产品分类：

 

 

Abstract ： The uncertainties of the integrated wind power consequently increase the operation risk of power systems measured by the index of loss of load probability (LOLP). The battery energy storage system (BESS) is incorporated into wind farms as another kind of reserve for reducing the operation risk of power systems. A special operation strategy is designed for BESS and any incomplete charge-discharge cycle is avoided. The charge/discharge sequences of BESS are chronological, so the sequential Monte Carlo simulation (SMCS) is adopted to assess the operation risk. Simulation results on a case system demonstrate that BESS can reduce the operation risk. In addition, another significant finding is that the charge-discharge cycle consumed for the operation risk reduction is tiny compared with BESS circle life. Thus, BESS still has the capability of achieving other purposes simultaneously. Key words: wind farm； battery energy storage system (BESS) ； operation risk； sequential Monte Carlo simulation (SMCS) ; loss of load probability (LOLP) ； cycle life consumption； spinning reserve

0 Introduction

 

In recent decades， wind power is showing a rapid development all around the world due to escalating fuel prices as well as increasing public concern on environment protection⑴.An energy policy called the renewable portfolio standard (RPS) is widely accepted by more and more countries and regions'-2-1 . RPS is indeed a promise of producing a specified percentage of the total power generation from renewable sources within a certain date. As a result, more and more large-scaled wind farms will be integrated into power systems. Unfortunately, wind power depends on climatic conditions and tends to be intermittent.

 

In many utilities, the intermittent wind power is considered as a negative load that cannot be dispatched and is excluded from the process of unit commitment (UC). On the basis of load and wind power predictions, conventional generator units are scheduled for serving the net load (i.e., load demand minus wind power) at the minimum cost by means of UC tool. Due to the limitation of present prediction tools’ the wind power prediction is far more difficult than the load prediction, and the prediction error increase dramatically with the prediction time horizon1-3-1 . Under this condition, the uncertainties of the net load increase due to the uncertainties of the wind power, so the operation risk of

Manuscript received September 25，2012.

 

This work is supported by National Key Technology Research and Development Program of China (No. 2011BAA07B07) and National Natural Science Foundation of China (No. 51077041).

power systems consequently increase. In this paper, the operation risk refers to the probability in which the scheduled generating capacity will fail to carry the net load over a scheduled time horizon (24 h or less). Loss of load probability (LOLP) is a widely used reliability index and quantifies the probability of the load loss resulted from insufficient generation capacity. So it is utilized to measure the operation risk in this work. Scheduling sufficient spinning reserve (SR) capacity is considered as an effective means for reducing the operation risk, but it comes at some cost, because additional generator units might be synchronized and other units might operate at their suboptimal output pointsw .

 

Since the early 1980s, the utility-scaled battery energy storage system (BESS) has been applied widely[s-6] . It has been used for load leveling, peak shaving, voltage-frequency stabilizing, and other purposes. In recent years, BESS has been incorporated into wind farms for mitigating the disadvantages resulted from the integrated wind power. In Refs. [7-10]，BESS has been incorporated into wind farms as an energy storage medium for dispatching the wind power. By controlling the charging/discharging behavior of BESS, the wind power can then be dispatched by system operators to some extent. The wind power is also intermittent. Its reliability contribution to generation systems is consequently far less than that of conventional generator units. In Ref. [11]，the case study on a test system reported in Ref. [12] demonstrates that the reliability contribution of a 90 MW wind farm is only equal to that of a 10 MW

 

conventional generator unit. Refs. [ 13-14] incorporated BESS into wind farms for mitigating the intermittency of wind power, so the reliability contribution of wind power can be dramatically improved1-13'14-1 .

 

The highlight of the present work is to utilize BESS to reduce the operation risk of power systems. BESS is designed to discharge for serving some load in case of power shortage resulted from unpredictable events, such as generator outages, sudden load/wind power changes or a combination of both. In this instance, BESS can be considered as another kind of reserve. Regarding expensive BESS, an important consideration for design is the finite number of charge-discharge cycles that BESS can undertake over its lifetime1-15^ . A special operation strategy is designed to make full use of the battery. A new index, called the expected cycle life consumption (ECLC), is proposed to measure the cycle life consumption of BESS during the scheduled time horizon. Sequential Monte Carlo simulation (SMCS), described in Ref. [16] , is adopted to assess the system operation risk, in which, both stochastic factors associated with power systems and effects of BESS are considered. Simulations on a case system are done to justify the feasibility of the designed BESS.

 

1 BESS Integrated as Reserve

 

1.1 System Configuration

 

In Refs. [7-8]，BESS is incorporated into a system at the point of common coupling (PCC) for dispatching wind power by adjusting the charging/discharging state of BESS. In the present work, BESS designed for reducing the operation risk of power systems is also incorporated into a wind farm according to the same technical scheme in Refs. [7-8]，as shown in Fig. 1.

BESS Conventer

Fig. 1 BESS integrated into a wind farm

In Fig. 1, Pwt and Pht are wind power output and charging/discharging power of BESS at hour t respectively (l^t^T, T is the total number of hours over the scheduled time horizon). If BESS discharges at hour t, Py,t is positive; if BESS charges at hour t, Pb>, is negative. Pdit represents the output power from the wind farm and the BESS power station. Assuming that power losses of converters are ignored, it can be computed as

 

Pd’t = Pw,t + Pb,t (1)

 

In many utilities, the wind power is considered as a

negative load when scheduling power systems. The precision of wind power prediction is far less than that of load prediction, so the integration of wind power increases the operation risk of power systems. In case of an emergency that SR fails to compensate power shortage entirely, the energy stored in BESS discharges to serve some load to reduce the operation risk of power systems. In this instance, BESS can be considered as another kind of reserve. Subsequently, SR provided by operating generator units starts to charge BESS. Under this condition, Pd(f can be negative, i. e. , the wind farm and BESS power station absorb energy from the grid. After the charging period, the recharged BESS has a capability of compensating unpredictable power shortage again and discharges as needed.

 

1 _ 2 Technical Characteristics of BESS

 

BESS is mainly comprised of batteries, the control and power conversion system (C-PCS)，and the rest of the plant. The rest of the plant is designed to provide good protection for batteries and C-PCS1-6-1 . Batteries are the most important and expensive component of BESS. The performances of BESS depend on the battery characteristics.

 

Regarding BESS, an important characteristic is the finite number of charge-discharge cycles that the battery can undertake over its lifetime[15] . The number of such cycles, called the cycle life, is of nonlinear and complex relationships with various factors, such as the depth of discharge (doov ) that the battery has to undergo, the ambient temperature and the charging/discharging rate of the battery. For example, the cycle life of a certain type of battery, as reported in Ref. [17]，decreases dramatically from 4 200 for ^dod — 20% to 2 000 for 80%) should be avoided as it may lead to permanent physical damage to the battery and an exceedingly low cycle life[18] . In the present work, if the battery is fully charged, its state of charge (SOC) is 1. Hence, when the battery is subsequently fully discharged, i.e., its ^dod reaches the allowable maximum value, denoted as dmax，thus SOC of the battery is 1 — dm&x. In order to obtain a judicious balance between the discharging level of battery energy and the battery lifetime, a reasonable value of dmax is 0.8[9] -

 

Due to the nonlinear relationship among SOC, charging/ discharging rate, internal losses of battery, and other factors, only nonlinear model[19] can accurately represent battery charging/discharging behaviors. Because the scheduled time horizon is short-term, a simplified linear battery model can be adopted, in which a fixed charging/ discharging efficiency, denoted as rj, is utilized to express internal losses of the battery.