Dynamic economic emission dispatch with load dema nd management for the load demand of electric vehicles during crest shaving and valley filling in smart cities environment
Introduction
In operating power-generating systems (PGSs), the economic dispatch (ED) of active output power is a critical task in the optimum scheduling of a group of online generators; this task enables the minimization of the total operating fuel cost (OFC) on the basis of several practical equality and inequality operating power constraints enforced by PGSs and the generators [[1], [2], [3], [4], [5]]. Given the growing widespread concern over a sustainable environment, emissions such as carbon dioxide (CO2), nitrogen oxide (NOx) and sulfur oxide (SOx), should be considered [[6], [7], [8], [9], [10]]. When ED of power is added to the OFC as another objective function, the economic emission dispatch (EED) becomes a multi-objective optimization problem [[11], [12], [13], [14], [15]]. The EED is aimed at simultaneously minimizing the OFC and emissions for a specific period of time, for example, 1 h. However, this mode of scheduling may be valuable only for an hour depending on the changes in load demands. For this problem, dynamic EED (DEED) is aimed at simultaneously minimizing the OFC and emissions under daily load demand changes (24 h). In such a case, the DEED is considered a multi-objective problem.
Currently, as a necessary design plan of PGS operation infrastructure in smart cities environment, electric vehicles (EVs) have attracted considerable interest due to fossil fuel shortage, wide environmental changes, and the fast development of communication technology in smart cities [[16], [17], [18], [19], [20]]. In the United Kingdom, in 2030 60% of vehicles will be EVs [21]. With such increases, the load demand of EVs increases and becomes an additional burden on PGSs due to the random battery charging of a large-scale penetration of EVs during the day [22,23]. A balance between the system generation and load demand is essential and should be made by adjusting the load demand throughout the day. Thus, applying the load demand management (LDM) [[24], [25], [26], [27], [28]] on the load demand of EVs during day is an essential task. The LDM monitors and controls the consumers’ energy consumption. In other words, control on the load demand during the day. The LDM strategy aims to reduce the energy consumption during the crest region (highest load demand areas during the day peak hours) and shifting the load demand to the valley region (lowest load demand areas during the day off-peak hours). Then, the system generation is partially extracted by the LDM. Thus, the maximum power delivered is directly affected by daily load demand changes. The LDM radically contributes in the design of PGS capacity to ensure stable and reliable operation, and to mitigate demand on generation and then reducing the volume of PGS capacity to satisfy economical and environmental benefits throughout flexible power-grid.
The load demand of EVs should be scheduled with consideration of the optimum dispatch of active output power and emissions satisfy several benefits in the load demand distribution and system generation, as follows, it adjusts the battery charging of EVs during day. In addition, it reduces the maximum load demand and improves the load curve, i.e., the shape of the load profile, as well as reducing the maximum active output power delivered by PGSs. It also allows provides financial benefits to consumers and utilities through dynamic pricing strategies throughout the valley region (off-peak period), thereby encouraging owners of EVs to charge their EVs in this region. Finally, it improves the stability and reliability of PGS.
Several applications of LDM for EVs have been achieved in the recent years, for example, demand response, mitigation of environmental deterioration, energy efficiency and conservation, reducing capital investments, flexible load profile, load shifting and load expansion, and vehicle-to-grid (V2G) that involves the control on the generated power export from EV battery back into the PGS [[29], [30], [31]].
One of the key drivers for growing commercialization of EVs now and into the future can be summarized, as follows. A key solution to reducing carbon emissions is the electrification of the transportation sectors. Using EVs improves the load profile through utilizing the valley regions and provides the potential to reduce the electricity costs for participating consumers. The fast development of cyber-physical power system technologies in the recent years means that smart sensors, meters, actuators, communications, security, and high-levels of safety also contribute to the increased demand for the purchase of EVs. Investment in the smart cities is another driver supporting the uptake of EVs.
Several studies investigated the combination between problems regarding LDM and DEED problem but limited. In Ref. [16] an integration approach was used between EVs and V2G technology. The authors in Ref. [16] proved that the V2G provides crest load shaving. In Refs. [32,33] different approaches were used to cover the load demand of EVs by using the grid-to-vehicle (G2V) technology, which is the control on the generated power export from the PGS back into the EV battery. The V2G technology involves filling the valley regions. In both Case Studies, peak shaving was not considered. In Ref. [34], another approach for crest shaving and valley filling (CSVF) EVs was used. The focus of this study is the minimization of the difference between the daily load demand and V2G availability. In Refs. [35,36], the V2G power dispatch and peak shaving were achieved throughout two approaches. All these case studies did not consider the EED problem when using the LDM strategy.
Currently, studies on the DEED problem that share the load demand of EVs while applying LDM are very limited. In Ref. [37], the EED problem with the battery charging of EVs using renewable energy sources (RESs) for a specific hour in day was considered. In Ref. [38], a scheduling strategy for the battery charging of EVs was used while the DEED problem has been solved. The DEED problem including the EVs was introduced in Ref. [39]. However, the optimum dispatch of power and emission outcomes were achieved without the use of the control strategy of battery charging and discharging of EVs during day. In Ref. [40], the scheduling of optimum charging was proposed for the minimization of CO2 emission in consideration of V2G technology for the EVs in the crest region. In Ref. [41] a bi-directional dispatch model was used for the EVs. The proposed model was used for the minimization of total OFC of a group of online generators. In Refs. [42,43], the unit commitment was used for the scheduling of the charging of EVs, and OFC was minimized by scheduling the start-up and shut-down of online generators. However, the focus of in these studies was the allocation of load demand corresponding to the committed online generators for the minimization of the total OFC. In Ref. [44], the smart charging and V2G technology were applied to the EVs to obtain economic and environmental benefits. However, the LDM is not applied as well as the multi-objective function is solved individually. A mixed-integer linear program was applied to determine the power flow and CO2 emission during day. However, the OFC is not included [45]. In Ref. [46], a robust optimization approach with Monte Carlo method were applied for the optimum dispatch of 300 EVs. The CO2 emission was included. However, the LDM is not included on the load demand of EVs. The authors in Ref. [47] proposed a model for V2G and G2V that can bring advantages to the PGS from the mitigation of the sharp deviations of renewable energy resources in which the OFC and emissions for transportation sector by using ε-constraint method integrated with grey wolf optimizer and particle swarm optimization (PSO) algorithm. However, solving such a multi-objective problem with a large penetration of EVs by using a combination of three techniques requires a long execution time to obtain optimum solution. In Ref. [48], the authors addressed the integration of RESs, e.g., solar and/or wind, with a large penetration of EVs in Galapagos Islands power-grid by proposing taxing on the consumers that utilizing TGU PGS due to CO2 emissions. However, low pricing is considered when using RESs. In Ref. [49], a large-scale of EVs is integrated with V2G technology throughout a combination of RESs using independent system operator. However, one a challenge is managing the active and reactive power flow into the power-grid. The outcomes in Ref. [49] demonstrate that the V2G integration with the EV's battery can mitigate uncertainty of RESs on PGS operation, and subsequently reduce the OFC by controlling the power flow through on-load tab changing and phase shifting transformers.
The need for more sustainable power generation and usage as well as the desire for smarter cities means the popularity of EVs are continuing to grow. This continues to increase load demand that will continue to burden PGS infrastructure. To address this, this study presents a novel approach that both applies LDM to the random demand of the crest and valley regions of large-scale EV load while simultaneously solving the multi-objective DEED problem that has several equality and inequality operating power constraints. The LDM is applied by shifting the load demand of EVs from the crest region using crest shaving to the valley region using valley filling. In addition, the trade-off solution for the OFC in ($) and emissions in (kg) in day is found. Combining LDM of the EVs load demand in the CSVF regions and solving the multi-objective DEED problem simultaneously provides important insight to help inform decisions around future PGS infrastructure. Achieving this can save the OFC in ($) and to reduce emissions in (kg), as well as volume of the investment during the maximum load demand.
The multi-objective problem has non-linear characteristics, and to solve it we propose the use of novel orthogonal particle swarm optimization (OPSO) algorithm. The OPSO algorithm is an evolutionary computation technique in which the population, i.e., swarm, comprises m particles, which looks for the optimum solution in a d-dimensional search space, where m > d. The m particles are separated into two clusters: an active cluster of d best particles and a passive cluster of (m ‒ d) particles. The diversity of the population is boosted by this procedure. The orthogonal diagonalization (OD) process is applied on the position vectors of d best particles. Then, the d orthogonal vectors in the active cluster draw an optimum solution of a multi-objective function. In every iteration, the velocity and position vectors of the d best particles in the active cluster are updated. Thus, only one guide is utilized. Meanwhile, the (m ‒ d) particles are left unchanged. As a result, the d best particles in the active cluster move toward the optimum solution in a d-dimensional search space steadily.
In the recent years, the optimization methods in Refs. [[50], [51], [52]] have been proposed with the name is called “orthogonal”. It is orthogonal learning particle swarm optimization (OLPSO) algorithms, first was used for a global optimization and the second was used for a local optimization [50], orthogonal global-best-guided artificial bee colony algorithm [51], and orthogonal genetic algorithm with quantization [52]. In these methods, a different strategy, namely, “orthogonal experimental design (OED)” has been applied. The OED is applied to obtain a set of possible optimum solutions. The OED allows the inputs interact among them in which the output process can be optimized. It works on a predefined orthogonal array table of N factors with Q levels per factor. However, the disadvantages of the OED are that it only convenes when there is weak or no interaction among the factors. Also, for large-scale problems such as the once considered herein the table of the orthogonal array that contains variables becomes complex and the orthogonality may not be able to recognized meaning that the optimum solution is not obtained. Thus, the algorithm is not able to obtain the best solution.
In the recent works, high-dimensional unimodal and multimodal benchmark functions taken from the congress on evolutionary computation (CEC) 2005, 2008, 2009, 2013, 2014 and 2015 have been solved by the OPSO algorithm [53]. The OPSO algorithm is effective in obtaining the OFC for different sizes of PGSs (small-, medium-, and large-scale thermal-generating units under several operating power constraints [[54], [55], [56], [57]].
The rest of this study is organized as follows. In Section 2, the formulation of the DEED problem with several equality and inequality operating power constraints is shown. In Section 3, the procedure of applying OPSO algorithm to the LDM and DEED problem is reported in consideration of EVs. The statistical simulation, analysis, and discussion are presented in Section 4. In Section 5, the conclusion of this study is provided.
Section snippets
Formulation of the DEED problem with operating power constraints
In this study, three challenges are handled simultaneously, which are the key components in this study. First, solve the multi-objective DEED optimization problem, i.e., total OFC and CO2, NOx and SOx emissions during a specified dispatch interval to find the trade-off feasible solution throughout the operation of online TGUs. Second, solve the operating power constraints imposed by the PGS, TGUs and EVs. Third, apply the LDM on the load demand of 30,000 EVs. In this section, the three key
Procedure of applying OPSO algorithm to LDM and DEED problem in consideration of EVs
The procedure of applying the OPSO algorithm to multi-objective DEED problem and the LDM strategy on the load demand of 30,000 EVs is provided in this section.
Statistical simulation, analysis and discussion
Two Case Studies are used for solving the DEED problem without/with applying the LDM on the load demand of EVs throughout the CSVF regions during day under several inequality and equality operating power constraints by using the OPSO algorithm. A total of 10 TGUs of PGSs are used in these two Case Studies taken from Ref. [71], as follows:
Case Study #1: Solving the DEED problem without considering LDM on the load demand of EVs.
Case Study #2: Solving the DEED problem while applying LDM on the
Conclusion
In this study, the dynamic economic emission dispatch (DEED) problem under several inequality and equality operating power constraints were solved simultaneously while applying the load demand management (LDM) on the load demand of 30,000 EVs in the crest shaving and valley filling (CSVF) regions by using an evolutionary computation technique is called OPSO algorithm. Applying the LDM on this a large-scale of EVs has made the crest- and-valley converge becomes large due to adjusting the power
References (72)
- et al.
A switched dynamical system approach towards the economic dispatch of renewable hybrid power systems
Int J Electr Power Energy Syst
(2018) - et al.
Solving the combined heat and power economic dispatch problems by an improved genetic algorithm and a new constraint handling strategy
Appl Energy
(2019) - et al.
A CVaR-robust-based multi-objective optimization model and three-stage solution algorithm for a virtual power plant considering uncertainties and carbon emission allowances
Int J Electr Power Energy Syst
(2019) - et al.
Dynamic economic and emission dispatch model considering wind power under Energy Market Reform: a case study
Int J Electr Power Energy Syst
(2019) - et al.
Full-infinite interval two-stage credibility constrained programming for electric power system management by considering carbon emission trading
Int J Electr Power Energy Syst
(2019) Energy consumption, CO2 emissions and economic growth in developed, emerging and Middle East and North Africa countries
Energy
(2019)- et al.
A dynamic economic emission dispatch considering wind power uncertainty incorporating energy storage system and demand side management
Renew Energy
(2016) - et al.
A low-carbon economic dispatch model incorporated with consumption-side emission penalty scheme
Appl Energy
(2019) - et al.
A multiobjective hybrid bat algorithm for combined economic/emission dispatch
Int J Electr Power Energy Syst
(2018) - et al.
Multiobjective economic-environmental power dispatch with stochastic wind-solar-small hydro power
Energy
(2018)