Викторина «ИИ в энергетике»

Ответы и пояснения появляются по мере прохождения теста.

1. Why evaluate an energy model under outages and weather extremes?
   • Rare conditions can affect whether predictions or actions remain safe and useful
   • They always appear equally in training data
   • Extreme cases are irrelevant to grid operations
   • A model can only use historical averages
2. A storage controller lowers simulated cost but violates a reserve requirement. What is the issue?
   • The simulation proves the controller is safe
   • The optimization objective missed a real operating constraint
   • Reserve requirements apply only to forecasting
   • Lower cost always means better grid performance
3. Which operating decision needs a different horizon than day-ahead forecasting?
   • A yearly equipment inventory
   • A static document search
   • Real-time balancing
   • A one-time model card review
4. How should AI’s energy effect be reported?
   • Count only the energy savings
   • Assume model serving uses no energy
   • Compare different periods without stating them
   • Measure the computation’s electricity use and any downstream savings at a defined boundary
5. What should an operator do when a forecast is uncertain or outside trained conditions?
   • Use the defined escalation or operating procedure rather than treating the forecast as certain
   • Follow it because optimization is automatic
   • Disable every safety rule
   • Assume the input is typical
6. Why version weather data and operating policies with the model?
   • They improve numerical precision automatically
   • They affect inputs and decision behavior and help reconstruct an operational result
   • They replace monitoring
   • They have no influence after training
7. A demand model performs well on average but misses a rare outage scenario. What does the average conceal?
   • A guaranteed reduction in electricity use
   • That the model has no input drift
   • An important low-frequency failure condition
   • The operating boundary is irrelevant
8. Which comparison can support a claim that an AI optimization saves energy?
   • Compare the model’s training time with a utility bill
   • Report only the best day
   • Use a simulation without stating assumptions
   • Measure computation and operating effects over the same boundary and period