Title of Research Project - Massachusetts Institute of ...



Grid-connected Photovoltaic Power Systems: Modeling And Topology Studies

MIT/MIST Collaborative Research Progress Report for Period September-2011 to February-2012

Principle Investigator at MIT: Prof. J. Kirtley and Prof. D. Perreault

Principle Investigator at Masdar Institute: Prof. V. Khadkikar and Prof. W. Xiao

Research Project Start Date: September 1, 2010

Introduction

Background

This research project focuses on the integration of PV solar power to the existing electric grid and addresses issues and possible solutions related with such interconnection. In this research work following aspects will primarily be considered:

• To increase the performance of photovoltaic systems through the use of innovative system structures, advanced control and optimal power converter interfaces.

• To develop an accurate computer simulation model of a grid connected PV plant with necessary controllers to perform maximum power extraction.

• To study the potential issues as well as benefits if one or more renewable systems (combination of solar and/or wind plants) are connected on the same distribution feeder and/or nearby feeders.

It is our intention that this project will help (i) to better understand the influence of solar plant local distribution network, (ii) improve the overall performance of solar plants.

Objective and Approach

In the third quarter of this research project (September-2011 to February-2012), following research objectives are being undertaken:

• Construction of a DAB converter prototype with the modulation strategies and test the benchmark prototype under different load case including rated load, heavy and light load cases. (PI: Dr. Xiao, MI and Dr. Perreault, MIT)

• Analysis of distributed power loss in DAB converter and elimination the reactive power. Analysis of the dc bus surge voltage and comparative evaluation of DC-Link Capacitors in DAB converter. (PI: Dr. Xiao, MI and Dr. Perreault, MIT)

• An LCL filter design to enhance the performance of PV system with an active damping controller. (PIs: Dr. Khadkikar, MI and Dr. Kirtley, MIT)

• New control scheme to coordinate proposed interline-photovoltaic (I-PV) system. (PIs: Dr. Khadkikar, MI and Dr. Kirtley, MIT)

Executive Summary

This research project is based on the grid connected photovoltaic (PV) solar power plant. Two teams are involved out of which one team (Dr. Perreault, MIT and Dr. Xiao, MI) is focusing on the device level. The other team (Dr. Kirtley, MIT and Dr. Khadkikar, MI) is studying the integration aspects of PV solar plant to the main grid. During the third quarter of the joint research work, the following aspects are accomplished or under progress:

• Regarding to the subproject “Design and prototyping of Dual Active Bridge Converter”, Two prototypes of DAB converter have been constructed. One prototype is used to evaluate the distributed power loss for the student thesis. The benchmark system has been tested under different load case including rated load, light load cases. Another prototype is used by Dr.Wen for loss reduction study. The modulation strategies have been implemented and the power loss analysis has been finished both for aspects of the hardware and software. The reactive power calculation has been analyzed and the way to eliminate reactive power in DAB converter is also investigated. The team has been working with several subtasks, such as reducing DC surge voltage/current, maximum power point tracking for PV systems, and high frequency AC power supply systems. This work results in several publications shown in the Publication section.

• Dr. Perreault, please add here

• PV system modeling and control: An LCL filter is designed to enhance the performance of PV system. Furthermore, in order to overcome the problem of possible resonance condition between L-C-L a new active damping technique is proposed. (Dr. Khadkikar and Dr. Kirtley)

• Multifunction PV solar power plant operation: In the previous report, we have proposed a new system configuration for a large-scale PV solar plant, called as Interline-PV (I-PV) system. A new droop control method, given a name P-Q-V droop controller, to regulate the grid voltage using the I-PV system is developed. (Dr. Khadkikar and Dr. Kirtley)

• A theoretical study of network reconfiguration of power distribution system is underway. The goal of this research is to propose a reliable and efficient topology of distribution network. Benefits, such as deferring the investment, reducing operation costs and improving energy efficiency of distribution systems, can be attained from the work. This work is carried out by Ms. Jiankang Wang under the supervision of Dr. Kirtley.

Research Tasks

Task-1: Benchmark System Test and Power Loss Distribution Analysis (PI: Dr. Xiao, MI)

An experimental prototype is implemented using DSP 2808 control board. Main parameters are shown as following: the input voltage Vin = 32V, the input voltage range 28V – 36V, the nominal output voltage Vo = 320V, the maximum deliver power Pmax = 250W and the switching frequency 200 kHz. This work is delayed by approximately by 2 to 3 months due to the procurement process and budget freezing in summer. The research and prototyping is accordingly delayed due to the order of key components such as magnetic core and many experimental apparatus including adjustable DC power supply, adjustable load and impedance analyzer. Different load case including rated load, heavy and light load cases will be tested for this prototype and some experimental results will be presented to compare with the theoretical analysis. During the waiting time, the MI team was working with other related study including the effect of parasitic inductance, PV system modeling, Maximum power point tracking for PV systems. DC link capacitor optimization, and high frequency AC power supply systems. The PV modeling work was published by IEEE. Trans. On Sustainable energy. The voltage surge reduction work is accepted for publication on the IET journal. Another two journal submissions are waiting for review results.

Task-2: An LCL filter design to enhance the performance of PV system with an active damping controller. (PIs: Dr. Khadkikar, MI and Dr. Kirtley, MIT)

This task is part of Objective-1 as mentioned in the original proposal. In this work we are developing an active damping method for an LCL filter based PV system to annul the effect of resonance on the system performance. This work has been delayed from the beginning of the project. However, the work has been progressing since from the hiring of Post-doc (Dr. Moin Hanif) for this project.

Task-3: Multifunction PV System: study potential benefits of one or more renewable energy systems connected two different feeders. (PIs: Dr. Khadkikar, MI and Dr. Kirtley, MIT)

In the last two reports, we have provided initial results on a new system configuration that we have proposed for a large-scale PV system. Here idea is to connect a large-scale PV system to two different feeders/networks. The newly proposed system thus can enable to manage power flow on the two different feeders through PV solar power plant inverters. The inverter modules in a PV power plant are configured such that the system is represented as a back-to-back inverter connected multi-line system, called as Interline-PV (I-PV) system. In this report, newly developed P-Q-V droop control method for I-PV system is presented. This task is mentioned as “Part A: Objective-2” in the original proposal.

Current Reporting Period Summary (March 2011 to August 2011)

Task 1: Weidong, please add here (Dr. Perreault and Dr. Xiao)

Research team members: MIT: Dr. Perreault; MI: Dr. Xiao, Dr.Wen, and Yosief Abraham (graduate student)

The high circulating current in the transformer is the drawback of the DAB converter. At light load, this circulating current increases the conduction losses and causes the converter to lose the natural zero-voltage switching feature. Thus, the test of the benchmark prototype under different load case including rated load, heavy and light load cases is necessary to comprehensively assess the performance of the prototype.

The simulation model based on PSIM and simulink has been created and compared with the experiment tests. Calculation and analysis of the DAB converter overall power losses will be performed. The mathematical model of the reactive power will be defined corresponding to different modulation strategies. Traditional phase-shift control strategies introduce large reactive power and contribute to large peak current and large system loss. Thus, the proposed control strategy will eliminate reactive power in DAB converter.

For DAB converter, the DC bus can experience large surge voltages due to the presence of parasitic inductance in addition to the hard-switching operations of voltage source inverters. The unexpected transient overvoltage can result in switching device failures. A surge voltage model need to be developed that takes into account the commutation loop inductances and the switching dynamics. An optimal design approach to lower the stray inductance will be presented. In DAB converter, sizing and selection of DC link capacitors involve tradeoffs among system performance including lifetime, reliability, cost, and power density. A comprehensive and comparative study on the DC-link capacitor applications and evaluations to meet the above requirements is investigated. The analysis should consider the facts of capacitor power loss, core temperature, lifetime, and the battery ripple current limit, which are critical for PV applications.

Please see Appendix-1 for more details.

Task 2: Development of PV Solar Plant Model (Dr. Kirtley and Dr. Khadkikar)

Research team members: MIT: Dr. Kirtley; MI: Dr. Khadkikar; MI Graduate students: Deema Al Baik and Ammar Elnosh; Post-doc: Dr. Moin Hanif

This work is intended to develop a generic PV solar power plant simulation model. The develop model will serve as a benchmark simulation model for future research work at MI and MIT.

Two MI graduate students, Ms. Deema Al Baik and Mr. Ammar Elnosh, are working on the PV solar plant model development. Ammar is focusing on developing an improved maximum power point tracking (MPPT) algorithm including PV solar array modeling. The MPPT work is ongoing and we expect to complete it in next 2-3 months. The details on MPPT will be provided in the next report.

Deema is also working on PV solar plant model development. However, her goal is on the AC side of the network. Part of her work was reported in the last report. Currently she is working on LCL filter design with an improved active damping method.

Dr. Moin Hanif has joined the MI research group in the month October-2011. He is also working on developing an active damping controller for LCL filter. Summary of ongoing work is given in Appendix-2.

Task 3: Multifunction PV System (Dr. Kirtley and Dr. Khadkikar)

Research team members: MIT: Dr. Kirtley; MI: Dr. Khadkikar; MI Graduate student: Ahmed Moawwad

At the beginning of this project, we (Dr. Khadkikar and Dr. Kirtley) have proposed a new system configuration for a large-scale PV plants, given a name as “Interline PV (I-PV)” system. One of the graduate students at Masdar Institute (Ahmed Moawwad under supervision of Dr. Khadkikar) is working on detailed study on the proposed I-PV system. We have accomplished significant advancement in this research work and a journal paper has been submitted. A new droop controlled method, called as P-Q-V droop method is proposed to regulate the grid voltage at point of common coupling. The research findings are submitted for the possible publication in IEEE Transactions on Power Delivery. The submitted paper details are given in Appendix-3.

Task 4: High Efficiency DC-DC Converters

To connect photovoltaic panels to the grid, interface circuitry is needed. In the architecture pursued in this project, DC/DC converters are used to boost voltage of individual photovoltaic panels to a high dc-link voltage, and one or more inverters are used to convert DC to AC. These DC/DC converters have to be designed with very high efficiency. In Fig. 1, the block diagram of a grid connected PV system is shown. The focus of this project is on the DC/DC power converter. In conventional, hard switched power converters the overlap of current and voltage is large during switching resulting in significant power loss. Soft switching is achieved by resonant topologies which decrease switching losses by Zero Voltage Switching (ZVS) or Zero Current Switching (ZCS). However, the magnetic loss increases with the additional magnetics needed for soft switching and a tradeoff of these two losses is necessary to achieve minimum overall loss.

[pic]

Figure 1: Block diagram of a grid-connected PV system

Details of this work are in Appendix-4.

Unscheduled Task: Distribution System Reconfiguration

A strong motivation of the research topic is generated by the critical role of distribution network in the whole power system and the challenges it has been facing during the recent several decades. As the final stage of electricity delivery to end users, it is consist of over 65% of US power network [32]-[34]. Renewable energies installed in the form of Distributed Generation (DG) and demand response show their influence on the whole power system through the distribution system. The aging infrastructure, yearly increasing demand, and upcoming new technologies of distribution systems raise the question of how to operate and reinforce the distribution network in a reliable and economic way to fulfill all the requirements.

The problem is traditionally addressing by reinforcing distribution network based on the so-called “fit-and-forgot” policy [35]. The “fit and forget” approach implies the design of the distribution system so as to meet technical constraints in the most onerous conditions (e.g. full generation/minimum load or no generation/full load) even if such situations have a small probability of occurrence. One advantage of this approach is that control problems were solved at the planning stage. However, under rapidly growing demand and targets of integrating DG, this practice of passive operation can be very costly or it will limit the capacity of distributed generation that can be connected to an existing system [36].

In contrast, network reconfiguration, which alters the network topological structure by changing the open/close status of the sectionalizer and tie switches, actively maximizes the use of existing circuits [37].With reconfiguration, distribution systems can keep the redundancies of extra feeders for contingent cases in long-term and operate in optimal radial structure. In addition to existing control on generation and demand, reconfiguration adds another control dimension by making network structure a dispatchable resource.

In spite of many studies on operation strategies of reconfiguration, there are few of them considering reconfiguration in the planning stage [38], [39]. However, since reconfiguration can be resource of improving operation, considering its role in planning has a great potential to save capital investment in generators, transformers and etc. In addition, the reconfiguration flexibility of a distribution system is determined at its planning stage. Planning reconfiguration capability together with other power distribution components may save cost and improve efficiency of operation.

During this reporting period, my study mainly includes using reconfiguration to (1) guarantee the required reliability while reducing the capital investment in distribution system planning; (2) mitigating the deterioration of voltage profile caused by DG insertion. Comprehensive literature review about reconfiguration algorithms and distribution system automation was conducted. Methodologies and results of the studied two problems are summarized in the following sections. Simulations have been conducted to verify the results.

2 Objectives and Methodologies

2.1 Objectives

The overall objectives of this research can be summarized in the following points:

• Proposing and demonstrating a planning scheme that fulfills the long-term and short-term requirements of distribution system.

• Proposing a closed-loop planning scheme that consider and optimize the operation flexibility induced by reconfigurable network structure.

• Investigating into the controllability of distribution system induced by reconfiguration.

These objectives will be addressed at the close of this research. The challenges to modern distribution networks can be interpreted as the conflicts of its short- and long-term requirements. In short-term, we want distribution networks to operate efficiently with fewer redundancies planned as possible. In long-term, however, the redundancies are needed to ensure reliable performance of networks with uncertainties induced from DG and time-varying loads. Deploying reconfiguration can exploit the existing circuits and mitigate many operation problems of distribution systems; meanwhile, it retains multiple configurations that can satisfy various system conditions. Due to these benefits, some investment may be unnecessary given reconfigurable networks. And these benefits are expected to be maximized in both short- and long-term if networks’ reconfiguration ability is well planned. Therefore, exploring reconfiguration’s effects on distribution system operation and optimizing it at the planning stage will close the loop.

2.2 Methodologies

During this reporting period, my study mainly includes using reconfiguration to (1) guarantee the required reliability while reducing the capital investment in distribution system planning; (2) mitigating the deterioration of voltage profile caused by DG insertion. Their methodologies are discussed separately in the following subsections.

2.2.1 Optimizing Reconfiguration’s Impacts on Distribution System Planning

This section briefly states the methodology of the first problem, using reconfiguration to ensure required reliability while reducing distribution system planning cost.

In last reporting period, I have shown that reconfiguration can provide larger contingency support neighborhood and thus save huge investment on transformers’ capacity and reduce the no-load loss of substations. A paper was published under the topic [40].

Following the study, customer service quality indicated by reliability of distribution systems planned is taken into account in the reporting period. Reconfiguration generates a larger contingency support neighborhood that in turn pushes upward the utilization rates of equipments. However, the probability that an outage will occur among the designated support units is increased at the same time [41].

Suppose that the area of the power system being considered has 88.5% loadings on all transformers, instead of 66%. In that case, when any transformer fails, and if the utility is to keep within a 133% overload limit, a failed unit's load has to be spread over two neighboring transformers, not just one. The size of the "contingency support neighborhood" for each unit in the system has increased by a factor of fifty percent. In a system loaded to 66%, there is only one major target. In a system or area of a system loaded to 88.5%, it occurs if a unit and either one of two neighbors is out. In an area of a system loaded to over 100% (as some aging areas are) it occurs whenever the unit and any one of three designated neighbors is out. But there are still Single Contingency Policy (SCP) (i.e. N-1) neighborhoods: each can tolerate only one equipment outage and still fully meet their required ability to serve demand. A second outage will very likely lead to interruption of service to some customers. In these larger neighborhoods, there are more targets for that second outage to hit. 'Trouble" that leads to an inability to serve customer demand is more likely to occur.

For this reason, reliability measured by System Average Interruption Duration Index (SAIDI) is used to constrained the distribution system plan proposed in study of last period [9]. LP models are developed to find the optimal decisions for the above problems.

The main implications of this study are: (1) enabling reconfigurable feeders can increase reliability with lower capital investment and operational cost when comparing to reinforcing transformer capacity; (2) the reliability of a set of well interconnected substations depends on the max-shortage of one of its substations, failure rates of units in the support neighborhood and size of the support neighborhood. The two implications disprove the statements in previous works [42], [43]. The approach used and the results obtained are currently being drafted in a paper.

2.2.2 Optimizing Reconfiguration’s Impacts on Distribution System Integrated with Distributed Generation

This section summarizes the methodology and results of the second problem, using reconfiguration to improve voltage profiles of distribution systems integrated with Distributed Generation (DG).

This study was initiated in July, 2011 and falls into two sub-problems: (1) quantifying the dependency of voltage profiles on DG’s insertion in terms of system voltage level, location, power capacity and dispersion level, and (2) demonstrating reconfiguration’s improvement on systems’ voltage profiles in theory and practice/ simulation.

When generators are connected to a feeder, voltage profile is likely to increase. According to ANSI, voltage of distribution systems are limited to [pic] in normal operation. Overvoltage of the standard will decrease lifetime and even cause failure of equipments [44] The generator voltage [pic] can be approximately given by [pic], where [pic] is the substation bus voltage, [pic] and [pic] are the active and reactive power injected in the generator bus, and [pic]and [pic]are the feeder resistance and reactance. To mitigate the voltage rise, traditional methods include to [14]:

• Reduce the primary substation voltage

• Allow DG to import reactive power

• Install auto transformers, or voltage regulators

• Increase the conductor size of feeders

• Constrain the generator at times of low demand

Compared the aforementioned methods, our study employs reconfiguration to changes the topology of distribution network, and thus to change[pic]and [pic] to affect the voltage profile. This method is verified with high implementation easiness, two to ten times low cost and 10-30% high effectiveness.

Until the reporting period, a DC circuit model, which may be extended to balanced 3-phase model easily, is built to for sub-problem (1), and simulations are conducted for sub-problem (2). The main implications of the study are:

1. Voltage profile of a feeder can be calculated with area criteria [46].

2. DG has greater impact on voltage profile of a feeder when inserted at farther end of the feeder. Therefore, to mitigate overvoltage induced:

a) Curtailment of output of DG’s at farther end of a feeder should be firstly considered.

b) A demand response program should be designed to give stronger incentive on customers at farther end of a feeder.

3. The maximum value of voltage on a feeder can be calculated with the proposed analytical expression, and is determined at the position where current changes direction and supplement current density is negative.

4. Dispersion level of DG can have either positive or negative impact on voltage profiles, depending on DG’s location on a feeder.

A paper is drafted to include the results. Theoretic part of sub-problem (2) is under development.

This study is a part of reconfiguration’s impacts on distribution system operation, which aims at investigating the relation of system topological flexibility and reconfiguration benefits. Results of the study are the basis of proposing a closed-loop planning scheme, which is the final purpose of my thesis.

3 Conclusion

Deploying reconfiguration can exploit the existing circuits and mitigate many operation problems of distribution systems; meanwhile, it retains multiple configurations that can satisfy various system conditions. Due to these benefits, some investment may be unnecessary given reconfigurable networks. And these benefits are expected to be maximized in both short- and long-term if networks’ reconfiguration ability is well planned.

Future Work (March 2012 to August 2012)

Task 1: Comprehensive evaluation of the Benchmark prototype and analysis of the power loss and reactive power

Research team members: MIT: Dr. Perreault; MI: Dr. Xiao, Dr. Wen, and Imran Syed (graduate student)

This part of the project involves three main areas which include:

• Testing

• Power loss analysis

• Reactive power analysis

• Testing: Dr. Huiqing Wen (MI) and Imran Syed (MI) under supervision of Dr. Xiao (MI). The study will focus on the DC microgrid applications to adopt on-site PV generation and battery backup systems. The pilot project will be located in the field station of Masdar Institute. Detailed work can be found in Apendix 1-1.

• Power loss analysis: Dr. Huiqing Wen (MI) and Imran Syed (MI) under supervision of Dr. Xiao (MI).The Masdar team will focus on the embedded software implementation

• Reactive power analysis: Dr. Huiqing Wen (MI) and Imran Syed (MI) under supervision of Dr. Xiao (MI).

Task 2: Novel modulation strategy investigation for DAB converter

Research team members: MIT: Dr. Perreault; MI: Dr. Xiao, Dr. Wen, and Imran Syed (graduate student)

This work focuses a simplified dual-phase-shift (SDPS) control strategy for DAB converter in whole operation range is analysis. The work includes analysis the analytical expression of the average output power, the reactive power, the rms and peak current based on the switching strategy. The soft-switching conditions will be analyzed and compared with the traditional PS control. The DAB converter power loss will be calculated and the algorithm to minimize the total power loss will be discussed. Simulations and experiments are expected to be carried out to verify the analysis. Details can be found in appendix 1-2.

Task 3: Development of PV Solar Plant Model (Dr. Kirtley and Dr. Khadkikar)

Research team members: MIT: Dr. Kirtley; MI: Dr. Khadkikar; MI Graduate students: Ammar Elnosh and Deema Al Baik, Post-doc: Dr. Moin Hanif

This work is delayed by 4 to 6 months. However, there is a steady progress since the hiring of Dr. Moin Hanif. The project objective-3 in the original proposal is to validate some of the research findings using a hardware PV system. We expect the installation of a 5kW PV system at Masdar Institute (using actual PV panels) will be done in the month of March/April – 2012.

In the next stage of the project we plan to finalize the following research aspects:

• Development of improved MPPT technique (student involved: Ammar)

• Active damping controller to improve the harmonics generated by the PV solar plant (student involved: Deema and Dr. Moin)

• Perform the initial experimental studies to validate some of the research objectives. (Dr. Moin and Dr. Khadkikar)

Task 4: Multifunction PV System (Dr. Kirtley and Dr. Khadkikar)

Research team members: MIT: Dr. Kirtley; MI: Dr. Khadkikar; MI Graduate student: Ahmed Moawwad

This task is almost completed. So far, two conference papers are resulted from this research work. A journal paper is in under review. Furthermore, we will be concluding this task by providing an additional functionality where an unbalance in the grid voltages will be compensated using I-PV system. An additional conference is expected from this last piece of the research work.

Publications/Presentations 

1. V. Khadkikar and J. Kirtley, “Interline Photovoltaic (I-PV) Power System – A Novel Concept of Power Flow Control and Management”. In the conference proceedings of IEEE PES General Meeting, 24-28 July, 2011, pages 1-6. (Included in Report-I)

2. A. Moawwad, V. Khadkikar, and J. Kirtley, “Photovoltaic Power Plant as FACTS Devices in Multi-Feeder Systems”. In the conference proceedings of Industrial Electronics conference (IECON-2011), 7-10 Nov. 2011, pages 1-6. (Included in Report-II)

3. D. Al-Baik and V. Khadkikar, "Effect of variable PV power on the grid power factor under different load conditions". In the conference proceedings of IEEE Electric Power and Energy Conversion Systems (EPECS), 15-17 Nov. 2011, pages 1-5. (Included in Report-II)

Publications/Presentations 

[1] Y. Mahmoud, W. Xiao, and H. H. Zeineldin, "A Simple Approach to Modeling and Simulation of Photovoltaic Modules," IEEE Trans. Sustainable Energy, vol. 3, pp. 185-186, 2012. (Published).

[2] H.Wen, W. Xiao, H. Li, and X. Wen, "Analysis and Minimization of DC Bus Surge Voltage for Electric Vehicle Applications," Electrical Systems in Transportation, IET, 2011. (Accepted)

[3] H. Wen, W. Xiao, X. Wen, and P.R. Armstrong, "Analysis and Minimization of DC-Link Capacitance for Electric Vehicle Application," IEEE TRANSACTION ON VEHICULAR TECHNOLOGY, 2011. (The Second Review)

[4] H.Wen, W. Xiao, and Z. LU, "Current-Fed High-Frequency AC Distributed Power System for Medium-High Voltage Gate Driving Applications," IEEE Transactions on Industrial Electronics, 2011. (Under Review)

[5] Y. Mahmoud, W. Xiao, and H. H. Zeineldin, “A New Parameterization Method for Single Diode photovoltaic Models”, submitted to IEEE Trans. Sustainable Energy in Jan 2012, (Under Review).

[6] H. Wen, W. Xiao, Xuhui Wen, “Comparative Evaluation of DC-Link Capacitors for Electric Vehicle Application," accepted for the presentation in the 21th IEEE International Symposium on Industrial Electronics (ISIE) which will take place in Hangzhou, Zhejiang, China on May 28-31, 2012.

[7] Huiqing Wen, Weidong Xiao, Han Li, and Xuhui Wen, “Analysis and Minimization of DC Bus Surge Voltage for Electric Vehicle Applications," accepted for the presentation in the 21th IEEE International Symposium on Industrial Electronics (ISIE) which will take place in Hangzhou, Zhejiang, China on May 28-31, 2012.

[8] W. Xiao, H. Wen, and H.H. Zeineldin, “Affine Parameterization and Anti-Windup Approaches for Controlling DC-DC Converters”, accepted for the presentation in the 21th IEEE International Symposium on Industrial Electronics (ISIE) which will take place in Hangzhou, Zhejiang, China on May 28-31, 2012.

[9] Y. H. Abraham, X. Weidong, W. Huiqing, and V. Khadkikar, "Estimating power losses in Dual Active Bridge DC-DC converter," in Electric Power and Energy Conversion Systems (EPECS), 2011 2nd International Conference on, 2011, pp. 1-5 (Published).

[10] W. Xiao, A. Elnosh, V. Khadkikar, and H. Zeineldin, "Overview of maximum power point tracking technologies for photovoltaic power systems," in 37th Annual Conference on IEEE Industrial Electronics Society, 2011, pp. 3900 (Published).

Appendices (If Absolutely Necessary)

Appendix-1: Benchmark System Test and Power Loss Distribution Analysis (PI: Dr. W Xiao)

The first prototype of DAB converter has been constructed in the Laboratory of MIST. Tests have been done for different phase shift angle and load. An experimental prototype was implemented using DSP 2808 control board. Main parameters are shown in table I and figure 1 shows the schematic of a single phase DAB converter.

Table I

Circuit Parameters of the Dual Active Bridge DC-DC Converter

|Rated Power |[pic] |300 |

|Rated Input Voltage |[pic] |32V |

|Rated output Voltage |[pic] |320V |

|Input Capacitor |[pic] |100uF |

|OutputCapacitor |[pic] |22uF |

|Transformer turn ratio |[pic] |1:6 |

|Inductor |[pic] |2.5uH |

|Frequancy |[pic] |200KHz |

|conversion ratio |[pic] |1 |

[pic]

Fig. 1 Circuit of single phase dual active bridge DC/DC converter

A simulation model for a microgrid has been created using PSIM/Simulink. Ideal voltage and current waveforms using the Phase-Shifted modulation for the boost and buck mode of operation have been simulated and analyzed. Different load case including rated load, heavy and light load cases are test for this prototype and some experimental results are presented to compare with the theoretical analysis. Figure 2 and figure 3 show the ideal voltages and current waveforms using the PS-M for the hard switching mode and the soft switching mode of operation.

[pic]

Fig.2 Ideal voltages and current waveforms using the PS-M for the hard switching mode of operation.

[pic]

Fig.3 Ideal voltages and current waveforms using the PS-M for the soft switching mode of operation.

[pic]

Fig.4 Ideal voltages and current waveforms using the PSPM for [pic]

[pic]

Fig.5 Ideal voltages and current waveforms using the PSPM for [pic]

[pic]

Fig.6 Ideal voltages and current waveforms using the PSPM for [pic].

[pic]

Fig.7 Ideal voltages and current waveforms using the PSPM for [pic].

[pic]

Fig.8 Ideal voltages and current waveforms using the PSPM for [pic].

[pic]

Fig.9 Ideal voltages and current waveforms using the PSPM for [pic].

The high circulating current in the transformer is the drawback of the DAB converter. At light load this circulating current will increase the conduction losses and causes the converter to lose the natural zero-voltage switching feature. Thus the test of the benchmark prototype under different load case including rated load, heavy and light load cases is necessary to comprehensively assess the performance of the prototype. The simulated waveform is shown in the following Figures including traditional PS-M and other modulation strategies.

[pic]

Fig. 10 Waveforms using the conventional control strategy for phase angle 90°

[pic]

Fig.11 Waveforms using the conventional control strategy for phase angle 45°

[pic]

Fig.12 Waveforms using the conventional control strategy for phase angle 30°

[pic]

Fig.13 Waveforms using the conventional control strategy for phase angle 15°

[pic]

(a)

[pic]

(b)

Fig.14 Waveforms using the novel modulation strategy

Test Setup is shown as following:

[pic] [pic]

a) Dab prototype (b) DSP control board

[pic] [pic]

(c) Power supply (d) Load

Fig.15 Experiment Test Setup

The efficiency of the DAB depends on power transfer , the input and output voltages and the difference between them for example if the voltage [pic] drops along with discharge of the energy storage device , power loss increases at a given power transfer . Experimental Waveforms are shown as following: The power transferred at the given value [pic] and phase shift angle of 45o for input voltage [pic]and output voltage [pic] , is found as 135.5 W theoretically .

[pic]

Fig.16 Experiment Result1(Vout=106.7W,d=0.8182)

[pic]

Fig.17 Experiment Result2(Vout=91.8W,d=0.8734)

Appendeix 1-2: Analysis the distribution of power loss in DAB converter and elimination the reactive power. Analysis of the dc bus surge voltage and comparative evaluation of DC-Link Capacitors in DAB converter. (Dr. Perreault and Dr. Xiao)

Research team members: MIT: Dr. Perreault; MI: Dr. Xiao, Dr. Wen, and Yosief Abraham (graduate student)

The expression of leakage inductor current for each time interval has also obtained. Soft-switching constraints for both bridges are also derived. Power losses in MOSFETs, diodes and transformer are calculated. The power losses in DAB can be classified as the switching and conduction losses. The switching losses in semiconductor devices are due to continuous switching (on and off) transitions during which a device is simultaneously exposed to high voltage and current. Although the DAB naturally operates in zero voltage switching (ZVS), it operates in ZVS only during “ON” switching transition. The conduction losses in DAB can further be classified as the conduction losses that occur in the semiconductor devices (MOSFETs) and the transformer and inductor losses. The switching losses in MOSFETs can be calculated from the following formula:

[pic] (1)

Where fsw is the switching frequency, VDS and IDS are the voltage and current at the time of switching, respectively, and ton and toff are time intervals during switching ON and OFF, respectively. Note that the power loss Psw represents the amount of power loss for one switch per leg during the ON switching transition. Similarly, the losses for the OFF transition can be calculated. Hence to calculate the total switching losses for bridge 1, for example, we can multiply Psw by 4 assuming the bridge has either on state or off state losses. The switching transition time can be calculated with the help of the device data sheet. To calculate the conduction losses, the losses due to the R_DS(on) and the voltage drop across the anti-parallel diode are based on the data sheet. For example, for bridge 1, on resistance of the MOSFET, R_DS(on) =8.2mΩ and the anti-parallel diode forward voltage drop, V_f=1.2V. Hence the conduction losses can be calculated in as:

[pic] (2)

[pic] (3)

The RMS current values for both the body diodes and the MOSFETs are calculated from the inductor current waveform depicted through the respective components. Fig.4 shows the different conduction time intervals for the switch .The dead time has a life time D1 hence the body diode conducts during this time interval contributing to the power losses , in the remaining segments(D2, D3, D4) the MOSFET conducts. The RMS currents for both the body- diode and MOSFET are estimated using (4) and (5).

[pic] (4)

[pic] (5)

[pic]

Fig.18. The different conduction time intervals of the inductor current through the bridges

The reactive power is defined corresponding to different modulation strategies and the corresponding equations are derived. Traditional phase-shift control will induce large reactive power and contribute to large peak current and large system loss. The control strategy to eliminate reactive power in DAB converter is investigated.

Considerably minimized converter power loss can be achieved with the use of alternative modulation strategies, which are basically originated from the variation the duty ratio of the H-bridge output voltage and the phase shift angle. A phase-shift plus pulse width modulation (PSPWM) control strategy is presented in [3] and one more degree of control freedom is added to extend the ZVS range. A optimal selection of phase shift angle and modulation duty ratio, useful to minimize overall converter loss, is analyzed in [4]. But the duty ratio of the gate signals is variable and calculated online, besides, the calculation of duty ratio is dependent on the plow flow direction and buck/boost operation modes. All these factors result in the complexity of the PSPWM control. A double-phase-shift (DPS) control is proposed to reduce the reactive converter power [5]. This control includes two phase-shift angles, which are phase-shift between the primary and secondary side of the transformer and phase-shift between the gate signals of the diagonal devices of the same side. But it doesn’t full consider the efficiency improvement, besides, the analysis of reactive power is not completed due to the complexity. The experimental comparison of PS and DPS control is presented in [6]. The efficiency improvement is not good as expectation in [5] due to the lack of systematic analysis of power loss distribution. In [7], the phase-shift trajectories for the minimal reactive power, the minimal rms and peak current are analyzed. But due to more control parameters and complex operation modes, the expression of reactive power is extreme complicated and the optimal design is not easy to implement. Thus, a simplified dual-phase-shift (SDPS) control strategy for DAB converter in whole operation range is analysis. The analytical expression of the average output power, the reactive power, the rms and peak current are derived based on the switching strategy. The soft-switching conditions are analyzed and compared with the traditional PS control. The DAB converter power loss is calculated and the algorithm to minimize the total power loss is proposed. Simulations and experiments are carried out to verify the analysis.

In Fig.16 the experimental inductor current wave form, [pic], output Voltage Vo and output current Io are shown .The experimental results have shown quite a big gap from theoretical calculations mainly due to the exclusion of the inductor and transformer losses .As it is shown in Fig. 5 the efficiency for the experimental converter at 45o is 91% , the theoretical efficiency at this particular operating point is calculated to be 95.19%.. The conduction losses [pic] and the switching losses [pic].The power transfer at the mentioned phase shift angle is [pic]. In this analysis the switching loss is seen to be a slightly higher than the conduction losses.

Fig. 19 and Fig. 20 shows the SDPS control strategy for the case of the phase-shift ratio D ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download