Perturb-and-Observe Maximum-Power-Point Tracking
Perturb-and-observe (P&O) maximum-power-point tracking is a control method that repeatedly nudges a solar panel’s operating voltage, measures the resulting power, and keeps nudging in the direction that increases power. A solar inverter uses it to extract nearly the most available energy as sunlight, temperature, and shading change.
The problem: a panel does not have one fixed output
A photovoltaic panel can produce different combinations of voltage and current. Connect it to a particular load and the load determines the operating point: perhaps the panel supplies high current at a lower voltage, or lower current at a higher voltage. The product of those values is electrical power:
[ P = V \times I ]
For a given amount of sunlight and a given panel temperature, one voltage produces more power than any other. This is the maximum power point (MPP). It is not usually the panel’s rated open-circuit voltage, where no current flows, or its short-circuit current, where the voltage is nearly zero.
Without tracking, the panel can operate away from the MPP. A fixed-voltage load, a battery, or an inverter whose input behavior does not adapt may force the panel to a point that leaves usable energy behind. The MPP also moves: stronger sunlight generally raises available current, while temperature changes can shift the useful voltage substantially. Partial shading can make the curve more complicated still.
The inverter therefore needs a way to choose the panel-side voltage while still delivering energy to the grid, a battery, or another downstream circuit. P&O provides that choice without requiring a detailed mathematical model of every panel.
The shape the controller is following
Imagine measuring panel power while sweeping the panel voltage from low to high. Power first rises, reaches a peak, and then falls. On the left side of the peak, increasing voltage increases power. On the right side, increasing voltage decreases power.
In calculus terms, the slope of the power-versus-voltage curve is positive before the MPP and negative after it. At the peak, the slope is approximately zero:
[ \frac{dP}{dV} \approx 0 ]
P&O does not normally calculate that derivative directly. Instead, it makes a small, deliberate change to voltage—a perturbation—and observes the measured power afterward.
If the last voltage change increased power, keep moving in the same direction. If it decreased power, reverse direction.
That simple rule causes the operating point to climb toward the peak from either side.
How one control cycle works
A typical inverter has voltage and current sensors on its photovoltaic input. Its controller samples both and computes input power. It also remembers the previous voltage and power, along with the direction of its most recent voltage adjustment.
A cycle looks like this:
- Measure the current panel voltage
Vand currentI. - Compute the present power
P = V × I. - Compare
Pwith the previous measured power. - If power rose after the previous voltage perturbation, retain that perturbation direction.
- If power fell, reverse the direction.
- Ask the inverter’s power converter to move the panel voltage by a small step.
- Wait for the electrical response to settle, then repeat.
The power converter is the inverter stage that controls how much energy flows from the panel into the rest of the system. Depending on the hardware, it may be a DC-to-DC converter, an inverter bridge, or coordinated control of both. The P&O algorithm does not directly command “more sunlight”; it changes the electrical load presented to the panel so that the panel voltage moves toward the requested value.
For example, suppose the panel is operating at 30 V and producing 200 W. The controller raises its voltage target slightly, and the next settled measurement is 30.2 V and 202 W. Because power increased, it raises the target again. If the next measurement is 204 W, it continues. Once it passes the peak, a further increase might produce 203 W. The controller detects the decrease and starts stepping voltage downward instead.
Why it never sits perfectly still
Near the MPP, any small move in either direction tends to reduce power. P&O therefore reverses direction repeatedly, producing a small oscillation around the peak. That is normal behavior, not necessarily a fault. The size of this oscillation is controlled by the voltage step size.
A larger step finds a newly moved MPP quickly, but it wastes more power while bouncing around the peak and can create larger electrical disturbances. A smaller step reduces steady-state loss but takes longer to respond. Real implementations choose a step and sampling interval that account for panel capacitance, converter dynamics, sensor noise, and the time needed for the input voltage and current to settle.
The controller must also avoid treating transient measurements as meaningful observations. If it changes voltage and samples immediately, the measured power may reflect converter dynamics rather than the panel’s new operating point. A practical implementation waits, filters measurements, or otherwise synchronizes the perturbation and observation.
The important limitation: sunlight can change during the experiment
P&O assumes that the main reason power changed was its own voltage perturbation. That assumption can fail when irradiance—the sunlight power arriving at the panel—changes quickly.
Suppose the controller raises voltage, but a cloud passes at the same time. Power may fall even though the voltage move was in the correct direction. Basic P&O interprets the fall as evidence that it should reverse, potentially moving away from the new MPP. Conversely, increasing sunlight can look like a successful perturbation.
Controllers mitigate this with filtering, suitable step timing, adaptive step sizes, or related tracking methods. Rapidly changing partial shade can also create several local peaks in the power curve; basic P&O may settle on one peak rather than discovering the globally highest one.
Where this appears in an inverter
If a log reports MPPT voltage, tracking, MPP lost, or an input operating point that oscillates around a target, it is describing this control loop or one of its supervisory states. The voltage is the panel-side value the tracker is trying to position; the current is measured both to calculate power and to reveal whether the panel can supply the requested operating point.
P&O is popular because it is model-light: it needs reliable voltage and current measurements and control over the panel’s electrical load, not a precise panel model. Its trade-offs—oscillation, response speed, sensor quality, and confusion during rapid irradiance changes—are direct consequences of learning the peak by deliberately moving around it.