Sliding mode control of drum water level in an industrial boiler unit with time varying parameters: A comparison with H∞-robust control approach
Highlights
► Implementation of two control strategies for desired performance of drum water level. ► Robust sliding mode and H∞ controllers are designed for time varying dynamic system. ► Results are compared for regulation and various desired commands of drum water level. ► Sliding mode control (SMC) leads to more smooth, rapid and robust time responses. ► SMC leads to less oscillatory behavior of control efforts and less energy consumption.
Introduction
Due to existence of high interaction between dynamic variables, uncertainties and load disturbances, boiler unit is one of the complex components of steam power plants. Moreover, boiler unit plays a critical role because the steam mass rate, its temperature and pressure affect the performance of power plant. Although the steam production is varied during plant operation, output variables must be maintained at their respected values. Therefore, regulation of water level of drum and tracking the load variation commands of drum pressure and power output are expected from a boiler–turbine system. However, the physical constraints exerted on the actuators must be satisfied by the control signals [1], [2].
Several dynamic models of the boiler system have been developed. In the early works, dynamic modelling of a boiler–turbine unit based on data logs, parameter estimation [3], [4], system identification [5], and simplification of nonlinear models [6] have been done. Also, several simulation packages such as SYNSIM for steam plants [7], and simulation of large boilers with natural recirculation [8] have been carried out. Using basic conservation rules, a model for water level dynamics in natural circulation of drum-type boilers has been developed [9]. Using physical principles and neural networks, dynamic nonlinear modelling of power plant has been investigated [10], [11]. In addition, other various nonlinear models of boiler–turbine units have been presented [12], [13], [14].
Effective control systems must be developed to achieve the appropriate performance of the boiler–turbine unit for large changes of the operating conditions. Various control methods have been used for boiler or boiler–turbine controller design. In the early works, linear optimal regulators for performance control of boiler–turbine units have been designed [15], [16]. Also, multivariable long-range predictive control based on local model networks [17], fuzzy based control systems for thermal power plants [18] and neuro-fuzzy network modelling and PI control of a steam–boiler system [19] have been presented. A loop-by-loop approach for water circulation control in once through boiler start up [20] and life extending control of boiler–turbine units by model predictive methods [21] have been investigated.
In other works, by constituting a linear parameter varying model of nonlinear boiler–turbine unit, gain scheduled optimal control and approximate dynamic feedback linearization have been applied [1], [2]. For robust performance of boiler–turbine units, locally robust intelligent supervisory system [22], backstepping-based nonlinear adaptive control [23], sliding mode and H∞ robust controllers [24], [25], [26] have been designed. Also, the optimum robust minimum-order observer has been designed for robust state estimation of the process and improving its performance [27]. However, limited number of works has been devoted to focus on the control of drum water level. For instance, designing the controller based on adaptive Grey predictor algorithm [28], cascade generalized predictive control strategy [29] and gain scheduled optimal control for adjusting the drum level have been investigated [1].
Stable control of drum water level is of great importance for economic operation of power plant steam generator systems. Sometimes, poor control of drum water level causes the emergency shutdowns in power plants. Also, to prevent over-heating of drum components or flooding of steam lines the amount of water in the steam drum must be kept constant. False water level, parameter mismatch due to variant working conditions and model uncertainties are some problems during boiler unit operation. Under such realistic conditions, designing a robust controller which overcomes these uncertain disturbances affecting water level of drum, has not been investigated in the previous works.
In this paper, sliding mode and H∞ robust control strategies are implemented to achieve desired tracking of drum water level under transient conditions; in the presence of model uncertainties. Two transfer functions with time varying parameters between drum water level, feed-water and steam mass rates are considered. The H∞ optimal robust controller is designed based on μ-synthesis with DK-iteration algorithm. μ-Analysis is carried out to investigate the nominal performance, robust stability/performance of the uncertain model. Results are compared for different desired commands of drum water level, including a sequence of steps ramps, and a combination of them (for both control strategies).
Section snippets
Process description and dynamic model of the system
As shown in Fig. 1 [26], [27], in a water-tube boiler, preheated water is fed into the steam drum and flows through the down-comers into the mud drum. Passing through the risers, water is heated and changed to saturation condition. This saturated mixer of steam and water enters the steam drum. There, steam is separated from water and flows into the primary and secondary super-heaters. Then, steam is more heated and is fed into the header. There is a spray attemperator between two super-heaters
Design of sliding mode control strategy
A simple approach to robust control is the sliding mode methodology in which the nth-order problem is replaced by an equivalent first-order system. Consider the nonlinear dynamic system with a single-input [31]:where the scalar y is the output (in this case, drum water level: ), is the control input and is the state vector. and are nonlinear functions of time and states (for simplicity, state vector is denoted by ). These functions
Design of H∞-robust control strategy
A control system is robust if it is insensitive to differences between the actual system and the model of the system which was used to design the controller. These differences are referred to as model uncertainty. The H∞ robust control technique is used to check if the design specifications are satisfied even for the worst-case uncertainty.
Simulation of the control design, results and discussion
In this section, to investigate the effect of proposed control approaches, real dynamic model of the power plant Qingdao in China; presented by Nanhua et al. [28] is considered as the case study. For dynamic system of Eq. (1), nominal values of the realistic parameters are given in Table 1 while real time varying characteristics of the parameters and are shown in Fig. 5. Time response of the open-loop system to different initial disturbances is shown in Fig. 6 (if the system
Conclusions
Stable control of drum water level is of great importance for economic operation of power plant steam generator systems. Sometimes, poor control of drum water level causes the emergency shutdowns in power plants. Also, to prevent over-heating of drum components or flooding of steam lines, the amount of water in steam drum must be kept constant. In this paper, robust sliding mode and control strategies are implemented to achieve desired regulation and tracking of drum water level, in the
References (32)
- et al.
Gain-scheduled control for boiler–turbine dynamics with actuator saturation
International Journal of Process Control
(2004) - et al.
Dynamic simulation of large boilers with natural recirculation
Journal of Computers & Chemical Engineering
(1999) - et al.
A model on water level dynamics in natural circulation drum-type boilers
International Communications in Heat and Mass Transfer
(2005) - et al.
Dynamic nonlinear modelling of power plant by physical principles and neural networks
Journal of Electrical Power & Energy Systems
(2000) - et al.
Neural modelling of steam boilers
Journal of Energy Conversion and Management
(2007) - et al.
Linear control of a boiler–turbine unit: analysis and design
ISA Transactions
(2008) - et al.
Drum-boiler dynamics
Journal of Automatica
(2000) - et al.
A local model networks based multivariable long-range predictive control strategy for thermal power plants
Journal of Automatica
(1998) - et al.
A fuzzy logic controller application for thermal power plants
Journal of Energy Conversion & Management
(2006) - et al.
Neurofuzzy network modelling and control of steam pressure in 300 MW steam–boiler system
Journal of Engineering Applications of Artificial Intelligence
(2003)
Water circulation control during once-through boiler start-up
Journal of Control Engineering Practice
Life extending control of boiler–turbine systems via model predictive methods
Journal of Control Engineering Practice
An intelligent supervisory system for drum type boilers during severe disturbances
Journal of Electrical Power & Energy Systems
Backstepping-based nonlinear adaptive control for coal-fired utility boiler–turbine units
Journal of Applied Energy
On the multivariable robust control of a boiler turbine system
Improving boiler unit performance using an optimum robust minimum-order observer
Journal of Energy Conversion & Management
Cited by (50)
Data-driven adaptive-critic optimal output regulation towards water level control of boiler-turbine systems
2022, Expert Systems with ApplicationsModel-based dynamic sliding mode control and adaptive Kalman filter design for boiler-turbine energy conversion system
2022, Journal of Process ControlA chattering-free sliding mode control strategy for modular high-temperature gas-cooled reactors
2019, Annals of Nuclear EnergyCitation Excerpt :For example, TS fuzzy technique is utilized for the nonlinear modeling of a boiler-turbine system (Zhang et al., 2017) and a hydro-turbine governing system (Wang et al., 2016). SMC has various attractive advantages, such as fast responses and strong robustness (Vadim, 2003; Moradi et al., 2012; Sharifi and Moradi, 2017; Xue et al., 2017; Hua et al., 2015; Ghabraei et al., 2015; Ghabraei et al., 2018; Reza Ansarifar and Saadatzi, 2015a,b; Ansarifar et al., 2015). Being of these advantages, SMC has been successfully applied to process control, such as the tracking control of an influenza epidemic (Sharifi and Moradi, 2017), and the load-tracking control of pressurized water reactors (Reza Ansarifar and Saadatzi, 2015a,b).