Integrated Diagnostic Framework for Process and Sensor Faults in Chemical Industry.

Area of Science:

• Chemical Engineering
• Process Control
• Fault Detection and Diagnosis

Background:

• Distinguishing between process and sensor faults is critical for safe and efficient operation of nonlinear chemical processes.
• Existing methods may face challenges in accurately and rapidly identifying fault types.

Purpose of the Study:

• To propose an integrated fault diagnosis framework capable of differentiating process faults from sensor faults in nonlinear chemical systems.
• To enhance the speed and accuracy of fault detection.

Main Methods:

• Embedding a cycle temporal algorithm within dynamic kernel principal component analysis (DKPCA).
• Utilizing a reconstruction-based contribution graph for diagnosing fault variables.
• Combining DKPCA with contribution analysis to differentiate fault types based on their characteristics.

Main Results:

• The integrated framework demonstrated improved fault detection speed and accuracy.
• Successful application and validation on the Tennessee Eastman process and acid gas absorption process.
• Effective distinction between process and sensor faults was achieved.

Conclusions:

• The proposed integrated fault diagnosis framework is effective for distinguishing process and sensor faults in nonlinear chemical processes.
• The combination of cycle temporal algorithm, DKPCA, and contribution graphs offers a robust solution for fault diagnosis.