3.2. Integrated Assessment Method of Emergency Plan
According to the proposed integrated assessment method of emergency plan, the four kinds of emergency plans discussed previously are evaluated, and the optimal plan is selected. The calculation process and algorithm are shown in
Figure 4 and the methodology steps together with the results are explained as follows.
Step 2: Calculation of objective weight μj for the indicator
The data in
Table 4 was normalized and then the normalized matrix
Y are obtained as shown in
Equation (S1) (Electronic Supplementary Material). The entropy of each indicator is obtained according to Equations (2) to (3). The entropy
ej and entropy weight of the indicators
μj are obtained according to Equations (4) to (5). The results are shown in
Table 5.
Step 3: Calculation of subjective weight λj for the indicator
We calculate the subjective weight
λj by using the AHP method for taking the subjective attributes of the data into consideration. After expert consultation, the judgment matrices of four kinds of emergency plans are obtained. Take the emergency plan one as an example for detailed introduction. Based on the principles of
Table 1 and on experts’ experience and knowledge, the calculations of the local weights of events B
1–B
4, resistance risk event C
11–C
13, timeliness event C
21–C
23, economy events C
31–C
33, and feasibility events C
41–C
44 were shown as the values in
Tables S6–S10 (in Electronic Supplementary Material). The local weight and global weight of each evaluation indicator could be achieved according to Equations (1) to (4) in Long’s paper (2016).
Step 4: Calculation of integration weighting ωj
The integration weighting
ωj of each indicator is calculated according to Equations (6) to (9). The results are shown in
Table 6.
Step 5: Building a weighted normalized decision matrix
Step 6: Calculate the relative closeness to the ideal solution
According to Equations (13) to (14), the distance from each alternative to the positive ideal solution and the negative ideal solution are calculated as shown in
Equations (S3) and (S4) (in Electronic Supplementary Material). Then, the relative closeness to the ideal solution of each emergency plan are calculated based on Equation (15), the result is shown in Equation (22).
Step 7: Calculate the inhomogeneity
The larger the discrete degree of an indicator is, the larger influence it may have in the decision-making process, and thus, the smaller the information entropy, the reliability of the scheme, and vice versa. The Shannon entropy of each emergency plan is calculated according to Equations (16) to (17), and the result is as shown in Equation (23).
Step 8: Determining the optimal emergency plan
The larger D is, the better the corresponding emergency plan is. Therefore, according to Equation (21), the optimal emergency plan is the emergency plan 3.
3.3. Discussion
This work presents an approach for assessing the emergency plan by combining improved TOPSIS, Shannon entropy and Coordinated development degree model. The Shannon entropy method was used to analyze different types of index values. TOPSIS is used to calculate the relative closeness to the ideal solution. Coordinated development degree model is applied to express the relationship between the relative closeness and inhomogeneity of the emergency plan, but their potential in emergency plan assessment has rarely been investigated in the literature.
The assessment results of two kinds of emergency plan by different methods such as TOPSIS and the integrated emergency plan are presented in
Table 7. We find out the differences between the two methods in the emergency plans and the results are shown in
Table 8. EP 2 and EP 4 gain the same ranking in both assessment methods. At the same time, the rankings of EP 1 and EP 3 are reversed by two methods. The optimal plan will change when we take into account the internal inhomogeneity of the indicators in one evaluated object.
The major driving factors of the water transfer project contingency plan can also be determined by the integrated assessment method. The weights of the indicators determined by the subjective weighting method and the objective weighting method are quite different, and the main driving factors cannot be determined. Therefore, we determine the main drivers based on the rankings of the indicators that affect the emergency plan. As shown in
Figure 5, it can be seen that the ranking of index weights determined by subjective weighting method such as indicators C21, C32 and C33 are the main influencing factors; while the index weighting order determined by objective weighting method can be seen, indicators C23, C41 and C43 are the main influencing factors. The rankings of indicators determined by objective weighting method and subjective weighting method are not the same. The reasons why the discrepancies occurred may be the incompleteness of the data and the limitations of the evaluation experts. The combined integration weight will effectively make up for these deficiencies, and improve the weight of each index to be more reasonable, in order to show the status of this indicator in the evaluation of the whole event more accurately. In the combined integration weight ranking, indicators C33, C31, C21 and C43 are in a prominent position. These rankings indicate that indicator C33 (implementation effect) plays an important and decisive role in the evaluation of emergency plans. The main factors affecting indicator C33 are water quality level after disposal, the quantity of abandoned water and damage degree caused in the water transfer project. The level of water quality after disposal indicates the degree of pollution impact, and the quantity of abandoned water and damage degree caused in the water transfer project represent the economic losses. In the realistic practice, the main considerations for decision makers when they need to determine the emergency plan are the effect of planned disposal and the economic loss that may be caused. Therefore, the combined integration weight method proposed in this paper is not only reasonable, but also relatively scientific.
In the process of evaluating the four emergency plans, we consulted five relevant experts. Their evaluation results for each plan are shown in
Figure 6 by different colors of the line. It can be seen from the evaluation results that the index fluctuation is small, except for individual indicators. The small fluctuations in the evaluation results of the indicators are mainly due to the fact that the evaluation experts we chosen are both familiar with environmental and water transport engineering, as well as, mainly because the knowledge level and the engineering experience between them are not too much different. Therefore, it is necessary to further improve the expert database in order to deal with other more different situations. Experts are not limited to the related scientific majors, and we should consult more staff member in many aspects, such as skilled technicians and managers in relevant projects. Comparing the four figures in
Figure 6, it can be seen that the indicator C43 (contingency response to accident evolution) in the emergency plans 1 and 2 is relatively low. The main reason is that these two plans did not consider controlling pollutants in the accident pool as soon as possible, leading to an increase in the diffusion range of pollutants. Moreover, in relative terms, the variability and resilience of emergency plans 1 and 2 are both poor. That is why we did not take these two plans into the consideration of our actual project.
In the emergency plan 4, the value of indicator C23 (supplies and speed of restoring water delivery) is slightly lower. This is mainly because in emergency plan 4, in order to ensure the safety of the water transfer project, the water level before the sluice gates is required to be allowed to increase by 30 cm, resulting in a complicated control of the gates and a long time being required for the water level to stabilize. However, in actual projects, if a major pollution event occurs, the scope of pollution should be reduced at first, then secondly, the safety of the water delivery project should be considered. Therefore, comprehensive consideration shows that contingency plan 4 is not very reasonable.
In emergency plan 3, the evaluation shows that each index is relatively uniform, and there is no excessive fluctuation. There is no doubt that the experts have high recognition of emergency plan 3, so emergency plan 3 is considered to be reasonable.
In summary, although the emergency plan 1 and 2 are relatively simple to operate, (1) the gate closing time is too fast, causing great damage to the channel; (2) the pollutants are disposed in the channel, and the cost is relatively high; (3) the polluted water after the disposal treatment still below the standard level and we cannot abandon the water directly, (4) the flexibility of the program is weak. While emergency plan 3 looks rather cumbersome, the gates on the Middle Route of the South-to-North Water Transfer Project are all automated, so they’re simple to operate in an actual project. The emergency plan 3 is the recommended solution in the decision support system [
19]. Therefore, the integrated assessment method proposed in this paper is relatively reasonable.
In this study, we introduced an improved integrated assessment method that can evaluate the emergency plan more objectively and reasonably. The main advantages of the proposed method are as follows: (1) The relative closeness to the ideal solution using TOPSIS and the inhomogeneity of all indicators based on the Shannon entropy method are synthetically considered to make the evaluation results closer to the actual situation. (2) The entropy method that considers the objective attributes to the data and the AHP method that takes into account the subjective characteristics of the data are combined, and the final determined index weight is more reasonable.
However, the proposed method also has certain deficiencies. (1) The optimal emergency plan is still limited by the experts’ experience and knowledge to some extent. Therefore, we will further improve the expert database to reduce the impact of professional knowledge and experience. (2) Although the indicator system can reflect the relationship between human society and the environment, it was imperfectly scientific and reasonable. Therefore, we will explore the multi-faceted factors affecting the effectiveness of the emergency plan and establish a more perfect indicator system in further research.