CN103345662A

CN103345662A - Commerce efficiency comprehensive evaluation method based on small-wave network method

Info

Publication number: CN103345662A
Application number: CN2013102940337A
Authority: CN
Inventors: 李慧玲; 王鑫; 张国忠; 刘爱民; 段丽荣; 辛五一; 张建华; 曾博; 许健; 杨煦; 刘大川
Original assignee: ZHANGJIAKOU POWER SUPPLY Co OF JIBEI ELECTRIC POWER CO Ltd; State Grid Corp of China SGCC
Current assignee: ZHANGJIAKOU POWER SUPPLY Co OF JIBEI ELECTRIC POWER CO Ltd; State Grid Corp of China SGCC
Priority date: 2013-07-12
Filing date: 2013-07-12
Publication date: 2013-10-09

Abstract

The present invention relates to a comprehensive evaluation method of commercial energy efficiency based on the wavelet network method. The method comprises the following steps: 1) establishing a comprehensive evaluation index system for energy efficiency projects; 2) calculating the individual index values of the comprehensive evaluation index system; 3) using the wavelet network The method determines the optimal weight of each index; 4) Calculates the comprehensive evaluation value of each alternative plan, and outputs the optimal plan. The present invention uses the wavelet network method to iteratively calculate the network parameters wi _j , r _j , b _j , a _j , so as to find the parameter values closest to the expert scores of each training program, and finally obtain the comprehensive evaluation results of each energy efficiency program. The different characteristics of different schemes comprehensively and systematically measure the advantages and disadvantages of each scheme, and give quantitative evaluation results, so that project decision makers can directly compare the overall advantages and disadvantages of each scheme, so as to select energy efficiency schemes. The wavelet network method is applied to the comprehensive evaluation of commercial energy efficiency, with small calculation errors and fast convergence speed.

Description

A Comprehensive Evaluation Method of Commercial Energy Efficiency Based on Wavelet Network Method

技术领域technical field

本发明涉及能效项目综合评价方法领域，特别是涉及一种基于小波网络法的商业能效综合评价方法。The invention relates to the field of comprehensive evaluation methods for energy efficiency projects, in particular to a comprehensive evaluation method for commercial energy efficiency based on wavelet network method.

背景技术Background technique

对能效项目进行综合评价的目的是为了衡量拟实施项目对投资方、电网公司、终端用户及全社会的预期贡献效果并明确项目所能带来的相关经济、环境及社会效益。评价指标是进行上述科学评价的基础。The purpose of comprehensive evaluation of energy efficiency projects is to measure the expected contribution effect of the proposed project to investors, power grid companies, end users and the whole society, and to clarify the relevant economic, environmental and social benefits that the project can bring. Evaluation indicators are the basis for the above-mentioned scientific evaluation.

现阶段在工业、建筑领域的能效项目评估方法相对比较成熟，中小型企业中的常用生产设备,如锅炉及蒸汽系统,水泵、风机和空气压缩机等电机系统,照明系统,暖通空调系统,都建立了相应的效率计算和节能效果分析模型。但这些计算模型仅仅从设备的节电水平和经济程度出发，忽略了能效提高带来的环境效益和社会效益，评价内容缺乏完整性。At present, the evaluation methods of energy efficiency projects in the industrial and building fields are relatively mature. Commonly used production equipment in small and medium-sized enterprises, such as boilers and steam systems, motor systems such as water pumps, fans and air compressors, lighting systems, HVAC systems, The corresponding efficiency calculation and energy-saving effect analysis models have been established. However, these calculation models only start from the power saving level and economical degree of the equipment, ignoring the environmental and social benefits brought about by the improvement of energy efficiency, and the evaluation content lacks integrity.

与此类似，现有的商业领域的能效评价指标往往只局限于各终端用电设备的节电能力，如空调，电冰箱，洗衣机，变压器的能效评价，并没有考虑能效项目实施后对投资方、电网公司的经济、环境和社会效益的影响，或是未设定相应的量化值，使得各能效方案没有统一的评判标准，各种独立的能效指标反应出的某项目的能效特点不能提供项目决策者和投资者选择投资方案的有效建议。Similar to this, the existing energy efficiency evaluation indicators in the commercial field are often limited to the power-saving capabilities of each end-use equipment, such as the energy efficiency evaluation of air conditioners, refrigerators, washing machines, and transformers, without considering the impact on investors after the implementation of energy efficiency projects , the economic, environmental and social benefits of the power grid company, or the corresponding quantitative value has not been set, so that there is no unified evaluation standard for each energy efficiency scheme, and the energy efficiency characteristics of a certain project reflected by various independent energy efficiency indicators cannot provide project information. Effective advice for decision makers and investors in choosing investment options.

发明内容Contents of the invention

本发明首次提出了能够对不同能效方案进行全面、科学、客观评价的综合能效指标体系，采用小波网络法对网络参数wi_j、r_j、b_j、a_j进行迭代计算，从而找到与各训练方案的专家打分最接近的参数值，最终得到各能效方案的综合评价结果，综合不同方案的不同特点，全面系统的衡量各方案的优劣，给出量化的评价结果，使得项目决策者能够直截了当的比较出各个方案的总体优劣情况，从而进行能效方案选择。小波网络法应用于商业能效综合评价，计算误差小，收敛速度快。The present invention proposes for the first time a comprehensive energy efficiency index system capable of comprehensively, scientifically and objectively evaluating different energy efficiency schemes. The wavelet network method is used to iteratively calculate the network parameters wi _j , r _j , b _j , a _j , so as to find the The experts of the scheme score the closest parameter value, and finally get the comprehensive evaluation result of each energy efficiency scheme, integrate the different characteristics of different schemes, comprehensively and systematically measure the pros and cons of each scheme, and give quantitative evaluation results, so that project decision makers can be straightforward Compare the overall advantages and disadvantages of each scheme, so as to select the energy efficiency scheme. The wavelet network method is applied to the comprehensive evaluation of commercial energy efficiency, with small calculation errors and fast convergence speed.

本发明为了得到科学、客观、正确的评价结果以指导做出针对当前能效限制因素的最佳工程决策，紧紧抓住了能效工程项目的目的来设计评价指标，把每一项指标的选取都围绕在能效项目所产生的经济性、收益性以及外部社会效益等方面进行考虑。另外，指标的设计全部具有实际应用性，这样有助于数据的采集和量化，从而把概念化的评价指标转换成我们所容易理解及获得的指标进行处理。In order to obtain scientific, objective and correct evaluation results to guide the best engineering decision-making for the current energy efficiency limiting factors, the present invention firmly grasps the purpose of energy efficiency engineering projects to design evaluation indicators, and selects each index Consider the economics, profitability and external social benefits generated by energy efficiency projects. In addition, the design of the indicators is all practical, which is helpful for the collection and quantification of data, so that the conceptual evaluation indicators can be converted into indicators that we can easily understand and obtain for processing.

小波网络（wavelet network，WN）是小波理论与神经网络相结合的产物，它是小波分解与前馈神经网络的融合。由于兼备了小波变换良好的时频局域化性质与神经网络自学习功能,小波网络已逐步成为评价、预测领域所采用的新方法。对于复杂对象系统的多属性综合评价，在统一指标类型的基础上，利用评价指标的无量纲数据，通过小波网络的学习，得到专家知识，建立由评价指标属性值到输出综合评价值的非线性映射关系。在对其他类似问题进行评价时，输入待评价对象的指标数据向量，即可经网络计算得到其综合评价值，从而达到自动运行，快速评价及决策支持的目的。Wavelet network (wavelet network, WN) is the product of the combination of wavelet theory and neural network, which is the fusion of wavelet decomposition and feedforward neural network. Due to the good time-frequency localization properties of wavelet transform and the self-learning function of neural network, wavelet network has gradually become a new method in the field of evaluation and prediction. For the multi-attribute comprehensive evaluation of complex object systems, on the basis of unifying the index type, using the dimensionless data of the evaluation index, through the learning of the wavelet network, the expert knowledge is obtained, and the nonlinearity from the attribute value of the evaluation index to the output comprehensive evaluation value is established. Mapping relations. When evaluating other similar problems, the index data vector of the object to be evaluated can be input, and its comprehensive evaluation value can be obtained through network calculation, so as to achieve the purpose of automatic operation, rapid evaluation and decision support.

本发明的目的通过以下技术方案来实现：The purpose of the present invention is achieved through the following technical solutions:

一种基于小波网络法的商业能效综合评价方法包括以下步骤：A comprehensive evaluation method of commercial energy efficiency based on wavelet network method includes the following steps:

1）建立能效项目综合评价指标体系；1) Establish a comprehensive evaluation index system for energy efficiency projects;

2）计算综合评价指标体系各单项指标值；2) Calculate the individual index values of the comprehensive evaluation index system;

3）运用小波网络的方法确定各指标的最佳权重；3) Use the method of wavelet network to determine the optimal weight of each index;

4）计算各备选方案的综合评价值，输出最优方案。4) Calculate the comprehensive evaluation value of each alternative plan, and output the optimal plan.

所述步骤（1）建立能效项目综合评价指标体系包括一级指标及其相应的二级指标，每一个二级指标又各自扩展成体现本级指标具体内容的三级指标，即形成了自上而下支配的三级层次结构。The step (1) establishes a comprehensive evaluation index system for energy efficiency projects, including first-level indicators and their corresponding second-level indicators. while the lower governs the three-level hierarchy.

所述一级指标内容包括：电网运行、环境效应、经济效益及资源利用效率四个方面。The first-level indicators include four aspects: power grid operation, environmental effects, economic benefits, and resource utilization efficiency.

所述二级指标由电网可靠性评价及负荷整形能力评价两方面内容组成；所述一级指标中的环境效应部分包括污染物排放和碳排放两方面内容；所述一级指标中的经济效益部分由主要包括财务评价、投资收益评价和全寿命周期成本分析三个方面内容组成；所述一级指标中的资源利用效率部分包括节能效果及电网侧资源使用两方面内容。The second-level indicators are composed of power grid reliability evaluation and load shaping capability evaluation; the environmental effect part of the first-level indicators includes pollutant emissions and carbon emissions; the economic benefits of the first-level indicators The part consists of three aspects: financial evaluation, investment return evaluation, and life cycle cost analysis; the resource utilization efficiency part of the first-level indicators includes two aspects: energy saving effect and power grid side resource use.

所述建立的综合评价指标体系的二级指标的可靠性评价包含系统平均缺电电量ASCI变化率及系统总电量不足量ENS变化率两项指标；所述二级指标中的负荷整形能力部分包含日负荷率、日峰谷差率、年负荷率及年最大峰谷差率变化四项指标；所述二级指标中的污染物排放部分包含氮氧化物减排量、硫氧化物减排量、粉尘减排量以及PM2.5减排量四项指标；所述二级指标中的碳排放一层对应为CO₂减排量指标；所述二级指标中的财务评价部分由净现值、内部收益率以及投资回收期三项指标进行系统分析计算；所述二级指标中的投资收益部分选取收益费用比率PCR及可避免用电成本AC2项指标以衡量能效项目的投资收益情况；所述二级指标中的全寿命周期成本部分采用寿期成本下降率作为衡量能效项目全寿命周期经济性的评价指标；所述二级指标中的节能效果部分扩展为终端年节电量及用电收益效率差值两项指标。The reliability evaluation of the secondary indicators of the established comprehensive evaluation index system includes two indicators: the ASCI change rate of the system's average power shortage and the ENS change rate of the system's total power shortage; the load shaping capability part of the second-level indicators includes The four indicators of daily load rate, daily peak-to-valley difference rate, annual load rate and annual maximum peak-to-valley difference rate change; the pollutant emission part in the second-level indicators includes nitrogen oxide emission reduction and sulfur oxide emission reduction , dust emission reduction and PM2.5 emission reduction four indicators; the first level of carbon emission in the second-level indicators corresponds to the _CO2 emission reduction indicator; the financial evaluation part of the second-level indicators is determined by the net present value Systematic analysis and calculation of the three indicators of internal rate of return, internal rate of return, and investment payback period; the investment income part of the second-level indicators selects the income expense ratio PCR and the avoidable electricity cost AC2 indicators to measure the investment income of energy efficiency projects; The life-cycle cost part of the above-mentioned second-level indicators adopts the life-cycle cost reduction rate as an evaluation index to measure the life-cycle economics of energy efficiency projects; the energy-saving effect part of the above-mentioned second-level indicators is extended to terminal annual electricity saving and electricity consumption income Two indicators of efficiency difference.

所述评价指标体系中三级指标权重值采用小波网络的方法求解，然后采用线性加权综合法对备选方案进行综合性评分，选取综合评分最高者为最优方案，具体实施步骤为：The three-level index weight values in the evaluation index system are solved by the wavelet network method, and then the linear weighted comprehensive method is used to comprehensively score the alternatives, and the one with the highest comprehensive score is selected as the optimal solution. The specific implementation steps are:

1)收集商业能效项目的基础资料，计算综合评价指标体系中的三级指标值；1) Collect the basic data of commercial energy efficiency projects, and calculate the three-level index values in the comprehensive evaluation index system;

2)根据专家意见，给可选能效方案初步打分，得到各方案的专家评议分数集合

2) According to expert opinions, give preliminary scores to the optional energy efficiency schemes, and obtain the set of expert evaluation scores for each scheme

3)将步骤1）中计算出的各值属性值向量{x_k(i)}转化为指标属性一致的数据{r_k(i)}；3) Transform the attribute value vector {x _k (i)} of each value calculated in step 1) into data {r _k (i)} with consistent index attributes;

将三种类型的指标做如下处理：The three types of indicators are handled as follows:

式中：r_ij为第i个方案中第j个指标值，r_ij ^*为r_ij经处理的值；m为规划方案个数；r_j为第j个指标的理想数值；In the formula: r _ij is the j index value in the i plan, r _ij ^* is the processed value of r _ij ; m is the number of planning plans; r _j is the ideal value of the j index;

4)赋予小波网络参数w_ij、r_j、b_j、a_j的随机初始值并赋予最大计算次数N；4) Give wavelet network parameters w _ij , r _j , b _j , a _j random initial values and give the maximum number of calculations N;

5)将各评价方案的一致化指标值r_k(i)输入到小波网络计算公式中，求得相应的综合评价值y_i并计算相应的能量误差E；5) Input the consistent index value r _k (i) of each evaluation scheme into the wavelet network calculation formula to obtain the corresponding comprehensive evaluation value y _i and calculate the corresponding energy error E;

小波网络计算公式为：The calculation formula of wavelet network is:

${y the y}_{k k} = = {Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} h h [[\frac{{Σ Σ}_{i i = = 11}^{m m} {w w}_{ij ij} {r r}_{k k} ((i i)) - - {b b}_{j j}}{{a a}_{j j}}]]$

网络的误差能量函数为：The error energy function of the network is:

$E E. = = \frac{11}{22} {Σ Σ}_{k k = = 11}^{P P} {(({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k}))}^{22}$

6）计算小波网络的梯度向量；6) Calculate the gradient vector of the wavelet network;

令make

${λ λ}_{k k} ((j j)) = = \frac{{Σ Σ}_{i i = = 11}^{m m} {w w}_{ij ij} {r r}_{k k} ((i i)) - - {b b}_{j j}}{{a a}_{j j}}$

有：have:

$g g (({w w}_{ij ij})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {w w}_{ij ij}} = = {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{{r r}_{k k} ((i i))}{{a a}_{j j}}))$

$g g (({r r}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {r r}_{j j}} = = {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) h h (({λ λ}_{k k} ((j j))))$

$g g (({a a}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {a a}_{j j}} = = - - {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{{λ λ}_{k k} ((i i))}{{a a}_{j j}}))$

$g g (({b b}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {b b}_{j j}} = = - - {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{11}{{a a}_{j j}}))$

其中，in,

$\frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} = = - - cos cos ((1.75 1.75 {λ λ}_{k k} ((j j)))) exp exp ((- - \frac{{λ λ}_{k k}^{22} ((j j))}{22})) {λ λ}_{k k} ((j j)) - - 1.75 1.75 sin sin ((1.75 1.75 {λ λ}_{k k} ((j j)))) exp exp ((- - \frac{{λ λ}_{k k}^{22} ((j j))}{22}))$

7)采用共轭梯度法调整网络参数，t为迭代次数：令7) Use the conjugate gradient method to adjust the network parameters, t is the number of iterations: let

${S S}_{t t} (({w w}_{ij ij})) = = \{\begin{matrix} - - {g g}_{t t} (({w w}_{ij ij})),, & t t = = 11 \\ - - {g g}_{t t} (({w w}_{ij ij})) + + \frac{| | | | {g g}_{t t} (({w w}_{ij ij})) | | | |}{| | | | {g g}_{t t - - 11} (({w w}_{ij ij})) | | | |} {S S}_{t t - - 11} (({w w}_{ij ij})),, & t t > > 11 \end{matrix}$

同理，可计算S_t(r_j)、S_t(a_j)、S_t(b_j)；Similarly, S _t (r _j ), S _t (a _j ), S _t (b _j ) can be calculated;

则网络参数的调节如下：Then the network parameters are adjusted as follows:

w_(t)ij＝w_(t-1)ij+αS_t-1(w_ij)w _(t)ij ＝w _(t-1)ij +αS _t-1 (w _ij )

r_(t)j＝r_(t-1)j+βS_t-1(r_j)r _(t)j ＝r _(t-1)j +βS _t-1 (r _j )

a_(t)j＝a_(t-1)j+γS_t-1(a_j)a _(t)j ＝a _(t-1)j +γS _t-1 (a _j )

b_(t)j＝b_(t-1)j+ηS_t-1(b_j)b _(t)j ＝b _(t-1)j +ηS _t-1 (b _j )

8)返回步骤4)，直到网络的误差能量函数值不大于给定值ε或计算次数超过最大计算次数N为止。8) Return to step 4) until the error energy function value of the network is not greater than the given value ε or the number of calculations exceeds the maximum number of calculations N.

9)根据步骤8）中确定的指标权重值，采用线性加权的方法计算各备选方案的综合评价值，选取分数最高方案为最优方案。9) According to the index weight value determined in step 8), use the linear weighting method to calculate the comprehensive evaluation value of each alternative plan, and select the plan with the highest score as the optimal plan.

本发明的优点在于：The advantages of the present invention are:

1)本发明遵循评价指标体系的设置原则及基本方法，结合我国实际能效工程的特点，从电网运行、环境效应、经济效益及资源利用效率四个方面出发，通过综合考虑，首先提出一套定量与定性分析相结合的居民及商业能效项目评价指标体系，全面、客观、准确地反应各能效方案的各个方面，充分兼顾了能效项目为电力用户、电力公司、能源服务公司（ESCO）以及全社会等不同参与主体带来的利益情况，可全面衡量预期产出（包括直接和间接效益）是否能够达到投资决策者的预期效果。1) The present invention follows the setting principles and basic methods of the evaluation index system, combines the characteristics of my country's actual energy efficiency projects, starts from four aspects of power grid operation, environmental effects, economic benefits and resource utilization efficiency, and through comprehensive consideration, first proposes a set of quantitative The residential and commercial energy efficiency project evaluation index system combined with qualitative analysis comprehensively, objectively and accurately reflects all aspects of each energy efficiency program, fully taking into account the energy efficiency projects for power users, power companies, energy service companies (ESCO) and the whole society. It can comprehensively measure whether the expected output (including direct and indirect benefits) can achieve the expected effect of investment decision makers.

2）本发明克服了传统评价方法的主观性，通过小波网络的方法，找到了与专家经验相适应的各底层指标权重值，使得评价结果既符合专家经验，又通过归一化将不同指标反应的内容量化到具体数值，对于不同的备选方案，项目投资者可以直接从最终的综合评分情况得出最优方案。2) The present invention overcomes the subjectivity of the traditional evaluation method. Through the method of wavelet network, the weight value of each underlying index that is suitable for the expert experience is found, so that the evaluation result not only conforms to the expert experience, but also reflects different indicators through normalization. The content of the project is quantified to a specific value. For different alternatives, project investors can directly obtain the optimal solution from the final comprehensive score.

3)本发明促使参与能效计划的电力用户通过采用科学管理方法和先进的技术设备节约电量和减少电力需求，从而期望在全寿命周期内获得最大的电费节约；能效项目实施能有效地改变原有负荷特性，实现“削峰”以降低高峰时段负荷水平，提高电网负荷率，优化电网运行，提高电力系统运行的稳定性、可靠性和经济性。终端电力用户能够以更少的能源投入满足其自身对电力效用的需求，能源使用量的降低将推动国家能源供应的高质量和安全性。3) The present invention encourages power users participating in the energy efficiency plan to save electricity and reduce power demand by adopting scientific management methods and advanced technical equipment, thereby expecting to obtain the greatest electricity cost savings in the entire life cycle; the implementation of energy efficiency projects can effectively change the original Load characteristics, realize "peak shaving" to reduce the load level during peak hours, increase the load rate of the power grid, optimize the operation of the power grid, and improve the stability, reliability and economy of the power system operation. End power users can meet their own demand for power utility with less energy input, and the reduction of energy usage will promote the high quality and security of national energy supply.

附图说明Description of drawings

图1是小波网络法步骤示意图；Fig. 1 is the step schematic diagram of wavelet network method;

图2是具体实施步骤；Fig. 2 is concrete implementation steps;

图3是综合评价指标体系。Figure 3 is the comprehensive evaluation index system.

具体实施方式Detailed ways

以下结合附图对本发明进行详细说明，图1是小波网络法步骤示意图,图1中将各属性值向量{x_k(i)}（共m个）进行指标一致化处理，得到指标属性一致的数据{r_k(i)},将各评价方案的一致化指标值r_k(i)输入到小波网络计算公式中，求得相应的综合评价值y_i并计算相应的能量误差E。The present invention will be described in detail below in conjunction with the accompanying drawings. Fig. 1 is a schematic diagram of the steps of the wavelet network method. In Fig. 1, each attribute value vector {x _k (i)} (a total of m) is subjected to index consistency processing, and the index attributes are consistent. For the data {r _k (i)}, the consistency index value r _k (i) of each evaluation scheme is input into the wavelet network calculation formula to obtain the corresponding comprehensive evaluation value y _i and calculate the corresponding energy error E.

如图2本发明基于小波网络法的商业能效项目评价方法，其中能效项目的综合评价方法应该至少包括：建立能效指标评价体系、计算综合评价指标体系各单项指标值、运用小波网络的方法确定各指标的最佳权重、计算各备选方案的综合评价值、输出最优方案等几部分内容：As shown in Fig. 2, the commercial energy efficiency project evaluation method based on the wavelet network method of the present invention, wherein the comprehensive evaluation method of the energy efficiency project should at least include: establishing an energy efficiency index evaluation system, calculating the individual index values of the comprehensive evaluation index system, and using the wavelet network method to determine each The optimal weight of the index, the calculation of the comprehensive evaluation value of each alternative, and the output of the optimal solution include several parts:

一、综合评价指标体系的建立。1. The establishment of a comprehensive evaluation index system.

结合我国实际能效工程的特点，本发明从电网运行、环境效应、经济效益及资源利用效率等四个方面出发，通过综合考虑，首先提出一套定量与定性分析相结合的居民及商业能效项目评价指标体系，如下表所示。如图3该体系包括4个一级指标，9个二级指标和20个三级指标。其中，然后在此基础上，再对每一类指标作简要分析及说明。Combined with the characteristics of my country's actual energy efficiency projects, this invention starts from four aspects of power grid operation, environmental effects, economic benefits and resource utilization efficiency, and through comprehensive consideration, firstly proposes a set of residential and commercial energy efficiency project evaluations that combine quantitative and qualitative analysis The index system is shown in the table below. As shown in Figure 3, the system includes 4 first-level indicators, 9 second-level indicators and 20 third-level indicators. Among them, and then on this basis, make a brief analysis and explanation of each type of indicators.

商业能效项目评价指标体系Evaluation Index System for Commercial Energy Efficiency Projects

1.电网运行1. Grid operation

本发明主要从可靠性及负荷整形能力2个方面对能效工程对电网运行的影响进行评价分析。The present invention mainly evaluates and analyzes the influence of the energy efficiency project on the operation of the power grid from two aspects of reliability and load shaping ability.

针对能效项目对电网运行可靠性水平的影响，本发明中主要分析系统平均缺电电量（ASCI）变化率及系统总电量不足量(ENS)变化率两项指标。前者表征在规定的时间内，系统中平均每个用户缺损电量的变化程度；后者则是指系统在规定时间内总的电量供给不足量变化比例。In view of the impact of energy efficiency projects on the reliability level of power grid operation, the present invention mainly analyzes the two indicators of the change rate of the system average power deficit (ASCI) and the total system power deficit (ENS). The former represents the change degree of the average power deficit of each user in the system within the specified time; the latter refers to the change ratio of the total power supply shortage of the system within the specified time.

（1）系统平均缺电电量变化率(1) Change rate of system average power shortage

系统平均缺电电量（ASCI）的计算公式如下：The calculation formula of the system average power shortage (ASCI) is as follows:

${P P}_{ASCI ASCI} = = \frac{{P P}_{ENS ENS}}{Σ Σ {N N}_{i i}} = = \frac{Σ Σ {t t}_{ui ui} {P P}_{i i}}{Σ Σ {N N}_{i i}} - - - - - - ((11 - - 11))$

其中，N_i为负荷点i的用户总数（用户），P_ENS为系统总电量不足量（WM·h/年）；Among them, N _i is the total number of users (users) at load point i, and P _ENS is the total power shortage of the system (WM·h/year);

基于式1-1，可以计算系统平均缺电电量变化率指标如下：Based on Equation 1-1, the system’s average power shortage indicator can be calculated as follows:

${I I}_{ASCI ASCI} = = \frac{{P P}_{ASCI ASCI}^{22} - - {P P}_{}^{ASCI ASCI}}{Δt Δt} - - - - - - ((11 - - 22))$

其中，Δt为一规定时间段（年），

分别为能效项目实施前后系统平均缺电电量值（MW·h/（用户·年））。Among them, Δt is a specified time period (year),

Respectively, the average power shortage value of the system before and after the implementation of the energy efficiency project (MW·h/(user·year)).

（2）系统总电量不足量变化率(2) The rate of change of the total power shortage of the system

系统总电量不足量(ENS)的计算公式如下：The calculation formula of the total system energy shortage (ENS) is as follows:

P_ENS＝Σt_uiP_i （1-3）P _ENS ＝Σt _ui P _i (1-3)

其中，P_i为负荷点i的平均负荷（kW），t_ui为负荷点i的年平均停电时间（h/年）。Among them, P _i is the average load of load point i (kW), and t _ui is the annual average power outage time of load point i (h/year).

基于式1-式4，可以计算系统总电量不足量变化率指标如下：Based on Equation 1-Equation 4, the change rate index of the total power shortage of the system can be calculated as follows:

${I I}_{ENS ENS} = = \frac{{P P}_{ENS ENS}^{22} - - {P P}_{}^{ENS ENS}}{Δt Δt} - - - - - - ((11 - - 44))$

其中，

分别为能效项目实施前后系统总电量不足量值（kW·h/年）。in,

Respectively, the insufficient value of the total system power before and after the implementation of the energy efficiency project (kW·h/year).

能效资源对负荷特性的影响效果主要体现在对负荷率及峰谷差的改变能力上。（1）日负荷率指日平均用电负荷与日最大负荷的比值，其计算如下：The impact of energy efficiency resources on load characteristics is mainly reflected in the ability to change the load rate and peak-to-valley difference. (1) The daily load rate refers to the ratio of the daily average power load to the daily maximum load, which is calculated as follows:

γ＝P_d.av/P_d.max （1-5）γ＝P _d.av /P _d.max (1-5)

其中，P_d.av为日平均负荷，P_d.max为日最大负荷；Among them, P _d.av is the daily average load, and P _d.max is the daily maximum load;

（2）日峰谷差率指日峰谷差与日最大负荷的比值，其计算如下：(2) The daily peak-to-valley difference ratio refers to the ratio of the daily peak-to-valley difference to the daily maximum load, which is calculated as follows:

$σ σ = = \frac{{P P}_{d d . . max max} - - {P P}_{d d . . min min}}{{P P}_{d d . . max max}} - - - - - - ((11 - - 66))$

其中，P_d.min为日最小负荷；Among them, P _d.min is the daily minimum load;

（3）年负荷率指全年平均负荷与年最大负荷的比值，其计算如下：(3) The annual load rate refers to the ratio of the annual average load to the annual maximum load, which is calculated as follows:

$δ δ = = \frac{{P P}_{y the y . . av av}}{{P P}_{y the y . . max max}} - - - - - - ((11 - - 77))$

其中，P_y.av为年平均负荷，P_y.max为年最大负荷；Among them, P _y.av is the annual average load, and P _y.max is the annual maximum load;

（4）年最大峰谷差率指全年日峰谷差率的平均值，其计算如下：(4) The annual maximum peak-to-valley difference rate refers to the average daily peak-to-valley difference rate throughout the year, which is calculated as follows:

$\overset{&OverBar; &OverBar;}{σ σ} = = \frac{Σ Σ {σ σ}_{i i}}{365365} - - - - - - ((11 - - 88))$

其中，σ_i为第i天的日峰谷差率。Among them, σ _i is the daily peak-to-valley difference rate on the i-th day.

2.环境效应2. Environmental effects

考虑到各类污染排放对社会影响程度以及我国节能减排目标等情况，本发明从污染物排放和碳排放两个方面对能效工程产生的环境效应实施评价分析。Considering the degree of impact of various types of pollution discharge on society and my country's energy conservation and emission reduction targets, the present invention evaluates and analyzes the environmental effects of energy efficiency projects from two aspects: pollutant discharge and carbon discharge.

其计算方法如下：Its calculation method is as follows:

（1）硫氧化合物减排量(1) Sulfur oxide emissions reduction

其中：in:

A_SO2——全寿命周期SO₂减排量，t；A _SO2 ——SO ₂ emission reduction in the whole life cycle, t;

λ_SO2——单位发电量SO₂排放强度,t/kWh;λ _SO2 - SO ₂ emission intensity per unit power generation, t/kWh;

W_zc——终端全寿命周期可避免电量，kWh;W _zc - the avoidable power in the whole life cycle of the terminal, kWh;

l——供电系统综合电量损失率，%。l——comprehensive power loss rate of the power supply system, %.

（2）氮氧化合物减排量(2) Nitrogen oxide emissions reduction

其中：in:

A_FC——全寿命周期粉尘减排量，t；A _FC - Dust emission reduction in the whole life cycle, t;

λ_FC——单位发电量粉尘排放强度，t/kWh;λ _FC —— dust emission intensity per unit power generation, t/kWh;

（3）粉尘减排量(3) Dust emission reduction

A_FC＝λ_FCW_zc/(1-l) （1-11） _AFC _＝ _λFCWzc /(1-l) (1-11)

其中：in:

（4）PM2.5减排量(4) PM2.5 emission reduction

A_pm2.5＝λ_pm2.5W_zc/(1-l) （1-12）A _pm2.5 = λ _pm2.5 W _zc /(1-l) (1-12)

其中：in:

Apm2.5——全寿命周期PM2.5颗粒减排量，t；Apm2.5—— PM2.5 particle emission reduction in the whole life cycle, t;

λpm2.5——单位发电量PM2.5颗粒排放强度,t/kWh;λpm2.5——PM2.5 particle emission intensity per unit power generation, t/kWh;

3.经济效益3. Economic benefits

本发明对能效项目的经济性评价主要包括财务评价、投资收益评价和全寿命周期成本分析三个方面。The economic evaluation of the energy efficiency project in the present invention mainly includes three aspects: financial evaluation, investment income evaluation and whole life cycle cost analysis.

（1）净现值（NPV）(1) Net Present Value (NPV)

在投资项目评价中，净现值指标NPV是最重要的指标之一。净现值是指按电力行业的基准收益率i_k，将能效项目计算期内各年的净现金流折现到投资期初的现值之和，其表达式为：In the evaluation of investment projects, the net present value indicator NPV is one of the most important indicators. The net present value refers to the _sum of discounting the net cash flow of each year in the calculation period of the energy efficiency project to the present value at the beginning of the investment period according to the benchmark rate of return ik of the power industry, and its expression is:

$NPV NPV = = {Σ Σ}_{t t = = 00}^{n no} {((CI CI - - CO CO))}_{t t} {((11 + + {i i}_{K K}))}^{- - t t} - - - - - - ((11 - - 1313))$

其中：in:

CI_t——现金流入量；CI _t - cash inflow;

CO_t——现金流出量；CO _t - cash outflow;

i_k——电力行业的基准收益率；i _k ——the benchmark rate of return of the electric power industry;

n——项目的寿命期。n - the lifetime of the project.

（2）内部收益率（IRR）(2) Internal rate of return (IRR)

内部收益率IRR是指使净现金流量的净现值等于零的折现率其表达式为：The internal rate of return (IRR) is the discount rate that makes the net present value of the net cash flow equal to zero. Its expression is:

${Σ Σ}_{t t = = 00}^{n no} {((CI CI - - CO CO))}_{t t} {((11 + + IRR IRR))}^{- - t t} = = 00 - - - - - - ((11 - - 1414))$

其中：in:

CI——现金流入；CI - cash inflow;

CO——现金流出；CO - cash outflow;

(CI-CO)t——第t年的净现金流量。(CI-CO)t——The net cash flow in year t.

（3）静态投资回收期（Pt）(3) Static payback period (Pt)

静态投资回收期Pt是在不考虑资金时间价值的条件下，以项目净收益抵偿项目全部投资所需要的时间，其表达式为：The static investment payback period Pt is the time required to offset the entire investment of the project with the net income of the project without considering the time value of funds, and its expression is:

${Σ Σ}_{t t = = 00}^{{P P}_{t t}} {((CI CI - - CO CO))}_{t t} = = 00 - - - - - - ((11 - - 1515))$

投资收益指的是企业进行投资所获得的经济利益。在本文提出的评价指标体系中，选取收益费用比率(PCR)及可避免用电成本(AC)等2项指标以衡量能效项目的投资收益情况，其计算公式分别如下：Investment income refers to the economic benefits obtained by enterprises from investment. In the evaluation index system proposed in this paper, two indicators, such as revenue cost ratio (PCR) and avoidable electricity cost (AC), are selected to measure the investment income of energy efficiency projects. The calculation formulas are as follows:

$PCR PCR = = \frac{{d d}_{z z,, b b}}{{d d}_{z z,, s the s}} * * 100100 % % - - - - - - ((11 - - 1616))$

$AC AC = = {Σ Σ}_{i i = = 11}^{n no} {W W}_{z z,, i i} {d d}_{i i} - - - - - - ((11 - - 1717))$

其中，d_z,b与d_z,s分别为终端用户单位用电可避免成本（元/(kWh)）；W_z,i及d_i分别为终端用户因能效改造在时段i的可避免电量（kWh）及其对应的售电电价（元/(kWh)）。n则为全寿命周期内的总时段数目。总体来讲，居民及商业用户能效项目全寿命周期费用（LCC）模型可表示为：Among them, d _{z, b} and d _{z, s} are the avoidable cost of end-user unit electricity consumption (yuan/(kWh)); W _{z, i} and d _i are the avoidable power consumption of end-users in time period i due to energy efficiency transformation (kWh) and its corresponding electricity sales price (yuan/(kWh)). n is the total number of time periods in the whole life cycle. Generally speaking, the life cycle cost (LCC) model of energy efficiency projects for residential and commercial users can be expressed as:

LCC＝C_I+C_O+C_M+C_F+C_D （1-18）LCC＝C _I +C _O +C _M +C _F +C _D (1-18)

其中：in:

LCC——设备全寿命周期成本；LCC——equipment life cycle cost;

CI——一次性投入成本，包括设计阶段的设计成本，建设阶段的设备采购成本、施工安装成本；CI—one-time input cost, including the design cost in the design phase, the equipment procurement cost in the construction phase, and the construction and installation cost;

CO——运行成本，包括设备损耗、运行人员培训成本以及能耗成本；CO——operating cost, including equipment loss, operating personnel training cost and energy consumption cost;

CM——维护成本；CM - maintenance cost;

CF——故障成本，包括停电损失、社会影响损失等；CF—failure cost, including power outage loss, social impact loss, etc.;

CD——废弃成本。CD - Obsolescence cost.

基于上式，进一步给出寿期成本下降率（LCCR）的计算公式如下：Based on the above formula, the formula for calculating the life cycle cost reduction rate (LCCR) is further given as follows:

$LCCR LCCR = = \frac{{LCC LCC}_{00} - - LCC LCC}{{LCC LCC}_{00}} * * 100100 % % - - - - - - ((11 - - 1919))$

其中，LCC0为普通用电设备全寿命周期成本费用；LCC为能效改造后相应设备的全寿命周期成本费用。Among them, LCC0 is the full life cycle cost of ordinary electrical equipment; LCC is the full life cycle cost of the corresponding equipment after energy efficiency transformation.

资源利用效率：Resource Utilization Efficiency:

考虑到能效项目对于提高终端用户及电网侧资源使用效率的重要意义，本文从节能效果及电网侧资源使用效率等2个方面对此评价研究。Considering the importance of energy efficiency projects for improving end-user and grid-side resource utilization efficiency, this paper evaluates and studies them from two aspects: energy-saving effect and grid-side resource utilization efficiency.

终端年节电量，是指各类能效设备投运前后各类终端用户总用电电量的差值，其计算公式如下：The terminal annual energy saving refers to the difference between the total electricity consumption of various end users before and after the operation of various energy efficiency equipment, and its calculation formula is as follows:

其中，为不同用户间用电规律差异系数，用于调整不同用户间的用电同时性及一致性；F为能效项目涉及的总用户数量；Q_t,i,0为终端用电环节i在时段t中的原用电电量；λ_i表示用电环节i的示性函数，当在该环节进行能效改造时，该变量取值为1，否则为0；α_EE,i为能效设备投运后用电环节i的降耗率；ζ_t为终端用户中各用电环节在时段t的同时率。此外，T为研究周期内总的时段数目；K则为终端用户总的用电环节数目。in, is the difference coefficient of electricity consumption among different users, which is used to adjust the simultaneity and consistency of electricity consumption among different users; _F is the total number of users involved in the energy efficiency project; λ _i represents the indicative function of power consumption link i, when the energy efficiency transformation is carried out _in this link, the value of this variable is 1, otherwise it is 0; The consumption reduction rate of the electricity link i; ζ _t is the simultaneous rate of each electricity link in the end user in the time period t. In addition, T is the total number of time periods in the research period; K is the total number of electricity consumption links of end users.

用电收益效率差值的计算公式如下：The formula for calculating the difference in electricity revenue efficiency is as follows:

$CE CE = = ((\frac{{Q Q}_{00} - - {Q Q}_{e e}}{Ep Ep})) * * 100100 % % - - - - - - ((11 - - 21 twenty one))$

其中Q₀及Q_e分别为能效设备投运前、后研究周期内终端用户的总用电量。Ep为经济产出总量。对于商业用户，Ep可取为研究周期内的营业额；对于居民用户，Ep可取为家庭成员数目所对应的GDP值。Among them, Q ₀ and Q _e are the total power consumption of end users in the research period before and after the energy efficiency equipment is put into operation, respectively. Ep is the total economic output. For commercial users, Ep can be taken as the turnover during the research period; for residential users, Ep can be taken as the GDP value corresponding to the number of family members.

针对能效投资对电网扩容的延迟作用，提出可免供电容量指标进行量化计算，其表达式如下Aiming at the delayed effect of energy efficiency investment on power grid expansion, a quantitative calculation of the exempt power supply capacity index is proposed, and its expression is as follows

其中，I_grid为电网企业因终端节电而减少的累计电网扩容容量（包括对应的备用容量）；

为电网备用容量的比例。Among them, I _grid is the cumulative grid expansion capacity (including the corresponding reserve capacity) reduced by the power grid enterprise due to terminal power saving;

is the ratio of grid reserve capacity.

在以上提出的综合评价指标体系中，充分兼顾了能效项目为电力用户、电力公司、能源服务公司（ESCO）以及全社会等不同参与主体带来的利益情况。在实际工程应用中，需要根据项目的实施目的及不同主体的受益条件，从中灵活选择所适合的指标，进一步实施计算分析。In the comprehensive evaluation index system proposed above, the benefits brought by energy efficiency projects to different participants such as power users, power companies, energy service companies (ESCOs) and the whole society are fully taken into account. In practical engineering applications, it is necessary to flexibly select the appropriate indicators according to the implementation purpose of the project and the benefit conditions of different subjects, and further implement calculation and analysis.

二、计算综合评价指标体系各三级指标值；2. Calculate the index values of the three levels of the comprehensive evaluation index system;

收集商业能效项目的基础资料，确定三级指标计算公式中相关计算参数，计算综合评价指标体系中的底层指标值。Collect the basic data of commercial energy efficiency projects, determine the relevant calculation parameters in the three-level index calculation formula, and calculate the bottom index value in the comprehensive evaluation index system.

三、运用小波网络的方法确定各指标的最佳权重；3. Use the method of wavelet network to determine the optimal weight of each index;

1)根据专家意见，给可选能效方案初步打分，得到各方案的专家评议分数集合 1) According to expert opinions, give preliminary scores to the optional energy efficiency schemes, and obtain the set of expert evaluation scores for each scheme

2)将步骤（1）中计算出的各属性值向量{x_k(i)}转化为指标属性一致的数据 ${r_{k} (i)} .$ 2) Transform the attribute value vector {x _k (i)} calculated in step (1) into data with consistent index attributes ${r_{k} (i)} .$

3)赋予小波网络参数w_ij、r_j、b_j、a_j的随机初始值并赋予最大计算次数N。3) Random initial values of wavelet network parameters w _ij , r _j , b _j , a _j are assigned and the maximum number of calculations N is assigned.

4)将各评价方案的一致化指标值r_k(i)输入到小波网络计算公式中，求得相应的综合评价值y_i并计算相应的能量误差E。4) Input the consistent index value r _k (i) of each evaluation scheme into the wavelet network calculation formula, obtain the corresponding comprehensive evaluation value y _i and calculate the corresponding energy error E.

小波网络计算公式为：The calculation formula of wavelet network is:

网络的误差能量函数为The error energy function of the network is

5)计算小波网络的梯度向量。5) Calculate the gradient vector of the wavelet network.

令make

有：have:

$g g (({b b}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {b b}_{j j}} = = - - {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{{λ λ}_{k k} ((i i))}{{a a}_{j j}}))$

其中，in,

6)采用共轭梯度法（Fletcher-Reeves公式）调整网络参数（t为迭代次数）：6) Use the conjugate gradient method (Fletcher-Reeves formula) to adjust the network parameters (t is the number of iterations):

令make

w_(t)ij＝w_(t-1)ij+αS_t-1(w_ij)w _(t)ij ＝w _(t-1)ij +αS _t-1 (w _ij )

r_(t)j＝r_(t-1)j+βS_t-1(r_j)r _(t)j ＝r _(t-1)j +βS _t-1 (r _j )

a_(t)j＝a_(t-1)j+γS_t-1(a_j)a _(t)j ＝a _(t-1)j +γS _t-1 (a _j )

b_(t)j＝b_(t-1)j+ηS_t-1(b_j)b _(t)j ＝b _(t-1)j +ηS _t-1 (b _j )

7)返回步骤(4)，直到网络的误差能量函数值不大于给定值ε或计算次数超过最大计算次数N为止。7) Return to step (4) until the error energy function value of the network is not greater than the given value ε or the number of calculations exceeds the maximum number of calculations N.

四、计算各备选方案的综合评价值，输出最优方案4. Calculate the comprehensive evaluation value of each alternative plan and output the optimal plan

依据之前确定的指标权重值，采用线性加权的方法计算各方案的综合评价值，选取分数最高方案为最优方案。According to the index weight value determined before, the comprehensive evaluation value of each scheme is calculated by linear weighting method, and the scheme with the highest score is selected as the optimal scheme.

如图3所示为多指标综合评价的小波网络结构图，图中输入层、隐层和输出层分别有m、n和1个单元节点。这里取Figure 3 shows the wavelet network structure diagram of multi-index comprehensive evaluation. In the figure, the input layer, hidden layer and output layer have m, n and 1 unit nodes respectively. Take here

${y the y}_{k k} = = {Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} h h [[\frac{{Σ Σ}_{i i = = 11}^{m m} {w w}_{ij ij} {r r}_{k k} ((i i)) {- - b b}_{j j}}{{a a}_{j j}}]]$

式中，x_k(i)、r_k(i)分别表示输入样本k的指标i的原始数据和无量纲化数据。wi_j、r_j表示权重系数，b_j、a_j分别表示小波基的平移因子和伸缩因子。此处的基本小波采用国外常使用的余弦调制的高斯波：Morlet母小波。其形式为In the formula, x _k (i) and r _k (i) represent the original data and dimensionless data of the index i of the input sample k, respectively. wi _j , r _j represent weight coefficients, b _j , a _j represent translation factor and expansion factor of wavelet base respectively. The basic wavelet here adopts the cosine-modulated Gaussian wave commonly used abroad: Morlet mother wavelet. Its form is

$h h ((t t)) = = cos cos ((1.75 1.75 t t)) exp exp ((- - \frac{{t t}^{22}}{22}))$

将复杂对象系统的评价样本k所对应的指标属性值向量{r_k(i)}作为小波网络的输入，与之对应的综合评价值作为网络的期望输出。定义网络的误差能量函数为：The index attribute value vector {r _k (i)} corresponding to the evaluation sample k of the complex object system is used as the input of the wavelet network, and the corresponding comprehensive evaluation value as the desired output of the network. Define the error energy function of the network as:

式中，y_k为评价样本k评价实际值，

为网络输出值。P为评价样本总数。通过网络参数wi_j、r_j、b_j、a_j的调整，使得网络的误差能量函数达到最低。In the formula, y _k is the actual evaluation value of evaluation sample k,

Output values for the network. P is the total number of evaluation samples. Through the adjustment of network parameters wi _j , r _j , b _j , a _j , the error energy function of the network can be minimized.

本发明将提出的基于小波网络法的商业能效项目的评价方法应用在实际能效项目评估中示例如下：The present invention applies the evaluation method of the commercial energy efficiency project based on the wavelet network method to the actual energy efficiency project evaluation as follows:

1.本发明以某商业区的能效改造为例，在规划区提出了8套能效改造方案，用P1～P8表示。1. The present invention takes the energy efficiency transformation of a commercial area as an example, and proposes 8 sets of energy efficiency transformation schemes in the planning area, represented by P1-P8.

2.计算综合评价指标体系各单项指标值2. Calculate the individual index values of the comprehensive evaluation index system

本次评价过程选取环境效应、经济效益、资源利用率三方面的内容作为比较对象，共涉及三级评价指标8个，用R1～R8表示。根据收集的相关数据和专家打分，并根据前述方法分别对能效改造方案的各单项指标进行无量纲处理，得到以下数据，具体如表1所示：The evaluation process selects environmental effects, economic benefits, and resource utilization as comparison objects, involving a total of 8 three-level evaluation indicators, represented by R1~R8. According to the collected relevant data and experts’ scoring, and according to the aforementioned method, the individual indicators of the energy efficiency transformation plan are processed dimensionlessly, and the following data are obtained, as shown in Table 1:

表1各能效方案评价数据Table 1 Evaluation data of each energy efficiency scheme

指标index R1R1 R2R2 R3R3 R4R4 R5R5 R6R6 R7R7 R8R8 得分Score P1P1 0.780.78 0.800.80 0.920.92 1.001.00 0.400.40 0.580.58 0.670.67 0.770.77 0.830.83 P2P2 0.880.88 0.670.67 1.001.00 0.560.56 0.690.69 0.290.29 0.500.50 0.910.91 0.780.78 P3P3 0.920.92 0.890.89 0.940.94 0.880.88 1.001.00 0.860.86 0.980.98 0.780.78 0.930.93 P4P4 1.001.00 0.890.89 0.970.97 0.660.66 0.870.87 0.960.96 0.790.79 0.950.95 0.920.92 P5P5 0.890.89 1.001.00 0.830.83 0.780.78 0.960.96 0.970.97 0.840.84 0.780.78 0.890.89 P6P6 0.830.83 0.780.78 0.790.79 0.680.68 0.930.93 0.890.89 0.940.94 0.880.88 0.830.83 P7P7 0.930.93 0.980.98 0.860.86 0.890.89 0.930.93 0.790.79 0.830.83 1.001.00 0.950.95 P8P8 0.870.87 0.950.95 0.890.89 0.930.93 0.920.92 1.001.00 0.970.97 0.960.96 0.940.94

3.运用基于小波网络法的商业能效综合评价方法确定各指标的最佳权重3. Use the comprehensive evaluation method of commercial energy efficiency based on the wavelet network method to determine the optimal weight of each index

以表1中前5组作为已经选择的能效改造项目的数据，作为训练集训练该小波网络，后3组为待评价对象。训练结果如表2所示，他们与期望的输出非常接近。The first 5 groups in Table 1 are used as the data of the selected energy efficiency transformation project as the training set to train the wavelet network, and the last 3 groups are the objects to be evaluated. The training results are shown in Table 2, and they are very close to the expected output.

表2训练结果Table 2 Training Results

方案编号Scheme number P1P1 P2P2 P3P3 P4P4 P5P5 期望输出expected output 0.830.83 0.780.78 0.930.93 0.920.92 0.890.89 训练结果training result 0.82770.8277 0.79390.7939 0.93120.9312 0.92420.9242 0.89210.8921 相对误差/%Relative error/% 0.2770.277 0.7820.782 0.1290.129 0.4570.457 0.2360.236

4.计算各备选方案的综合评价值，输出最优方案。4. Calculate the comprehensive evaluation value of each alternative plan, and output the optimal plan.

对未经训练的3组测试集仿真评价的结果如表3所示，其中能效改造项目按优到劣的顺序依次为p7,p8,p6。Table 3 shows the simulation evaluation results of the three untrained test sets, in which the energy efficiency improvement projects are p7, p8, and p6 in order from good to bad.

表3测试结果与专家打分比较Table 3 Comparison of test results and expert scores

方案编号Scheme number P6P6 P7P7 P8P8 期望输出expected output 0.830.83 0.950.95 0.940.94 训练结果training result 0.82770.8277 0.95390.9539 0.93120.9312 相对误差/%Relative error/% 0.2770.277 0.4110.411 0.9360.936

据此，得到最优能效方案为p7。Accordingly, the optimal energy efficiency scheme is obtained as p7.

整个过程用Matlab7.2软件编制小波网络模型，网络训练1863次，训练时间耗时2.37s，结果较为理想。The whole process uses Matlab7.2 software to compile the wavelet network model, the network is trained 1863 times, the training time takes 2.37s, and the result is ideal.

5.与前馈神经网络法的比较5. Comparison with Feedforward Neural Network Method

为了说明使用小波网络进行商业能效综合评价的优越性质，将小波网络与BP神经网络在同一迭代次数下的误差进行了比较，小波网络的平均误差为0.00478，最小误差为0.00277，最大误差为0.00936；传统BP网络平均误差为0.028176，最小误差为0.026317，最大误差为0.029628.显然，小波网络的精度明显优于BP神经网络，也更适合商业能效项目综合评价这类评价指标体系繁杂的领域In order to illustrate the superiority of using wavelet network for comprehensive evaluation of commercial energy efficiency, the errors of wavelet network and BP neural network under the same iteration number are compared. The average error of wavelet network is 0.00478, the minimum error is 0.00277, and the maximum error is 0.00936; The average error of the traditional BP network is 0.028176, the minimum error is 0.026317, and the maximum error is 0.029628. Obviously, the accuracy of the wavelet network is significantly better than that of the BP neural network, and it is also more suitable for the comprehensive evaluation of commercial energy efficiency projects.

用同样的样本数据对两种网络进行收敛性研究，取相对误差指标为0.5‰，小波网络经732次迭代收敛，BP神经网络经903次迭代收敛。证明小波网络的收敛性也优于BP神经网络。Using the same sample data to study the convergence of the two networks, the relative error index is taken as 0.5‰, the wavelet network converges after 732 iterations, and the BP neural network converges after 903 iterations. It is proved that the convergence of wavelet network is better than that of BP neural network.

应当理解，以上借助优选实施例对本发明的技术方案进行的详细说明是示意性的而非限制性的。本领域的普通技术人员在阅读本发明说明书的基础上可以对各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。It should be understood that the above detailed description of the technical solution of the present invention with the aid of preferred embodiments is illustrative rather than restrictive. Those skilled in the art can modify the technical solutions recorded in each embodiment on the basis of reading the description of the present invention, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the corresponding technical solutions Essentially deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A commercial energy efficiency comprehensive evaluation method based on wavelet network method, is characterized in that comprising the following steps:

(1) Establish a comprehensive evaluation index system for energy efficiency projects;

(2) Calculate the individual index values of the comprehensive evaluation index system;

(3) Use the method of wavelet network to determine the optimal weight of each index;

(4) Calculate the comprehensive evaluation value of each alternative plan, and output the optimal plan.

2. A comprehensive evaluation method for commercial energy efficiency based on wavelet network method according to claim 1, characterized in that: said step (1) establishes a comprehensive evaluation index system for energy efficiency projects, including first-level indicators and their corresponding second-level indicators , each second-level index is expanded into a third-level index that reflects the specific content of the index at this level, that is, a three-level hierarchical structure dominated from top to bottom is formed.

3. A comprehensive evaluation method for commercial energy efficiency based on wavelet network method according to claim 2, characterized in that: said primary index content includes four aspects: power grid operation, environmental effects, economic benefits and resource utilization efficiency.

4. A kind of comprehensive evaluation method of commercial energy efficiency based on wavelet network method according to claim 2, it is characterized in that: described secondary index is made up of power grid reliability evaluation and load shaping ability evaluation two aspects content; The environmental effect part of the index includes two aspects: pollutant discharge and carbon emission; the economic benefit part of the first-level index is mainly composed of three aspects: financial evaluation, investment return evaluation and life cycle cost analysis; The resource utilization efficiency part of the first-level indicators includes two aspects: energy-saving effect and resource utilization on the grid side.

5. A kind of commercial energy efficiency project evaluation method based on wavelet network method according to claim 2, it is characterized in that: the reliability evaluation of the secondary index of the comprehensive evaluation index system of the establishment includes the ASCI change of the average power shortage of the system The load shaping capability in the secondary indicators includes four indicators: daily load rate, daily peak-to-valley difference rate, annual load rate and annual maximum peak-to-valley difference rate change ; The pollutant discharge part in the second-level indicators includes four indicators of nitrogen oxide emission reduction, sulfur oxide emission reduction, dust emission reduction and PM2.5 emission reduction; the carbon in the second-level indicators The first level of emissions corresponds to the _CO2 emission reduction indicator; the financial evaluation part of the second level indicator is systematically analyzed and calculated by the three indicators of net present value, internal rate of return and investment payback period; the investment in the second level indicator For the income part, two indicators, the revenue-to-expense ratio PCR and the avoidable electricity cost AC, are selected to measure the investment income of energy efficiency projects; the life-cycle cost part of the second-level indicators uses the life-cycle cost reduction rate as a measure of the energy efficiency project’s entire life cycle. The evaluation index of economy; the energy-saving effect part of the above-mentioned secondary index is expanded into two indexes of terminal annual power saving and power consumption efficiency difference.

6. A kind of commercial energy efficiency project evaluation method based on wavelet network method according to claim 2, it is characterized in that: in the evaluation index system, the three-level index weight value adopts the method of wavelet network to solve energy efficiency, and then adopts linear weighted synthesis The method is used to comprehensively score the alternative schemes, and the one with the highest comprehensive score is selected as the optimal scheme. The specific implementation steps are as follows:

1) Collect the basic data of commercial energy efficiency projects, and calculate the three-level index values in the comprehensive evaluation index system;

3) Transform the attribute value vector {x _k (i)} of each value calculated in step 1) into data {r _k (i)} with consistent index attributes;

The three types of indicators are handled as follows:

In the formula: r _ij is the j index value in the i plan, r _ij ^* is the processed value of r _ij ; m is the number of planning plans; r _j is the ideal value of the j index;

4) Give wavelet network parameters wi _j , r _j , b _j , a _j random initial values and give the maximum number of calculations N;

5) Input the consistent index value {r _k (i)} of each evaluation scheme into the wavelet network calculation formula, obtain the corresponding comprehensive evaluation value y _i and calculate the corresponding energy error E;

The calculation formula of wavelet network is:

{y the y}_{k k} = = {Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} h h [[\frac{{Σ Σ}_{i i = = 11}^{m m} {w w}_{ij ij} {r r}_{k k} ((i i)) - - {b b}_{j j}}{{a a}_{j j}}]]

The error energy function of the network is:

E E. = = \frac{11}{22} {Σ Σ}_{k k = = 11}^{P P} {(({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k}))}^{22}

6) Calculate the gradient vector of the wavelet network;

make

{λ λ}_{k k} ((j j)) \frac{{Σ Σ}_{i i = = 11}^{m m} {w w}_{ij ij} {r r}_{k k} ((i i)) - - {b b}_{j j}}{{a a}_{j j}}

have:

g g (({w w}_{ij ij})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {w w}_{ij ij}} = = {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{{r r}_{k k} ((i i))}{{a a}_{j j}}))

g g (({r r}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {r r}_{j j}} = = {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) h h (({λ λ}_{k k} ((j j))))

g g (({a a}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {a a}_{j j}} = = - - {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{{λ λ}_{k k} ((i i))}{{a a}_{j j}}))

g g (({b b}_{j j})) = = \frac{&PartialD; &PartialD; E E.}{&PartialD; &PartialD; {b b}_{j j}} = = - - {Σ Σ}_{k k = = 11}^{P P} (({\overset{^^}{y the y}}_{k k} - - {y the y}_{k k})) (({Σ Σ}_{j j = = 11}^{n no} {r r}_{j j} \frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} \frac{11}{{a a}_{j j}}))

in,

\frac{&PartialD; &PartialD; h h}{&PartialD; &PartialD; {λ λ}_{k k} ((j j))} = = - - cos cos ((1.75 1.75 {λ λ}_{k k} ((j j)))) exp exp ((- - \frac{{λ λ}_{k k}^{22} ((j j))}{22})) {λ λ}_{k k} ((j j)) - - 1.75 1.75 sin sin ((1.75 1.75 {λ λ}_{k k} ((j j)))) exp exp ((- - \frac{{λ λ}_{k k}^{22} ((j j))}{22}))

7) Use the conjugate gradient method to adjust the network parameters, t is the number of iterations: let

{S S}_{t t} (({w w}_{ij ij})) = = \{\begin{matrix} - - {g g}_{t t} (({w w}_{ij ij})),, & t t = = 11 \\ - - {g g}_{t t} (({w w}_{ij ij})) + + \frac{| | | | {g g}_{t t} (({w w}_{ij ij})) | | | |}{| | | | {g g}_{t t - - 11} (({w w}_{ij ij})) | | | |} {S S}_{t t - - 11} (({w w}_{ij ij})),, & t t > > 11 \end{matrix}

Similarly, S _t (r _j ), S _t (a _j ), S _t (b _j ) can be calculated;

Then the network parameters are adjusted as follows:

w _(t)ij ＝w _(t-1)ij +αS _t-1 (w _ij )

r _(t)j ＝r _(t-1)j +βS _t-1 (r _j )

a _(t)j ＝a _(t-1)j +γS _t-1 (a _j )

b _(t)j ＝b _(t-1)j +ηS _t-1 (b _j )

8) Return to step 4) until the error energy function value of the network is not greater than the given value ε or the number of calculations exceeds the maximum number of calculations N.

9) According to the index weight value determined in step 8), the comprehensive evaluation value of each plan is calculated using the linear weighting method, and the plan with the highest score is selected as the optimal plan.