Individual/collective blade pitch control of floating wind turbine based on adaptive second order sliding mode

Highlights

• Design & application of adaptive super-twisting algorithm for floating wind turbine.

• Collective blade pitch control and individual blade pitch control are combined.

• Very reduced knowledge of system model is necessary.

• Comparison with standard controller GSPI.

Abstract

A new control strategy based on adaptive second order sliding mode approach is applied to a floating wind turbine system in the above rated region. This adaptive controller is well adapted to a highly nonlinear system as floating wind turbine, and can be easily implemented with very reduced knowledge of modeling. The proposed controller partially based on multi-blade coordinates transformation combines collective and individual collective blade pitch control, for power regulation, platform pitch motion reduction and reduction of blades fatigue load. The proposed controller is implemented on FAST simulator and shows high level of performances.

Introduction

Floating wind turbines (FWTs) allow the use of the huge wind resource in ocean area and are considered as a promising solution of renewable energy. However, some issues arise: first-of-all, unlike the onshore wind turbine, the floating platform introduces additional degrees of freedoms (DOFs) that have negative impacts, especially on the platform pitch motion. They also induce the issue of negative damping (Nielsen et al., 2007) that leads to system instability and degrades the power production. Furthermore, with the increasing capacity and flexibility of wind turbines, fatigue loads of the structure became more and more important especially for floating systems and affect the service life. Therefore, reducing the fatigue loads is a key-point (Bossanyi, 2003, Menezes et al., 2018) for large scale wind turbines. The control strategy must provide an efficient solution for such problems and appears crucial for wind turbine systems.

The main control objectives of FWT in the above rated region consist in maintaining the power output at rated value meanwhile avoiding the negative damping, i.e. reducing the platform pitch motion (Jonkman et al., 2009). Many works have been made based on the collective blade pitch (CBP) control strategy; in this case, all the blades of the wind turbine are controlled by a similar way. Among existing results, in (Jonkman, 2008), the famous baseline gain scheduling proportional integral (GSPI) control is proposed: platform pitch motion is successfully reduced but with a large power fluctuation. In (Wakui et al., 2017), a novel gain scheduling control strategy is developed improving the power regulation while keeping the same platform pitch motion as GSPI. Linear quadratic regulator control, model predictive control and feed-forward control (Namik et al., 2008, Schlipf et al., 2012, Schlipf et al., 2015) have been also applied to FWT systems. However, fatigue load reduction is not considered in those works. In terms of the fatigue load reduction, individual blade pitch (IBP) control (Bossanyi, 2003, Lio et al., 2018, Ossmann et al., 2017, Selvam et al., 2009) has been introduced: in this case, the blades are independently controlled. This approach has also been extended to floating ones (Cunha et al., 2014, Suemoto et al., 2017, Namik and Stol, 2014).

Nonetheless, in all these works, the control design is based on linearized models around an equilibrium point obtained by the FAST (Fatigue, Aerodynamics, Structures and Turbulence) software (Jonkman et al., 2005). These approaches based on linearization around an operating point, are in opposition with the use of FWTs in a large operating domain. Hence, in order to avoid this drawback, different models for different equilibrium points are required that induces design of different controllers producing a large effort of parameter tuning.

An other solution could be the use of nonlinear control strategies based on nonlinear models. However, given that the nonlinear models of FWT (Homer and Nagamune, 2018, Jonkman et al., 2005, Sandner et al., 2012) are not well adapted to the control design, there are few studies on the nonlinear control of FWT. As detailed just below, a solution consists in designing nonlinear control laws that are efficient on a large operating domain, but without precise models in order to reduce as much as possible the modeling effort. Sliding mode control (SMC) (Utkin, 1977) is a well-adapted solution, given its robustness versus uncertainties and perturbations.

Then, in the current work, SMC is applied for the floating wind turbines control. SMC requires very limited knowledge of system model (especially in its adaptive version) while keeping robust versus uncertainties and perturbations. Such control algorithms ensure that the sliding variable (that is defined from the control objectives) converges to a vicinity of the origin in a finite time. However, due to the discontinuous feature of “standard” SMC control, chattering phenomenon (i.e. high frequency oscillations of control input) appears. In order to reduce the chattering effect, super-twisting (STW) (Shtessel et al., 2014) control combined with gain adaptation (Plestan et al., 2010, Shtessel et al., 2012) is applied in this study. The STW is one of most famous second order sliding mode control that generates continuous control input and thereby, reduce the chattering. Furthermore, STW only requires the knowledge of sliding variable; then, it can be viewed as an output feedback control and is very simple to implement. Furthermore, adaptive gain allows to keep control accuracy versus perturbations and uncertainties, even in the case that the information of system model is very reduced. In fact, only the relative degree (Isidori, 1999) of the sliding variable is required. Hence, such algorithm is very well-adapted to the FWT control problem. Notice that, in authors’ previous works (Zhang et al., 2019a, Zhang et al., 2019b, Zhang and Plestan, 2020), super twisting control with gain adaptation based on CBP control technology has been successfully applied to FWT system. Here, the first novelty is based on the fact that IBP and CBP control structures are combined to control the power and to reduce the fatigue load of the wind turbine, especially the blade load reduction (among the structure loads, the blade root reduction is the most important, being the source of the loads for the rest of the structures (Jelavić et al., 2010)). An other novelty is the use of an adaptive second order sliding mode controller in the frame of IBP/CBP control strategy.

Section 2 introduces the model of the FWT. Section 3 states the control problem. Section 4 describes the STW control laws with adaptation laws based on both CBP and IBP control approaches, and their application to the FWT. Section 5 displays the results obtained by FAST/Simulink co-simulations, and their analysis.