A chance constraint microgrid energy management with phase balancing using electric vehicle demand aggregation

ABSTRACT

Microgrids are sustainable distribution systems which integrate plug-in electric vehicles (EVs) and renewable energy sources to minimize grid dependency. The uncertainty in the forecast data of renewables and erratic consumption of energy by consumers aggravates voltage stability in a three-phase microgrid. Voltage instability and ineluctable uncertainty in the system inevitably make energy management of resources infeasible. To address the challenges of voltage stability and uncertainty in microgrid energy management, a novel two-stage energy management framework using stochastic chance constraint model predictive control (MPC) is proposed to optimize operational costs with reduced reserve costs. In the primary stage, an optimal day-ahead resource dispatching strategy is modeled using robust chance-constrained to handle the uncertainty of the forecasted data. Wasserstein metric ambiguity set is developed against uncertainty risk operations for low conservativeness and tractable constraints. The secondary stage control, with a shorter timescale, regulates the deviations in the primary stage state parameters using the MPC technique. Also, a fuzzy rule-based unbalance voltage control for EV parking lots (EVPL) is proposed in the secondary stage by modifying each phase’s power using phase switches. The proposed energy management framework adopts a mixed-stage optimization structure, where the concerned problem is progressively optimized over diverse time scales. The results of the proposed strategy show that lowering the confidence interval and increasing sampled data reduces operating costs by 7.57% and 5.34%, while penalty costs are reduced by 51.9% and 74.9%, respectively. It is also observed that this strategy minimizes the voltage unbalance on a modified IEEE 123 bus system and validates with benchmark approaches.

KEYWORDS:

• Chance constraint
• electric vehicle parking lot
• microgrid energy management
• model predictive control
• phase balancing

Nomenclature

Acronyms:	=
ALD	=	Aggregated load demand
BESS	=	Battery Energy Storage System
CVaR	=	Conditional value-at-risk
DE	=	Diesel generator
DG	=	Distributed generator
EMS	=	Energy management systems
EV	=	Electric vehicle
EVPL	=	EV parking lots
HSS	=	Hydrogen storage systems
MGCC	=	Microgrid control center
MILP	=	Mixed integer linear programming
MLD	=	Mixed Logical Dynamical
MPC	=	Model predictive control
MT	=	Microturbines
PBR	=	Power balance ratio
PV	=	Photovoltaic panels
SOC	=	State of charge
UPR	=	Unbalanced power ratio
VU	=	Voltage unbalance
VUF	=	Voltage unbalance factor
WT	=	Wind turbines
Indices/Sets:	=
t	=	Time period
N	=	Control horizon
$N_{dg}$	=	Number of DG
$N_{fdg}$	=	Number of fuel DG
$N_{EV}$	=	Number of EVs
$N_{EVCS}^{\hspace{0.17em}}$	=	Number of vehicles at charging station
$N_{bus}$	=	Total buses in system
M	=	Sampled data
$\mathbb{P}$	=	Probability distribution
$E$	=	Expected value
I	=	Identity matrix
Parameters:	=
vw	=	Wind velocity (m/s)
TA	=	Ambient Temperature, (25°C)
IR	=	Irradiance of the PV, (W/m2)
Pt	=	Power at “t,”(kW)
$\delta$	=	Logical operator (ON/OFF)
η	=	Efficiency (%)
$V_{hydride}$	=	Metal Hydride Capacity (Nm)
$C_{bat}$	=	Battery Capacity (kWh)
$P_L^{\hspace{0.17em}},P_G^{\hspace{0.17em}}$	=	Active power of loads and generators
$Q_L^{\hspace{0.17em}},Q_G^{\hspace{0.17em}}$	=	Reactive power of loads and generators
$\theta$	=	Angle of the voltage
V	=	Voltage at a node
${\phi }_{price}$	=	Emission penalty cost
J	=	Jacobian matrix
Y	=	Admittance matrix
$I_i^p$	=	Current of the phase “p” at ith node
$V^{+},V^{-},V^o$	=	Positive, negative and zero voltage sequence values
ai,bi,ci	=	Cost of coefficients of DGs
cp	=	Penalty price ($)
$I_i^p$	=	Current in phase p at ith node
$S_i^p$	=	Apparent power in phase p at with node (KVA)
$C_{op}^t$	=	Operating cost ($)
$C_{op\mathrm{\_}e}^t$	=	Emission cost ($)
$C_{op\mathrm{\_}f}^t$	=	Fuel cost ($)
$C_{op\mathrm{\_}g}^t$	=	Grid cost ($)
$E_e^t$	=	Emissions at “t” (kg)
$\mathcal{B}$	=	Ambiguity set
Variables:	=
$P_{RE}^t$	=	Renewables power at t, (kW)
$P_{RE\varsigma }^t$	=	Renewable Forecast Error (kW)
$P_{dg}^t$	=	DG power at “t,” (kW)
$P_{BESS}^t$	=	Battery Energy Storage Power (kWh)
$P_{BT}^t$	=	Battery power output at t, (kW)
$P_{HSS}^t$	=	Hydrogen Storage System (kWh)
$P_{H2}^t$	=	Power from hydrogen (kW)
$P_{Elz}^t$	=	Power to Electrolyzer (kW)
$P_{EVCS}^t$	=	Power across the EV charging station (kW)
$P_{EV}^t$	=	Electric vehicle power (kW)
Icharge	=	Charge consumed (A)
$P_{\mu grid}^t$	=	Power of the microgrid (kW)
$E_{price}^t$	=	Electricity price (¢)
$P_{L,ph}^t$	=	Load Power (kW)
$P_{EVPL,ph}^t$	=	Phase power of EV parking lot (kW)
$P_{loss}^t$	=	Power loss at each phase(kW)
${\mu }_{ph}^k$	=	Unbalance factor
$C_p^t$	=	Penalty cost at t ($)

Data availability statement

The data used to support the findings of this study are available online openly, and any further information required is available from the corresponding author upon request https://open-power-system-data.org/.

Disclosure statement

No potential conflict of interest was reported by the authors.