IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY, VOL. 22, NO. 6, NOVEMBER 2014 2141
Power Optimization for Photovoltaic
Microconverters Using Multivariable
Newton-Based Extremum Seeking
Azad Ghaffari, Miroslav Krsti´ c, Fellow, IEEE , and Sridhar Seshagiri
Abstract—Extremum seeking (ES) is a real-time optimization
technique that has been applied to maximum power point
tracking (MPPT) design for photovoltaic (PV) microconverter
systems, where each PV module is coupled with its own
dc/dc converter. Most of the existing MPPT designs are scalar,
i.e., employ one MPPT loop around each converter, and all
designs, whether scalar or mutivariable, are gradient based.
The convergence rate of gradient-based designs depends on the
Hessian, which in turn is dependent on environmental conditions,
such as irradiance and temperature. Therefore, when applied
to large PV arrays, the variability in environmental conditions
and/or PV module degradation results in nonuniform transients
in the convergence to the maximum power point (MPP). Using a
multivariable gradient-based ES algorithm for the entire system
instead of a scalar one for each PV module, while decreasing the
sensitivity to the Hessian, does not eliminate this dependence. We
present a recently developed Newton-based ES algorithm that
simultaneously employs estimates of the gradient and Hessian in
the peak power tracking. The convergence rate of such a design
to the MPP is independent of the Hessian, with tunable transient
performance that is independent of environmental conditions. We
present simulation as well as the experimental results that show
the effectiveness of the proposed algorithm in comparison with
the existing scalar designs, and also to multivariable gradient-
based ES.
Index Terms— Dc/dc microconverters, maximum power point
tracking (MPPT), Newton-based extremum seeking (ES),
photovoltaic (PV) arrays.
I. I NTRODUCTION
M
AXIMUM power point tracking (MPPT) algorithms
for extracting the maximum achievable power from
a photovoltaic (PV) system have been studied by several
researchers [4], [15], [17]–[19], with detailed comparisons
presented in [5], [11], and [12]. Several recent works [2], [3],
[14], [16] have focused on the application of gradient-based
extremum seeking (ES) [1] to MPPT design.
To the best of our knowledge, there are a limited number
of multivariable MPPT schemes described in the literature,
Manuscript received June 3, 2013; revised October 4, 2013; accepted
January 3, 2014. Manuscript received in final form January 15, 2014. Date
of publication February 17, 2014; date of current version October 15, 2014.
Recommended by Associate Editor M. Guay.
A. Ghaffari and M. Krsti´ c are with the Department of Mechanical and
Aerospace Engineering, University of California, San Diego, La Jolla, CA
92093-0411 USA (e-mail: aghaffari@ucsd.edu; krstic@ucsd.edu).
S. Seshagiri is with the Department of Electrical and Computer Engineering,
San Diego State University, San Diego, CA 92182-1309 USA (e-mail:
seshagir@engineering.sdsu.edu).
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TCST.2014.2301172
among which we refer the reader to [15], [18], and [19].
Reference [19] uses a multivariable version of the popular
Perturb and Observe (P&O) algorithm. Unlike scalar designs
that require one current sensor for each module, the algorithm
only requires a single current sensor on the dc bus. The
operating point of the dc/dc converters is perturbed asynchro-
nously, to minimize the possibility of converter interaction
having a detrimental effect on the other modules. Closely
related to [19] is the work in [18], where extra variables
are employed in the classical P&O algorithm to overcome
the limitation of scalar designs, which the authors say fail
when the feasibility region is nonconvex. It is unclear how [18]
compares with distributed architectures, with respect to power
loss recovery in the case of module mismatch. Reference [15]
uses particle swarm optimization (PSO), which is an algorithm
that employs multiple agents to search for the peak power. This
paper does not describe the specific criteria used to select
the number of agents and parameters of the PSO, nor the
conditions on the voltage and power boundary limits to stop
the algorithm at maximum power point (MPP). In addition,
in a PV system with a higher number of PV modules, the
process of reinitialization and the tracking performance depend
strongly on variable conditions, such as environmental factors,
the nature of the PV modules, and the shading area. The
authors claim that the required number of sensors is reduced
to two, but to compute the pulse duration, the output voltage
of each boost converter needs to be monitored by a separate
sensor. The method also has an adaptation time of the order
of 1–2 s. By contrast, the response time in ES-based designs
is much smaller, of the order of 0.1 s.
Expansion of the conventional P&O MPPT methods to the
case of cascade PV modules with microconverters (one dc/dc
converter for each module), as presented in [19], requires a
step-by-step P&O, namely the core part of the algorithm is
a scalar P&O, which finds the MPP of each PV module at a
time. This results in a longer convergence time. Furthermore,
coupling effects between modules that are exacerbated due
to a partial shading and model mismatch cause nonuniform
transient responses under different environmental conditions.
This problem holds in the PSO method of [15]. Multivariable
ES, unlike the other MPPT approaches, treats the entire
cascade PV system as a whole and it simply fits every PV sys-
tem architecture without any need to redesign the control
loop.
In [9] and [10], we presented a multivariable gradient-based
ES MPPT design for the microconverter architecture, where
1063-6536 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.