Use this URL to cite or link to this record in EThOS:
Title: Design and implementation of fractional order PID controllers for process control systems
Author: Edet, Emmanuel
ISNI:       0000 0004 7960 8878
Awarding Body: University of Strathclyde
Current Institution: University of Strathclyde
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
New methods of designing Fractional Order PID (FOPID) controllers are developed for regulating product's purity in distillation columns. It is widely known that poorly tuned PID controllers lead to bad quality of products and accompanied reduction in profit margin in the process industry. In distillation columns, tight composition control of products with 98% purity level is not achievable with simple pressure controllers only due to sensitivity to disturbances. Therefore, FOPID controllers are proposed to solve these multivariable impurity problems. FOPID controllers have extra tuning parameters that can counteract effects of time delays in distillation columns if properly tuned. This property is exploited as a tool for improved performance. Original contributions of this thesis include the development of three new design methods for multivariable FOPID controllers and results are analysed using inverse maximum singular value of relevant sensitivity functions to assess robust stability. Several conventional PID controller design methods are also reviewed for the purpose of comparison. Thereafter, a decentralised FOPID control system is developed for multivariable systems based on plant's critical frequency point and results show improved performance over conventional PID controllers. The proposed critical frequency point method provides very easy-to-use tuning rules similar to Cohen-Coontuning rule for integer-order PID controllers. In addition, a new decentralised multivariable FOPID controller is also proposed based on Internal Model Control(IMC) method but settings are tuned using Biggest Log-magnitude Technique (BLT). This IMC-FOPID control design scheme overcomes the need for critical frequency point experiments. Another contribution of this thesis is the development of a novel discrete Fractional Order Predictive PI (FOPPI) control design scheme suitable for linear multivariable systems. Comparative study of these methods is presented. Simulation results prove that both top and bottoms products' purity of 98% are achievable with improved disturbance rejection when using the proposed FOPPI controller. In comparison, simpler FOPID control design schemes developed in continuous time are found to meet design objectives but at the expense of having a more conservative control action.
Supervisor: Katebi, Reza Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral