Robust converter-fed motor control based on active rejection of multiple disturbances

Highlights

• Robust controller is proposed for highly disturbed converter-fed motor system.

• Design is based on an idea of active disturbance rejection control (ADRC).

• Proposed solution can deal with complex disturbances, including harmonic signals

• Control action is expressed in a concise and practically appealing form.

• Proposed approach is validated with experiments and theoretical results.

Abstract

In this work, an advanced motion controller is proposed for buck converter-fed DC motor systems. The design is based on an idea of active disturbance rejection control (ADRC) with its key component being a custom observer capable of reconstructing various types of disturbances (including complex, harmonic signals). A special formulation of the proposed design allows the control action to be expressed in a concise and practically appealing form reducing its implementation requirements. The obtained experimental results show increased performance of the introduced approach over conventionally used methods in tracking precision and disturbance rejection, while keeping similar level of energy consumption. A stability analysis using theory of singular perturbation further supports the validity of proposed control approach.

Introduction

With the rapid development of power electronics devices, the use of DC-DC buck converters has become an interesting alternative to linear regulators for DC motor control (Sira-Ramirez & Silva-Ortigoza, 2006). The DC-DC buck converter-DC motor combination offers a ”smooth start” of the drive, which mitigates the unwanted effects in the armature circuit. In order to achieve high performance of motion control, this start has to be realized in engineering practice in the inevitable presence of various sources of uncertainties (Chen et al., 2020, Sira-Ramirez et al., 2013, Wang et al., 2017), including (i) unknown load torques; (ii) various parametric uncertainties including those coming from external voltage source, winding resistance, and load resistance; as well as (iii) unmodeled dynamics of the sensing device, equivalence series resistance, and direct current resistance in the converter system. This amount of uncertainty in conventional converter-fed motor systems often goes beyond the capability of a standard industrial off-the-shelf PID controller (Silva Ortigoza et al., 2017). As a result, numerous advanced control techniques have been proposed to date, including differential flatness-based controller (Linares-Flores & Sira-Ramirez, 2004b), neuroadaptive backstepping method (Nizami, Chakravarty, & Mahanta, 2017), generalized proportional–integral control (Linares-Flores & Sira-Ramirez, 2004a), hierarchical cascade-like scheme (Silva-Ortigoza, Hernandez-Guzman, Antonio-Cruz, & Munoz-Carrillo, 2015), nonlinear adaptive controller (Roy, Paul, Sarkar, Pervej, & Tumpa, 2017), and most recently sliding mode control (Rauf, Li, Madonski, & Yang, 2020). Even though these methods successfully address the problem of uncertainties in the buck converter-fed DC motor systems using different tools of modern control theory, they have not been widely adopted in engineering practice. There are several potential reasons for this (depending on the considered approach), for example impractical assumptions about availability of certain signals for controller synthesis, large number of tuning parameters, or the use of high-order plant models, which may lead to relatively complicated control rules.

That is why, the problem of high performance control of uncertain converter-fed motors has been recently investigated from the perspective of active disturbance rejection control (ADRC). This control concept has been introduced to the general public in Han (2009) and its latest developments have been recently summarized in surveys (Chen et al., 2016, Łakomy et al., 2020, Madonski and Herman, 2015, Sariyildiz et al., 2020). The interest in this particular class of techniques came from their recent applications to various motion (Łakomy and Michałek, 2020, Liu et al., 2020, Madonski et al., 2020), power (Lotfi et al., 2016, Wei et al., 2019, Wu et al., 2017), and process (Wu et al., 2019, Zheng et al., 2009, Zheng et al., 2018) control problems as well as successful transition of ADRC from academia to industry through its incorporation in embedded motion control chips from Texas Instruments (called InstaSPIN.¹ ) Through the utilization of profound implications of the integrator chain form (Chen et al., 2020, Gao, 2014) and the concept of real-time total disturbance reconstruction and attenuation, several ADRC schemes have been shown to give promising results in governing converter-driven motor systems, e.g. Linares-Flores et al., 2012, Sira-Ramirez and Oliver-Salazar, 2013 and Yang, Wu, Hu, and Li (2019).

The problem investigated in this work concerns two major limitations of the currently available ADRC methods for converter-fed DC motors. The first one comes from their expression in 2DOF output-based form (in which one degree-of-freedom deals with real-time disturbance reconstruction and rejection and the other for governing the resultant simplified plant model). Although such topology offers important robust and adaptive features, its hardware implementation may be problematic in the cases of motor angular velocity tracking, which nominally use high-order plant models for controller design. In scenarios where a reduced-order observer cannot be used (due to limited plant modeling/sensing capabilities), the conventional ADRC structure requires the availability of multiple consecutive time-derivatives of the output signal and the reference signal in order to synthesize the controller (Madonski, Shao, Zhang, Gao, Yang, and Li, 2019, Michalek, 2016). It is rarely the case in industrial applications of converter-fed drives that analytical forms of all these signals are available or that these signals are directly measurable.

The second limitation investigated in this work is the often use of disturbance observers in ADRC that are based solely on a polynomial representation of the total disturbance. This makes conventional ADRC-based approaches only practically capable of handling slowly varying disturbances (Du et al., 2020). Consequently, the conventional polynomial model limits the abilities to capture the fast varying harmonic disturbances (Sariyildiz et al., 2020), which are common in converter-driven DC motor systems. For example, harmonic currents can cause adverse effects in power systems such as overheating, interferences in sensitive communication equipment, capacitor blowing, motor mechanical vibration, excessive neutral currents, or resonances with the power grid. The harmonic disturbances thus negatively influence the tracking performance, justifying the need for their mitigation through the governing scheme.

Motivated by the above limitations, a new ADRC design is proposed in this work. Its goal is to retain the desired capabilities of standard 2DOF output-based ADRC, while minimizing its disadvantages related to impractical assumptions about signal availability and limited capabilities for attenuating complex multifarious disturbances. The proposed design utilizes a special state transformation and a dedicated observer. The transformation allows the control action to be expressed in a concise practically appealing form reducing its implementation time and requirements, thus making it more easily deployable in various industrial control platforms. The observer is used to virtually decompose the acting lumped disturbance into polynomial-like and sinusoidal-like signals. These two acting disturbance models allow to represent a majority of multifarious disturbances with a satisfactory level of approximation, then simultaneously reconstruct them with a single observer, and finally compensate their effect on the governed output signal in real-time.

In order to verify the efficacy of the proposed design, several experiments are conducted on a laboratory platform to evaluate the performance of the introduced approach. The proposed control technique is compared with some conventional solutions using several criteria like tracking precision, disturbance rejection, and energy consumption. Furthermore, the stability of the proposed control system is proved using singular perturbation theory.