Gas Network Leak Detection Using Sensor-Based Physical Models

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods are inadequate for detecting and quantifying leaks in complex gas networks between sources and consumers, as they are designed for long, straight pipelines and do not account for the complexities of compressor plants and consumer areas.

Innovation Solution

A method involving sensors to determine physical parameters, a training phase to establish a physical model using estimation algorithms, and an operational phase to predict leaks by comparing sensor measurements, allowing for quick detection and quantification of leaks in the gas network.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If methods designed for long, straight pipelines are used, then monitoring is simpler, but they are not suitable for complex gas networks with compressor plants and consumer areas

Engineering Contradiction:
Improveease of implementing monitoring methodVSAvoidapplicability to complex gas networks
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The complex gas network is divided into multiple segments or zones, each monitored by dedicated sensors. The network topology is segmented into manageable units that can be individually analyzed, allowing the monitoring system to handle complexity while maintaining ease of implementation through modular approaches.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The monitoring method transitions from static assumptions (constant flow, simple topology) to dynamic modeling that adapts to changing network conditions. The system continuously updates its understanding of network state, accommodating compressor operations, consumer variations, and topology changes, making it versatile for complex networks while maintaining computational tractability.

Inventive Principle:
Principle #15Dynamics

2Reliability

If physical models are trained during start-up phase, then the model adapts to current network conditions, but this requires interrupting the operational phase

Engineering Contradiction:
Improveaccuracy of leak detectionVSAvoiddowntime for retraining model
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The physical model is retrained periodically at scheduled intervals rather than continuously or on-demand. This periodic retraining maintains detection accuracy by updating the model with current network conditions while minimizing operational disruption, as retraining occurs only at predetermined times when interruption is most acceptable.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The system prepares for model retraining by scheduling it during planned maintenance windows or low-demand periods. By anticipating the need for model updates and planning ahead, the system minimizes unexpected operational interruptions while ensuring the model remains accurate for reliable leak detection.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If multiple sensors are deployed at different locations, then leak detection coverage is improved, but system complexity increases

Engineering Contradiction:
Improveleak detection accuracyVSAvoidnumber of sensors and data processing
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

Each sensor in the network is designed to perform multiple functions: measuring pressure, temperature, and flow conditions simultaneously. This multi-functionality allows comprehensive leak detection coverage with fewer sensor units, reducing overall system complexity while maintaining high measurement precision through the versatile capabilities of each sensor.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

Multiple sensor readings and data streams are merged and processed collectively to detect leaks. By combining information from various sensors and analyzing them together through integrated algorithms, the system achieves high detection accuracy while managing complexity through unified data processing rather than separate analysis for each sensor.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3903018B1Method for detecting leaks in a gas network under pressure or under vacuum and gas network
Publication Date: 2022.12.07 ATLAS COPCO AIRPOWER NV
  • EP3903018B1 patent drawingFigure 1
  • EP3903018B1 patent drawingFigure 2

AI summary

Method for detecting leaks (12) in a gas network comprising; - sources (6); - consumers (7); - sensors (9a, 9b, 9d); characterised in that the method comprises the following steps: - a training or estimation phase (15) determining a physical model between measurements of a first and a second set of sensors (9a, 9b, 9c, 9d); - an operational phase (18) where the established physical model between the measurements of the first and the second set of sensors (9a, 9b, 9c, 9d) is used to predict leaks (12) in the gas network (1); wherein the operational phase (18) comprises the following steps: ™ calculating the value of a second group of sensors (9a, 9b, 9c, 9d) from the readings from the first group of sensors (9a, 9b, 9c, 9d) using the physical,model; - determining the difference between the calculated values with the read values of the second group of sensors (9a, 9b, 9c, 9d); - determining, on the basis of a residual value analysis, whether there is a leak (12) in the gas network (1).