ECoalVis: Visual Analysis of Control Strategies in Coal-fired Power Plants

Lastest Update: 25th April 2024

Introduction

a novel interactive system visually analyze the control strategies of coal-fired power plants extracted from historical sensor data.

Motivation

• Environment
• Risky
• Limited datasets and strong assumptions
• Difficult to capture the potential spread of the impact of control strategies

Challenges

• Extraction of control strategies
• Cascading impact
• Time-lag-aware analysis

Requirement and design

• [R1] Extract the impact of control strategies with time series queries (forward)
  • Discretizing time series
  • Aligning sensor events
  • Matching partial strategy
• [R2] Identify responsible control strategies for anomalies in important sensors (backward)
  • Time-lag-aware breadth-first search
• [R3] Explore the spatial propagation of control strategy impact
  • relationship-oriented/Context-oriented
• [R4] Obtain the temporal cascading of control strategy impact
  • topology-temporal-separated
• [R5] Inspect the details of the sensor time series

Critical thinking

• Multi-dimensional visualization
• Link
• Extend to other domains with strong temporal correlation