1

Performance Evaluation of A PHEV Parking Station Using Particle Swarm Optimization Wencong Su1, Student Member, IEEE, and Mo-Yuen Chow2, Fellow, IEEE

Abstract-- There is expected to be a large penetration of Plugin Hybrid Electric Vehicles (PHEVs) into the market in the near future. As a result, many technical problems related to the impact of this technology on the power grid need to be addressed. The anticipating large penetration of PHEV into our societies will add a substantial energy load to power grids, as well as add substantial energy resources that can be utilized. There is also a need for in-depth study on PHEVs in term of Smart Grid environment. In this paper, we propose an algorithm for optimally managing a large number of PHEVs (i.e., 500) charging at a municipal parking station. We used Particle Swarm Optimization (PSO) to intelligently allocate energy to the PHEVs. We considered constraints such as energy price, remaining battery capacity, and remaining charging time. A mathematical framework for the objective function (i.e., maximizing the average State-of-Charge at the next time step) is also given. We characterized the performance of our PSO algorithm using a MATLAB simulation, and compared it with other techniques. Index Terms--PHEV, Smart Grid, Particle Swarm Optimization, Intelligent Energy Management

E

I. INTRODUCTION

CONOMIC and environmental incentives, as well as advances in technology, are reshaping the traditional view of power systems. Plug-in Hybrid Electric Vehicles (PHEVs) have received the increased attention because of their low pollution emissions and low cost per mileage. Ultimately, PHEVs will shift energy demands from crude oil to electricity for the individual transportation sector [1]. By drawing on and supplying power to the power grid, electric vehicles could displace the use of petroleum. This would reduce pollution and alleviate security issues related to oil extraction, importation, and combustion. Furthermore, powered parking structures have the ability to ease the pain of the energy crisis [2]. PHEVs could also improve the financial viability and technical performance of the electric utility industry, as well as serve as a source of revenue for their owners [3]. This work was supported in part by the National Science Foundation, Award number: EEC-08212121 and this work is a part of an ongoing project in collaboration of the FREEDM systems center (Future Renewable Electric Energy Delivery and Management) with ADAC (Advanced Diagnosis Automation and Control) Lab at North Carolina State University and ATEC (Advanced Transportation Energy Center). 1 Wencong Su is with the Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC, 27606, USA 2 Mo-Yuen Chow is with the Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC, 27606, USA Email: [email protected] [email protected]

978-1-4577-1002-5/11/$26.00 ©2011 IEEE

The Electric Power Research Institute (EPRI) projects that 62% of the entire U.S. vehicle fleet will consist of PHEVs by 2050 (moderate penetration scenario) [4]. Accordingly, there is a growing need to address the implications of this technology on the power grid. Large numbers of PHEVs have the potential to threaten the stability of the power system. For example, the load on the power grid will need to be managed very carefully in order to avoid interruption when several thousand PHEVs are introduced into the system over a short period of time (e.g., during the early morning hours when people arrive at work). Moreover, due to differences in the needs of the PHEVs parked in the deck at any given time, the demand pattern will also have a significant impact on the electricity market. In order to maximize customer satisfaction and minimize disturbances to the grid, a sophisticated controller will need to be designed in order to allocate power appropriately. This controller must take into consideration real-world constraints (i.e., communication and infrastructure variations among individual vehicles.) The controller must also accommodate for differences in arrival/departure times, as well as the number of PHEVs in the parking deck. The algorithm should also be robust to uncertainty, be capable of making decisions in real-time with limited communication bandwidth, and work seamlessly with existing utilities. We have previously studied a theoretical system of PHEVs in a municipal parking deck [5-7]. This initial work was validated by implementing the algorithm in Matlab/Simulink and Labview. Communication between the central controller, PHEV chargers, and vehicles was achieved using the ZigBee protocol [6]. Fig. 1 highlights the main problems we are currently trying to address with specific technology solutions. We are currently evaluating the performance different control strategies in a real-world PHEV parking deck operating under various energy constraints. Specifically, we are investigating the use of the Particle Swarm Optimization (PSO) method for developing real-time, large-scale optimizations for allocating power. PSO is an iterative stochastic optimization method that has recently been used to solve a variety of problems in different application areas [8]. The remainder of this paper is organized as follows: Section II will describe the specific problem that we are trying to solve. We will provide the optimization objective and constraints, as well as show the mathematical formulation of our algorithm. Section III will review the PSO method, as well as describe how the algorithm works for optimization

2

problems. The results of our simulation, as well as further analysis of the parking deck scenarios, are presented in Section IV. Finally, we summarize our paper and provide a brief discussion of future work in Section V.

pay at time step k; w i ( k ) is the charging weighting term of the i-th PHEV at time step k (this is a function of the energy price, the remaining charging time, and the current SOC); SoCi ( k + 1) is the state of charge of the i-th PHEV at time step k+1. The weighting term gives a reward proportional to the attributes of a specific PHEV. For example, if a vehicle has a lower initial SOC and less remaining charging time, but the driver is willing to pay a higher price, the controller allocates more power to this PHEV charger: (2) wi (k ) ∝ [Cr ,i (k ) + Di (k ) + 1/ Tr ,i (k )] Since the three terms Cr ,i (k ) , Di ( k ) , 1/ Tr ,i ( k ) are not on the same scale, we need to normalize all of the terms in order to assign similar importance to each: Cr ,i (k ) − Min[Cr ,: (k )] cr ,i (k ) = Max[Cr ,: (k )] − Min[Cr ,: (k )]

d r ,i ( k ) =

Fig. 1. Roadmap: Performance Evaluation of A PHEV Municipal Parking Deck

II. PROBLEM FORMULATION The objective function being considered in this paper is the maximization of the average SOC (State of Charge) for all vehicles at the next time step. We will consider the energy price, charging time, and current SOC in this model. In order to make our system more robust, we will also allow vehicles to leave prior to their expected departure time (i.e., the PHEV is unplugged abruptly). This would result in a serious failure in terms of optimal power allocation, and the PHEV battery may not be adequately charged (even if it has been plugged-in for a long time) [9]. Therefore, the proposed function aims at ensuring some fairness in the SOC-distribution at each time step. This will help to ensure that a reasonable level of battery power is attained, even in the event of an early departure). The objective function is defined as: Max J1(k ) = ∑ w i (k )SoCi ( k + 1) (1) i

wi (k ) = f (Cr ,i (k ), Tr ,i (k ), Di (k )) Cr ,i (k ) = (1 − SoCi (k )) * Ci where

Cr ,i (k ) is the remaining battery capacity required to

be filled for i-th PHEV at time step k; Ci is the rated battery capacity of the i-th PHEV; Tr ,i ( k ) is the remaining time for charging the i-th PHEV at time step k; Di ( k ) is the price difference between the real-time energy price and the price that a specific customer at the i-th PHEV charger is willing to

Dr ,i (k ) − Min[ Dr ,: (k )]

Max[ Dr ,: (k )] − Min[ Dr ,: (k )] 1 Tr ,i (k ) − Min[1 Tr ,: (k )] t r ,i ( k ) = Max[1 Tr ,: (k )] − Min[1 Tr ,: (k )] The parking deck operators may also have different interests and assign different importance factors to cr ,i ( k ) ,

ti (k ) , and di (k ) depending on their own preferences. Thus: wi (k ) = α1cr ,i (k ) + α 2ti (k ) + α 3 di (k ) The charging current is also assumed to be constant over Δt . [ SoCi (k + 1) − SoCi (k )] ⋅ Ci = Qi = I i (k )Δt (3)

SoCi (k + 1) = SoCi (k ) +

I i ( k ) Δt Ci

Where Δt is the user-defined sample time; I i ( k ) is the charging current over Δt . The battery model is considered to be a capacitor circuit obeying the equation: dV Ci i = I i dt Therefore, over a short period of time, we can approximate the voltage change to be linear until it saturates: V (k + 1) − Vi (k ) (4) Ci ⋅ i = Ii Δt I Δt Vi (k + 1) − Vi ( k ) = i Ci Since our decision variable is the power allocated to PHEVs, we replace I i ( k ) with Pi ( k ) :

I i (k ) =

Pi (k ) Pi (k ) = ' Vi (k ) 0.5 × [Vi (k + 1) + Vi (k )]

Substituting I i ( k ) into Equation (4) yields:

(5)

3

Vi (k + 1) =

4.

2 Pi (k ) Δt + Vi 2 (k ) Ci

SoCi (k + 1) = SoCi (k ) +

Update the velocities of the particles according to Equation (7). Clamp the velocities if | Vid (k ) |> Vmax (where Vmax is the velocity limit).

Pi (k )Δt 0.5 ⋅ Ci ⋅ [Vi (k + 1) + Vi (k )]

If Vmax is too small, the particles may become trapped in a local region. However, if Vmax is too

Finally, the objective function becomes: ⎧ ⎫ ⎪ ⎪ Pi (k )Δt ⎪ ⎪ J1(k ) = −∑ w i ⋅ ⎨SoCi (k ) + ⎬ 2Pi (k )Δt i 2 ⎪ 0.5 ⋅ Ci ⋅ [ + Vi (k ) + Vi (k )] ⎪ ⎪ ⎪ Ci ⎩ ⎭

large, the particles are more likely to pass a better region. In practice, Vmax is often set at 10%-20% of

(6)

which is subject to:

the range of the variables over a particular dimension. Update the positions of the particles according to Equation (8). Repeat steps 2-5 until the maximum number of iterations or the minimum error criteria is met.

5.

∑ Pi ( k ) ≤ Putility ( k ) ×η

6.

i

0 ≤ Pi (k ) ≤ Pi ,max ( k ) Possible real-world constraints could include the charging rate (i.e., slow, medium, and fast), the time that the PHEV is connected to the grid, the desired departure SOC, the maximum electricity price that a user is willing to pay, certain battery requirements etc. Furthermore, the available communication bandwidth could limit sampling time, which would have effects on the processing ability of the vehicle. The primary energy constraints being considered in this paper include the power available from the utility ( Putility ) and the maximum power ( Pi ,max ) that can be absorbed by a specific vehicle. The overall charging efficiency of the parking deck is described by η . From the system point of view, η is assumed to be constant at any time step. III.

PARTICLE SWARM OPTIMIZATION

In 1995, Kennedy and Eberhart presented a new, evolutionary computation algorithm called Particle Swarm Optimization (PSO) [10-11]. PSO is an iterative stochastic optimization method. It simulates the behavior of flocks of birds or schools of fish. In PSO, each solution is a "bird" (or, more generally, a “particle”) in the search space. All of the particles have (1) fitness values (which are evaluated by the fitness function to be optimized) and (2) velocities (which direct the flying of the particles). The particles fly through the search space by following the current optimum particles. At each iteration, each of the particles is updated by following the individual and group bests. Gradually, the particles tend toward the global “near-optima” region. Generally speaking, the principal steps in PSO can be summarized as follows: 1. Generate a group of random solutions (particles) in the feasible region. Since we normally have very little information about the global optima, these particles are scattered over the search space as uniformly as possible. 2. Evaluate the distance between the new solution and the desired solution based upon a fitness function. 3. Compare the fitness value at the current iteration with previous best, and update the individual best (pbest) and group best (gbest).

Vid (k + 1) = ω ⋅ Vid ( k ) +

(7)

α1 ⋅ rand1 ⋅ ( pbest id − X id (k )) + α 2 ⋅ rand 2 ⋅ (gbest − X id (k ))

(8)

X id (k + 1) = X id (k ) + Vid (k + 1)

Where Vid (k ) is the velocity of the individual particle at iteration k; rand1 and rand 2 are uniform random numbers between 0 and 1; X id ( k ) is the position of a particle at iteration k; pbest id is the best value achieved by the individual so far; gbest is the best value of the group; and

α2

α1

represent the cognitive constant and social constant

respectively. These constants represent the weights for the stochastic acceleration terms. In general, α1 + α 2 = 4 ; to begin,

α1 = α 2 = 2 . [12]. Finally, ω

is the inertia constant

that balances the information sharing between an exploratory mode versus an exploitative mode. In practice, ω often decreases linearly from 0.9 to 0.4. Initially, a higher value allows particles to move freely throughout the search space to find the global optima. Once a near-optimal region is reached, ω is decreased to a lower value in order to more narrowly define the search. Therefore, ω is given according to the following equation [13]:

ω = ωmax −

ωmax − ωmin itermax

⋅ iter

(9)

We selected PSO for solving the PHEV problem because it is very robust for solving complex, non-linear optimization problems. For example, the algorithm’s performance is not largely affected by limited computer memory, the size of the problem, or dimension limits. Since we are dealing with a large number of PHEVs (i.e., 500 or greater) connected to the grid service, PSO is particularly suitable for solving our problem.

4

In order to make real-time decisions related to the amount of power allocated to each vehicle, a large amount of raw data needs to be processed in a short amount of time. Compared to other approaches (i.e., the Interior Point Method and Genetic Algorithm), PSO is much easier to be implement because it relies on simple calculations. Moreover, there are fewer parameters to adjust with PSO, and it can achieve high quality solutions with a stable convergence rate. Energy scheduling at a PHEV deck is also subject to different constraints that limit the search space to a certain feasible region. PSO can easily handle the constraints separately, eliminating the need for additional parameters. Also, there is no limit to the number or format of the constraints [14]. Fig. 2 shows how we applied PSO to the PHEV problem. In our algorithm, the particle is reset to the previous position if the constraints do not hold true.

Point Method (IPM) and the Genetic Algorithm (GA) method. We ran each scenario 100 times for each algorithm. The particle movement found using the PSO method is described in Fig. 3. Initially, the particles were allowed to move freely in the search space in pursuit of the global optima. Gradually, they reached the near-optima region, but the diversity of the swarm is still maintained. Eventually, they converged on the global near-optima. However, because PSO is a stochastic method and the maximum number of internal iterations was limited, there existed the chance of generating different solutions at each run. Despite this potential for variation, the standard deviation was found to be less than 2.2% of the average, proving that the solutions are very similar. -0.55 Mean Score Best Score -0.6

-0.65

Score

-0.7

-0.75

-0.8

-0.85

-0.9

0

20

40

60

80 100 Generation

120

140

160

180

Fig. 3. Convergence Tendency of the Proposed Optimization Method

Fig. 2. Flowchart of PSO Implementation SIMULATION RESULTS AND ANALYSIS

All of the calculations were run on an Intel(R) Core™ i5 CPU [email protected], 6.00GB RAM, Microsoft 64 bit Windows 7 OS and Matlab 2009b. Parameter values were set as described as: SwarmSize=40; Maximum PHEV charger

limit Pi ,max ( k ) is 10KW; the system efficiency is 90%;

Putility ( k ) = 90% × n × Pi ,max ( k ) ; the scheduling period is 24 hours; the cognitive constant and social constant α1 = α 2 = 2 ; α1 = α 2 = α 3 = 1 in (3). Initially, we simulated our algorithm for 5 PHEVs. Then, we increased the number of PHEVs to 50 and 500. We compared the results of the PSO algorithm with the Interior

0.2 0.18 0.16 Percentage of Vehicles

IV.

The results of the test using 50 PHEVs are shown in Fig. 4 and Fig. 5. For this test, the initial SOC was defined as a continuous uniform random number between 0.2 and 0.6. The sample time was set to 1200 seconds (20 hours). The remaining charge time was defined as continuous random number between 0 and 6 hours. The price that the customers were willing to pay for electricity was defined as a continuous random number between $1 and $2. The battery capacity was assumed to be identical for all vehicles.

0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0.1

0.2

0.3

0.4 State of Charge

0.5

Fig. 4. Initial SOC for 50 PHEV case

0.6

0.7

5

The computational speed of the three algorithms is compared in Table I. Because IPM only searches for local minima in the search space, it performs faster than the other algorithms. However, PSO is less time-consuming than GA because it does not perform selection and crossover operations.

0.4 0.36 0.32

Percentage of Vehicles

0.28 0.24 0.2

TABLE I AVERAGE CPU TIME COMPARISON

0.16 0.12 0.08

50 PHEV

0.04 0 0.4

0.5

0.6

0.7 State of Charge

0.8

0.9

1

Fig. 5. Departure SOC for 50 PHEV case

The results of the test using 500 PHEVs are shown in Fig. 6 and Fig. 7. For this test, the initial SOC was defined as a continuous uniform random number between 0.1 and 0.75. The sample time was set to 1200 seconds (20 hours). The remaining charging time was defined as continuous random number between 6 and 24 hours. The price that the customers were willing to pay for electricity was defined as continuous random number between $1 and $2. The battery capacity of each vehicle was defined as a continuous uniform random number between 6 and 15 Ah.

Computation Time per 100 trails at time step k

0.12

Standard Deviation

PSO

15.31 sec

-0.8940

3.44e-4

IPM GA

7.57 sec 122.3 sec

-0.8943 -0.8935

1.5298e-7 7.8916e-4

PSO is also fairly immune to the size and non-linear nature of the problems being considered. It can reach the nearoptimal solution at a reasonable convergence rate (i.e., the execution time does not exponentially growing with respect to the number of PHEVs; see Table II). Furthermore, the experimental results show that the proposed method is capable of obtaining high quality solutions with stable convergence characteristics. TABLE II AVERAGE CPU TIME WITH 50, 500, AND 5000 PHEVS # of PHEVs 50 500 5000

0.14

Mean

Average CPU Time at time step k 0.1531 sec 0.3133 sec 1.5065 sec

Percentage of Vehicles

0.1

0.08

0.06

0.04

0.02

0 0.1

0.2

0.3

0.4 State of Charge

0.5

0.6

0.7

Fig. 6. Initial SOC for 500 PHEV case

Compared with the other techniques, PSO is also easier to implement, constraint handling is more straightforward, and there are fewer parameters to adjust. Additionally, because every particle remembers both its personal and neighborhood best values, PSO has a more effective memory capacity than GA. PSO is also more efficient at maintaining swarm diversity because the individual particles use information related to the most successful particle in order to improve themselves. In GA, the worst solutions are discarded (and only the good ones saved), causing the population to evolve around a subset of only the best individuals [10].

0.18

V. CONCLUSION AND FUTURE WORK

0.16

Percentage of Vehicles

0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0.4

0.5

0.6

0.7 0.8 State of Charge

0.9

Fig. 7. Departure SOC for 500 PHEV case

1

1.1

In this paper, we described the performance evaluation of a PHEV municipal parking deck from a mathematical perspective. In order to manage the energy allocated to the PHEVs in real-time, we have applied PSO. PSO uses the previously stored system data in order to solve optimization problems. Any change to the system requires saving the new data and re-executing the algorithm. In order to achieve effective optimization, the simulation parameters are required at every time step. This requires a large amount of raw data to be processed in a short period of time. By integrating PSO with other stochastic methods, online analysis can be achieved. In our model, we considered the constraints imposed by

6

energy cost, remaining battery capacity, and remaining charging time in our model. Our simulation results demonstrate that the algorithm converges to a solution in a reasonable amount of time, and is immune to the size and nonlinear nature of the problem. Furthermore, it is faster than GA and easier to implement than some of the more traditional methods. In the future, we will consider using more objective functions (minimizing the overall charging time, etc.) to satisfy both customer interests and the needs of the power grid. However, conflicts can arise when multiple objective functions are used. The simplest solution to this problem involves aggregating all of the objectives into a single function. In this case, the weights assigned to each can be fixed or dynamically changed during the optimization process (Weighted Aggregation Approach). The main disadvantage of this approach is that it is not always possible to find the appropriate weighted function. Tradeoffs between multiple objectives should also be considered, and it is often necessary to find multiple Pareto optimal solutions (Pareto front) [10, 15-17]. Once a method is agreed upon, we will test it against other techniques such as the Interior-point Method and the Estimation of Distribution Algorithm. VI.

ACKNOWLEDGEMENT

The work is sponsored by the FREEDM systems center (Future Renewable Electric Energy Delivery and Management) and ATEC (Advanced Transportation Energy Center) at North Carolina State University. The authors would like to thank all colleagues at ADAC (Advanced Diagnosis Automation and Control) Lab for helpful discussion. VII. REFERENCES [1] K. Parks, P. Denholm, and T. Markel, “Cost and Emissions Associated with Plug-in Hybrid Vehicle Charging in the Xcel Energy Colorado Service Territory”, Technical Report, National Renewable Energy Laboratory (NREL), May 2007 [2] J. Tomic, and W. Kempton, “Using fleets of electric-drive vehicles for grid support”, in Journal of Power Sources, 168(2), 459–468, 2007 [3] B. K. Sovacool, and R. F. Hirsh, “Beyond batteries: An examination of the benefits and barriers to plug-in hybrid electric vehicles (PHEVs) and a vehicle-to-grid (V2G) transition”, Energy Policy, Vol. 37, Issue 3, 1095-1103, March 2009. [4] M. Duvall and E. Knipping, “Environmental Assessment of Plug-in Hybrid Electric Vehicles”, EPRI, July 2007. [Online] [5] P. Kulshrestha, L. Wang, M.-Y. Chow, and S. Lukic, "Intelligent Energy Management System Simulator for PHEVs at Municipal Parking Deck in a Smart Grid Environment," in Proceedings of IEEE Power and Energy Society General Meeting, Calgary, Canada, 2009. [6] P. Kulshrestha, K. Swaminathan, M.-Y. Chow, and S. Lukic, "Evaluation of ZigBee Communication Platform for Controlling the Charging of PHEVs at a Municipal Parking Deck," in Proceedings of IEEE Vehicle Power and Propulsion Conference, Dearborn, Michigan, U.S.A, Sept 7-11, 2009 [7] W. Su, M.-Y. Chow, “An Intelligent Energy Management System for PHEVs Considering Demand Response,” in Proceedings of 2010 FREEDM Annual Conference, Tallahassee, Florida, U.S.A. [8] J. Kennedy and R. C. Eberhart, Swam Intelligence, San Mateo, CA: Morgan Kaufmann, 2001 [9] S. Han, S. Han, and K. Sezaki, “Development of an Optimal Vehicle-toGrid Aggregator for Frequency Regulation”, IEEE Transaction on Smart Grid, Vol.1, NO.1, June 2010

[10] Y. Del Valle; G. K. Venayagamoorthy; S. Mohagheghi; J. C. Hernandez; R. G. Harley; “Particle Swarm Optimization: Basic Concepts, Variants and Applications in Power Systems,” IEEE Transaction on Evolutionary Computation, Vol 12, Issue 2, April 2008 [11] J. Kennedy, and R. Eberhart, “Particle Swarm Optimization”, IEEE International Conference on Neural Networks, Perth, WA, U.S.A, Nov 1995 [12] R. Eberthart, Y. Shi, and J. Kennedy, Swam Intelligence, San Mateo, CA, Morgan Kaufmann, 2001 [13] M. A. Abido, “Particle Swarm Optimization for Multi-machine Power System Stabilizer Design”, in proceedings of IEEE PES Summer Meeting, Vol. 3. Pp. 1346-1351, 2001 [14] X. Hu, “Particle Swarm Optimization”, in Tutorial of the IEEE Swarm Intell. Symp., 2006 [15] C. A. Coello Coello and M. Salazar Lechuga. “MOPSO: A Proposal for Multiple Objective Particle Swarm Optimizations”. In Congress on Evolutionary Computation (CEC’2002), volume 2, pages 1051–1056, Piscataway, New Jersey, May 2002. [16] C. A. Coello Coello, Gregorio Toscano Pulido, and Maximino Salazar Lechuga, “Handling Multiple Objectives with Particle Swarm Optimization”, IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, VOL. 8, NO. 3, JUNE 2004 [17] Xiaohui Hu, and Russell Eberhart, “Multiobjective Optimization Using Dynamic Neighborhood Particle Swarm Optimization”, 2002

VIII.

BIOGRAPHIES

Wencong Su is currently working toward Ph.D. degree in the Department of Electrical and Computer Engineering at North Carolina State University. He received B.S. with distinction in Electrical Engineering from Clarkson University in 2008 followed by a M.S. in Electrical Engineering from Virginia Tech in 2009. He also worked as a R&D engineer intern at ABB U.S. Corporate Research Center in Raleigh, NC, from May 2009 to August 2009. His current research interests are Microgrid modeling and simulation, distributed control, Intelligent Energy Management System for Plug-in Hybrid Electric Vehicles, and Performance Evaluation of A PHEV Municipal Parking Deck. Mo-Yuen Chow received the B.S. degree from the University of Wisconsin, Madison, in 1982 and the M.Eng. and Ph.D. degrees from Cornell University, Ithaca, NY, in 1983 and 1987, respectively. Upon completion of the Ph.D. degree, he joined the Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, and has held the rank of Professor since 1999. Dr. Chow's research focuses on fault diagnosis and prognosis, distributed control, and computational intelligence. He has been applying his research to areas including mechatronics, power distribution systems, distributed generation, motors and robotics. Dr. Chow has established the Advanced Diagnosis and Control Laboratory at NC State University. He has published one book, several book chapters, and over one hundred journal and conference articles related to his research work. He is an IEEE Fellow, and has received the IEEE Region-3 Joseph M. Biedenbach Outstanding Engineering Educator Award the IEEE ENCS Outstanding Engineering Educator Award. He is the Editor-in-Chief of IEEE Trans. on Industrial Electronics, and an Associate Editor of IEEE Trans. on Mechatronics.