Intelligent hierarchical energy and power management to control the voltage and frequency of micro-grids based on power uncertainties and communication latency

https://doi.org/10.1016/j.epsr.2021.107567Get rights and content

Highlights

Abstract

This paper presents an intelligent energy management method to control the voltage and frequency at the primary and secondary control levels of micro-grids. The proposed model is based on the model predictive control (MPC) at the primary and intelligent neural network (INN) at the secondary. The purpose of this paper is to control the voltage and frequency based on the uncertainties of power generation resources and loads. In order to validate, the results of the proposed model are compared in different scenarios with other methods such as Fuzzy and PID controllers. The performance of the proposed method in primary and secondary as well as the coordination of this method with a high degree of sensitivity along with HESS and diesel generator units, in addition to increasing stability and reliability, protects the operation of the micro-grid against fluctuations based on uncertainties. The results show that this method, in addition to accuracy, has a higher operating speed than other controllers. As a result, the voltage and frequency restoration in the primary and the compensation of deviations from the reference in the secondary are associated with acceptable results. In order to validate the simulation results and experimental results are also presented.

Introduction

In recent years, the problems caused by fossil fuels in the production of electrical energy have created several environmental challenges. Therefore, researchers have proposed the idea of electrical micro-grids based on renewable energy. Micro-grids not only have the ability to cope with environmental challenges but also change the technical and economic structures of current networks [1, 2].

Although micro-grids have many advantages, the use of this idea in current distribution networks has faced operating conditions with various challenges such as: control challenges, security, reliability, flexibility, energy management and stability. Advances in the power electronics industry, as well as energy storages, have enabled micro-grids to provide attractive and reliable responses to these challenges. Therefore, it can be stated that power electronics and energy storage devices as two wings, play an essential role in the development of the micro-grid idea [3, 4, 5, 6, 7].

Control infrastructure is considered as one of the most important parts of micro-grids. Control factors in micro-grids are divided into three basic levels, which are: primary, secondary and tertiary. According to the various factors and objectives of operation as well as scheduling frameworks, each control level is responsible for meeting the set goals of the same level. Each level of control operates within its own time frame. Short-term objectives at the primary level, medium-term objectives at the secondary level and long-term objectives at the tertiary level are considered [8, 9, 10].

In recent years, with advances in the artificial intelligence industry and intelligent algorithms, micro-grids have seen significant advances in controlling key factors such as voltage, frequency, and power. Today, artificial intelligence structures and intelligent algorithms based on machine learning theories such as neural networks not only provide the optimal technical and economic utilization for micro-grid operators, but also the use of these structures will increase the degree of accuracy and reliability [11,12].

In order to ensure the performance of micro-grid, control methods and stability analysis are crucial. In micro-grid, by increasing the levels of connection to the main grid, increasing the presence of distributed generation resources, using energy storage devices, increasing use of power electronics, using communication infrastructure and considering electricity market factors increase the complexity of their operation. Therefore, the complexities of micro-grids must be met by providing secure and reliable solutions. In [12], in addition to examining the methods of artificial intelligence, it has also stated the features and limitations of this method. This article is a comprehensive overview of the applications of this method in security assessment, stability assessment and fault detection in micro-grids. On the other hand, the challenges facing this idea such as the high requirements on data, unbalanced learning, interpretable artificial intelligence, problems in learning transfer and the strength of artificial intelligence based on the quality of communication have been examined.

In [13], a control method based on the predictive method for voltage and frequency regulation in an island micro-grid is presented. The proposed method is considered with respect to the voltage and frequency fluctuations at the moment of islanding. In this paper, by connecting different micro-grids to each other, it is possible to control the voltage and frequency fluctuations with proper distribution of active and reactive powers. In this paper, in addition to comparing the proposed method with PI controllers, the proposed method is analyzed in three scenarios in order to validate it. The results show the optimal control of voltage, frequency and control errors reduction. One of the weaknesses of this work is that although it considers the battery storage as part of the control structure, its dynamic behavior has not been studied. Basically, because the voltage and frequency must be controlled over short periods of the time, the dynamic behavior of the battery is not able to control the voltage and frequency fluctuations in instantaneous disturbances.

Ref [14] presents technical and economic analysis of energy management based on intelligent methods and machine learning theory. The goals set in this reference are based on two main objectives, which include maximizing the profit from energy exchange and also minimizing fluctuations due to the exchange Power flow between the micro-grid and the main network. Therefore, the objective function is planned based on technical and economic factors. The intelligent methods described in this reference include Genetic algorithms, Dynamic programming, ANFIS, Fuzzy systems, Support vector machine, Neural networks. In addition to comparing methods, their characteristics and limitations are also examined.

In [15], the authors have proposed an adaptive neuro-fuzzy inference system (ANFIS) model based on general micro-grid droop characteristics to control the voltage and frequency. The proposed model, independent of the effect of line impedance, has been investigated in various scenarios. Voltage and frequency are two basic parameters in electrical networks that are mainly dependent on active and reactive power fluctuations. In micro-grids based on distributed generations, the use of voltage source converters with a specific droop characteristic is common. With changes and fluctuations in load, fluctuations occur between production and consumption. These changes cause deviations in the output voltage and frequency of the converters. If these changes are large, the stability of the micro-grid may be compromised. Based on the results obtained from this study, it is possible to observe the proper performance of the proposed model against various load fluctuations. The proposed intelligent model is based on a combined neural network and fuzzy algorithm. One of the strengths of the proposed model is the intelligence and micro-grid learning ability against fluctuations affecting voltage and frequency changes, which in turn increases the layers studied in the neural network structure. Although this structure has good results in island micro-grids, but the validation of this structure has not been analyzed in the grid connected operation mode as well as networked micro-grids. On the other hand, the absence of storage resources, which play an essential role in controlling the power fluctuations, is a significant part that challenges the proposed structure.

In [16], a new method for controlling voltage and frequency in AC micro-grid in the presence of wind farms based on synchronous generators is presented. The proposed technique allows the wind farm to be connected to the main grid using an HVDC rectifier. The control model presented in this paper allows the AC micro-grid, which consists of a wind farm and a rectifier, to be operated in two modes of voltage or current control. Among the strengths presented in the proposed model are endurance against load fluctuations in the island operation mode, resistant to capacitive bank switching, resistant to power generation constraints due to low wind speed and also AC rectifier switching. One of the major challenges facing this model is its efficiency in hybrid micro-grid structures based on DC power generation resources. This is important because the hybrid micro-grid structure in addition to many advantages, has a variety of control complexities.

In [17], the authors have presented a control model based on compensation of the voltage and frequency deviations in the secondary level of island micro-grids by considering the active power distribution. The proposed model based on multi-objective functions considers the effects of model uncertainty and changing the parameters of power generation units. It thus offers a certain degree of strength against all these factors. In the proposed model, in order to control voltage and frequency fluctuations, the information of distributed generation units and the resources that are in their neighborhood are considered. This method not only simplifies communications infrastructure, but also reduces investment and operating costs. The proposed design counteracts any fluctuations in the terminal voltage and frequency of DG units. This method also facilitates various errors based on the optimal division of active power in a predetermined ratio, despite the variations of variable load over time. The proposed control scheme is based on a nonlinear dynamic model of inverter-based DG units (which may not all be the same). One of the ambiguities in this research is the lack of validation of the proposed model in the grid connected mode. On the other hand, in the inverter-based micro-grids, the imbalance between production and consumption causes a deviation in the voltage and output frequency of these converters, which in the absence of a proper micro-grid control strategy will be unstable.

In [18], the authors have proposed a model for voltage and frequency control of an islanded micro-grid at the secondary level based on the finite time factor. The proposed model compensates for the voltage and frequency deviations from their reference values by controlling the active and reactive power flow. Therefore, it has presented its goals by properly distributing active and reactive power as well as using communication infrastructure based on local information and neighborhoods. Therefore, the communication structure used in this research is designed based on the neighbor-to-neighbor protocol. The strengths of this research based on the results are: the efficiency of the proposed model by considering the unequal impedances of the lines and also control of the load fluctuations and operation in the plug and-play mode.

In [19], a new multilayer architecture is designed for the control algorithm based on the large signal model, which enables the micro-grid to operate in a wide range of operating points. The objectives of the designed controllers include voltage and frequency amplitude regulation as well as the output power of distributed generation sources. In this paper, two levels of local and centralized control are considered. The operation scenario of these controllers is based on communication structures. The performance of the proposed method is such that when the central controllers are unable to operate due to errors in the communication networks, then the local controllers must control the critical micro-grid factors. The voltage and frequency control method is based on the modified droop method and is considered in the distribution network with high R/X rate. The inner layer of the proposed model is responsible for regulating voltage and frequency, while the outer layer is responsible for controlling the distribution and exchange of power. This model has also been analyzed for validation in three scenarios of islanded, grid connected and networked mode. One of the strengths of this paper is the reduction of uncertainties in the energy distribution management system by exchanging power in the networked mode micro-grid operation structure. One of the ambiguities in this article is the lack of accurate study of the impact of errors in the communications infrastructure and how the proposed model works with such errors. On the other hand, not examining the effect of communication infrastructure delays and the resulting uncertainties on the proposed method also poses a major challenge to this model.

In [20], a new demand response (DR) control method is proposed to control frequency fluctuations by considering the compensation of problems caused by communication delay and frequency deviation error detection. In this paper, a centralized and distributed control method based on flexible loads is programmed. The results of the proposed method show the reduction of the maximum frequency deviation.

As mentioned in the literature review section of previous research [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], the weaknesses of research have been categorized. Some of the most important research challenges related to micro-grids can be expressed as follows:

  • 1)

    As the number of micro-grids connected to the main grid increases, controlling and monitoring the power flow between the micro-grids and the main grid will become more complex. Therefore, lack of management and control of exchange power will lead to stability problems and reduced reliability in the operation of energy distribution networks in the presence of multiple micro-grids.

  • 2)

    Since the voltage and frequency are considered as one of the most important factors of operation and stability, control of these two factors in the period of milliseconds and seconds is essential. Due to the fact that the micro-grid has the feature of operation in two modes of connection to the main grid and islanded, so it is very important to provide an efficient model in controlling the voltage and frequency, especially in switching time intervals between these two modes of operation. At the same time, it should be noted that the operation method in these two cases is so different from each other that this should also be considered in providing a control method. For example, when the micro-grid is operated as a connection to the main network, the frequency deviation range of about 1% (0.5 Hz) is considered. While this deviation in islanded mode is equal to 3% (1.5 Hz) [8,11].

  • 3)

    Because synchronizing the voltage amplitude between the micro-grid and the main grid is complex, connecting two active electrical grids with different voltage amplitudes create a surge current between two grids. This in turn will lead to stability problems and reduced reliability.

  • 4)

    The connection of the micro-grid to the main grid will cause the dependence of the micro-grid power, while in the case of islands, the micro-grid power will be independent. Algorithm for changing power independence as well as power dependence between these two systems is another issue that should be considered in providing methods of energy management and micro-grid power control.

  • 5)

    Because the micro-grid includes various devices such as distributed generation sources, energy storage devices, power electronics, communication infrastructure and electrical loads with different characteristics, providing a control structure appropriate to the characteristics of the equipment in the micro-grid is difficult and vital. In other words, it can be said that due to the complexities of micro-grids, the presentation of control models in addition to high importance, has both potential and actual complexity.

  • 6)

    Distribution and management of power between power generation sources, energy storage and consumers due to their inequality coefficient of coincidence, controlling the voltage and frequency of the micro-grid becomes complex and difficult in different operating conditions.

  • 7)

    Problems due to the lack of fast and accurate error detection in the micro-grid, which exploits it in network connection modes and islanded mode leads to unstable conditions.

  • 8)

    With the increase of micro-grid penetration in current networks in order to monitor and control optimally at different levels of micro-grids, conventional methods do not have sufficient capability. Therefore, the use of intelligent and advanced methods is essential.

  • 9)

    Lack of fast and accurate investigation, prediction and monitoring of dynamic micro-grid behavior based on transient events, provide critical operation problems. Transient factors can be divided into the following two categories, which it is necessary to consider these factors in presenting control models: Normal transient factors: load step, motor starting, switching, load shedding, transformer inrush, etc. Unexpected disturbances: sudden power outage, short circuit, etc.

  • 10)

    Lack of investigation and consideration the uncertainties affecting the dynamic behavior of the micro-grid, which poses critical problems for the stability and management of the micro-grid.

Part 1: In this paper, an attempt has been made to consider various factors that affect the performance and operation of the micro-grid. The proposed method is based on the management and control of power and energy with the aim of controlling the voltage and frequency of the micro-grid. Then, by using predictive and intelligent models, increasing the level of accuracy and speed of the proposed method in detecting different events and providing an appropriate reaction based on uncertainties is presented. Meanwhile, with the optimal management of power distribution based on intelligent frameworks, on the one hand, power exchange with the main network and on the other hand, power management between energy production sources, storages and consumers is controlled. In this regard, by considering the predictive dynamic model of some equipment, the speed and accuracy of the proposed method in dealing with various events has increased.

Part 2: In this paper, the proposed model is based on controlling the operation objectives in primary and secondary micro-grids, using predictive model and intelligent algorithm. The two main factors in the primary and secondary levels are the voltage and frequency fluctuations control. These two factors are considered as two important parameters in the operation of the micro-grid. At the primary level, the control method must control the instantaneous fluctuations of voltage and frequency. At the secondary level, the deviation of these two parameters from their reference values is controlled by the proposed method. At the primary level, voltage and frequency parameters are controlled using a prediction model based on uncertainties as well as the use of optimization tools along with the optimal performance of energy storage units. At the secondary level, voltage and frequency deviation relative to reference values are controlled using an intelligent control method based on neural network algorithm and analysis of micro-grid data. The effective factors in the proposed model are the analysis of the dynamic behavior of the hybrid energy storage system (HESS) and the backup diesel generator in the active and reactive power distribution. In addition, the proposed model is based on the use of communications infrastructure. The simulation and experimental results of the proposed method along with control algorithms have been analyzed to evaluate the efficiency of this method. Fig. 1 shows the proposed control structure in this paper.

The main innovations of this article are summarized below:

  • 1

    Presenting a robust voltage and frequency control model based on predictive control structure at the primary level and INN algorithm at the secondary level by considering the uncertainty of distributed generation sources, loads, communication delays and the role of flexible loads in managing critical scenarios.

  • 2

    Presenting the optimal active and reactive power distribution model and intelligent energy management between power generation sources and HESS in order to respond appropriately to the control of voltage and frequency fluctuations in both primary and secondary levels by considering the charge, discharge and SOC thresholds of HESS.

  • 3

    Increasing the speed and accuracy of the criteria to compensate for instantaneous voltage and frequency fluctuations, as well as reducing the standard deviation from the reference values of voltage and frequency by intelligent and predicted structure.

  • 4

    Increasing detection speed as well as high accuracy in dealing with events affecting the operation of the micro-grid

The rest of the article includes sections: In section 2, modeling of micro-grid parameters along with the control method in primary and secondary is stated. Section 3 describes the simulation and implementation requirements. Section 4 describes the results. Finally, a conclusion is presented in Section 5.

Section snippets

Main grid model

Eqs. (1)-(2) show the power flow model between the micro-grid and the main network as well as the critical voltage prediction model of the micro-grid during operation based on the linear model by Taylor approximation [21], [22]. In Eq. 2, the sensitivity coefficientsvcriticaliPT, vcriticaliQT are calculated according to the Ref [22].[PgridminP(t)PgridmaxQgridminQ(t)Qgridmax],t(Pgrid(t))2+(Qgrid(t))2(Sgridmax)2V^criticali[k]=[Vcriticali[0]+vcriticalipDGT·ΔPDG[k]+vcriticaliQDGT

Simulation and performance requirements

In this section, the requirements and execution algorithms in the simulation section are presented. Thus, as shown in Fig. 5, the network is considered part of the distribution network 20 / 0.4KV Rajaee Port in Iran. In order to standardize the existing network and use control protocols, changes have been made in the network to improve the performance of the proposed method. The technical specifications of micro-grid is presented in Table 5.

Scenario 1: Voltage and frequency restoration based on micro-grid disconnection from main grid (load fluctuations)

Hypothesis:

  • Diesel generator start-up time is between 5 to 8 minutes and distributed generation sources along with HESS unit are able to supply the loads in this time.

  • The control method is examined in micro-grid number 1.

  • The HESS has 55% of the initial charge.

  • Disconnection and connection of power generation sources in minutes and the performance of controllers in seconds is considered.

At t=5s, the micro-grid is disconnected from the main grid and continues to operate in island mode. In this

Conclusion

In this paper, an intelligent method based on MPC and INN algorithm to control the voltage and frequency at the primary and secondary control levels is described. Accordingly, the voltage and frequency control at the primary is considered by the MPC based on the uncertainties of power generation sources and load fluctuations. At the primary level, the voltage and frequency control model is presented as a multi-objective optimization problem in a multi-time period according to the effective

Authorship Statement

All persons who meet authorship criteria are listed as authors, and all authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing, or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in the Journal of ‘Electric Power

CRediT authorship contribution statement

Reza Sepehrzad: Conceptualization, Methodology, Software, Investigation, Validation, Writing – original draft, Formal analysis. Amirhossein Mahmoodi: Writing – review & editing, Software, Formal analysis. Seyedeh Yosra Ghalebi: Writing – review & editing, Visualization, Investigation, Formal analysis. Ali Reza Moridi: Software, Investigation, Writing – review & editing. Ali Reza Seifi: Supervision, Project administration, Investigation, Conceptualization, Methodology.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Reference (39)

  • Bing Yan et al.

    Markovian-based stochastic operation optimization of multiple distributed energy systems with renewables in a local energy community

    Electric Power Syst. Res.

    (2020)
  • M. Eslahi et al.

    Day-Ahead scheduling of centralized energy storage system in electrical networks by proposed stochastic MILP-Based bi-objective optimization approach

    Electric Power Syst. Res.

    (2021)
  • Ayse Selin Kocaman et al.

    A stochastic framework to evaluate the impact of agricultural load flexibility on the sizing of renewable energy systems

    Renew. Energy

    (2020)
  • M. Khalid et al.

    An optimal operation of wind energy storage system for frequency control based on model predictive control

    Renew. Energy

    (2012)
  • Roozbeh Sangi et al.

    A novel hybrid agent-based model predictive control for advanced building energy systems

    Energy Convers. Manage.

    (2018)
  • Yao Li et al.

    An MPC based optimized control approach for EV-based voltage regulation in distribution grid

    Electric Power Syst. Res.

    (2019)
  • N. Hatziargyriou et al.

    Micro-grids

    IEEE Power Energy Mag

    (2007)
  • Ahmed Saad et al.

    Data-centric hierarchical distributed model predictive control for smart grid energy management

    IEEE Trans. Ind. Inf.

    (2018)
  • Frate, Guido Francesco, Lorenzo Ferrari, and Umberto Desideri. "Energy storage for grid-scale applications: Technology...
  • Cited by (25)

    View all citing articles on Scopus
    View full text