Solar Performance Monitoring: How We Tell Whether a PV System Is Working Properly

Author: Steve Fairless
Published: 28th May 2026 · Updated: 2nd September 2026

Good solar monitoring is not about checking whether today's generation number looks high or low. It is about comparing the system with its own design, weather, season and historical behaviour so that genuine faults stand out from normal variation. Solar output changes every day with irradiance, temperature, cloud, shading and time of year. A useful monitoring process therefore looks for patterns rather than reacting to one number.

At Sustainable Energy Engineering, we use monitoring as part of both handover and long-term system care. We want customers to understand the difference between kW and kWh, how to recognise an inverter or string outage, what battery flows mean and when a change in generation deserves investigation.

Monitoring is particularly valuable because some PV faults do not stop the whole system. One string, MPPT or inverter can underperform while the rest of the array continues generating, so the electricity bill may not make the loss obvious.

Updated: 2nd September 2026.

Start With the Design Baseline

An MCS performance estimate is not a guaranteed generation contract, but it gives a structured baseline. MCS MIS 3002 Issue 6.0 uses installed PV capacity, roof orientation, inclination, postcode region and a shade factor to establish expected annual output. Where storage is included, the handover information also records estimated self-consumption and grid independence.

That baseline is useful because it tells us what the system was designed to do. Monitoring without the design context can encourage misleading comparisons: two 5 kWp systems can produce different annual energy because one faces south, another east-west, one is shaded and another is not.

Why One Cloudy Day Tells You Very Little

PV generation is weather-dependent. A dark summer day can produce less than a bright spring day, and a cold clear day can produce strong instantaneous power because module voltage and efficiency respond to temperature as well as sunlight.

We therefore look at longer windows: week against week when weather is broadly similar, month against the same month in previous years, and annual generation against the modelled range. A single poor day is rarely diagnostic on its own.

PVGIS Is Useful for Weather and Location Context

The European Commission Joint Research Centre's PVGIS provides solar-resource and PV-generation modelling for different locations and system configurations. It is useful for understanding the seasonal shape we should expect and for checking whether a site's long-term generation is broadly plausible.

PVGIS is not a substitute for the actual system design or monitoring. The installed roof, shading, equipment, downtime and maintenance history all affect real performance. We use modelling as a reference point, then investigate the difference rather than assuming the model or the inverter dashboard must be right.

What a Healthy Daily Generation Curve Can Look Like

On an unshaded single-orientation array, a clear day often produces a smooth rise toward midday and fall later in the day. East-west systems produce a broader curve, and shading can create repeated dips at predictable times. Clouds can create rapid peaks and troughs.

That means the shape is often more informative than the total. A regular morning drop at the same time every clear day may be a chimney shadow. A sudden permanent loss of the afternoon half of the curve can indicate an MPPT or string issue. We look for changes from the system's own established pattern.

Clipping Is Not Automatically an Inverter Fault

If DC array capacity is larger than the inverter's AC output, the inverter can reach its rated power during strong solar conditions. A flat top in the power curve for part of a bright day can therefore be a normal result of intentional DC-to-AC sizing rather than evidence of failure.

The question is whether the design predicted that behaviour and whether the annual energy gained from the larger array justifies the limited periods of clipping. We do not diagnose clipping from a screenshot without knowing the array and inverter ratings.

How We Spot a String or MPPT Problem

Some monitoring platforms show separate MPPT voltages, currents or power. Others only show total inverter output. Where separate data is available, a sudden imbalance between roof faces or strings can help narrow a fault.

Without detailed string monitoring, the pattern can still be visible in daily generation. A system may continue producing at 50% or 70% of normal output because one electrical group is missing. This is one reason customers should not rely only on whether the inverter display says “generating”.

Communication Failure and Generation Failure Are Different

A monitoring app can go offline because Wi-Fi, Ethernet, cloud services or the communications gateway has failed while the inverter continues generating normally. The reverse is also possible: the app can remain reachable while the inverter reports an electrical fault.

We distinguish data loss from energy loss. Generation-meter readings, inverter status and on-site electrical information can help confirm whether the problem is communications or the PV system itself.

Battery Monitoring Needs a Different Set of Questions

With storage, generation alone is not enough. We look at when the battery charges, where that energy came from, how deeply it cycles, whether it is hitting charge or discharge power limits, how much reserve is held and whether the tariff schedule matches the customer's objective.

A battery that sits full before midday may have insufficient usable capacity for the available solar surplus. A battery that never fills may be too large for the array or may be intentionally keeping headroom for cheap grid charging. A battery that imports during a high household load may be limited by kW output even if it still contains plenty of kWh.

Self-Consumption Is a Financial Metric, Not a Quality Score

A high self-consumption percentage can be positive because more solar is used on site, but it is not always the goal to maximise that percentage. A deliberately larger array may export more while still delivering greater total savings and generation.

Likewise, a household with a strong export tariff may choose different battery behaviour from one focused entirely on avoiding imports. We measure success against the intended financial and energy strategy, not one universal dashboard percentage.

Seasonal Shading Can Change Without a System Fault

Trees grow, neighbouring vegetation changes and the sun's path is lower in winter. New buildings, roof alterations or even a moved aerial can create new shadows. A repeated seasonal reduction should therefore be checked against the physical site as well as the inverter.

If performance has changed after years of stable output, we compare monitoring curves with current roof conditions before assuming module degradation.

When Monitoring Should Trigger a Service Visit

Some warnings can be resolved through communications or a known grid event. Others need physical inspection or testing. We recommend investigation where there is repeated inverter shutdown, recurring isolation or ground-fault warnings, unexplained loss of a string or MPPT, sustained generation reduction, unusual battery behaviour, heat or damage around equipment, or missing production that cannot be explained by weather.

Our solar servicing and fault diagnosis uses monitoring history as one part of the diagnostic evidence. The exact fault code and timestamp can save significant time on site.

Commercial Monitoring Needs Clear Responsibility

On a commercial site, the question is not simply whether monitoring exists; it is who is expected to act on it. An alarm that sits unread for six weeks has little value. We prefer a clear process for who receives alerts, how performance is reviewed and when an issue is escalated.

For commercial solar systems, it can also be useful to compare multiple inverters, roof zones or buildings. Normalised yield helps identify an underperforming section even when the largest array still has the highest raw kWh total.

What MCS Handover and Maintenance Mean for Monitoring

MCS MIS 3002 includes commissioning, handover and maintenance requirements. Energy Saving Trust likewise recommends that installers explain the main inverter fault signals and troubleshooting information. We see monitoring education as part of that handover: the customer should know what normal looks like and what information to save when something goes wrong.

A Simple Monthly Solar Monitoring Routine

Our View: Monitoring Should Make the System Easier to Own

A good dashboard does not need the owner to become a solar engineer. It should make normal generation understandable, reveal faults early and provide enough information for a service engineer to start diagnosis quickly.

We use the system's design, monitoring history and real site conditions together. That is more reliable than judging success from a single high-generation screenshot or comparing one roof with a neighbour's different system.