Abstract:Infrastructure monitoring is critical for safe operations and sustainability. Water distribution networks (WDNs) are large-scale networked critical systems with complex cascade dynamics which are difficult to predict. Ubiquitous monitoring is expensive and a key challenge is to infer the contaminant dynamics from partial sparse monitoring data. Existing approaches use multi-objective optimisation to find the minimum set of essential monitoring points, but lack performance guarantees and a theoretical framework.
Here, we first develop Graph Fourier Transform (GFT) operators to compress networked contamination spreading dynamics to identify the essential principle data collection points with inference performance guarantees. We then build autoencoder (AE) inspired neural networks (NN) to generalize the GFT sampling process and under-sample further from the initial sampling set, allowing a very small set of data points to largely reconstruct the contamination dynamics over real and artificial WDNs. Various sources of the contamination are tested and we obtain high accuracy reconstruction using around 5-10% of the sample set. This general approach of compression and under-sampled recovery via neural networks can be applied to a wide range of networked infrastructures to enable digital twins.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is how to reconstruct the contaminant propagation dynamics in Water Distribution Networks (WDNs) through sparsely sampled data. Specifically, the research objective is to accurately predict and monitor the pollution diffusion in the water network using limited sensor data without relying on large - scale monitoring devices. The following are the core problems of this research and their solutions:
### Research Background and Problem Description
1. **Importance of Infrastructure Monitoring**
- Water distribution networks are critical infrastructures, and their safe operation is of utmost importance.
- Contaminants in water may come from natural disasters, industrial crimes or terrorist attacks, etc., all of which pose a serious threat to the safety of the water supply system.
2. **Limitations of Existing Methods**
- **High Cost of Comprehensive Monitoring**: Installing sensors at every node is not only expensive but also difficult to maintain, especially in underground pipeline systems.
- **Lack of Theoretical Guarantee in Multi - objective Optimization Methods**: Although existing optimization methods can find the smallest set of monitoring points, they cannot provide performance guarantees and lack theoretical framework support.
### Methods Proposed in the Paper
To solve the above problems, the paper proposes the following innovative methods:
1. **Graph Fourier Transform (GFT)**
- Use GFT to compress the pollution propagation dynamics in the network, identify key data collection points, and ensure reconstruction performance.
- GFT can transform the dynamic signals in complex networks to the frequency domain, thereby finding the most important nodes for monitoring.
2. **Autoencoder - inspired Neural Networks (AE - NN)**
- Build an AE - based neural network to further under - sample from the initial sampling set, so that the pollution dynamics of the entire network can be roughly reconstructed with only a small number of data points.
- This method can achieve high - precision reconstruction of pollution dynamics in actual and simulated water supply networks, using only about 5 - 10% of the sample set.
3. **Application of the Integrated Framework**
- Propose a general compression and under - sampling recovery method, which is applicable to various network infrastructures with complex flow dynamics, such as water supply networks, power networks, etc., to realize the application of digital twin technology.
### Main Contributions
1. **Reconstruction of Pollution Dynamics under Sparse Data**
- Use GFT to determine the minimum number of sampling nodes, and reconstruct the pollution dynamics through inverse GFT or neural networks.
- Introduce neural networks to handle more sparsely sampled data, thereby further reducing the number of required monitoring nodes.
2. **Combination of Theory and Practice**
- Provide theoretical performance guarantees and verify the effectiveness of the method through experiments.
- Tests were carried out under different pollution sources (known and unknown), and high reconstruction accuracy was achieved in all cases.
### Conclusion
The method proposed in the paper not only solves the limitations of existing methods but also provides an efficient and economical solution for the safety monitoring of water supply networks. By combining GFT and neural networks, researchers can still maintain high pollution - dynamic reconstruction accuracy while significantly reducing the number of sensors, which is of great significance for improving the safety and response speed of water supply systems.