Performance evaluation of the fast model predictive control scheme on a CO₂ capture plant through absorption/stripping system

Abstract

The Classical Model Predictive Control (CMPC) has the drawback of slow response in complex dynamic systems. In this work, the Fast Model Predictive Control (FMPC), which accelerates the computation time through the fragmental solution of a complex quadratic program (QP), is investigated as a possible alternative to control the standard CO₂ capture plant using MEA with various step changes. Aspen PLUS® and MATLAB® ^a are utilized to implement the control strategy. The study concluded that the FMPC controller has an average settling time of 51.42 s for all step changes, which is 74.8% faster than the CMPC. The average IAE value for FMPC was approximately 0.1307 which is 59 times smaller than the CMPC controller. Additionally, the ISE and ITSE values demonstrated much improved outcomes for the FMPC controller. The offsets for the FMPC are maintained at negligible levels through suitable tunning since offsets are the main hurdle observed when the FMPC controller is implemented on chemical process systems.

Introduction

Clean, efficient, and economical energy has been a long-standing desire to address the huge challenges posed by equipment safety, climate change, and global warming. Natural gas is a fossil fuel that comprises a significant amount of methane, as well as contaminants such as mercury, water vapours, organic sulphur, inert gases, carbon dioxide (CO₂), and hydrogen sulphide (Alcheikhhamdon and Hoorfar, 2016). Due to the depletion of pure natural gas reserves, the existence of such pollutants presents a challenge. Contaminated natural gas engenders serious risks to human health and equipment safety, including lung and skin illness, corrosion, fouling, and unexpected shutdowns (Sanni et al., 2020). The high CO₂ content of natural gas diminishes the heating value and burning rate of the gas mixture, resulting in decreased engine output (Ayandotun et al., 2012). Thus, efficient control strategies are paramount to ensure the sweet gas stream CO2 compositions are always maintained within the desired values (Lee and Kim, 2020, Park et al., 2021b).

The major and frequently employed CO₂ capture technologies include absorption/stripping (Gutierrez et al., 2018), adsorption (Tao et al., 2019), cryogenic distillation (Babar et al., 2019), membrane separation (Harrigan et al., 2020), chemical looping combustion (CLC) (Khan et al., 2020), biological separation (Matito-Martos et al., 2020), hydrate separation (Kida et al., 2020) etc. Above all, chemical absorption with a monoethanolamine (MEA) solvent is the utmost established and commonly used CO₂ capture technology at the industrial scale due to its high CO₂ capture efficiency, fast reactivity, low cost, abundant solvent, uncomplicated retrofitting, and capability to handle gases having low CO₂ partial pressure (Akinola et al., 2018; Tay et al., 2017). However, the complexity of the capturing system creates significant difficulties in the form of variable interactions, nonlinearities, constraints in the capture plant (Li et al., 2018). As a result, the development of an effective, adaptive, and quick responding control strategy capable of overcoming the inherent challenges in CO₂ capture plants is critical for cost-effective CO₂ capture.

The fundamental objectives of the control strategy are to maintain the CO₂ capture rate at set points without offsets and to prevent the stripper temperature from exceeding the desired value, which might accelerate the degradation of the MEA solvent (Mechleri et al., 2017). Most of the control strategies are based on PID and MPC-based controllers to ensure the controllability and flexibility of the CO₂ capture plant. The PI-based decentralized control scheme has been implemented on an MEA-based CO₂ capture plant integrated with a power plant. In a study by Nittaya et al. (Nittaya et al., 2014a), the controller performance is evaluated by introducing sinusoidal changes in the gas composition and CO₂ capture rate with the control objective of CO₂ capture rate and purity at 87% and 95%, respectively. Another work by Cristea et al. (Cristea et al., 2020) implemented a decentralized PI controller with the control purpose of maintaining CO₂ capture rate and energy performance while rejecting disturbances and tracking set points. Depending on the amount of disturbance/change, the controller acquired the new setpoints in 2–6 h. However, due to the delayed response time, high overshoot, and large offsets, a more effective control technique capable of resolving such issues is required.

The large variety of implementations of MPC-based control schemes on absorption/stripping systems has been noticed in the literature due to the benefits of handling uncertainties, variable interactions, processing with multiple-input multiple-output (MIMO) systems (Cormos et al., 2015, He et al., 2016, Rúa et al., 2021). In MPC, a model is utilized to predict the future behavior of the system. The basic schematic diagram of the MPC is illustrated in Fig. 1 (Tahir et al., 2018).

Cormos et al. (Cormos et al., 2015) implemented MPC control scheme in MATLAB® on a CO₂ capture plant based on an absorption/stripping system integrated with a power plant. Step, ramp, and sinusoidal changes in inputs are introduced to control the CO₂ removal efficiency. The amount of liquid flow is adjusted to maintain the CO₂ removal efficiency to 87%. In comparison to a PI-based control scheme, the MPC controller has expressed better performance in reducing overshoot and controlling the constraints. He et al. (He et al., 2016) also designed an MPC control structure for a CO₂ capture system to evaluate its controllability and flexibility. The performance of the MPC controller is determined under sudden changes for load and setpoint tracking scenarios. The results illustrated that the MPC scheme exhibits minimum variation from the set points (CO₂ removal/product purity) as compared to the PI controller. However, the response of such conventional/classical MPC-based (CMPC) controllers is still on the slow side compared to more advanced MPC-based control strategies. It can be improved by fine-tuning the input/output weights or by using fast-responding control techniques such as the Fast MPC (FMPC) control scheme.

The control algorithm of CMPC is solved in the form of a Quadratic Program (QP). The QP comprises a system matrix with complex dynamics of the CO₂ capture plant. The solution of such a complex QP requires a large settling time, which results in a slow output response especially for complicated systems having large dimensions (Qi et al., 2015). The computational burden of the controller can be reduced through an advanced control strategy termed as FMPC controller, which solves the QP fragmentally into small fractions and accelerates calculation time in control action through online optimization (Blanchard and Adegbege, 2017; Kouzoupis et al., 2015; Mönnigmann and Kastsian, 2011; Quirynen et al., 2015). The FMPC-based approaches have been widely applied to traffic, electrical and electronic systems including management of freeway traffic congestion (Han et al., 2017), energy storage system (Ni et al., 2021), turbine generator system (Hou et al., 2020), multilevel power converters (Ramirez et al., 2020), Jump Markov Systems (Tonne and Stursberg, 2017), solar systems (Pipino et al., 2020) etc. Qi et al. (Qi et al., 2015) established a FMPC control approach on a DC motor to demonstrate the MIMO feasibility of a fast dynamic system with 1-bit signal control. On another work, (Schindele and Aschemann, 2011) a nonlinear FMPC method was developed for the two primary axes of an overhead travelling crane, ensuring required trajectory tracking and active dampening of the crane load and oscillations, respectively.

However, FMPC approach has not been implemented on CO₂ capture plants due to the challenges of identified model’s accuracy and presence of offsets. Since the key variables to be controlled in the system is the amount of CO₂ in the outlet gas stream, which is critical to ensure that it is always at the desired specifications under any disturbances/set-point changes, a fast-acting controller is clearly beneficial. Therefore, in this paper, the implementation of the FMPC on a standard CO₂ capture plant via absorption/stripping system is investigated and gauged against the CMPC. Based on the authors’ knowledge, this is the first study related to the implementation of the FMPC controller on a CO₂ capture plant based on an absorption/stripping system.