Evaluating China's biomass power production investment based on a policy benefit real options model

Energy 73 (2014) 751e761

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Evaluating China's biomass power production investment based on a policy benefit real options model Xingwei Wang a, *, Yanpeng Cai b, c, **, Chao Dai d a

Key Laboratory for Water and Sediment Sciences of Ministry of Education, School of Environment, Beijing Normal University, Beijing 100875, China State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China c Institute for Energy, Environment and Sustainable Communities, University of Regina, 120, 2 Research Drive, Regina, Saskatchewan S4S 7H9, Canada d College of Environmental Science and Engineering, Peking University, Beijing 100871, China b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 April 2014 Received in revised form 20 June 2014 Accepted 21 June 2014 Available online 14 July 2014

In this study, a policy benefit real options model was developed to evaluate biomass power production investment in China. A method based on the cumulative probability was proposed using binomial decision tree calculations for the exercising of options in order to evaluate the optimal investment timing. Two scenarios were analyzed to identify the optimal investment strategy with/without the consideration of revenue from certified emission reduction (CER). Uncertainties in straw purchased price, government incentives, and technological improvements were considered. The results showed that it was not optimal for immediate investment in biomass power production in China. Given full government subsidy, the thresholds of straw purchased price for scenarios 1 and 2 are 213.55 and 218.87 RMB/ton, respectively, while the current straw purchased price in Chinese market is 220 RMB/ton. The investment of biomass power production would be executed at 2022 and 2028 with/without the consideration of revenue from CER in the current situation in China if there are no government incentive to encourage motivation, respectively. The conclusion could provide useful information for power enterprise decision-makers on whether and when to invest a biomass power production in China in an uncertain environment. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Biomass power production Economic evaluation Policy benefit model Binomial tree model Real options approach Uncertainty

1. Introduction Utilization of biomass has attracted increasing attention across the world due to growing concerns over depletion of conventional energy reserves, as well as the growth of associated environmental and greenhouse gas emissions [1e4]. Biomass is one of the most important renewable energy resources in a prospective renewable and sustainable energy future since it is considered carbon-neutral [5,6]. This is especially crucial for many newly prosperous countries such as China. According to this country's “The 12th Five Year Plan for Renewable Energy Development”, electricity generated by biomass will have reached a total installed capacity of 13 GW by 2015. This value will have been double by 2020, supposing to account for 4% of the total energy consumption [7,8]. In order to achieve this goal, China's National Development and Reform

* Corresponding author. Tel./fax: þ86 10 58809850. ** Corresponding author. State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China. E-mail addresses: [email protected] (X. Wang), yanpeng.cai@bnu. edu.cn (Y. Cai). http://dx.doi.org/10.1016/j.energy.2014.06.080 0360-5442/© 2014 Elsevier Ltd. All rights reserved.

Commission (NDRC) and many other relevant governmental agencies have developed a series of policies and regulations, such as Temporary Measures for Revenue Allocation of Addition Price on Electricity of Renewable Energy Resource, Interim Measures On Renewable Power Surcharge Collection and Allocation, and Temporary Measures for Management of Subsidy Fund of Utilizing Straw Energy Resources [8]. However, the process for identifying potential strategies and decisions related to biomass power production are complex due to the diversity in specific biomass features and the differences in biomass-based technologies [9]. For example, the outdated generation technology, high cost of straw collection, storage, and transportation still hinder the development of biomass power generation in China [10]. Decision-makers and energy managers are thus facing numerous challenges in generating biomass utilization strategies and policies. Particularly, whether and when should power enterprise invest a biomass power production plays a vital role in biomass related decision-making in a region. Therefore, it is desired to provide a comprehensive analysis and evaluation of biomass power generation under a complicated and volatile environment in China. Previously, numerous studies were undertaken to investigate and evaluate biomass power generation in China based on various

752

X. Wang et al. / Energy 73 (2014) 751e761

methods. For example, Zhao and Yan [11] assessed the factors of strengths, weaknesses, opportunities, and threats of the biomass power generation industry in China through using the SWOT (Strength, Weakness, Opportunities and Threat, in which “S” and “W” are internal factors, while “O” and “T” are external factors) analysis method. Wu et al. [12] analyzed economic characteristics of biomass gasification and power generation in China regarding the associated costs for investments, electricity generation, and waste treatment. Liu et al. [13] systematically analyzed temporal and spatial patterns of crop stalk resources, evaluated potential bioenergy of straw resources, and explored possible pathways of identifying straw-based energy strategies in Inner Mongolia, China. Zhang et al. [14] estimated the cost of straw-based power generation through the adoption of the method of life cycle analysis. The results showed that straw cost took the largest share among the operation cost. Also, the basic causes of the high cost included many factors such as straw characteristic, mismatch between demand and supply, immature technology, inappropriate project planning and low motivation of farmers in selling straw. Sun et al. [15] displayed a spatial planning framework to identify the appropriate development areas for biomass utilization at a regional scale. The analyses showed a clear picture of how to identify the locations of biomass power plants to minimize the cost and maximize the supply security of feedstock. The methods discussed above analyzed the current situation of China's biomass power production development and conducted corresponding economic evaluation. However, they were mainly based on the conventional cost-profit evaluation approach when conducting economic evaluation, such as discounted cash flow method (DCF) with the criterion of net present value (NPV). These methods have the defects when dealing with biomass power production investment evaluation due to the inherent characteristics of biomass power production investment in China: (1) the investment costs are mostly irreversible; (2) the timing of biomass power production is at the discretion of the firm; (3) the high uncertainty of payoff; (4) the scarce cumulative information, technology gap, and high investment risk; (5) the strong influence of national policy on biomass power production investment [16,17]. The reason is that the decision under multiple NPV criteria is normally based on static data and information, and cannot provide updated information for supporting dynamic decision-making regarding the investment that need to consider uncertainties in the future [18]. Comparatively, a number of studies were conducted on the evaluation of renewable energy generation project planning using real options approach (ROA). For instance, Boomsma et al. [19] adopted an ROA to analyze investment timing and capacity choice for renewable energy projects under different support schemes. Kumbaroglu et al. [20] presented a policy planning model, which integrated the learning curve information of renewable power generation technologies into a dynamic programming model that featured real options analysis. The model evaluated investment alternatives in a recursive manner and had the ability to delay an irreversible investment outlay that could subsequently affect the prospects for the diffusion of different power generation technologies. Kjærland [21] applied ROA to evaluate potential hydropower investment opportunities in Norway. The approach explained investment behaviors in a way that was not captured by NPV approach. Moreover, the analysis showed that such an ROA based approach could give insight into the value of investment opportunities and aggregate investment behavior in this industry. Lee and Shih [22] presented a policy benefit evaluation model, which incorporated cost efficiency curve information on renewable power generation technologies into an ROA framework. The method was used to quantitatively evaluate the policy value provided through developing renewable energy in the face of

uncertain fossil fuel prices and policy-related factors. The results demonstrated that the renewable energy development policy with internalized CO2 emission costs was an appropriate policy from ~ a and Mutale [18] sustainability point of view. Martínez-Cesen proposed an advanced ROA based methodology for supporting renewable energy generation projects planning. The results showed high-expected profits for projects could be achieved through an advanced ROA based approach. Reuter, Fuss and colleagues [23] employed an ROA method to investigate specific policy characteristic of renewable energies and their associated uncertainties in a stylized setting through explicitly taking into account market effects of investment decisions. Detert and Kotani [24] analyzed the changing investment environment for renewable energy with ROA and explored its potential in developing economies through studying the case of Mongolia under uncertain coal prices. The aforementioned studies discussed the optimal strategy for renewable energy project investment in an uncertain environment through using ROA. Compared with DCF approach, the real options approach (ROA) can possibly postpone judgment on an investment to an appropriate time and thus is suitable for the evaluation of biomass power production with considerable uncertainties. It has the following intrinsic properties to deal with: (a) the irreversibility of the investment, (b) the uncertainty in cash flows of the future investment, and (c) the timing of the investment flexibility [25]. The decision rule of real option method offers enhanced flexibility in decision-making. The investor could postpone judgment on an investment and wait for favorable circumstance and thus provide new opportunities [26], which is in line with the actual management. However, due to the specific characteristics of biomass resource in China, such as scattered distribution, plenty of varieties, obvious seasonality, difficult to collect, storage and transport, and low effective utilization [27], few previous studies were conducted for handling complexities and uncertainties in China's biomass power production investment evaluation and then provided decision-support for the investor through ROA. Therefore, the objective of this study is to establish a policy benefit real options evaluation model to analyze China's biomass power production investment. A method based on cumulative probability will be proposed through using binomial decision tree calculations for exercising the option in order to evaluate the appropriate investment timing. Uncertainties in straw purchased price, government incentives, and technological improvements will be considered. The results will be particularly useful in providing appropriate investment timing for investors under uncertainty, and also provide information for power enterprises' biomass power production investment evaluation and related policy-making in China. Renewable energy utilization 0.40%

Directcombusting by farmer 49%

Animal feed 25%

Edible mushroom 1%

Abandon 13%

Industry 5% Recycled 7%

Fig. 1. The structure consumption of straw in China.

X. Wang et al. / Energy 73 (2014) 751e761

2. Overview of biomass power production in China The biomass resources in China are abundant, widely distributed and have large various and output [8]. The yield of straw has increased at a rate of 1.4% annually [28]. The current amount of biomass resources in China is 540 million tons of standard coal equivalent (TCEs), the available amount is around 280 million TCEs [29]. The large amount of biomass resources provides great potential for biomass power development. However, the effective utilization rate of biomass resource is low (Fig. 1). The amount of renewable energy utilization is 2.75 million tons, merely accounting for 0.4% of the total energy consumption. A large proportion of them are used as fuel by direct-combusting. The efficiency of direct-combusting is extremely low, approximately 5e8% [30,31]. Moreover, a considerable quantity of straw is discarded or directcombusted in farm field, which not only lead to a waste of biomass resource, but also bring a series of environmental problems. In order to effectively utilize biomass resource, a series of laws and regulations, as well as the detailed implementation rules are established and greatly improve the enthusiasm on the development of biomass power generation industry, such as ‘The Renewable Energy Law Of The People's Republic Of China’, ‘The Tentative Management Measures for Allocation of Price and Expenses for Generating Electricity by Renewable Energy’, and ‘The Medium and Long Term Development Plan for Renewable Energy’, which is of significance for reducing pollution and promoting biomass power production development. Fig. 2 displays the process from straw collection to power and heat generation. Generally speaking, this process contains four stages, including field harvesting, collecting, conversion, and marketing stages. Briefly described, by the time of crop harvest, straw is moved, pre-processed and delivered to biomass power plant where it is burned to produce electricity as the main product and heat as a co-product. The ash content is applied as appropriate to agricultural soils to recycle nutrients particularly phosphorus and potassium. Collecting straw from fields is challenging for many biomass power plants in China due to the specific characteristics of various biomass resource: (a) unlike the other renewable energy sources, the available straw are spatially scattered and loose. The plant operators need to sign contracts with a large number of farmers to secure their straw supply due to a small piece of field per family owned in China; (b) vehicles usually cannot travel directly to fields to collect straw due to poor road conditions in the countryside,

753

which need more enhanced manpower; (c) farmers are reluctant to travel long distances to deliver straws from their fields to collection stations [7,32]. Moreover, the continuous operation of a biomass power production plant requires adequate straw supply. However, the acquisition of straw is rather seasonal, and is unrealistic to collect straw continuously all the year. The biomass power plant has to collect straw to support operation of relevant plants at least for half a year. Generally, local farmers can sell straws to biomass power plant directly after the harvest of grain, cotton and other primary products. The straw purchased price increased from 140 RMB/ ton in 2006 to about 220 RMB/ton in 2013 [14,33]. Despite of the yearly growth in the price, many farmers would like to burn the straw in field rather than sell them due to high handling and shipping costs, and shortage of rural manpower. According to investigation, the furthest distance that the farmers would transport straws to collection station or biomass power plant is approximately 20 km. Greater than this distance, the revenue of selling straw could not afford handling and shipping costs. The NDPC's edict “Notice on Construction Management of Biomass Power Generation Projects” (No. 1803) in August 2010 pointed out: “biomass power plant should be located in the grain producing areas where having abundant straws; only one biomass power plant is allowed to establish in one county or within a radius of 100 km.” The estimation and evaluation of straw potential based on land surface survey and statistical data [34]. In this study, four collection stations would be considered, and each collection station is constructed within a radius of 20 km. The average distance from collection station to biomass power plant is 60 km with consideration of the actual road conditions, which is shown in Fig. 3. Farmers can send crop straws to the collection station or power plant according to the actual situation, such as distance, road conditions, and so on. Straw transportation requires many truck movements and the use of considerable quantities of petroleum-based fuel. The transportation cost accounts for a significant proportion of the total expense, especially in a long distance [35]. When crop straws are transported to collection stations, some pretreatment should be conducted. Those processes contain quality detecting, weighing, crushing, baling, carrying, stacking, storing, unstacking, and entrucking. The sold straws can bring additional economic benefit for farmers. In this research, the collection station is assumed to be responsible for collecting and simple pretreatment, not responsible for storing. All of the pretreatment biomass resource should be transported to power plant in time.

Fig. 2. The full chain of biomass power project.

754

X. Wang et al. / Energy 73 (2014) 751e761

Fig. 3. The distribution of collection station around biomass power plant.

In a biomass power plant, extra facilities with special devices for straws storage have to be constructed due to the reason that the availability of straws may be subject to seasonal fluctuation or restriction. All of these facilities must be well built against rain, moisture, fire, and lightning. This would lead to a high cost in design, construction and maintenance of biomass power plant, which is also the reason of the high investment cost for biomass power plant [14,15]. The high costs in construction and operation has resulted in a low commercial profit to biomass power production in China, most enterprises have to rely on the government subsidies. These policies of tax preferences and financial subsidies help reduce the cost of biomass power production which in turn bridges the profit gap between the biomass power and the traditional energy resources. The benefits can then support the sustainable development of biomass power generation industry. 3. Modeling formulation Consider a biomass power production investment evaluation wherein a decision maker is responsible for reevaluating the updated information according to the change of the situation and making corresponding investment decision on whether and when to invest a biomass power production within a multi-period horizon [36e38]. From the perspective of biomass power generation enterprises, the decision maker can formulate the problem as maximizing the sum of expected discounted profits over the planning period [39,55e57]. The total investment cost (cash outflows) mainly comprises initial investment, operation and maintenance costs, straw purchase cost, transportation cost, and sales taxes and addition [14]. The revenue (cash inflows) mainly consists of product sale (electricity, heat, and by-product ash content),

government subsidy, and the potential revenue from certified GHGemission reduction. The cost of biomass power generation in a given supply-area S can be expressed as follows:

Net benefits ¼ ETS $QS þ HSS $PHS þ TCBS $j$PACS þ CER$Pc $BI  IS  Co&m  FSPS $TCBS  TRS  STA (1) where ETS denotes electricity tariff of biomass power generation in supply-area S, (RMB/kW h); QS is the annual amount of electricity generated by biomass power plant in supply-area S, (kW h); HSS is the annual amount of heat generated by biomass power plant in supply-area S, (PJ); PHS is the price of heat supply in supply-area S, (RMB/PJ); TCBS is the annual amount of straw resource that consumed by biomass power plant in supply-area S, (ton); j is the proportion of ash content after biomass resource combusting, (%); PACS is the price of ash content in supply-area S, (RMB/ton); CER (certified emission reduction) is the annual amount of certified emission reduction, (ton); Pc is the carbon price in international carbon trading market, (RMB/ton); BI is a binary variable, representing whether or not considering the potential revenue from certified emission reduction; IS is the initial investment cost in supply-area S, (RMB); Co&m is the annual operation and maintenance costs over the lifetime of installation, (RMB); FSPS is the straw purchase price in supply-area S, (RMB/ton); TRS is the annual transportation cost from the collecting station to power plant, (RMB); STA is the sales tax and addition of biomass power production, (RMB), which is in accordance with the relevant provisions of Chinese tax policies, output tax rate of the project is 17%, input

X. Wang et al. / Energy 73 (2014) 751e761

755

Fig. 4. The structure chart of heat balance in biomass power plant.

tax rate is deducted according to raw and auxiliary materials and fuel, including straw for fuel 13% and the rest 17% [14]. The annual straw availability of supply-area can be evaluated as follows:

TCBi ¼ z$b1 $b2 $b3 $q$pR2i

(2)

where TCBi denotes the annual straw availability in collection station i; x is supply security coefficient of straw that taking into account hazards from climatic, social origin or others; b1 is the percentage of cultivated area in the total land area within a radius of Ri; b2 is the coefficient of straw collection; b3 is the coefficient of straw availability; q is the annual straw yield per cultivated area (t/ km2). Ri the collection radius of collection station i. Nowadays, comprehensive utilization is the trend of biomass power generation technology in China. The main technology is combined heat and power generation (CHP), and the total efficiency can reach 90% [40]. According to many practical projects, a biomass power generation plant generally supplies steam and waste heat to its adjacent areas to generate additional income. Specially, the feed-in tariff can be set according to the government policy on biomass power. The straw energy content can be taken into account the energetic method. Also, the total revenue of electricity sale based on input capacity and annual running time can be calculated as follows [17,41]:

TETS;t ¼ ETS;t  f  IC  RTt  ð1  aÞ

(3)

where TETS,t denotes the total revenue of electricity sale in period t in supply-area S, (RMB); ETS,t denotes unit feed-in tariff at period t in supply-area S, (RMB/kW h); IC represents input capacity, (MW); RTt strands for running time in period t, (h). a is the house-service consumption rate, (%); 4 is energy utilization efficiency, f ¼ h1  h2  h3  h4  h5 , among of them, h1 for straw storage efficiency, h2 for boiler efficiency, h3 for insulation efficiency of heatsupply piping, h4 for stream turbine efficiency, h5 for external supply efficiency, h1 ¼ 1  HL1 =B0 , h2 ¼ 1  HL2 =B0 , h3 ¼ ðGi  Qh Þ= ðGp  Qg Þ, h4 ¼ 1  HL4 =ðGi  Qh Þ, h5 ¼ ðQe þ E0 Þ=ðQe þ Eg Þ, B0 represents the straw energy content entered the biomass plant; B denotes the straw energy content entered the boiler burning; Gp represents boiler duty; Gi is the stream quantity of turbine inlet; Qg is the water heat of boiler feed; Qh is the stream heat of self-use; Eg is the electricity generation; Ez is power consumption of biomass power plant; E0 is the electricity of external supply; Qe is the heat of external supply; HL1 is the energy loss of straw storage; HL2 is the boiler energy loss; HL3 is the pipeline energy loss; HL4 is the energy loss of steam turbine generator, which is displayed in Fig. 4.

Since the amount of carbon dioxide needed in the growth of biomass material is equal to that emitted in the burning of the material. That is to say, net emission of carbon dioxide in biomass power generation is near to zero [11,42e45]. So certified emission reduction in biomass power plant is calculated according to the local power grid emission factor, which is described as follows:

CERt ¼ f  IC  RTt  EF

(4)

where CERt denotes the amount of certified emission reduction for biomass power plant in period t, (ton); EF stands for emission factor in local power grid, (t CO2/kW h). Because the development of biomass power is relatively late, relevant data of straw purchased price is limited. Moreover, the market of straw purchased price is full of uncertainty in the future. It will fluctuate with the prospect of biomass power generation industry. In this study, we assume straw purchased price follow a non-stationary stochastic process and are governed by a Geometric Brownian Motion (GBM) as follows [18,19,21,22,46]:

dFSPS ¼ m1 FSPS dt þ s1 FSPS dw

(5)

where FSPS is straw purchased price in supply-area S, (RMB/ton); pffiffiffiffiffi dw is independent increments of Wiener process du ¼ εt dt , where εt is a normally distributed random variable with mean 0 and standard deviation 1; m1 , and s1 represent the drift and variance parameters of straw purchased price. The total investment cost is one of the important factors affecting the scale development of biomass power generation. In this study, we assume the impact from technological improve on capital cost is strengthened with technology improvement, and obeys

IS ðtÞ ¼ IC  UIS ð0Þ  eat

(6)

where IS ðtÞ refers to the total capital cost in supply-area S in period t, (RMB); UIS(0) denotes the unit investment cost for biomass power plant in the base period in supply-area S, (RMB/kW); a is a parameter reflecting the effect of technology improvement on capital cost. The straw transportation cost mainly depends on the power plant size, straw availability, average transportation distance, straw density, carrying capacity, and the traveling speed [47]. In order to simply the transportation model, the average fuel consumption is adopted in this study, which would not differentiate full-load vehicle and empty vehicle in the process of straw transportation, which is described as follows:

756

X. Wang et al. / Energy 73 (2014) 751e761

TCBS  ð1  qÞ Q

where CRi denotes quantity of fuel consumed by transportation vehicle, (L/km); DIS is the average transportation distance in supply-area S, (km); PF is the price of fuel, (RMB/L); Q is the maximum load of transportation vehicle, (ton); q is the proportion of straw resource that transported directly to power plant, (%).

Thus, if the current straw purchased price is S, it will be uS, dS at the end of the next time step. Assuming that the final value of expected return of biomass power production was not relevant to the movement order (up-movement and down-movement), meaning u  d ¼ 1. The total investment value of biomass power production is studied based on straw purchase price under various stochastic processes. Let Dt be the time step in the model, r be the risk-free interest rate then,

3.1. Net present value of biomass power production investment

p¼

TRS ¼ 2$CR  DIS  PF 

(7)

If we assume the biomass power plant would be constructed at t ¼ t0, and project construction phase is one year, and the lifetime is t1, in other words, the running time of power plant is t ¼ t0 þ 1 until the end of the power plant lifetime. The next 20 years will be investigated for biomass power production investment in this study. Theoretically, the investor can reevaluate biomass power production in each year and then make a decision on whether or not to invest biomass power production. Here we assume the decision must be made at the beginning of each year. The government subsidy factor is kð0  k  1Þ, and the remaining total deployment investment is g, (%). Then, the net present value of power plant is formulated as follows:

NPVt0 ¼

t1 X

ðETS $QS þ HSS $PHS þ TCBS $j$PACS

t¼t0 þ2

þ CER$Pc $BIÞ$ð1 þ r0 Þt1 t 

t1 X

ðCo&m þ FSPS $TCBS

t¼t0 þ2

 þ TRS þ STAÞð1 þ r0 Þt1 t  ð1 þ r0 Þt0 $ IC  UIS ð0Þ   eat $ð1  gÞ$ð1  kÞ

erDt  d ud

(10)

pffiffiffiffi es Dt ,

where u ¼ d ¼ 1=u. The net present value (NPV) of a biomass power production on each node at each period can be obtained by Eq. (9), based on the calculation of straw purchased price under various stochastic processes. From the perspective of biomass power production investment containing option to defer, the investment value (IV) at each node is calculated as follows:

  IVi;j ¼ max IVi;j ; 0

(11)

If the NPV of biomass power production investment is negative at node (i, j), the investor will quit the project and the investment value is zero; if the NPV of biomass power production investment is positive, the investment will occur and the investment value is the NPV at node (i, j). At each node, the enterprise can reevaluate the biomass power production to decide whether to continue or abandon the investment. Thus, the total investment value containing real option value (TV) is recalculated step-by-step backward from the last time step to the current step as in Eq. (12).

o n   TVij ¼ max IVði;jÞ ; P  IViþ1;j þ ð1  PÞ  IViþ1;jþ1 erDt

¼ ðETS $QS þ HSS $PHS þ TCBS $j$PACS þ CER$Pc $BI  Co&m  FSPS $TCBS  TRS  STAÞ

(12)

t1 t0 1

ð1 þ r0 Þ 1 r0 ð1 þ r0 Þt1 t0

 ð1 þ r0 Þt1 $IC  UIS ð0Þ$eat1 $ð1  gÞ$ð1  kÞ (8) The above formula can be rewritten using continuously compounded interest as follows:

NPVt1 ¼ ðETS $QS þ HSS $PHS þ TCBS $j$PACS þ CER$Pc $BI er0  er0 ðt0 t1 Þ er0  1 ðr0 aÞt1  IC  UIS ð0Þ$e $ð1  gÞ$ð1  kÞ

The more economically favorable decision is determined from among the alternatives of construction of the biomass power production and deferment of the construction in the period of option to defer. Generally, the investment is treated as an American call option because it is exercised only if it brings benefit. The detailed decision rule is shown in Table 1. For a biomass power production based on real option theory, the total investment value (TV) should include two parts: net present value of traditional method (NPV) and real option value (ROV).

TV ¼ NPV þ ROV

 Co&m  FSPS $TCBS  TRS  STAÞ

(9)

3.2. Binomial tree model-based real options approach A binomial tree model is employed to calculate investment value of a biomass power production under real options approach. It is assumed that straw purchased price is allowed to go up or go down by a ratio with constant risk-free interest rate and volatility. Let Pu and Pd be the risk-neutral probabilities corresponding to the time when straw purchased price increases or decreases, respectively (the risk-neutral probability is the probability by which the risks in cash flow are adjusted toward neutral. This valuation of an option is treated as an American call option with dividend in finance) [17]. It means that straw purchased price has two possible values at each time step, u with probability Pu, d with probability Pd.

(13)

In general, the simulation methodology can be summarized as follows: Step 1: calculate the straw purchased price at each node of binomial tree model in the period of option to defer of biomass power production investment, the current straw purchased price is adopted as the initial straw purchased price, and the corresponding parameters are calculated from history data.

Table 1 The decision rule of biomass power project investment using delay real option. Traditional method

Investment value under real option rule

Decision

NPV > 0 NPV > 0

TV > NPV TV ¼ NPV

NPV  0 NPV < 0

TV > 0 TV ¼ 0

Executing the option to defer Abandon the option and invest immediately Executing the option to defer Abandon the investment

X. Wang et al. / Energy 73 (2014) 751e761

Step 2: calculate the NPV of biomass power production at each node of binomial tree model based on the straw purchased price obtained according to Eq. (9) in the period of option to defer of biomass power production investment. Step 3: make a decision at each node of the binomial tree model according to Eq. (11), which means the investor will quit the project if the NPV is negative. Step 4: calculate the biomass power production value containing real option value step-by-step from the last time step to the current step according to Eq. (12). The value at current step is the total investment value of biomass power production under the rule of real option. Step 5: calculate the real option value (ROV) according to Eq. (13). The investment timing of biomass power production is evaluated using a binomial tree calculation for option valuation. The decision tree, as a junction of two paths where either execution of investing on biomass power plant or its deferment, is chosen stepby-step, moving forward from the present year [17]. The probability of up-movement and down-movement to the next lattice is Pu and Pd. The cumulative probability of executing biomass power plant investment at year j is calculated as follows:

q¼



1 1  v pffiffiffiffiffiffi þ Dt ; 2 2 s

1 v ¼ u  s2 2

(14)

(d) (e) (f)

(g)

collection station/power plant is undertook by farmer according to analysis in Section 2; the crop straw purchased price in power plant is equal to that in the collection station; the agricultural acreage and straw species remain unchanged in supply-area S in planning period; the collection station only responsible for collecting and simple pretreatment, not responsible for storing, and all the biomass resource should transport to power plant in time; and the kind of biomass power plant in this study is direct-fired.

Two basic scenarios will be considered as described below: (i) the optimal investment decision is obtained without consideration of the potential revenue from certified emission reduction (Scenario 1); and (ii) the optimal investment decision is obtained with consideration of the potential revenue from certified emission reduction (Scenario 2). 3.4. Data collection and carbon price In this study, the biomass power plant with a 25 MW turbo-set driven by a boiler that the nominal capacity is 130 t/h, which adopt direct-combustion of straw resource, would be investigated. The average unit investment of the 25 MW project is 1.12  107 RMB/ MW. According to the practice, a 25 MW biomass power plant will require around 0.92  105 TCEs that it equivalent to 1.4 to 2.38  105 tons straws [48]. In terms of feed-in tariff for biomass

  4i;j ¼ 1  q $4i1;j1 þ q$4i1;j ; 41;1 ¼ 1; when executing the option to defer 4i;j ¼ 0; when abandon the option and invest immediately

4j ¼ 1 

j X

4i;j

(16)

i¼0

where q is the up-movement probability; u is the expected rate of change of the underlying asset; 4i;j denotes the probability of continuous occurrence of deferring to option in the path to lattice (i, j); 4j represents the cumulative probability of executing biomass power production investment at year j.

3.3. Assumptions and scenarios In order to simplify the model, several assumptions are made as follows: (a) it is assumed that the biomass resource in this study only refers to crop straw; (b) since investment in biomass power plant is financially unviable in the current situation, some kind of government incentive to encourage motivation is necessary. We assume two policy tools that the government can apply to assure the required return on investment in biomass power production in order to trigger off its development in China, i.e., electricity tariff, subsidization of capital cost; (c) the transportation cost in the model only refers to the distance between collection station and power plant, since the transportation cost of crop straw delivered from field to

757

(15)

power plant, the NDPC released the “Notice on Improvement of Power Pricing Policy for Agriculture and Forestry Biomass Generated” in July 2010, which regulated that the on-grid price for power generated from agriculture and forestry biomass is 0.75 RMB/kW h [49]. This value is adopted as feed-in tariff. Generally, the ash content from boiler and residua would be collected. As by-product of biomass power production, it has many uses which are chemical fertilizer, construction materials and adsorbent for the removal of various pollutants, especially after chemical and physical activation. The price of ash content is 300 RMB/ton based on our investigation. The average fuel (diesel) cost in 2012 is 7.5 RMB/L, and the transportation fuel consumption rate is 0.06 L/t km for medium-sized tractors [14,49]. Biomass power generation is attractive due to minimal carbon dioxide emission. Carbon dioxide emission reductions bring not only considerable environmental benefits to China, but also great economic benefits to biomass power generation enterprises [8]. Based on the previous research [25,50e53], it is reasonable to assume that carbon price process is a stochastic process with a certain amount of volatility and drift [54]. Carbon price follows a nonstationary stochastic process and is governed by a Geometric Brownian Motion (GBM) as follows:

dPc ¼ m2 Pc dt þ s2 Pc dw

(17)

where Pc is the carbon price for existing thermal power, (RMB/ kW h);pffiffiffiffiffi dw is independent increments of Wiener process du ¼ εt dt , where εt is a normally distributed random variable

758

X. Wang et al. / Energy 73 (2014) 751e761

with mean 0 and standard deviation 1; m2 and s2 represent the drift and variance parameters of the carbon price, respectively. The parameters (including m2 , s2 ) are estimated based on EUETS carbon price data collecting from 2009 to 2012 (monthly) according to the following step (because the carbon emission trading market is a new market, there is not enough historical data for the making of carbon price projections): op

Ut ¼

Ptþ1 op

Pt

;

ðt ¼ 0; 1; …; nÞ

(18)

where Ptop denotes the monthly carbon price in period t. The mean and variance of volatility of carbon price can be obtained through Eq. (19).

8 n > 1X > > U ¼ ðUt  1Þ > > n t¼0 < " # n > X > 2 > 2 > S2 ¼ 1 ðU  1Þ  nU > t : n  1 t¼0

(19)

The drift and variance parameters can be obtained by solving the following equation set.

8 > U > > S > > : s ¼ pffiffiffiffiffiffi Dt

(20)

4. Results and discussion Table 2 displays the results of NPV and real option value under two basic scenarios. The decision result is obtained according to the rule listed in Table 1. The NPV of a biomass power production are 185.18  106 RMB and 145.76  106 RMB for Scenarios 1 and 2, respectively, while the TV for both scenarios are 74.54  106 RMB and 77.24  106 RMB, respectively. The ROV for Scenarios 1 and 2 are 259.72  106 RMB and 223.00  106 RMB, respectively. The option to delay is executed for both the Scenarios 1 and 2 under the current condition according to the decision rule of ROA. The conclusion has profound significance in terms of guidance for decision-makers, but is not explicit. In order to make clear the gap between straw purchase price needed for an immediate investment of biomass power production and current price in Chinese market, the following studies are required. Biomass power production investment is executed immediately based on the two conditions: (1) NPV > 0; (2) the total investment value is equal to NPV. In order to quantify this gap, the threshold of straw purchase price above which it is optimal to invest immediately is investigated under various conditions. Fig. 5 illustrates the total investment value and the corresponding threshold of straw purchased price with various government subsidies under Scenario 1, on the assumption that the

Table 2 The investment decision of biomass power project using real option model under two scenarios. Scenario

Scenario 1 Scenario 2

Unit: 106 RMB

Decision

NPV

TV

ROV

185.18 145.76

74.54 77.24

259.72 223.00

Executing the option to defer Executing the option to defer

lifetime of a power plant is 20 years. As expected, the threshold of straw purchased price ascends noticeably with the increase of government subsidy. The total investment value also increases with government subsidy rising. For example, the threshold of straw purchased price is 133.64 RMB/ton when there is no government subsidy, while this value increases to 213.55 RMB/ton when the government gives a full grant for biomass power production. The total investment value is 74.54  106 RMB with no government subsidy, and 117.09  106 RMB with full government subsidy. The reason is mainly due to the fact that the government subsidy is considered as cash inflow and can partially offset the high investment cost of biomass power production. Moreover, the investor can afford higher straw purchased price with the government subsidy rising. Nevertheless, the current straw purchased price is as high as 220 RMB/ton in Chinese market, that is to say, even if given a full government subsidy; it is still not optimal for immediate investment due to the large cost on straw collection and transportation in China's farm field. Fig. 6 displays the total investment value and the corresponding threshold of straw purchased price under various government subsidies under Scenario 2. Compared with Scenario 1, the total investment value and the threshold of straw purchased price increase. For example, the total investment value is 77.24  106 RMB with no government subsidy, and 123.49  106 RMB with full government subsidy. The threshold of straw purchased price is 142.62 RMB/ton when there is no government subsidy and 218.87 RMB/ton when the government subsidy reaches to 1. The reason is due to the fact that the potential revenue from certified GHG-emission reduction can increase the cash inflow of biomass power production and narrow the gap between straw purchased price needed for investing immediately and the current straw purchased price in the market. Fig. 7 illustrates the cumulative probability of investing biomass power production in planning period, assuming the year when the cumulative probability arrives at a 4j ¼ 0:2 as the target year for executing the investment option. According to the above assumption, the investment of biomass power production will be executed at 2028 and 2022 for Scenarios 1 and 2 in the current situation in China if there is no government incentive to encourage motivation, respectively. Although the option to defer is executed under the current market, the revenue from GHG-certified emission reduction is conducive to biomass power production deployment in the long run. This can be mainly attributed to the revenue of CER partially offsetting the high straw purchased and transportation cost of biomass power production investment. From the perspective of the real option method, the investment value of biomass power production would increase with the uncertainties. The volatility of straw purchased price is an important influencing factor for biomass power production investment and represents the uncertainty of the straw supply market. Fig. 8 presents the influence of the volatility of straw purchased price on threshold price for the two scenarios, with various proportions of government subsidy. The threshold of straw purchased price is 133.64 RMB/ton for the Scenario 1 compared with 142.62 RMB/ton for Scenario 2 when the volatility of crop straw purchased price is 0.2 without a government subsidy. However, this value increases to 151.28 and 159.42 RMB/ton, 168.16 and 175.66 RMB/ton when the proportion of government subsidy reaches 0.2 and 0.4 for the two scenarios, respectively. The thresholds of straw purchased price increase with the government subsidy rising, and decrease with the increase of the volatility of straw purchased price. The reason is that an increasing in the volatility of straw purchased price means that the uncertainty of the investment value increase, which could potentially decrease the total cash outflows. All these measures are advantageous to biomass power production deployment, and could

14000

FSP

12000

200

10000

150

8000

100

6000 4000

50

2000

0

6

TV

(10 RMB)

250

759

Total investment value

Threshold of straw purchased price (RMB/ton)

X. Wang et al. / Energy 73 (2014) 751e761

0 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

Government subsidy factor (k)

14000

FSP

12000

200

10000

150

8000

100

6000 4000

50

2000

0

(10 RMB)

TV

6

250

Total investment value

Threshold of straw purchased price (RMB/ton)

Fig. 5. The total investment value and threshold of straw purchased price under various proportions of government subsidies k for Scenario 1.

0 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

Government subsidy factor (k) Fig. 6. The total investment value and threshold of straw purchased price under various proportions of government subsidies k for Scenario 2.

narrow the gap between the straw purchased price needed for biomass power production investment of both scenarios and current price in the market.

5. Conclusion

Cumulative probability

In this study, a policy profit real options model was developed to evaluate the investment of biomass power production from the perspective of power generation enterprise. Uncertainties in straw purchased price, government incentives, and technological improvements were considered. Using current market data, it was not

0.40 Scenario 2 Scenario 1 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 2013 2015 2017 2019 2021 2023 2025 2027 2029 2031 2033 Year

Fig. 7. The cumulative probability of investing biomass power project immediately in planning period for two scenarios.

optimal for immediate investment in biomass power production. Given full government subsidy, the thresholds of straw purchased price for Scenarios 1 and 2 were 213.55 and 218.87 RMB/ton, respectively. The current straw purchased price in Chinese market was 220 RMB/ton. The rise in government subsidy was advantageous to biomass power production investment and could narrow the gap between straw purchased price needed for immediate investment and current price in the market since it was considered to increase the cash inflow. The investor could afford an increased straw purchased price with the consideration of the revenue from CER. The optimal investment timing of biomass power production was evaluated by binomial tree analysis for the two scenarios. Although the option to defer was executed under current market circumstance, the revenue from CER was conducive to biomass power production investment in the long run. The investment of biomass power production would be executed at 2028 and 2022 for Scenarios 1 and 2 in the current situation in China if there is no government incentive to encourage motivation, respectively. The results obtained indicate that the gap still exists between the straw purchased price needed for immediate investment of biomass power production and current price in the market, even if the government gives a full grant. The results obtained would be particularly useful in providing optimal timing for investment decisions under uncertainty, and also provide information for power enterprises' biomass power production evaluation and related policy-making in an uncertain environment. The methods in this study can be extended to any

X. Wang et al. / Energy 73 (2014) 751e761

Thresho ld o f stra w purcha sed price (RMB/ton)

760

200 180 160 140 120 100 80 60 40 20 0

k=0

k = 0.2

S1

S2 σ = 0.2

k = 0.4

S1

k = 0.6

S2

S1

σ = 0.4

S2 σ = 0.6

The volatility of straw purchased price Fig. 8. The influence of the volatility of straw purchased price on threshold price with various proportions of government subsidy.

similar utilities considering investment in alternative opportunities under uncertainty. Acknowledgements This work was supported by the National Science Foundation for Innovative Research Group (No. 51121003), and the special fund of State Key Lab of Water Environment Simulation (11Z01ESPCN). The authors would like to extend special thanks to the editor and the anonymous reviewers for their constructive comments and suggestions in improving the quality of this paper. References [1] Kotowicz J, Sobolewski A, Iluk T. Energetic analysis of a system integrated with biomass gasification. Energy 2013;52:265e78. n F. Assessment of CCHP systems based on [2] Maraver D, Sin A, Royo J, Sebastia biomass combustion for small-scale applications through a review of the technology and analysis of energy efficiency parameters. Appl Energy 2013;102:1303e13. [3] Molino A, Giordano G, Motola V, Fiorenza G, Nanna F, Braccio G. Electricity production by biomass steam gasification using a high efficiency technology and low environmental impact. Fuel 2013;103:179e92. [4] Wang N, Chang Y-C. The evolution of low-carbon development strategies in China. Energy 2014;68:61e70. [5] Lee U, Balu E, Chung JN. An experimental evaluation of an integrated biomass gasification and power generation system for distributed power applications. Appl Energy 2013;101:699e708. [6] Kwon PS, Østergaard PA. Priority order in using biomass resources e energy systems analyses of future scenarios for Denmark. Energy 2013;63:86e94. [7] Yu H, Wang Q, Ileleji KE, Yu C, Luo Z, Cen K, et al. Design and analysis of geographic distribution of biomass power plant and satellite storages in China. Part 2: road delivery. Biomass Bioenergy 2012;46:785e92. [8] Zhao X, Wang J, Liu X, Feng T, Liu P. Focus on situation and policies for biomass power generation in China. Renew Sustain Energy Rev 2012;16(6):3722e9. [9] Shafie SM, Masjuki HH, Mahlia TMI. Life cycle assessment of rice straw-based power generation in Malaysia. Energy 2014;70:401e10. [10] Zhao X, Wang J, Liu X, Liu P. China's wind, biomass and solar power generation: what the situation tells us? Renew Sustain Energy Rev 2012;16(8): 6173e82. [11] Zhao Z, Yan H. Assessment of the biomass power generation industry in China. Renew Energy 2012;37(1):53e60. [12] Wu CZ, Huang H, Zheng SP, Yin XL. An economic analysis of biomass gasification and power generation in China. Bioresour Technol 2002;83(1):65e70. [13] Liu J, Wu J, Liu F, Han X. Quantitative assessment of bioenergy from crop stalk resources in Inner Mongolia, China. Appl Energy 2012;93:305e18. [14] Zhang Q, Zhou D, Zhou P, Ding H. Cost analysis of straw-based power generation in Jiangsu Province, China. Appl Energy 2013;102:785e93. [15] Sun Y, Wang R, Liu J, Xiao L, Lin Y, Kao W. Spatial planning framework for biomass resources for power production at regional level: a case study for Fujian Province, China. Appl Energy 2013;106:391e406. [16] Yang M, Blyth W, Bradley R, Bunn D, Clarke C, Wilson T. Evaluating the power investment options with uncertainty in climate policy. Energy Econ 2008;30(4):1933e50.

[17] Zhang X, Wang X, Chen J, Xie X, Wang K, Wei Y. A novel modeling based real option approach for CCS investment evaluation under multiple uncertainties. Appl Energy 2014;113:1059e67. ~ a EA, Mutale J. Application of an advanced real options [18] Martínez-Cesen approach for renewable energy generation projects planning. Renew Sustain Energy Rev 2011;15(4):2087e94. [19] Boomsma TK, Meade N, Fleten S-E. Renewable energy investments under different support schemes: a real options approach. Eur J Operational Res 2012;220(1):225e37. [20] Kumbaroglu G, Madlener R, Demirel M. A real options evaluation model for the diffusion prospects of new renewable power generation technologies. Energy Econ 2008;30(4):1882e908. [21] Kjærland F. A real option analysis of investments in hydropowerdthe case of Norway. Energy Policy 2007;35(11):5901e8. [22] Lee S-C, Shih L-H. Renewable energy policy evaluation using real option model e the case of Taiwan. Energy Econ 2010;32:S67e78. [23] Reuter WH, Fuss S, Szolgayov a J, Obersteiner M. Investment in wind power and pumped storage in a real options model. Renew Sustain Energy Rev 2012;16(4):2242e8. [24] Detert N, Kotani K. Real options approach to renewable energy investments in Mongolia. Energy Policy 2013;56:136e50. [25] Zhou W, Zhu B, Fuss S, Szolgayova J, Obersteiner M, Fei W. Uncertainty modeling of CCS investment strategy in China's power sector. Appl Energy 2010;87(7):2392e400. [26] Kato M, Zhou Y. A basic study of optimal investment of power sources considering environmental measures: economic evaluation of CCS through a real options approach. Electr Eng Jpn 2011;174(3):9e17. [27] Chen L, Li X, Wen W, Jia J, Li G, Deng F. The status, predicament and countermeasures of biomass secondary energy production in China. Renew Sustain Energy Rev 2012;16(8):6212e9. [28] Zeng X, Ma Y, Ma L. Utilization of straw in biomass energy in China. Renew Sustain Energy Rev 2007;11(5):976e87. [29] Chen Y, Hu W, Sweeney S. Resource availability for household biogas production in rural China. Renew Sustain Energy Rev 2013;25:655e9. [30] Zhang PD, Yang YL, Li GQ, Li XR. Energy potentiality of crop straw resources in China. Renew Energy Resour 2008;25(6):80e3 [in Chinese]. [31] Gao LW, Ma L, Zhang WF, Wang FH, Ma WQ, Zhang FS. Estimation of nutrient resource quantity of crop straw and its utilization situation in China. Trans CSAE 2009;25(7):173e9 [in Chinese]. [32] Yu H, Wang Q, Ileleji KE, Yu C, Luo Z, Cen K, et al. Design and analysis of geographic distribution of biomass power plant and satellite storages in China. Part 1: straight-line delivery. Biomass Bioenergy 2012;46:773e84. [33] Yu XD, Fan FM. Simulation analysis on fuel collection, processing, storage and transportation used in straw power plant. Trans CSAE 2009;25(10):215e9 [in Chinese]. [34] Long H, Li X, Wang H, Jia J. Biomass resources and their bioenergy potential estimation: a review. Renew Sustain Energy Rev 2013;26:344e52. [35] McIlveen-Wright DR, Huang Y, Rezvani S, Redpath D, Anderson M, Dave A, et al. A technical and economic analysis of three large scale biomass combustion plants in the UK. Appl Energy 2013;112:396e404. [36] Cai YP, Huang GH, Yang ZF, Lin QG, Tan Q. Community-scale renewable energy systems planning under uncertainty-An interval chance-constrained programming approach. Renew Sustain Energy Rev 2009;13(4):721e35. [37] Cai YP, Huang GH, Tan Q, Yang ZF. An integrated approach for climate-change impact analysis and adaptation planning under multi-level uncertainties. Part Methodol Renew Sustain Energy Rev 2011;15(6):2779e90. [38] Cai YP, Huang GH, Tan Q, Liu L. An integrated approach for climate-change impact analysis and adaptation planning under multi-level uncertainties. Part II. Case study. Renew Sustain Energy Rev 2011;15(6):3051e73.

X. Wang et al. / Energy 73 (2014) 751e761 [39] Wang X, Cai Y, Chen J, Dai C. A grey-forecasting interval-parameter mixedinteger programming approach for integrated electric-environmental managementda case study of Beijing. Energy 2013;63:334e44. [40] Zuwała J. Life cycle approach for energy and environmental analysis of biomass and coal co-firing in CHP plant with backpressure turbine. J Clean Prod 2012;35:164e75. [41] Osmani A, Zhang J. Optimal grid design and logistic planning for wind and biomass based renewable electricity supply chains under uncertainties. Energy 2014;70:514e28. [42] Alnatheer O. Environmental benefits of energy efficiency and renewable energy in Saudi Arabia's electric sector. Energy Policy 2006;34(1):2e10. [43] Nguyen TLT, Hermansen JE, Mogensen L. Environmental performance of crop residues as an energy source for electricity production: the case of wheat straw in Denmark. Appl Energy 2013;104:633e41. [44] Ajanovic A, Haas R. CO2-reduction potentials and costs of biomass-based alternative energy carriers in Austria. Energy 2014;69:120e31. [45] Bentsen NS, Jack MW, Felby C, Thorsen BJ. Allocation of biomass resources for minimising energy system greenhouse gas emissions. Energy 2014;69:506e15. [46] Lin TT, Ko C-C, Yeh H-N. Applying real options in investment decisions relating to environmental pollution. Energy Policy 2007;35(4):2426e32. [47] Shabani N, Sowlati T. A mixed integer non-linear programming model for tactical value chain optimization of a wood biomass power plant. Appl Energy 2013;104:353e61. [48] Fang JT, Song JH, Zhan ZG, Li J. Study on combustion adjustment characteristics of 50MW CFB biomass-fired boiler in Zhanjiang biomass power plant. Adv Mater Res 2013;608:242e8.

761

[49] Jiang DM, Zhu PX. Comprehensive utilization of rural straw resource in Jiangsu Province-economic analysis on biomass power generation in Yancheng. Chin J Eco-Agriculture 2008;16(5):1208e12. [50] Zhu L, Fan Y. A real options-based CCS investment evaluation model: case study of China's power generation sector. Appl Energy 2011;88(12): 4320e33. [51] Abadie LM, Chamorro JM. European CO2 prices and carbon capture investments. Energy Econ 2008;30(6):2992e3015. [52] Fuss S, Szolgayova J, Obersteiner M, Gusti M. Investment under market and climate policy uncertainty. Appl Energy 2008;85(8):708e21. [53] Heydari S, Ovenden N, Siddiqui A. Real options analysis of investment in carbon capture and sequestration technology. Comput Manag Sci 2012;9(1): 109e38. [54] Insley MC. On the option to invest in pollution control under a regime of conomique tradable emissions allowances. Can J Economics/Revue Can d'e 2003;36(4):860e83. [55] Cai YP, Huang GH, Tan Q, Yang ZF. Identification of optimal strategies for energy management systems planning under multiple uncertainties. Appl Energy 2009;86(4):480e95. [56] Tan Q, Huang GH, Cai YP. A superiority-inferiority-based inexact fuzzy stochastic programming approach for solid waste management under uncertainty. Environ Model Assess 2010;15(5):381e96. [57] Cai YP, Huang GH, Yang ZF, Sun W, Chen B. Investigation of public’s perception towards rural sustainable development based on a two-level expert system. Expert Syst Appl 2009;36(5):8910e24.