Energy-Efficient Heterogeneous Wireless Sensor Deployment with Multiple Objectives for Structural Health Monitoring

Chengyin Liu, Zhaoshuo Jiang, Fei Wang, Hui Chen2016 · DOI

TL;DR

This paper proposes a novel framework for optimally deploying heterogeneous wireless sensors (accelerometers and strain gauges) for structural health monitoring (SHM). It tackles the challenge of balancing modal information quality with energy efficiency by developing a multi-objective optimization function and using a Particle Swarm Optimization (PSO) algorithm. The framework also incorporates a clustering approach to further reduce energy consumption in the network.

What Problem Does This Paper Address?

Structural health monitoring (SHM) systems are crucial for assessing the integrity of civil structures. Traditional wired sensors are expensive and difficult to maintain. Wireless Sensor Networks (WSNs) offer a more scalable and cost-effective solution, but deploying them for SHM presents unique challenges, especially with heterogeneous sensor types (e.g., accelerometers and strain gauges). These challenges include:

  • Heterogeneity: Different sensor types produce data that may not be easily combined using traditional methods that rely on sensor data orthogonality.
  • Energy Constraints: Wireless sensor nodes are typically battery-powered, making energy efficiency a critical design consideration.
  • Optimal Deployment: Determining the ideal number and placement of sensors to maximize the quality of modal information (essential for diagnosing structural health) while minimizing energy consumption is a complex optimization problem.
  • Network Topology: Efficiently routing data from numerous sensors to a central point without draining batteries requires careful consideration of network topology.

This paper aims to address these issues by proposing an integrated framework for optimal heterogeneous sensor deployment that considers both modal information quality and energy efficiency, along with an effective network topology for energy conservation.

Methods

The paper introduces a two-pronged approach:

  1. Optimal Sensor Placement (OSD) Framework:
    • Placement Quality Metric: To address the challenge of heterogeneous data, the authors propose a placement quality index that combines the Modal Clarity Index (MCI) and Modal Relative Error (MRE). These metrics do not rely on the orthogonality of modal vectors, making them suitable for mixed sensor data.
  2. Energy Consumption Model: A detailed energy consumption model for both accelerometer and strain gauge nodes is developed, considering data acquisition, processing, transmission, and reception. This model is integrated into a multi-layer spherical network model.
  3. Multi-Objective Optimization Function: A combined objective function is formulated to simultaneously maximize modal identification accuracy (using the placement quality index) and minimize network energy consumption. This function is weighted to allow for trade-offs between these two objectives.
  4. Particle Swarm Optimization (PSO): The PSO algorithm is employed to efficiently solve the computationally challenging multi-objective optimization problem, especially for systems with a large number of degrees of freedom (DOFs). PSO is chosen for its fast convergence and ability to handle complex optimization problems.
  5. Network Topology Optimization (Clustering):
    • Two-Phase Energy Consumption: The energy consumption for a cluster-tree network is modeled in two phases: network initialization and stable data transmission.
  6. Optimal Clustering: The paper derives an expression for the optimal number of cluster heads (CHs) to minimize total energy consumption. This involves calculating the expected distance between CHs and member nodes and then optimizing the total energy consumption with respect to the number of CHs.

Datasets

Not discussed in the paper. The study uses a numerical simulation based on a real-world bridge.

Key Insights and Ideas

  • Heterogeneous Data Handling: The use of MCI and MRE for placement quality is a crucial step in enabling effective OSD for heterogeneous WSNs, overcoming limitations of traditional methods.
  • Integrated Optimization: Combining modal information quality and energy consumption into a single multi-objective function provides a holistic approach to sensor network design for SHM.
  • PSO for Complex Problems: PSO proves to be an effective algorithm for solving the complex, multi-objective sensor placement problem, offering a good balance between performance and computational cost.
  • Clustering for Energy Savings: The cluster-tree topology significantly reduces energy consumption compared to a flat topology by minimizing data transmission distances, even if it increases the coverage radius. This highlights the importance of network topology in energy-efficient WSN design.
  • Trade-offs are Inherent: The study explicitly acknowledges and quantifies the trade-offs between modal identification accuracy and energy consumption, providing a framework to find an optimal balance.

Key Results

  • The proposed OSD framework, using PSO, successfully determined an optimal sensor layout for a four-span bridge.
  • The optimal configuration involved 14 wireless accelerometers and 10 wireless strain gauges, with a coverage range of approximately 96.7 m.
  • The sensor placement was found to be non-symmetric in the longitudinal direction, indicating that optimal placement is structure-specific.
  • The cluster-tree topology demonstrated a significant energy saving of 33.3% compared to a flat topology.
  • The cluster-tree topology also increased the coverage radius by 28.3%, suggesting that energy savings are achieved through more efficient data routing rather than simply reducing the network's reach.

Limitations and Future Work

  • Simplified Network Model: The paper uses a spherical network model and assumes ideal communication without errors or delays. Real-world WSNs often experience more complex communication patterns and potential failures.
  • Member Node Selection: While the paper determines the optimal number of clusters and CHs, the selection of specific member nodes within each cluster for optimal data transmission is noted as a complex problem that is "out of the scope of this paper."
  • Computational Cost: The paper mentions that further optimization of data packet transmission paths could lead to even greater energy savings but would require more computation time.
  • Future Work: The authors suggest that further determination of specific member nodes within clusters and more sophisticated optimization of data packet transmission paths could lead to further improvements in energy efficiency. The framework is also applicable to other monitoring scenarios using heterogeneous sensors.

Summary updated

BibTeX
@misc{liu2016energy,
  author = {Chengyin Liu and Zhaoshuo Jiang and Fei Wang and Hui Chen},
  title = {{Energy-Efficient Heterogeneous Wireless Sensor Deployment with Multiple Objectives for Structural Health Monitoring}},
  year = {2016},
  doi = {10.3390/s16111865},
}
Figure