Smart Bi-Directional Inverter Control and PV-ESS

Integration for Net Zero Energy Residential Buildings

Sandipan Patra1*, Mohamed Bahloul1, Rohit Trivedi1 and Shafi Khadem1
1
International Energy Research Centre, Tyndall National Institute, Cork, Ireland
*[email protected], [email protected], [email protected] and [email protected]

Keywords: Net Zero Energy Building, Energy Security, Intelligent Bi-directional Inverter, Household Inverter,
Model Predictive Control

Abstract
Climate change, sustainable energy goals, and declining investment costs have increased the use of solar photovoltaic generation
in residential buildings. According to research, residential buildings account for around 40% of the world's overall energy use.
Most of this utilised energy is derived from fossil fuels, a cause of environmental degradation. The net zero energy building
(NZEB) provides a solution to this worrisome pollution. Rooftop solar panels coupled with home-scale storage units are
becoming more common in residential areas, and they provide prosumer families with tremendous potential to profit from
selling their excess energy to the grid or neighbours. Consequently, the primary purpose of this research paper is to integrate
solar (photovoltaic) PV with a household energy storage system (ESS) and the utility grid with NZEB's electrical load. Power
electronic converters are necessary to interface renewable energy sources (RESs) with the grid and the NZEB in order to
accomplish this objective. This research work aims to develop an intelligent next-generation bi-directional inverter controller
to meet the consumer flexibility demand under different operational constraints. A model predictive current control topology
for the inverter has been proposed to achieve the bi-directional power flow. Fourth-order Newton Gregory (N–G), backward
differencing polynomial has been used for prediction purposes.

1 Introduction HP) will likely experience overloading or power deficiency for

their dynamic operation due to load and generation
Increasing concern about global warming has spurred action to uncertainties during any power outage contingency. Therefore,
reduce CO2 emissions by increasing energy efficiency and integrating renewable energy resources (RESs) and dynamic
decreasing the use of fossil fuels. The residential and transport loads like HP and EV would create a more complicated system
sector in Ireland generated about 70 per cent of CO2 emissions where multiple sources and loads of power are required to be
in 2021 [1], [2], with optimistic scenarios predicting a 90 per coordinated. Moreover, during the normal operation/grid-
cent reduction in emissions by 2050 [3]. Incorporating connected mode, consumers would have enough flexibility to
distributed renewable energy production, such as solar participate in the future energy market.
photovoltaic systems on rooftops, and improving energy In recent years, remote microgrids/ energy communities have
efficiency in existing buildings, such as implementing heat witnessed substantial renewable energy penetration. Because
pumps (HP) or adopting electric vehicles, are only a few diesel generators are costly to transport to distant sites, low-
examples of possible futures. In this aspect, Irish green smart cost PV and wind energy have made sense in terms of lowering
buildings with rooftop PV, HP and electric vehicles (EV) will fuel costs and CO2 emissions [5], [6]. Renewables and
play a significant role in helping Ireland achieve its energy conventional generation may coexist even in the most
targets and limiting Ireland’s exposure to the financial and vulnerable grid networks by using an inverter as a voltage
societal costs associated with missed targets. The Irish energy source and adding synchronous generator (SG)-compatible
sector will need to use innovative means to achieve its responses through a virtual synchronous machine (VSM)
sustainable targets. controller and energy storage [7]–[9]. Energy resources in the
Greener houses are a component of Ireland's effort to combat immediate vicinity might be used to isolate network segments
climate change and satisfy international and EU obligations to and increase overall system resiliency in an emergency.
minimise air pollution and carbon footprint. The Climate Limited energy storage (low inertia) and low fault current
Action Plan 2021 outlines how Ireland will achieve this (some current limiting at 110% of their rating) combined with
objective. It contains goals to make Irish houses more eco- superior real and reactive power management have instead
friendly, such as increasing the Building Energy Rating (BER) made the system weaker in large numbers, making it a danger
of 500,000 current dwellings to a B2 rating by 2030. to the whole grid. In addition, there are bidirectional power
A solar PV system comprises solar panels, an inverter that flows from renewable generating near consumption points, as
converts the electricity produced (from DC to AC), and a opposed to one-way power flows from central generators to
battery to store energy for later use in bigger systems. A typical loads. The high penetration of distributed energy resources
solar PV system may save consumers up to €400 a year in (DERs) necessitates the creation of new grid management
electricity bills [4]. However, modern green homes with techniques. It is also possible for the voltage and frequency of
renewables, storage, and eco-friendly appliances (like EVs and a grid to vary, including voltage flicker and low-frequency

Authorized licensed use limited to: VIT University. Downloaded on September 21,2023 at 15:57:27 UTC from IEEE Xplore. Restrictions apply.
nadir values, if the grid rigidity is too low. Due to the fact that operation for a smart home inverter are more challenging,
grid-following inverters don't contribute to grid inertia or especially in the absence of grid backup during the islanded
system strength on their own, the grid becomes weaker as their mode. Further, an enhanced control topology is required for
popularity increases. Pure voltage regulator inner loops make the bidirectional power flow and for a smooth transition of the
grid-forming inverters too inflexible to work with other smart home inverter from island to grid-connected mode,
resources and the existing grid [10]. Grid-tied inverter control which demands communication with delays smaller than
methods [11] are critical for the efficient operation and 1msec if the controller successfully synchronises with the grid.
deployment of distributed energy resources and distributed The converter should be configured to automatically change
energy storage. Throughout the previous decade, significant the operating mode, i.e., inverter or rectifier in power factor
efforts and publications have been made in this area. correction mode, according to an instruction issued by the
Despite the significant computational overhead in actual home energy management system and the battery's state of
implementation, the model predictive control (MPC) has charge.
attracted a lot of interest in the power converter and drives the The main contribution of this study is to propose a bi-
research community owing to its ability to handle multi- directional AC/DC converter control topology for low-scale
system constraints and nonlinearities [12]. The MPC household applications. Fourth-order Newton Gregory (N–G)
incorporates a sophisticated mathematical model for backward differencing polynomial-based model predictive
predicting future states, which ultimately aids in the broad use controller has been proposed for the bi-directional inverter
of power electronics and promotes automatic digital control. control. The proposed Fourth-order Newton Gregory
Modern power electronics-based applications need MPC to backward differencing polynomial-based prediction model is
discover a solution in the range of tens to hundreds of very efficient for future state prediction to design the proposed
microseconds, a feat made possible by improvements in high- MPC controller. This methodology also helps reduce the
speed digital microcontrollers [12]–[14]. However the computation burden without compromising prediction
technology has great potential, but more work is needed to get accuracy [26]. Different test scenario has been presented to
it to the commercial and industrial levels [15]. verify the efficacy of the proposed controller. The rest of the
article is constructed as follows: Section 2 discusses the
Several MPC methods like finite control set MPC (FCS-MPC)
architecture of the proposed inverter topology, and Section 3
[16], optimal switching vector MPC (OWV-MPC) [17],
provides the model predictive control design method for the
optimal switching sequence MPC (OWS-MPC) [18], [19],
proposed configuration. On that basis, the switching sequence
continuous control set MPC (CCS-MPC) [20]–[22], and
generation is discussed in section 4. Further, the simulation
explicit MPC (EX-MPC) [23], [24] etc. has been already
results are given in section 5.
discussed in literature for different power electronics
applications. The FCS-MPC, OWV-MPC and OWS-MPC 2 Design of Bi-Directional Converter
methods are classified based on the characterisation of solving
optimisation problems. The switching sequence generated for The proposed system comprises a PV system mounted on the
these methods is based on minimising the selected cost roof of a modest residential building (10 kW load max) and
function. In contrast, the CCS-MPC and EX-MPC produce the linked to the network through a bi-directional inverter, as
continuous signal (duty ratio) and use a modulator to generate illustrated in Fig. l. The proposed bi-directional inverter would
the power converter's switching sequence. While the CCS- increase users' flexibility, allowing them to sell surplus energy
MPC and EX-MPC do guarantee a constant switching to the system operator or neighbour when they participate in
frequency for the converter, the MPC is notoriously difficult any kind of energy trading mechanism. One of the most
to formulate when dealing with nonlinear and constraint-based important components in this system is the voltage source bi-
systems. It's also possible that these control techniques will fail directional inverter, piques the author's curiosity throughout
if several samples are expected, which might be inefficient in the article.
the event of a significant nonlinearity or faulty situation.
Because the proposed techniques heavily rely on offline
computations for parameter modifications, they may be
useless for a stable operation. However, FCS-MPC, OWV-
MPC, and OWS-MPC use the power converter's discretisation
function to their advantage, eliminating the need for a separate
modulator to provide the necessary switching sequence [25].
And for these reasons, they are the most popular and used
MPC techniques for power electronic applications [25].
In order to provide security of power supply (during the
islanded mode) and reduce the cost of power (during grid- Fig. 1. Structure of the NZEB under study
connected mode), ES units are required to be introduced in the
smart home. An advanced inverter controller is required to Fig. 2 depicts a simplified schematic of the proposed bi-
manage the ES and RES to prioritise the renewables, directional ac–dc converter based on twin buck converters.
coordinate them with the ES and ensure reliable power that Four power switches S1, S2, S3 and S4, with parallelly connected
meets the quality requirements. Due to the fast intermittency diodes, are used, as well as two inductors L1 and L2, filter
of loads and low inertia-based generations, stability and capacitor C and two split DC-bus capacitors C1 and C2 are also

Authorized licensed use limited to: VIT University. Downloaded on September 21,2023 at 15:57:27 UTC from IEEE Xplore. Restrictions apply.
utilised. The converter works as a rectifier when the power is The voltage and current flow equations from Fig. 4, can be
transferred from the AC grid to the DC source such as battery. obtained as-
Alternatively, it works as an inverter when the power is ܸ஺஼ ൌ ܸ௅ଵ ൅ ܸ௅ଶ ൅ ܸூே௏ (1)
transferred from DC source to AC grid or load. The voltage
‫ܫ‬௚௥௜ௗ ൌ ‫ܫ‬஼ ൅ ‫ܫ‬௅ଵ (2)
across the capacitors is kept well-balanced through the
proposed controller to maintain stable operation. Further eqn. (1) and (2) can be written in terms of VAC as-
ௗூ೒ೝ೔೏ ௗ మ ௏ಲ಴
ܸ஺஼ ൌ ‫ܮ‬ െ ‫ܥܮ‬ ൅ ܸூே௏ (3)
ௗ௧ ௗ௧ మ
S1 S3
where, L =L1+L2,
C1 L1 Iinv
VBat Igrid

Igrid IL1 L1
C2 L2 Irec VAC
S4 S2
IC
VL1

Vinv
VAC C Inverter
VL2
Fig. 2. Schematic of the proposed bi-directional inverter

In order to operate the bi-directional power flow (for active L2

power), when the voltage and the current are in phase, the
circuit operates as inverter, and in the case of 180° phase
difference, it works as a rectifier (conventional for controller Fig. 4. Simplified circuit of the proposed system
design). Fig. 3 depicts the four states of operation for the Discretising equation (3) using the forward Euler method, one
rectification mode, which is identical to the inverter mode can get-
except that the voltage and current are in phase with each other
ூ೒ೝ೔೏ ሺ௞ାଵሻିூ೒ೝ೔೏ ሺ௞ሻ
in the inverter mode. It would be interesting to note that the ܸ஺஼ ሺ݇ሻ ൌ ‫ܮ‬ ൅ ܸூே௏ ሺ݇ሻ െ
DC bus capacitor C2 always charges during the positive half ்
௏ಲ಴ ሺ௞ାଵሻିଶ௏ಲ಴ ሺ௞ሻା௏ಲ಴ ሺ௞ିଵሻ
cycle, whereas C1 always charges during the negative half ‫ܥܮ‬ ்మ
(4)
cycle.
In order to find the inverter current for the (k+1) instant,
D2 D2
equation 4 can be written as-
S1
L1 S1 L1
C1 C1 ்
VAC
VBat
VAC
VBat ‫ܫ‬௚௥௜ௗ ሺ݇ ൅ ͳሻ ൌ ‫ܫ‬௚௥௜ௗ ሺ݇ሻ ൅ ൫ܸ஺஼ ሺ݇ሻ െ ܸூே௏ ሺ݇ሻ൯ ൅
௅
஼
C2 C2 ሺܸ஺஼ ሺ݇ ൅ ͳሻ െ ʹܸ஺஼ ሺ݇ሻ ൅ ܸ஺஼ ሺ݇ െ ͳሻሻ (5)
D1 S2 D1 S2
்

where VAC(k+1) is the predictive grid voltage. By using fourth-

order Newton Gregory (N–G), backward differencing
S1 D2 S1 D2 polynomial, the predictive grid voltage could be defined as-
C1 C1
VAC
VBat VAC
VBat ܸ஺஼ ሺ݇ ൅ ͳሻ ൌ ͷܸ஺஼ ሺ݇ሻ െ ͳͲܸ஺஼ ሺ݇ െ ͳሻ ൅ ͳͲܸ஺஼ ሺ݇ െ ʹሻ െ
C2 C2 ͷܸ஺஼ ሺ݇ െ ͵ሻ ൅ ܸ஺஼ ሺ݇ െ Ͷሻ (6)
D1 S2 L2 D1 S2 L2
Thus the predictive grid current and voltage can be easily
defined by using equations (5) and (6).
Fig. 3. Operation modes of the inverter.
4 Generation of Gate Pules patterns
3 Design of Model Predictive Controller
The error between the predicted grid current reference and
This section uses a model predictive controller for the measured grid current should be minimised by optimally
proposed bi-directional converter. The operation of bi- selecting the gate pulse patterns of the inverter. During each
directional power flow is dependent on the grid current sampling period (25 μs), the gate pattern is updated according
reference. Therefore, the main control objective during the to the cost function defined as -
rectifier mode is to maintain the DC bus voltage and the grid
‫כ‬
current reference. In contrast, during the inverter mode, the ݃ሺ݇ ൅ ͳሻ ൌ ห‫ܫ‬௚௥௜ௗ ሺ݇ ൅ ͳሻ െ ‫ܫ‬௚௥௜ௗ ሺ݇ ൅ ͳሻห (7)
control objective is to maintain only the grid current reference.
In this operation, the DC bus voltage is controlled through the The main objective of this cost function is to minimise the
battery management system, which is not the focus of this inverter current tracking error. The gate pattern is selected in
research work. The overall control objectives are achieved every sampling period according to the operational mode
through controlling the grid current references and optimally (inverter/rectifier) and cost function. The phenomenon is
selecting the appropriate gate pulse patterns. Fig.4 shows a illustrated in Fig. 5. In order to achieve the smallest tracking
simplified diagram of the proposed system, as shown in Fig.2. error, the gate pulse pattern [0 1 0 0] is selected.

Authorized licensed use limited to: VIT University. Downloaded on September 21,2023 at 15:57:27 UTC from IEEE Xplore. Restrictions apply.
 simulation time of 0.35 seconds, it abruptly rose to 20 A.
k-2 k-1 k k+1 According to Fig. 8, the proposed controller may be evaluated
*
I INV based on its ability to monitor the reference signal, which is
1000 shown in the above figure. The predicted signal always kept
pace with the reference signal, even during times of
IINV 0100 uncertainty.

0010

Fig. 5. Optimal gate switching pattern to reduce grid current

tracking error
The overall control block diagram is shown in Fig. 6. As
described in section 2, the circuit operates as a grid tie inverter
when the voltage and current signal are in phase (i.e., VAC൏0),
whereas it will act as an active rectifier when the voltage and
current signal are in opposite phase (i.e., VAC൐0). Moreover, it
can be seen that the predictive value of grid current reference
directly influences grid voltage, which can contain the
harmonics. Therefore, a phase look loop is employed to obtain Fig. 8. Reference current tracking performance
only the fundamental component of the grid voltage.
Fig. 9 shows the voltage and current waveforms over the
Switching Cost Function
Pattern Selection Minimization
I g* r id transition from the inverter mode to the rectifier mode. Initially,
Igrid(k+1)
when the inverter and the grid have been in phase, it operates
Mode Prediction
tion model as an inverter, and in this test scenario, at t = 0.35 sec, the
Selection (Eqn 5 & 6)
S1 S3 current reference has been changed to operate as a rectifier. At
L1 PLL that point, the inverter switches to rectifier mode, and the power
C1 VAC(k), Igrid(k) has been drawn from the grid to charge the battery and meet the
VBat household load demand, as depicted in Fig. 10.
C2 S4 L2
S2

Fig. 6. The model predictive control scheme for bi-directional

Inverter
5 Simulation Validation
The proposed model predictive current control scheme was
simulated in MATLAB/Simulink environment. The peak value
of the AC voltage was set to 400 V, the L1 and L2 inductance to
7.5 mH each, the filter capacitor 40 ɊF, the DC link capacitor
4000 ɊF, the cumulative load to 5 kW, and the sampling period
to 60 Ɋs. Fig. 9.Transition from inverter mode to rectifier mode

Fig. 10. Bi-directional power transfer

Fig. 7. Regular operation of inverter with load change
The system was first evaluated for abrupt load variations, as The final test case scenario demonstrates the transition from
shown in Fig. 7. At the beginning of the simulation, the system grid-connected mode to islanded mode. The simulation results
is programmed to operate at a current reference of 15A; at the are depicted in Fig. 11. As shown in figure 10, the transition

Authorized licensed use limited to: VIT University. Downloaded on September 21,2023 at 15:57:27 UTC from IEEE Xplore. Restrictions apply.
from grid-connected mode to islanded mode produces a voltage 123757, Jul. 2022.
drop as the inverter operates in the current control mode. But, [6] M. Fathi, R. Khezri, A. Yazdani, and A. Mahmoudi,
Due to the support of BESS to maintain the DC bus voltage, the “Comparative study of metaheuristic algorithms for
voltage drop (less than 6%) is in the allowable range for the optimal sizing of standalone microgrids in a remote
islanded mode of operation. area community,” Neural Comput. Appl., vol. 34, no.
7, pp. 5181–5199, Apr. 2022.
[7] Y. Yang et al., “A New Virtual Inductance Control
Method for Frequency Stabilization of Grid Forming
Virtual Synchronous Generators,” IEEE Trans. Ind.
Electron., 2022.
[8] A. Suvorov, A. Askarov, Y. Bay, and A. Gusev,
“Freely Configurable Model of Virtual Synchronous
Generator for Grid-Forming Converter,” SSRN
Electron. J., 2022.
[9] P. R. Chowdhury, E. Shoubaki, and M. Manjrekar,
“Modified Voltage Controlled-Virtual Synchronous
Machine Controller in the application of Parallel
Connected Grid-Forming Inverters,” Proc. 2021 IEEE
12th Int. Symp. Power Electron. Distrib. Gener. Syst.
PEDG 2021, Jun. 2021.
Fig. 11. The transition from grid-connected mode to islanded [10] D. B. Rathnayake et al., “Grid Forming Inverter
mode Modeling, Control, and Applications,” IEEE Access,
vol. 9, pp. 114781–114807, 2021.
6 Conclusions [11] Q. Liu, T. Caldognetto, and S. Buso, “Review and
Comparison of Grid-Tied Inverter Controllers in
The proposed model predictive current controller for the bi- Microgrids,” IEEE Trans. Power Electron., vol. 35,
directional household inverter has been proposed and tested in no. 7, pp. 7624–7639, Jul. 2020.
this research work. This method maximises the benefits of [12] S. Vazquez et al., “Model predictive control: A review
MPC and improves the practical engineering application of the of its applications in power electronics,” IEEE Ind.
MPC algorithm. The Matlab-based simulation results with Electron. Mag., vol. 8, no. 1, pp. 16–31, 2014.
different case studies prove the efficacy of the proposed [13] W. R. Sultana, S. K. Sahoo, S. Sukchai, S. Yamuna,
controller. Furthermore, the smooth transition results and the and D. Venkatesh, “A review on state of art
necessity for just one set of sensors make the home-size grid-
development of model predictive control for
tied inverter with bi-directional power flow desirable for low-
renewable energy applications,” Renew. Sustain.
cost systems. In addition, further research is required in order
to improve the voltage stability in the islanded mode of Energy Rev., vol. 76, pp. 391–406, Sep. 2017.
operation. [14] J. Holtz, “Advanced PWM and Predictive Control-An
Overview,” IEEE Trans. Ind. Electron., vol. 63, no. 6,
7 Acknowledgment pp. 3837–3844, Jun. 2016.
[15] G. A. Papafotiou, G. D. Demetriades, and V. G.
This work is a part of MiFIC and TRIDENT project, and Agelidis, “Technology readiness assessment of model
the authors in IERC thankfully acknowledge the support from predictive control in medium- and high-voltage power
the Department of the Environment, Climate and electronics,” IEEE Trans. Ind. Electron., vol. 63, no.
Communications and DTIF from the Department of Enterprise, 9, pp. 5807–5815, Sep. 2016.
Trade and Employment. [16] P. Karamanakos and T. Geyer, “Guidelines for the
8 References Design of Finite Control Set Model Predictive
Controllers,” IEEE Trans. Power Electron., vol. 35,
[1] Sustainable Energy Authority of Ireland, “ENERGY no. 7, pp. 7434–7450, Jul. 2020.
IN IRELAND 2021 Report,” 2021. [17] L. M. A. Caseiro, A. M. S. Mendes, and S. M. A. Cruz,
[2] “News Releases 2021 | Environmental Protection “Dynamically Weighted Optimal Switching Vector
Agency.” [Online]. Available: Model Predictive Control of Power Converters,” IEEE
https://www.epa.ie/news-releases/news-releases- Trans. Ind. Electron., vol. 66, no. 2, pp. 1235–1245,
2021/ireland-will-not-meet-its-2020-greenhouse-gas- Feb. 2019.
emissions-reduction-targets-action-is-needed-now-to- [18] P. J. Gomez, L. Galvan, E. Galvan, J. M. Carrasco, and
meet-2030-eu-targets.php. [Accessed: 13-May-2022]. S. Vazquez, “Optimal Switching Sequence Model
[3] C. and C. Department of the Environment, “Climate Predictive Control for Single-Phase Cascaded H-
Action Plan 2021,” 2021. Bridge,” IECON Proc. (Industrial Electron. Conf.,
[4] SEAI, “RESIDENTIAL ENERGY Roadmap,” 2020. vol. 2021-October, Oct. 2021.
[5] M. Saffari and M. McPherson, “Assessment of [19] A. Mora, R. Cárdenas-Dobson, R. P. Aguilera, A.
Canada’s electricity system potential for variable Angulo, F. Donoso, and J. Rodriguez,
renewable energy integration,” Energy, vol. 250, p. “Computationally Efficient Cascaded Optimal

Authorized licensed use limited to: VIT University. Downloaded on September 21,2023 at 15:57:27 UTC from IEEE Xplore. Restrictions apply.
 Switching Sequence MPC for Grid-Connected Three- demosaicing using WRGB based Newton Gregory
Level NPC Converters,” IEEE Trans. Power interpolation method,” Measurement, vol. 135, pp.
Electron., vol. 34, no. 12, pp. 12464–12475, Dec. 935–942, Mar. 2019.
2019. [27] R. W. H. Richard W. Hamming, Numerical Methods
[20] C. Bordons and C. Montero, “Basic Principles of MPC for Scientists and Engineers. New York: McGraw-
for Power Converters: Bridging the Gap between Hill, 1992.
Theory and Practice,” IEEE Ind. Electron. Mag., vol.
9, no. 3, pp. 31–43, Sep. 2015. Appendix
[21] A. A. Ahmed, B. K. Koh, and Y. Il Lee, “A
Comparison of Finite Control Set and Continuous Backward differencing polynomial equation of fourth-order
Control Set Model Predictive Control Schemes for Newton Gregory (N–G) [27] for predictive modeling are as
Speed Control of Induction Motors,” IEEE Trans. Ind. follows:
Informatics, vol. 14, no. 4, pp. 1334–1346, Apr. 2018. An interpolation polynomial may be based on previous y values
[22] A. T. Nguyen, S. W. Ryu, A. U. Rehman, H. H. Choi, and extrapolated to yn+1. Here is the fourth-order N–G
and J. W. Jung, “Improved Continuous Control Set backward differencing polynomial-
Model Predictive Control for Three-Phase CVCF
Inverters: Fuzzy Logic Approach,” IEEE Access, vol. ‫ݏ‬ ‫ ʹݏ‬൅‫ݏ‬ ‫ ͵ݏ‬൅͵‫ ʹݏ‬൅ʹ‫ݏ‬
‫݊ݕ‬൅‫ ݄ݏ‬ൌ ‫ ݊ݕ‬൅ ‫ ݊ݕ ͳ׏‬൅ ‫ ݊ݕ ʹ׏‬൅ ‫ ݊ݕ ͵׏‬൅
9, pp. 75158–75168, 2021. ͳǨ ʹǨ ͵Ǩ
‫ݏ‬Ͷ ൅Ͷ‫ ͵ݏ‬൅ͳͳ‫ ʹݏ‬൅͸‫ݏ‬
[23] S. Ruan, Y. Ma, N. Yang, C. Xiang, and X. Li, “Real- ‫׏‬Ͷ ‫݊ݕ‬
time energy-saving control for HEVs in car-following ͶǨ

scenario with a double explicit MPC approach,” where s is the fraction of h to which the polynomial is being
Energy, vol. 247, p. 123265, May 2022. extrapolated, and ‫׏‬1=yn-yn-1, ‫׏‬2= yn-2yn-1+yn-2 etc. from N–G
[24] M. Kvasnica, P. Bakaráč, and M. Klaučo, backward differencing. Setting s=1, yields the following
“Complexity reduction in explicit MPC: A equation:
reachability approach,” Syst. Control Lett., vol. 124,
‫݃݊ݕ‬
݊൅ͳ
ൌ ͷ‫ ݊ݕ‬െ ͳͲ‫݊ݕ‬െͳ ൅ ͳͲ‫݊ݕ‬െʹ െ ͷ‫݊ݕ‬െ͵ ൅ ‫݊ݕ‬െͶ
pp. 19–26, Feb. 2019.
[25] S. Vazquez, J. Rodriguez, M. Rivera, L. G. Franquelo, where ‫݃݊ݕ‬
݊൅ͳ
is the result of extrapolating the N–G interpolation
and M. Norambuena, “Model Predictive Control for
polynomial.
Power Converters and Drives: Advances and Trends,”
IEEE Trans. Ind. Electron., vol. 64, no. 2, pp. 935– A Lagrange interpolation polynomial of the same order
947, Feb. 2017. extrapolated to yn+1 yielded this formula.
[26] K. Elayaperumal Paul and V. Sathianathan
Saraswathibai, “Maximum accurate medical image

Authorized licensed use limited to: VIT University. Downloaded on September 21,2023 at 15:57:27 UTC from IEEE Xplore. Restrictions apply.