneously.
pdb5627 correctly states that there must be equilibrium. You can for instance charge batteries and/or reduce the power produced by power plants that can be regulated, as described by VoltageSpike and PStechPaul.
These are ways to actively attempt to create an equilibrium, and it's how grid operators maintain a stable frequency.
But, the grid frequency varies, and that's because the sum of production and the sum of loads (as we think of them) aren't equal at all times.
When someone turns off a motor, or there's a gust of wind, you'll have excess energy that must be compensated for. What happens in real-time is that the grid frequency goes up or down. The solutions VoltageSpike and PStechPaul mentions are ways to avoid that the frequency continues to rise, or continues to fall above or below certain given limits.
The figure below shows the grid frequency in the Norwegian/European grid a few minutes ago. As you can see, it goes up and down constantly.
The x-axis is 90 seconds and the y-axis goes from 50.000 Hz to 50.010 Hz.
In the 4 minutes from 12:09:30 to 12:13:30, the grid frequency went from 50.0011 to 50.0384
Obviously, 0.0373 Hz is not much, unless you remember that this is the frequency of every single synchronous motor and generator in the entire grid2. The amount of inertia and rotating mass that's present in the grid is enormous, so a small change in frequency means a lot of excess energy has been stored (or used) in the rotating masses in the grid.
Go check it out yourself. :) (At this very moment the frequency is 49.973 Hz.)
1 If you haven't heard of active and reactive power then you can disregard the word active.
2 We usually consider the grid frequency to be identical in the entire interconnected grid. This is not really true since there's a delay and damping factor. A large motor being turned off in Sweden will cause a frequency spike and a temporary frequency rise nearby (until the load loss has been compensated for). The frequency in nearby stations will also increase, but the effect in Spain is obviously negligible. The effect from one single motor will only have a limited reach, but a cloud momentarily covering a 1 GW solar plant will have a greater reach.
To maintain a steady-state equilibrium, the power generated and consumed must be kept equal. This can be shown be an equation like the following: $$\sum{P_{gen}} - \sum{P_{load}} = 0$$
You could also think of it as a simple sum of all power injections to the grid having to sum to zero, where generators are positive and loads are negative and energy storage devices could be either positive or negative: $$\sum{P_i} = 0$$
Most loads and some generators are non-controllable and don't contribute to load following. The generators and loads that do participate in load following have to adjust their output or consumption to compensate for the aggregate changes to load and non-controlled generation. There must be some device(s) that are responsive to changes to the load balance, otherwise equilibrium cannot be maintained.
Note that not all renewables are intermittent / non-controllable. For example, hydropower, geothermal, and biomass plants are often able to regulate water or steam flow to control their output. Finding ways for more types of renewable generation to be able to participate in load following / frequency control is an area of active research.
For renewable and sustainable energy generation, excess production can be stored in batteries or other systems on-site, or if they are grid-tied, they can transfer it there, and draw back from the grid when needed. For an isolated micro-grid, without local storage, excess production can be "wasted" by simply regulating the output voltage, but then there will be brownouts and blackouts when the wind isn't blowing or the sun isn't shining. So a practical micro-grid would have local storage, which, conveniently, can be the battery packs in electric vehicles.
It's not matched by the renewables, the matching comes from the base load. I've seen the actual graphs from a coal plant that makes up the difference from a local wind farm. So the coal plant changes it's generation to make up the difference
What really happens is there is an central office monitors the grid and asks the control operators of the plant to produce more or less depending on what they need to keep the grid stable.
