In November 2025, Metroscope gathered more than 40 performance experts from 10 nuclear sites at Gravelines CNPE. Across discussions, plant teams shared how issues that once required weeks of investigation are now resolved within days.
The objective was to understand how investigations actually happen on site. Operators explained how they detect performance losses, validate a diagnosis, decide when to act, and assess maintenance benefits. The discussion focused on what it really takes to move from detection to resolution.
From weak signals to costly delays
Across sites, teams described a consistent pattern: most costly faults start as weak and ambiguous signals. Detecting them is necessary, but not sufficient. What creates value is the ability to move from signal to decision in a structured way.
First, the signal must be quantified. A deviation only becomes actionable when its impact is expressed in operational terms, typically in megawatts or economic loss. This allows teams to understand what is at stake and justify the effort required to investigate and resolve it.
> At Penly, a small leak on a reheater bypass valve led to a significant production loss. The deviation in steam flow was detectable, but it did not trigger action until its impact was quantified. Without this step, the issue would likely have remained unaddressed.
Second, the quantified signal must be shared and aligned across teams. Decisions often involve multiple stakeholders with different priorities. Expressing the impact in a common metric enables clearer discussions and faster prioritization.
> At Chooz, an issue on the condenser cleaning system was initially deprioritized. Once the loss was quantified, the team could justify addressing it during the next outage. The result was confirmed after restart, reinforcing the decision process.
Third, teams must trust the diagnostic and act without extended validation cycles. This depends on both the reliability of the tools and prior experience with similar cases.
> At Tihange, issues that previously required weeks of investigation are now resolved within two days. During restart phases, where multiple deviations appear simultaneously, the ability to quickly isolate, assess, and act is critical.
Cross-functional collaboration supports all three steps. At Gravelines, resolving a reheater valve leakage required combining diagnostics, field observations, and structured exchanges between teams. This alignment reduced uncertainty and accelerated the decision.
These cases show that moving from weak signals to action depends on three combined conditions: quantifying impact, aligning stakeholders, and trusting the diagnostic. When these are in place, teams reduce hesitation and shorten investigation time.
A shared foundation for better decisions
Beyond individual cases, these discussions highlighted both what is required to improve decision-making and why dedicated moments of structured experience sharing are essential.
They identified key areas of monitoring improvement useful in Metroscope’s product development:
- Expanding diagnostic coverage to capture a wide range of failure modes.
- Improving data reliability as low-quality or inconsistent data increase uncertainty and slow down decisions.
- Making insights usable beyond specialist teams by simplifying diagnostics interpretation.
These elements directly impact investigation time and the confidence associated with decisions.
Open discussion, where operators compare practices and discuss real situations, helps teams converge on what works. Less experienced teams gain faster access to proven approaches, while more mature teams refine their methods by confronting different perspectives.
Metroscope's Experience Sharing event (MEX 2025) built on this dynamic. By grounding discussions in real cases, it provided participants with concrete ways to improve their practices, while clarifying the conditions required to move effectively from detection to decision. These moments of collective reflection play a direct role in improving how decisions are made on site, by turning individual experience into shared, actionable knowledge, and by continuously informing how monitoring solutions must evolve.
