电磁层析成像（EMT）技术详解

Introduction / 引言

In industrial process monitoring, “seeing” the fluid state inside pipelines or vessels has long been a goal pursued by engineers. Multiphase flow—the phenomenon where two or more phases (gas-liquid, liquid-liquid, solid-liquid, etc.) mix and flow—is widespread in oil extraction, metallurgical smelting, chemical production, and other fields. Due to the complexity of multiphase flow, traditional single-point measurement methods (such as pressure sensors, flow meters) struggle to provide global spatial distribution information. The emergence of tomography technology has made it possible to “see through” the inside of pipelines. Among these, Electromagnetic Tomography (EMT), with its unique advantages of non-invasiveness, no radiation, and sensitivity to conductive/magnetic materials, is gradually becoming a powerful tool in the field of multiphase flow monitoring.

I. Basic Principles of EMT / 一、EMT的基本原理

1.1 Electromagnetic Induction and Sensitivity Distribution / 电磁感应与灵敏度分布

The core physical foundation of electromagnetic tomography is Faraday’s Law of Electromagnetic Induction. Unlike CT using X-ray attenuation, ECT (Electrical Capacitance Tomography) using capacitance differences, or ERT (Electrical Resistance Tomography) using resistance differences, EMT obtains internal information through the interaction between alternating magnetic fields and eddy currents within the measured object.

Specifically, the EMT system’s excitation coil applies an alternating magnetic field (typically frequency between 10 kHz and 10 MHz) outside the pipeline. This alternating magnetic field penetrates the pipeline wall and induces eddy currents in conductive or magnetic media within the measured region. These eddy currents generate secondary magnetic fields that superimpose on the original excitation field, causing changes in the spatial magnetic field distribution. By detecting these changes, the distribution of conductivity (σ) and permeability (μ) inside the pipeline can be reconstructed.

1.2 Coil Array Structure / 线圈阵列结构

A typical EMT sensor consists of a coil array arranged around the external pipeline being measured. Common array configurations include:

• 8-coil system: The most basic structure, suitable for laboratory research
• 16-coil system: More common in industrial prototypes, moderate spatial resolution
• 32-coil and above: High-resolution systems for fine imaging

These coils are arranged evenly around the circumference and alternately serve as excitation coils and detection coils during operation. Taking a 16-coil system as an example, typically using an “adjacent excitation” strategy: when coil 1 excites, the remaining 15 coils successively detect; then coil 2 excites, the rest detect… and so on. Such a complete scanning cycle can collect hundreds of independent measurement values, constituting the projection data needed for imaging.

Since the sensitivity distribution of different excitation-detection coil pairs varies—detection regions closer to the excitation coil have higher sensitivity, while regions farther away have lower sensitivity—the final reconstructed image has relatively lower spatial resolution in the pipeline center region compared to edge regions. This is EMT’s inherent “soft-field” characteristic.

1.3 Image Reconstruction / 图像重建

EMT image reconstruction is a typical inverse problem. The forward problem describes “given internal conductivity distribution, find external measurement signals,” while the inverse problem reverses this—“given external measurement signals, find internal conductivity distribution.”

Commonly used reconstruction algorithms include:

• Linear Back Projection (LBP): Fastest computation speed but limited accuracy, often used for real-time preview
• Tikhonov Regularization Method: Adding regularization terms to suppress noise, one of the most commonly used algorithms in industrial applications
• Iterative Algorithms (such as Landweber iteration): Higher accuracy but large computational cost, requiring trade-offs between real-time performance and image quality
• Deep Learning Methods: An emerging direction in recent years, using neural networks to learn the mapping from measurement data to images, demonstrating potential superior to traditional algorithms in simulation environments

II. Comparison of EMT with ECT and ERT / 二、EMT与ECT、ERT的对比

In the electrical tomography “family,” EMT, ECT, and ERT each have their strengths. The table below compares the three:

Characteristic	EMT	ECT (Electrical Capacitance Tomography)	ERT (Electrical Resistance Tomography)
Sensitive Physical Quantity	Conductivity σ, Permeability μ	Permittivity ε	Conductivity σ
Applicable Media	Conductive/magnetic substances	Insulating/low-conductivity substances	Conductive substances
Excitation Method	Alternating magnetic field	Alternating electric field	Alternating current/voltage
Sensor Type	Coil array	Electrode array	Electrode array
Non-invasiveness	★★★★★ (Completely non-contact)	★★★☆☆ (Electrodes need wall contact)	★★☆☆☆ (Current injection required)
For Liquid Metals	Highly sensitive	Not applicable	Not applicable (electrode polarization)
Safety	No radiation, no ionization	Safe	Safe

From the table, EMT’s most outstanding advantage is completely non-contact measurement. ECT and ERT electrodes need direct contact with the measured medium or close fitting to the pipeline inner wall, which is difficult to achieve in high-temperature, high-pressure, or corrosive environments. EMT coils are installed outside the pipeline without contact with the medium, naturally suitable for harsh operating conditions.

Additionally, EMT has natural sensitivity to high-conductivity media—especially liquid metals—which is difficult for ECT and ERT to match.

III. Applications of EMT in Multiphase Flow Detection / 三、EMT在多相流检测中的应用

3.1 Liquid Metal Flow Measurement / 液态金属流量测量

Liquid metals play important roles in nuclear industry (such as lead-bismuth alloy-cooled fast reactors) and metallurgical industry (such as molten steel during continuous casting). However, the high temperature, high conductivity, and strong corrosiveness of liquid metals pose severe challenges to traditional contact measurement methods.

EMT demonstrates unique advantages in this regard. Since liquid metal conductivity is far higher than bubbles, inclusions, or solidified shells, EMT can clearly distinguish gas-liquid interfaces within pipelines, achieving bubble distribution measurement, flow pattern identification, and void fraction measurement. In continuous casting processes, EMT can be used to monitor molten steel flow patterns within the mold, helping optimize parameter design of submerged entry nozzles, thereby reducing slag entrainment and surface defects.

The Leo-group team at the University of Leeds in the UK has done pioneering work in liquid metal EMT. Their developed EMT system has been successfully applied to experimental research on sodium-cooled fast reactors and lead-bismuth alloy loops, achieving visualized monitoring of liquid metal two-phase flow.

3.2 Metallurgical Process Monitoring / 冶金过程监测

In the metallurgical industry, EMT application scenarios are extremely rich:

• Continuous casting process: Monitoring molten steel flow within the mold and shell thickness distribution, providing real-time feedback for speed control and secondary cooling
• Electromagnetic stirring process: Combining EMT’s working principle with electromagnetic stirring (EMS) technology enables online evaluation of stirring effects
• Blast furnace process: Monitoring gas-solid two-phase flow states in the tuyere raceway, optimizing coal injection and blast parameters

3.3 Other Industrial Applications / 其他工业应用

• Hydraulic transport: Concentration and velocity distribution measurement of solid-liquid two-phase flows such as ore slurry and mud slurry
• Chemical processes: Real-time monitoring of multiphase mixing uniformity in reactors
• Biomedicine: Although still in exploratory stages, EMT shows potential application prospects in brain neural activity monitoring (using blood conductivity changes)

IV. Technical Challenges and Development Prospects / 四、技术难点与发展前景

4.1 Main Technical Challenges / 主要技术挑战

Although EMT has broad prospects, it still faces several key challenges in moving toward large-scale industrial application:

(1) Soft-field effect and image blurring. Unlike CT’s “hard-field” characteristics, EMT’s sensitivity distribution is highly coupled with the measured medium’s spatial distribution—the presence of media changes the field distribution, which in turn changes the sensitivity distribution. This makes the inverse problem highly nonlinear, and reconstructed image spatial resolution is limited (typically only 5%~10% of pipeline diameter), making it difficult to resolve fine structures.

(2) Difficulty in 3D expansion. Most current EMT systems use 2D cross-sectional imaging, but actual industrial processes often involve 3D flows. 3D EMT requires significantly increasing coil count and measurement channels, with signal processing and image reconstruction computational complexity rising sharply.

(3) Environmental interference. Electromagnetic interference present in industrial sites (motors, frequency converters, etc.) and eddy current effects from metal structures (pipeline flanges, support frames) seriously affect EMT measurement accuracy. Extracting valid signals in strong interference environments is an urgent engineering challenge.

(4) Calibration and quantification. EMT measurement signals are influenced by multiple factors including medium conductivity, temperature, and sensor geometry. Achieving precise quantitative measurement (rather than only qualitative imaging) requires complex calibration processes.

4.2 Development Prospects / 发展前景

Looking forward, EMT technology is expected to make breakthroughs in the following directions:

• Multimodal fusion: Combining EMT with ECT, ERT, or ultrasound tomography (UTT) to leverage multiple sensitive mechanisms for complementarity, improving imaging quality and quantitative accuracy
• AI-driven reconstruction: Deep learning, physics-informed neural networks (PINN), and other AI methods are expected to significantly improve inverse problem solution speed and accuracy, making high-resolution real-time imaging possible
• Flexible sensors: Wearable/attachable coil arrays based on flexible electronics technology reduce sensor constraints on pipeline geometry
• Chip-based integration: With ASIC and SoC technology development, EMT data acquisition systems are evolving toward miniaturization, low power consumption, and high integration

Conclusion / 结语

Electromagnetic Tomography, with its unique advantages of complete non-contact and high sensitivity to conductive/magnetic substances, demonstrates irreplaceable value in fields such as liquid metal monitoring and metallurgical process control. Although technical challenges such as limited spatial resolution and difficult 3D imaging are difficult to thoroughly solve in the short term, with continued progress in directions including multimodal fusion and AI-driven reconstruction, EMT is moving from laboratory to industrial site. For engineers focused on advanced detection technology, EMT is a frontier direction worth tracking.