Research on Optimal Operation of Active Distribution Network with Energy Storage Facilities

. With a large amount of intermittent distributed generation(DG) integrating into active distribution network(ADN), the flexibility requirements during the operation process should be paid more attention. Therefore, a dynamic optimal operation model is built. In this model, the maximization of economic benefits during the production cycle is set as the optimization objective, and the constrains of security and flexibility are considered. Then based on this model, the security, economy and flexibility requirements of power system are satisfied by coordination and interaction of different flexible sources. Finally, a case study is given to illustrate the validity and effectiveness of the models.


Introduction
An important task of power system operation is the optimization of generation scheduling and load allocation. According to difference of time scale, the optimal operation of power system includes static optimization and dynamic optimization [1][2]. Static optimization focus on the optimal objective at a certain moment. However, dynamic optimization focus on the multi-period optimization of generation sources based on the day-ahead dispatching framework. Dynamic optimization takes the coupling relationship of adjacent periods into account. Therefore, it is more suitable to the optimization of ADN operation [3][4][5].
For ADN, when it is an independent economic entity in a certain market environment, the ultimate objective of ADN operation is to make the benefit maximization under the precondition of system security and reliability. Usually, ADN operates under two business modes. In the first mode, ADN is a pure network operator, and only provide electricity distribution service. In the second mode, ADN is not only a network opera-tor, but also own some DGs as generation operator.
In this paper, a dynamic optimal operation model of ADN is built for the second business mode. The maximization of economic benefits during the production cycle is set as the optimization objective, and the constrains about security, flexibility etc. are considered. Then based on this model, coordination and interaction act between the network, generation and load, in order to meet of the security, economy and flexibility requirements [6][7].

Objective of Optimal Operation Mode
When ADN operate under the second business mode, the maximization of economic benefits during the production cycle is set as the optimization objective, as shown below: Where, T is the amount of time intervals; ( ) RE t is the operation benefit of ADN at the t-th time interval; ( ) CO t is the operation cost of ADN at the t-th time interval.
When the environment effectiveness is considered, the comprehensive benefit of ADN include two parts: one is the electricity benefit, the other is environment benefit. The electricity benefit depends on the amount of electricity sales and the corresponding electricity price. The environment benefit depends on the reduction amount of Carbon dioxide emission and the corresponding carbon-trade price. The reduction amount of Carbon dioxide emission is equal to the amount of Carbon dioxide discharged by coal-fired thermal power unit, which generated the same amount of electricity, as shown below: Where,

( )
REE t is the electricity benefit of ADN at the t-th time interval; ( ) REC t is the environment benefit of ADN at the t-th time interval.
The operation cost of ADN are comprised of electricity purchase cost, grid-loss cost, charge/discharge cost of energy storage(ES) facilities etc. as shown below:

Equality Constrains of ADN Operation
The equality constraints of ADN operation consider the factors such as benefit, cost, balance between power supply and demand.

Electricity benefit
The electricity benefit of ADN at the t-th time interval is shown as below: Where,

( )
TIN t is the electricity sale price at the t-th time interval; ( ) LO t is the electricity sale price at the tth time interval; t Δ is time length of each interval.

Environment benefit
The environment benefit of ADN at the t-th time interval is shown as below: Where, CP is the carbon-trade price; ( ) EMC t is the reduction amount of Carbon dioxide emission at the t-th time interval; ( ) ERE t is the amount of electricity produced by renewable generation source at the t-th time interval; CE is the conversion coefficient between carbon emission and coal-fired by thermal power plant.

Electricity purchase cost
The electricity purchase cost of ADN at the t-th time interval is shown as below: Where,

( )
CPI i is electricity feed-in tariff of the i-th DG; ( ) , CPO i t is the feed-in power output of the i-th DG at the t-th time interval.

Grid-loss cost
The grid-loss cost of ADN at the t-th time interval is shown as below: Where,

( )
TIN t is electricity sales price of ADN at the t-th time interval; GLF is the grid-loss coefficient, which is the proportion of grid-loss to total load.

Charge/Discharge-loss cost of ES
The charge/discharge-loss cost of ES at the t-th time interval is shown as below: power of ES at the t-th time interval, if charge negative, else positive; ESC is loss cost of per unit charge/discharge electricity.

Charge/Discharge energy balance of ES
The charge/discharge energy balance of ES at the t-th time interval is shown as below:

Power balance of ADN
The power balance of ADN at the t-th time interval is shown as below: Where,

Inequality Constrains of ADN
When the grid structure has been determined, the operation of ADN should consider not only the grid transmission constraint, but also the constraints such as: upper and lower limitation of power output, flexible adjust ability, ramping ability, energy storage.

Active power output constraint
The active power output constraint of DG is shown as below: Where: For the dispatch DG, the active power output is affected by rated capacity, minimal technical output, etc. for the intermittent DG such as wind power generation, photovoltaic(PV) power generation, all the electricity produced by the these DGs are fully accommodated by ADN. For micro-grid(MG), the exchanging power between MG and main network should be maintained below a certain proportion of MG's rated capacity.

Flexibility Constraint
The flexibility constraint is to solve the problem caused by some uncertainty conditions, such as: change of solar irradiation, emergency, disturbance. The flexibility of ADN is depended not only on the inherent property of ADN, but also on the current operation status. Here, the flexibility constraint of ADN is shown as below [8][9][10]: Where, / _ ( ) flex re t + − is the flexibility requirements of ADN at the t-th time interval, "+"means upward adjustment; "-"means downward adjustment; ( ) ( ) LO t is the load power of ADN at the t-th time interval.

Ramp-rate Constraint
The dispatchable DGs should meet the ramp-rate constraint during the adjustment process, and include upward adjustment and downward adjustment, as shown below: Where, ( , ) CSP i t is the active power output of the ith DG at the t-th time interval; ( , 1) CSP i t − is the active power output of the i-th DG at the (t-1)-th time interval; i r is ramp-rate of the i-th DG; max t is the allowable maximal adjustment time.

Energy Storage Constraint
The energy status of ES should be in the rated energy range, shown as below: min_ ( ) max_ SOC SOC t SOC ≤ ≤ (17) Where, max_ SOC is upper limitation of energy status; min_ SOC is lower limitation of energy status; ( ) SOC t is the energy status of ES at the t-th time interval, usually between 20%~80% of the rated energy.

Case Study
A typical daily scenario of ADN is used for case study. In which, the maximal load is 438 MW; the minimal load is 223 MW; the installed capacity of dispatchable DGs is 50MW; the installed capacity of distributed PV is 100MW; the installed capacity of distributed wind generation(WG) is 40MW; the installed ES is 50MW/100MWh; the injection power from external network (EN) is 420MW. The feed-in electricity price of ADN is 0.35 ¥/kWh between 1:00-8:00 and is 0.6 ¥/kWh between 9:00-24:00. The electricity sales price of EN is 0.35 ¥/kWh between 1:00-8:00 and is 0.46 ¥/kWh between 9:00-24:00. The property parameters of different DGs are shown in Table 2. Although the installed capacity of MG is 60MW, the feed-in power is limited to be lower than 20% of the rated capacity. For ES, the upper/lower limitation of active power output are ±50 MW; the upper/lower limitation of SOC are 80/20 MWh; the ramp rate is 25 MW/min; the feed-in electricity price is set equal to energy loss cost: 0.05 ¥/kWh; the initial value of ES energy status is 20 MWh. For ADN, the grid-loss coefficient is 0.05; the flexible reserve capacity is 0.1; the maximum adjustment time is 5min. All the model and algorithm are programed on GAMS software. The optimal results for ADN operation are shown as Table 3  The results are shown in table 2. During the 1th interval to the 8th interval and the 23rd interval to 24th interval, the feed-in electricity price of dispatchable DGs is higher than the wholesa le price of EN, so the ADN firstly purchase electricity from EN. When during the 9th interval to the 22nd interval. the feed-in electricity price of dispatchable DGs is lower than the wholesale price of EN, so the ADN firstly purchase electricity from DG and MG. Through the optimization of whole process, the ADN get the maximal benefit: 432460 ¥.

Conclusions
Along with a lot of intermittent DGs integrated into ADN, the flexibility problem has aroused more and more concern. Flexible adjust sources such as ES, MG and other dispatchable DGs, are widely applied in ADN, in order to meet the flexibility requirement brought by intermittent DGs, such as PV and WG.
Then a dynamic optimal operation model of ADN is built in this paper. The maximization of economic benefits during the production cycle is set as the optimization objective, and the constrains about security, flexibility etc. are considered. Then the case study of ADN optimal operation is done based on a typical daily scenario. The security, economy and flexibility requirements of power system are satisfied by coordination and interaction of different flexible sources. This work is supported by science and technology project of state grid corporation of China: research and application demonstration of the aggregation effect of multipoint distributed energy storage in power grid.