CN103745111A - Method of predicting driving range of all-electric passenger vehicles - Google Patents

Abstract

The invention discloses a method of predicting driving range of all-electric passenger vehicles. The method includes the steps of S1, modeling by means of fuzzy clustering; S2, modeling for energy consumption of an all-electric passenger vehicle by a fuzzy time sequence analysis algorithm; S3, building a fuzzy time sequence model; S4, dividing energy consumption into super-high, high, normal, low and super-low states for description; S5, subjecting an influencing factor predicted value, deducted through a fuzzy time sequence, to weight similarity matching so as to obtain an energy consumption match value, and operating the energy consumption match value and a section energy consumption weight to finally obtain a weighted energy consumption predicted value; S6, subtracting the weighted energy consumption predicted value by an energy consumption residue value to obtain an energy consumption residue predicted value, and converting the energy consumption residue predicted value into a driving range; S7, ending the algorithm until the section energy consumption residue predicted value is smaller than or equal to a lowest energy consumption required value. By accurately predicting the driving range, the problem that predicting the driving range for all-electric passenger vehicles at preset can be solved.

Description

Translated fromChinese

纯电动乘用车续驶里程预测方法Prediction method for driving range of pure electric passenger vehicles

技术领域technical field

本发明涉及电动乘用车的技术领域，特别涉及一种纯电动乘用车续驶里程预测方法。The invention relates to the technical field of electric passenger vehicles, in particular to a method for predicting the driving range of pure electric passenger vehicles.

背景技术Background technique

纯电动乘用车虽然目前在国家节能减排政策下得到较快发展，但由于受其续驶里程的制约，目前仍难以大规模推广，仍处于试点运营状况，尤其是目前充电站规划和实施仍然难以符合实际推广要求的情况下。续驶里程主要受电池容量和运营状况限制，在目前电池技术储能和可靠性仍难以取得突破的情况下，研究者们提出了各种分析与评估续驶里程的方法。在目前阶段，对续驶里程的估算方法大多依赖于对SOC的更准确估算，从而转换成相对准确的续驶里程，因而目前的研究热点集中在如何准确预测SOC上。比较具有代表性的算法，如针对电动汽车用锂离子电池组的能修正初始误差的荷电状态估算方法；针对磷酸铁锂电池给出了其改进的PNGV模型，采用扩展卡尔曼滤波算法完成了SOC的准确估计；对于磷酸铁锂电池，安时积分SOC估算方法中初始SOC的影响最大，应该建立初始SOC的修正算法,对电池的SOC累积误差进行清除以提高精度；采用电化学阻抗谱来分析等效电路模型参数，以研究电池的电压特性和动态功率特性，通过综合分析实际充放电条件的主要特征来提取电池典型的参数辨识工况，并利用粒子群优化算法分析模型参数；此外，修订的伏安法、Kalman算法、神经网络等算法的应用也比较广泛。在进一步提高SOC估算精度情况下，基于锂离子电池电化学模型提出电动公交车续驶里程预测方法、纯电动汽车续驶里程RBF神经网络预测算法、通过分析纯电动汽车行驶中主电路负载电流变化建立其续驶里程计算模型、利用BP神经网络预测电动汽车续驶里程的方法等。就研究者在公开发表信息内容上看，续驶里程的预测的依据主要采用SOC参数作为主要参考，大多采用等速法以及工况法进行预测，所得到延长续驶里程的结论大多倾向于选用低阻力轮胎、进行车身的流线型改进、减轻空车重量、选用高能量电池、动力传动系合理匹配等从电动汽车本身架构以及零配件选用上。然而，影响纯电动乘用车续驶里程的因素，即能耗影响因素还包括时段（早晚高峰）、路况、驾驶员驾驶习惯等因素，为此，本发明利用在已经正常营运的纯电动乘用车行车过程中所产生的数据，结合交通运行指标体系，基于模糊聚类算法和模糊时间序列算法建立纯电动乘用车续驶里程预测模型，为纯电动乘用车充电站规划等提供决策支持。Although pure electric passenger vehicles have developed rapidly under the national energy-saving and emission-reduction policies, due to the constraints of their driving range, it is still difficult to promote them on a large scale, and they are still in pilot operation, especially the current planning and implementation of charging stations. It is still difficult to meet the actual promotion requirements. The driving range is mainly limited by the battery capacity and operating conditions. In the current situation where battery technology energy storage and reliability are still difficult to achieve breakthroughs, researchers have proposed various methods for analyzing and evaluating the driving range. At the current stage, most of the estimation methods for driving range rely on a more accurate estimation of SOC, which can be converted into a relatively accurate driving range. Therefore, the current research hotspots are focused on how to accurately predict SOC. More representative algorithms, such as the state of charge estimation method that can correct the initial error for lithium-ion battery packs for electric vehicles; for lithium iron phosphate batteries, an improved PNGV model is given, and the extended Kalman filter algorithm is used to complete Accurate estimation of SOC; for lithium iron phosphate batteries, the initial SOC has the greatest influence in the ampere-hour integral SOC estimation method, and a correction algorithm for the initial SOC should be established to clear the cumulative error of the battery's SOC to improve accuracy; use electrochemical impedance spectroscopy to Analyze the parameters of the equivalent circuit model to study the voltage characteristics and dynamic power characteristics of the battery, extract the typical parameters of the battery to identify the working conditions by comprehensively analyzing the main characteristics of the actual charging and discharging conditions, and use the particle swarm optimization algorithm to analyze the model parameters; in addition, The revised voltammetry, Kalman algorithm, neural network and other algorithms are also widely used. In the case of further improving the SOC estimation accuracy, based on the electrochemical model of lithium-ion batteries, a prediction method for the driving range of electric buses, a RBF neural network prediction algorithm for the driving range of pure electric vehicles, and an analysis of the main circuit load current change during the driving of pure electric vehicles are proposed. Establish its mileage calculation model, use BP neural network to predict the mileage of electric vehicles, etc. As far as researchers have published information, the basis for the prediction of driving range is mainly based on SOC parameters as the main reference, and most of them use constant velocity method and working condition method for prediction. Low-resistance tires, streamlined improvement of the body, reduction of empty weight, selection of high-energy batteries, reasonable matching of the power train, etc., are from the structure of the electric vehicle itself and the selection of spare parts. However, the factors that affect the mileage of pure electric passenger vehicles, that is, the factors that affect energy consumption also include factors such as time period (morning and evening peak hours), road conditions, and driving habits of drivers. The data generated during the driving process, combined with the traffic operation index system, based on the fuzzy clustering algorithm and the fuzzy time series algorithm, establishes a prediction model for the mileage of pure electric passenger vehicles, and provides decision-making for the planning of charging stations for pure electric passenger vehicles. support.

发明内容Contents of the invention

本发明的目的在于克服现有技术的缺点与不足，提供一种一种纯电动乘用车续驶里程预测方法。The object of the present invention is to overcome the shortcomings and deficiencies of the prior art, and provide a method for predicting the driving range of pure electric passenger vehicles.

本发明的目的通过下述技术方案实现：The object of the present invention is achieved through the following technical solutions:

纯电动乘用车续驶里程预测方法，包括下述步骤：A method for predicting the mileage of a pure electric passenger car, comprising the following steps:

S1、以路段能耗为基础，纯电动乘用车营运过程中所有影响因素作为聚类的维度，采用模糊聚类的方式进行建模；S1. Based on the energy consumption of the road section, all the influencing factors in the operation process of pure electric passenger vehicles are used as the clustering dimension, and the fuzzy clustering method is used for modeling;

S2、通过模糊聚类中心对路段之间的能耗及影响因素的能耗情况进行描述以及标识后，通过模糊时间序列分析算法对纯电动乘用车能耗状况进行建模；S2. After describing and marking the energy consumption between road sections and the energy consumption of influencing factors through the fuzzy clustering center, model the energy consumption of pure electric passenger vehicles through the fuzzy time series analysis algorithm;

S3、引进隶属度权值矩阵，将观测值在各模糊集上的隶属度的值作为用模糊矩阵进行预测的权值，建立模糊时间序列模型，反映观测值与各模糊子区间之间的联系的同时，进行预测时候不再需要附加预测规则，从而实现预测的准确率的提升；S3. Introduce the membership degree weight matrix, use the membership degree value of the observation value on each fuzzy set as the weight value for prediction with the fuzzy matrix, establish a fuzzy time series model, and reflect the connection between the observation value and each fuzzy subinterval At the same time, it is no longer necessary to add prediction rules when making predictions, so as to improve the accuracy of predictions;

S4、对于路段能耗，根据实际观测值与聚类中心的实际情况，按照正态分布算法，将能耗划分为超高、较高、正常、较低、超低五个能耗状态描述；S4. For the energy consumption of road sections, according to the actual observation value and the actual situation of the cluster center, according to the normal distribution algorithm, the energy consumption is divided into five energy consumption state descriptions: super high, high, normal, low, and super low;

S5、路段能耗及其影响因素经过模糊聚类后，得到能耗聚类中心及其对应路段的实际行驶长度，结合经过模糊时间序列所推算的影响因素预测值，经过加权相似性匹配后，得到能耗匹配值，能耗匹配值与路段能耗权值运算最终得到加权能耗预测值；S5. After fuzzy clustering of road energy consumption and its influencing factors, the energy consumption cluster center and the actual driving length of the corresponding road section are obtained, combined with the predicted value of the influencing factors calculated through the fuzzy time series, after weighted similarity matching, The energy consumption matching value is obtained, and the calculation of the energy consumption matching value and the energy consumption weight value of the road section finally obtains the weighted energy consumption prediction value;

S6、纯电动乘用车营运续驶里程预测的方法是，在得到加权能耗预测值后，用剩余能耗值减去加权能耗预测值，得到预测剩余能耗值，再换算成续驶里程；S6. The method for predicting the mileage of pure electric passenger vehicles in operation is: after obtaining the weighted energy consumption forecast value, subtract the weighted energy consumption forecast value from the remaining energy consumption value to obtain the predicted remaining energy consumption value, and then convert it into continuous driving mileage;

S7、在第n+1个路段中，如果预测剩余能耗值满足最低能耗要求值的情况下，再对n+2个路段进行能耗预测，得到其加权能耗预测值后，以第n+1个路段的预测剩余能耗值减去第n+2个路段的预测剩余能耗值，后再次与最低能耗要求值进行比较，直至路段的预测剩余能耗值少于或者等于最低能耗要求值结束算法。S7. In the n+1th road section, if the predicted remaining energy consumption value meets the minimum energy consumption requirement value, then perform energy consumption prediction on the n+2 road section, and obtain its weighted energy consumption prediction value, then use the first The predicted remaining energy consumption value of the n+1 section minus the predicted remaining energy consumption value of the n+2th section is compared with the minimum energy consumption requirement value again until the predicted remaining energy consumption value of the section is less than or equal to the minimum energy consumption value. The energy consumption requirement value ends the algorithm.

优选的，步骤S1之前，还包括步骤S0：建立能耗模型以及等级划分，以已经规划好的行驶线路为基础，将线路按照路段进行划分，每一个路段的线路长度、驾驶员驾驶模型、电流、电压、路况均为影响路段能耗的因素；根据交通运行规律，以每周相同时间的行驶工况和影响能耗的因素为基础、路段平均的能耗平均值为能耗参考，对纯电动乘用车的能耗情况进行划分。Preferably, before step S1, step S0 is also included: establishing an energy consumption model and class division, based on the already planned driving route, dividing the route according to road sections, the line length of each road section, the driver's driving model, the current , voltage, and road conditions are all factors affecting the energy consumption of the road section; according to the traffic operation rules, based on the driving conditions at the same time of the week and the factors affecting energy consumption, the average energy consumption of the road section is the energy consumption reference. The energy consumption of electric passenger vehicles is divided.

优选的，以纯电动乘用车能耗数据为对象的模糊聚类算法具体步骤如下：Preferably, the specific steps of the fuzzy clustering algorithm based on the energy consumption data of pure electric passenger vehicles are as follows:

选取随机值ε>0，选定并初始化聚类中心V⁽⁰⁾，使之具有能耗标识，令s=0；Select a random value ε>0, select and initialize the cluster center V⁽⁰⁾ so that it has an energy consumption label, let s=0;

第一步：确定参数b以及初始化模糊分类矩阵U⁽⁰⁾；The first step: determine parameter b and initialize fuzzy classification matrix U⁽⁰⁾ ;

第二步：更新U^(s)为U^(s+1)；i=1,…,c;j=1,…,N，并按

进行迭代，并计算U^(s)时的

：$v_i^{\left(s\right)}=\frac{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left[\hat{P}\left({\mathrm{\ω}}_i\right|x_j,\widehat{\mathrm{\θ}})\right]}^b{x}_j}{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left[\hat{P}\left({\mathrm{\ω}}_i\right|x_j,\widehat{\mathrm{\θ}})\right]}^b},$i=1,2,…,c；Step 2: Update U^(s) to U^(s+1) ; i=1,…,c;j=1,…,N, and press

Iterate and calculate U^(s) when

:$v_i^{\left(\mathrm{the\; s}\right)}=\frac{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left[\hat{P}\left({\mathrm{\ω}}_i\right|x_j,\widehat{\mathrm{\θ}})\right]}^b{x}_j}{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left[\hat{P}\left({\mathrm{\ω}}_i\right|x_j,\widehat{\mathrm{\θ}})\right]}^b},$ i=1,2,...,c;

第三步：以矩阵范数比较U^(s)和U^(s+1)，如果||U^(s)-U^(s+1)||<ε，迭代停止；否则，s=s+1，返回第二步。Step 3: Compare U^(s) and U^(s+1) by matrix norm, if ||U^(s) -U^(s+1) ||<ε, the iteration stops; otherwise, s=s+1 , return to the second step.

优选的，步骤S2中，路段能耗样本中，标识属性能耗数据、路网运行指标的影响因素、聚类中心样本之间具有时间序列的特征，影响因素的时间序列特征直接影响了路段能耗值，进而对规划线路的能耗分配构成直接影响，为此，建立基于路段样本的时间序列模型，适合于建立续驶里程预测模型。Preferably, in step S2, in the energy consumption samples of road sections, there are characteristics of time series among the identification attribute energy consumption data, influencing factors of road network operation indicators, and cluster center samples, and the time series characteristics of influencing factors directly affect the energy consumption of road sections. Therefore, the establishment of a time series model based on road section samples is suitable for establishing a mileage prediction model.

优选的，步骤S4具体为：Preferably, step S4 is specifically:

设能耗论域U={u₁,u₂,…,u_n},n=5，论域U的模糊集A_i可以表示为$A_i=\frac{f_{A_i}\left(u_1\right)}{u_1}+\frac{f_{A_i}\left(u_2\right)}{u_2}+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{f_{A_i}\left(u_n\right)}{u_n}=\underset{U}{\&Integral;}\frac{f_{A_i}\left(u_k\right)}{u_k},\&ForAll;u_k\&Element;U,n=5,k=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,5,$其中

是模糊集A_i的隶属函数（可选择三角形、梯形等，根据纯电动乘用车的特性，本文选择等腰梯形作为模糊函数），符号“+”表示连接符

且

u_k为模糊集A_i的一个元素，是元素u_k属于模糊集A_i的隶属度，且隶属函数满足$f_{A_i}\left(u_k\right)\&Element;\left[\mathrm{0,1}\right],$1≤k≤n。Suppose the domain of energy consumption U={u₁ ,u₂ ,…,u_n },n=5, the fuzzy set A_i of domain U can be expressed as$A_i=\frac{f_{A_i}\left(u_1\right)}{u_1}+\frac{f_{A_i}\left(u_2\right)}{u_2}+\&CenterDot;\&Center\; Dot;\&CenterDot;+\frac{f_{A_i}\left(u_{\mathrm{no}}\right)}{u_{\mathrm{no}}}=\underset{u}{\&Integral;}\frac{f_{A_i}\left(u_k\right)}{u_k},\&ForAll;u_k\&Element;u,\mathrm{no}=5,k=\mathrm{1,2},\&Center\; Dot;\&Center\; Dot;\&Center\; Dot;,5,$ in

is the membership function of the fuzzy set A_i (triangular, trapezoidal, etc. can be selected. According to the characteristics of pure electric passenger vehicles, this paper chooses isosceles trapezoidal as the fuzzy function), and the symbol "+" indicates the connector

and

u_k is an element of fuzzy set A_i , is the membership degree of element u_k belonging to the fuzzy set A_i , and the membership function satisfy$f_{A_i}\left(u_k\right)\&Element;\left[\mathrm{0,1}\right],$ 1≤k≤n.

优选的，存在一种模糊关系集R(t-1,t)，并且满足

则F(t)可由F(t-1)推导得出；其中，

为一种关系运算算子，若F(t-1)=A_i，F(t)=A_i，则两个连续数据F(t-1)与F(t)之间的模糊逻辑关系（FLR）可表示为F(t-1)→F(t)，F(t-1)被称为模糊逻辑关系的左边关系，F(t)则为模糊逻辑关系的右边关系；根据纯电动乘用车的实际情况，所有路段能耗的模糊逻辑关系均为唯一对应，可定义模糊逻辑组（FLG），并表示为：F(t-1)→F(t),t=1,2,…,17；对于唯一顺序的模糊时间序列，每个路段的能耗权值表示如下：Preferably, there is a fuzzy relation set R(t-1,t), and it satisfies

Then F(t) can be derived from F(t-1); among them,

is a relational operator, if F(t-1)=A_i , F(t)=A_i , then the fuzzy logic relationship between two continuous data F(t-1) and F(t) ( FLR) can be expressed as F(t-1)→F(t), F(t-1) is called the left relation of the fuzzy logic relation, and F(t) is the right relation of the fuzzy logic relation; In the actual situation of vehicles, the fuzzy logic relationship of energy consumption of all road sections is uniquely corresponding, and the fuzzy logic group (FLG) can be defined and expressed as: F(t-1)→F(t),t=1,2, ...,17; for a unique sequence of fuzzy time series, the energy consumption weight of each road segment is expressed as follows:

$\mathrm{<span><span>W</span>W</span>}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>=</span>=</span><span><span>[</span>[</span>{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>,</span>,</span>{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>,</span>,</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span>{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>]</span>]</span>$

$<span><span>=</span>=</span><span><span>[</span>[</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>,</span>,</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>]</span>]</span>$

$<span><span>=</span>=</span><span><span>[</span>[</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>]</span>]</span>$

其中w_n(t)是其中一个路段能耗的权值；c_n为该路段所对应能耗等级模糊隶属度

所对应的去模糊化正态分布值；E_n(t)是该路段对应的聚类中心所对应的能耗值。Among them, w_n (t) is the weight of energy consumption of one of the road sections; c_n is the fuzzy membership degree of energy consumption level corresponding to the road section

The corresponding defuzzification normal distribution value; E_n (t) is the energy consumption value corresponding to the cluster center corresponding to this section.

优选的，步骤S7的过程中，续驶里程可根据预测剩余能耗值满足最低能耗要求值的情况进行累加，直到算法结束，其累计续驶里程总和即为纯电动乘用车的最大营运续驶里程，即：Preferably, in the process of step S7, the driving mileage can be accumulated according to the predicted remaining energy consumption value meeting the minimum energy consumption requirement value, until the end of the algorithm, the sum of the accumulated driving mileage is the maximum operating value of the pure electric passenger car Driving distance, namely:

$\mathrm{<span><span>L</span>L</span>}<span><span>=</span>=</span>\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>k</span>k</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>{\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>,</span>,</span>{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>-</span>-</span>{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>-</span>-</span>\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>\&GreaterEqual;</span>\&Greater\; Equal;</span>\mathrm{<span><span>0</span>0</span>}<span><span>.</span>.</span>$

本发明相对于现有技术具有如下的优点及效果：Compared with the prior art, the present invention has the following advantages and effects:

1、充分考虑影响影响纯电动乘用车能耗的各类因素，并且将各类因素进行量化计算，跳出目前以电池剩余能量为核心的预测算法，提出了综合考虑各种因素的数学模型。1. Fully consider various factors that affect the energy consumption of pure electric passenger vehicles, and quantify and calculate various factors, jump out of the current prediction algorithm that focuses on the remaining energy of the battery, and propose a mathematical model that comprehensively considers various factors.

2、将驾驶员的驾驶行为转变车辆状态行为予以量化描述，即以加减速、滑行、制动、匀速行驶等进行量化描述。2. Quantitatively describe the driver's driving behavior changes and vehicle state behaviors, that is, quantitatively describe acceleration and deceleration, coasting, braking, and constant speed driving.

3、根据纯电动乘用车驾驶情况，必须先定义好目的地，然后根据可选择路径中的路段交通情况等进行能耗预测，从而实现续驶里程的预测。3. According to the driving situation of pure electric passenger vehicles, the destination must be defined first, and then the energy consumption prediction is performed according to the traffic conditions of the road sections in the optional route, so as to realize the prediction of the driving range.

4、由于纯电动乘用车运行里程影响因素较多，利用交通运行指标预测能耗的时间序列方法并不合适，特别是行驶路段中不断区域之间的交通运行指标差别较大的情况下。本发明提供的方法从研究各种影响因素的模糊聚类方法出发，预测各种影响因素与路段交通工况，结合模糊聚类中心与路段的加权能耗系数，可克服目前模糊时间序列在瞬间突变区域的预测不准确问题。4. Since there are many factors affecting the operating mileage of pure electric passenger vehicles, the time series method of predicting energy consumption using traffic operating indicators is not suitable, especially when there are large differences in traffic operating indicators between continuous areas in the driving section. The method provided by the invention starts from the fuzzy clustering method of studying various influencing factors, predicts various influencing factors and road section traffic conditions, and combines the fuzzy clustering center and the weighted energy consumption coefficient of the road section, which can overcome the current fuzzy time series in an instant Inaccurate prediction of mutation regions.

5、通过规划好的行驶线路，并进行路段长度累加续驶里程的方法可以将续驶里程准确定位到具体路段上。在驾驶过程中若遇到线路调整等，其能耗等也可以随之进行重新计算，这对促进PND、CPND、车载信息服务等行业其较大作用，特别是纯电动乘用车车载信息服务方面，提供简单、快捷、稳定的位置服务。5. Through the planned driving route and the method of accumulating the driving mileage of the length of the road section, the driving mileage can be accurately positioned on the specific road section. If there is line adjustment during driving, its energy consumption can also be recalculated accordingly, which will play a major role in promoting PND, CPND, vehicle information service and other industries, especially pure electric passenger vehicle vehicle information service On the one hand, it provides simple, fast and stable location services.

6、通过准确预测续驶里程，可解决目前纯电动乘用车难以实现里程预测的问题。由于纯电动乘用车续驶里程预测的不准确性，充电站等分布比较单一，直接影响了购车者对纯电动乘用车的购买欲望。根据实际情况建立预测模型，准确预测各种交通工况下的续驶里程，可为充电站规划和推动纯电动乘用车发展提供决策支持。6. By accurately predicting the mileage, it can solve the problem that it is difficult to predict the mileage of pure electric passenger vehicles. Due to the inaccuracy of the prediction of the driving range of pure electric passenger vehicles, the distribution of charging stations is relatively simple, which directly affects the purchase desire of car buyers for pure electric passenger vehicles. Establish a prediction model based on the actual situation to accurately predict the mileage under various traffic conditions, which can provide decision support for the planning of charging stations and the promotion of the development of pure electric passenger vehicles.

附图说明Description of drawings

图1是本发明的流程图。Fig. 1 is a flow chart of the present invention.

具体实施方式Detailed ways

下面结合实施例及附图对本发明作进一步详细的描述，但本发明的实施方式不限于此。The present invention will be further described in detail below in conjunction with the embodiments and the accompanying drawings, but the embodiments of the present invention are not limited thereto.

实施例Example

如图1所示，本法发明纯电动乘用车续驶里程预测方法，包括下述步骤：As shown in Figure 1, the method for predicting the driving range of a pure electric passenger car according to the present method comprises the following steps:

1.1能耗模型及等级划分1.1 Energy consumption model and classification

不同的路段能耗的差异性导致了纯电动乘用车的能耗必须进行细分并建立不同的模型方可准确对续驶里程进行预测。城市区域行驶路段的工况受路网影响较大，因此纯电动乘用车在不同工况下驾驶其续驶里程必须根据具体的能耗影响因素的具体情况进行预测。传统的根据剩余能量的简单预测方式在实际应用中其准确性难以达到实际使用要求。The difference in energy consumption of different road sections leads to the need to subdivide the energy consumption of pure electric passenger vehicles and establish different models to accurately predict the driving range. The working conditions of road sections in urban areas are greatly affected by the road network, so the driving range of pure electric passenger vehicles under different working conditions must be predicted according to the specific conditions of specific energy consumption factors. The accuracy of the traditional simple prediction method based on the remaining energy is difficult to meet the actual use requirements in practical applications.

纯电动乘用车的能耗模型是根据城市每一个路段的能耗情况进行续驶里程计算，并累加需要行驶的不同路段总和进行计算，因此，纯电动乘用车续驶里程的预测在实际使用中，必须首先输入目的地，规划线路，方可进行预测。The energy consumption model of pure electric passenger vehicles is based on the energy consumption of each road section in the city to calculate the driving mileage, and calculate the sum of different road sections that need to be driven. Therefore, the prediction of pure electric passenger vehicle driving mileage is practical In use, you must first input the destination and plan the route before making predictions.

以已经规划好的行驶线路为基础，将线路按照路段进行划分，每一个路段的线路长度、驾驶员驾驶模型、电流、电压、路况（交通运行指标）等均为影响路段能耗的因素。根据交通运行规律，以每周相同时间的行驶工况和影响能耗的因素为基础、路段平均的能耗平均值为能耗参考，对纯电动乘用车的能耗情况进行划分。表1至表5为能耗划分的情况，其数据为城市中心区某工作日的能耗划分情况。能耗划分的具体边界值可根据实际情况进行调整。Based on the planned driving route, the route is divided into road sections. The length of each road section, driver's driving model, current, voltage, road conditions (traffic operation indicators), etc. are all factors that affect the energy consumption of the road section. According to the law of traffic operation, based on the driving conditions and factors affecting energy consumption at the same time every week, the average energy consumption of the road section is used as the energy consumption reference, and the energy consumption of pure electric passenger vehicles is divided. Tables 1 to 5 show the division of energy consumption, and the data are the division of energy consumption on a working day in the city center. The specific boundary value of energy consumption division can be adjusted according to the actual situation.

表1影响因素能耗分级（电池部分）Table 1 Influencing factors Energy consumption classification (battery part)

Table1The Grading of Energy Consumption Influence Factors（Battery）Table1 The Grading of Energy Consumption Influence Factors (Battery)

表2影响因素能耗分级（交通运行指标部分）Table 2 Energy consumption classification of influencing factors (traffic operation index part)

Table2The Grading of Energy Consumption Influence Factors（Traffic Performance Index）Table2 The Grading of Energy Consumption Influence Factors (Traffic Performance Index)

表3影响因素能耗分级（驾驶员行为参数部分）Table 3 Energy consumption classification of influencing factors (driver behavior parameters part)

Table3The Grading of Energy Consumption Influence Factors（Driving Characters）Table3 The Grading of Energy Consumption Influence Factors (Driving Characters)

表4影响因素能耗分级（驾驶员里程参数部分）Table 4 Influencing factors energy consumption classification (driver mileage parameter part)

Table4The Grading of Energy Consumption Influence Factors（Driving Actual Service Life Characters）Table4 The Grading of Energy Consumption Influence Factors (Driving Actual Service Life Characters)

表5影响因素能耗分级（其他部分）Table 5 Influencing Factors Energy Consumption Classification (Other Parts)

Table5The Grading of Energy Consumption Influence Factors（Other）Table5 The Grading of Energy Consumption Influence Factors (Other)

1.2续驶里程预测算法1.2 Driving range prediction algorithm

在已经对纯电动乘用车能耗的主要影响因素进行量化后，可建立基于路段的线路能耗模型。After the main factors affecting the energy consumption of pure electric passenger vehicles have been quantified, a section-based line energy consumption model can be established.

以路段能耗为基础，纯电动乘用车营运过程中所有影响因素作为聚类的维度，采用模糊聚类的方式进行建模。为更好地对能耗进行定性描述，能耗等级以及影响因素等级已经按照实际情况根据专家意见以及实际营运均值数据等进行划分，而模糊聚类中心的数据具有不确定性，同时聚类中心所表达的模型不一定能够准确表达典型能耗状况，因此本文选取典型能耗样本进行标识，使其成为初始化的带标识的聚类中心，并对其他样本数据完成聚类运算，达到更好地完成能耗数据划分的目的，从而建立基于模糊聚类的时间序列能耗分析模型。Based on the energy consumption of the road section, all the influencing factors in the operation process of pure electric passenger vehicles are used as the clustering dimension, and the fuzzy clustering method is used for modeling. In order to better qualitatively describe energy consumption, the energy consumption level and the level of influencing factors have been divided according to the actual situation according to expert opinions and actual operating average data, while the data of the fuzzy clustering center is uncertain, and the clustering center The model expressed may not be able to accurately express the typical energy consumption situation. Therefore, this paper selects typical energy consumption samples for identification, making it an initialized cluster center with identification, and completes clustering operations on other sample data to achieve better energy consumption. Complete the purpose of energy consumption data division, so as to establish a time series energy consumption analysis model based on fuzzy clustering.

聚类算法的每一步迭代中，每一个样本点都被认为是完全属于某一类别。模糊聚类算法放松这一条件，假设是模糊隶属（Fuzzy Membership）于某一类。从根本上说，这种隶属度函数等价于式

中的

其中，

是隶属函数的参数向量。In each iteration of the clustering algorithm, each sample point is considered to belong to a certain category. The fuzzy clustering algorithm relaxes this condition, assuming that fuzzy membership (Fuzzy Membership) belongs to a certain category. Fundamentally, this membership function is equivalent to the formula

middle

in,

is the parameter vector of the membership function.

对于纯电动乘用车能耗影响因素数据集X={x₁,x₂,…,x_N}，N为数据集中元素的个数，c是聚类中心数，最小化全局代价函数为：For the data set X={x₁ ,x₂ ,…,x_N } of factors affecting energy consumption of pure electric passenger vehicles, N is the number of elements in the data set, c is the number of cluster centers, and the minimized global cost function is:

${\mathrm{<span><span>J</span>J</span>}}_{\mathrm{<span><span>fuz</span>fuz</span>}}<span><span>(</span>(</span>\mathrm{<span><span>U</span>u</span>}<span><span>,</span>,</span>\mathrm{<span><span>V</span>V</span>}<span><span>)</span>)</span><span><span>=</span>=</span>\underset{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>c</span>c</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}\underset{\mathrm{<span><span>j</span>j</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{<span><span>[</span>[</span>\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>P</span>P</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>\&omega;</span>\&omega;</span>}}_{\mathrm{<span><span>i</span>i</span>}}<span><span>|</span>|</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>\&theta;</span>\&theta;</span>}}<span><span>)</span>)</span><span><span>]</span>]</span>}^{\mathrm{<span><span>b</span>b</span>}}{<span><span>|</span>|</span><span><span>|</span>|</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>-</span>-</span>{\mathrm{<span><span>\&mu;</span>\&mu;</span>}}_{\mathrm{<span><span>i</span>i</span>}}<span><span>|</span>|</span><span><span>|</span>|</span>}^{\mathrm{<span><span>2</span>2</span>}}$

$\mathrm{<span><span>s</span>the\; s</span>}<span><span>.</span>.</span>\mathrm{<span><span>t</span>t</span>}<span><span>.</span>.</span>\left\{\begin{array}{c}\underset{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>c</span>c</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>P</span>P</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>\&omega;</span>\&omega;</span>}}_{\mathrm{<span><span>i</span>i</span>}}<span><span>|</span>|</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>\&theta;</span>\&theta;</span>}}<span><span>)</span>)</span><span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span><span><span>\&ForAll;</span>\&ForAll;</span>\mathrm{<span><span>j</span>j</span>}\\ \stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>P</span>P</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>\&omega;</span>\&omega;</span>}}_{\mathrm{<span><span>i</span>i</span>}}<span><span>|</span>|</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>\&theta;</span>\&theta;</span>}}<span><span>)</span>)</span><span><span>\&Element;</span>\&Element;</span><span><span>[</span>[</span>\mathrm{<span><span>0,1</span>0,1</span>}<span><span>]</span>]</span><span><span>,</span>,</span><span><span>\&ForAll;</span>\&ForAll;</span>\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}\\ \underset{\mathrm{<span><span>j</span>j</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>N</span>N</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>P</span>P</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>\&omega;</span>\&omega;</span>}}_{\mathrm{<span><span>i</span>i</span>}}<span><span>|</span>|</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>\&theta;</span>\&theta;</span>}}<span><span>)</span>)</span><span><span>></span>></span>\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span><span><span>\&ForAll;</span>\&ForAll;</span>\mathrm{<span><span>i</span>i</span>}\end{array}\right.$

其中，V={v₁,v₂,…,v_c}，v_i为ω_i类的中心矢量，μ_i是正态分布的均值，即聚类中心，b是一个用来控制不同类别的混合程度的自由参数，即权值b∈(1,∞)。每个样本点的聚类隶属度函数是归一化的，即Among them, V={v₁ ,v₂ ,…,v_c }, v_i is the center vector of class ω_i , μ_i is the mean value of normal distribution, that is, the cluster center, b is a variable used to control different classes The free parameter of the degree of mixing is the weight b∈(1,∞). The cluster membership function of each sample point is normalized, that is

$\underset{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>0</span>0</span>}}{\overset{\mathrm{<span><span>c</span>c</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>P</span>P</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>\&omega;</span>\&omega;</span>}}_{\mathrm{<span><span>i</span>i</span>}}<span><span>|</span>|</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>)</span>)</span><span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span>\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>,</span>,</span>\mathrm{<span><span>c</span>c</span>}<span><span>;</span>;</span>\mathrm{<span><span>j</span>j</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>,</span>,</span>\mathrm{<span><span>N</span>N</span>}$

令

表示先验类别概率

求解

和

即可推出${\mathrm{\&mu;}}_i=\frac{{\mathrm{\&Sigma;}}_{j=1}^n{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j)\right]}^b{x}_j}{{\mathrm{\&Sigma;}}_{j=1}^n{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j)\right]}^b},$$\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j)=\frac{{(1/{\left|\right|x_j-{\mathrm{\&mu;}}_i\left|\right|}^2)}^{1/(1-b)}}{{\mathrm{\&Sigma;}}_{r=1}^c{(1/{\left|\right|x_j-{\mathrm{\&mu;}}_i\left|\right|}^2)}^{1/(1-b)}}$使J_fuz(U,V)达到最小的解。make

Denotes the prior category probabilities

solve

and

ready to launch${\mathrm{\&mu;}}_i=\frac{{\mathrm{\&Sigma;}}_{j=1}^{\mathrm{no}}{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j)\right]}^b{x}_j}{{\mathrm{\&Sigma;}}_{j=1}^{\mathrm{no}}{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j)\right]}^b},$$\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j)=\frac{{(1/{\left|\right|x_j-{\mathrm{\&mu;}}_i\left|\right|}^2)}^{1/(1-b)}}{{\mathrm{\&Sigma;}}_{r=1}^c{(1/{\left|\right|x_j-{\mathrm{\&mu;}}_i\left|\right|}^2)}^{1/(1-b)}}$ Make J_fuz (U,V) reach the minimum solution.

以纯电动乘用车能耗数据为对象的模糊聚类算法具体步骤如下：The specific steps of the fuzzy clustering algorithm based on the energy consumption data of pure electric passenger vehicles are as follows:

选取随机值ε>0，选定并初始化聚类中心V⁽⁰⁾，使之具有能耗标识，令s=0。Select a random value ε>0, select and initialize the cluster center V⁽⁰⁾ to make it have an energy consumption label, and set s=0.

第一步：确定参数b以及初始化模糊分类矩阵U⁽⁰⁾；The first step: determine parameter b and initialize fuzzy classification matrix U⁽⁰⁾ ;

第二步：更新U^(s)为U^(s+1)。i=1,…,c;j=1,…,N，并按

进行迭代，并计算U^(s)时的

：Step 2: Update U^(s) to U^(s+1) . i=1,…,c;j=1,…,N, and press

Iterate and calculate U^(s) when

:

$v_i^{\left(s\right)}=\frac{\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j,\widehat{\mathrm{\&theta;}})\right]}^b{x}_j}{\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j,\widehat{\mathrm{\&theta;}})\right]}^b},$i=1,2,…,c；$v_i^{\left(\mathrm{the\; s}\right)}=\frac{\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j,\widehat{\mathrm{\&theta;}})\right]}^b{x}_j}{\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[\hat{P}\left({\mathrm{\&omega;}}_i\right|x_j,\widehat{\mathrm{\&theta;}})\right]}^b},$ i=1,2,...,c;

第三步：以矩阵范数比较U^(s)和U^(s+1)，如果||U^(s)-U^(s+1)||<ε，迭代停止；否则，s=s+1，返回第二步。Step 3: Compare U^(s) and U^(s+1) by matrix norm, if ||U^(s) -U^(s+1) ||<ε, the iteration stops; otherwise, s=s+1 , return to the second step.

通过模糊聚类中心对路段之间的能耗及影响因素的能耗情况进行描述以及标识后，可通过模糊时间序列分析算法对纯电动乘用车能耗状况进行建模。路段能耗样本中，标识属性能耗数据、路网运行指标等影响因素、聚类中心样本之间具有时间序列的特征，影响因素的时间序列特征直接影响了路段能耗值，进而对规划线路的能耗分配构成直接影响，为此，建立基于路段样本的时间序列模型，适合于建立续驶里程预测模型。After describing and marking the energy consumption between road sections and the energy consumption of influencing factors through the fuzzy clustering center, the energy consumption of pure electric passenger vehicles can be modeled through the fuzzy time series analysis algorithm. In the energy consumption samples of road sections, there are time series characteristics between the identification attribute energy consumption data, road network operation indicators and other influencing factors, and cluster center samples. Therefore, the establishment of a time series model based on road section samples is suitable for establishing a mileage prediction model.

引进隶属度权值矩阵，将观测值在各模糊集上的隶属度的值作为用模糊矩阵进行预测的权值，建立模糊时间序列模型，反映观测值与各模糊子区间之间的联系的同时，进行预测时候不再需要附加预测规则，从而实现预测的准确率的提升。需要说明的是，影响因素的权值矩阵反映影响因素的关联性和影响程度，而总体能耗的权值矩阵反映的则是相邻路段的能耗效率状况。Introduce the weight matrix of membership degree, use the value of the membership degree of the observation value on each fuzzy set as the weight value for prediction with the fuzzy matrix, establish a fuzzy time series model, reflect the relationship between the observation value and each fuzzy subinterval, and at the same time , it is no longer necessary to add prediction rules when making predictions, so as to improve the accuracy of predictions. It should be noted that the weight matrix of the influencing factors reflects the relevance and degree of influence of the influencing factors, while the weight matrix of the overall energy consumption reflects the energy efficiency of adjacent road sections.

对于路段能耗，根据实际观测值与聚类中心的实际情况，按照正态分布算法，将能耗划分为超高、较高、正常、较低、超低五个能耗状态描述。设能耗论域U={u₁,u₂,…,u_n},n=5，论域U的模糊集A_i可以表示为$A_i=\frac{f_{A_i}\left(u_1\right)}{u_1}+\frac{f_{A_i}\left(u_2\right)}{u_2}+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{f_{A_i}\left(u_n\right)}{u_n}=\underset{U}{\&Integral;}\frac{f_{A_i}\left(u_k\right)}{u_k},\&ForAll;u_k\&Element;U,n=5,k=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,5,$其中是模糊集A_i的隶属函数（可选择三角形、梯形等，根据纯电动乘用车的特性，本文选择等腰梯形作为模糊函数），符号“+”表示连接符

且

u_k为模糊集A_i的一个元素，

是元素u_k属于模糊集A_i的隶属度，且隶属函数满足$f_{A_i}\left(u_k\right)\&Element;\left[\mathrm{0,1}\right],$1≤k≤n。For the energy consumption of road sections, according to the actual observation value and the actual situation of the cluster center, according to the normal distribution algorithm, the energy consumption is divided into five energy consumption state descriptions: super high, high, normal, low, and super low. Suppose the domain of energy consumption U={u₁ ,u₂ ,…,u_n },n=5, the fuzzy set A_i of domain U can be expressed as$A_i=\frac{f_{A_i}\left(u_1\right)}{u_1}+\frac{f_{A_i}\left(u_2\right)}{u_2}+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{f_{A_i}\left(u_{\mathrm{no}}\right)}{u_{\mathrm{no}}}=\underset{u}{\&Integral;}\frac{f_{A_i}\left(u_k\right)}{u_k},\&ForAll;u_k\&Element;u,\mathrm{no}=5,k=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,5,$ in is the membership function of the fuzzy set A_i (triangular, trapezoidal, etc. can be selected. According to the characteristics of pure electric passenger vehicles, this paper chooses isosceles trapezoidal as the fuzzy function), and the symbol "+" indicates the connector

and

u_k is an element of fuzzy set A_i ,

is the membership degree of element u_k belonging to the fuzzy set A_i , and the membership function satisfy$f_{A_i}\left(u_k\right)\&Element;\left[\mathrm{0,1}\right],$ 1≤k≤n.

存在一种模糊关系集R(t-1,t)，并且满足

则F(t)可由F(t-1)推导得出。其中，为一种关系运算算子，若F(t-1)=A_i，F(t)=A_i，则两个连续数据F(t-1)与F(t)之间的模糊逻辑关系（FLR）可表示为F(t-1)→F(t)，F(t-1)被称为模糊逻辑关系的左边关系，F(t)则为模糊逻辑关系的右边关系。根据纯电动乘用车的实际情况，所有路段能耗的模糊逻辑关系均为唯一对应，可定义模糊逻辑组（FLG），并表示为：F(t-1)→F(t),t=1,2,…,17。对于唯一顺序的模糊时间序列，每个路段的能耗权值表示如下：There is a fuzzy relation set R(t-1,t), and it satisfies

Then F(t) can be derived from F(t-1). in, is a relational operator, if F(t-1)=A_i , F(t)=A_i , then the fuzzy logic relationship between two continuous data F(t-1) and F(t) ( FLR) can be expressed as F(t-1)→F(t), F(t-1) is called the left relation of the fuzzy logic relation, and F(t) is the right relation of the fuzzy logic relation. According to the actual situation of pure electric passenger vehicles, the fuzzy logic relationship of energy consumption of all road sections is uniquely corresponding, and the fuzzy logic group (FLG) can be defined and expressed as: F(t-1)→F(t),t= 1,2,...,17. For a unique sequence of fuzzy time series, the energy consumption weight of each road segment is expressed as follows:

$\mathrm{<span><span>W</span>W</span>}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>=</span>=</span><span><span>[</span>[</span>{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>,</span>,</span>{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>,</span>,</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span>{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>]</span>]</span>$

$<span><span>=</span>=</span><span><span>[</span>[</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>,</span>,</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>+</span>+</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>]</span>]</span>$

$<span><span>=</span>=</span><span><span>[</span>[</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>2</span>2</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>,</span>,</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span>\frac{{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}<span><span>]</span>]</span>$

其中w_n(t)是其中一个路段能耗的权值；c_n为该路段所对应能耗等级模糊隶属度

所对应的去模糊化正态分布值；E_n(t)是该路段对应的聚类中心所对应的能耗值。Among them, w_n (t) is the weight of energy consumption of one of the road sections; c_n is the fuzzy membership degree of energy consumption level corresponding to the road section

The corresponding defuzzification normal distribution value; E_n (t) is the energy consumption value corresponding to the cluster center corresponding to this section.

纯电动乘用车营运续驶里程预测的关键部分在于如何确定路段能耗及其影响因素对能耗的影响。续驶里程预测算法流程。路段能耗及其影响因素经过模糊聚类后，得到能耗聚类中心及其对应路段的实际行驶长度，结合经过模糊时间序列所推算的影响因素预测值，经过加权相似性匹配后，得到能耗匹配值，能耗匹配值与路段能耗权值运算最终得到加权能耗预测值，即完成了上文所表述的全部过程。The key part of predicting the operating mileage of pure electric passenger vehicles is how to determine the energy consumption of the road section and the impact of its influencing factors on energy consumption. Algorithm flow of driving mileage prediction. After fuzzy clustering of road energy consumption and its influencing factors, the energy consumption cluster center and the actual driving length of the corresponding road section are obtained. The energy consumption matching value, the energy consumption matching value and the energy consumption weight of the road section are calculated to finally obtain the weighted energy consumption prediction value, that is, the entire process described above is completed.

纯电动乘用车营运续驶里程预测的方法是，在得到加权能耗预测值后，用剩余能耗值减去加权能耗预测值，得到预测剩余能耗值，再换算成续驶里程。The method of predicting the mileage of pure electric passenger vehicles in operation is: after obtaining the weighted energy consumption forecast value, subtract the weighted energy consumption forecast value from the remaining energy consumption value to obtain the predicted remaining energy consumption value, and then convert it into the continuation mileage.

在第n+1个路段中，如果预测剩余能耗值满足最低能耗要求值的情况下，再对n+2个路段进行能耗预测，得到其加权能耗预测值后，以第n+1个路段的预测剩余能耗值减去第n+2个路段的预测剩余能耗值，后再次与最低能耗要求值进行比较，直至路段的预测剩余能耗值少于或者等于最低能耗要求值结束算法。在该过程中，续驶里程可根据预测剩余能耗值满足最低能耗要求值的情况进行累加，直到算法结束，其累计续驶里程总和即为纯电动乘用车的最大营运续驶里程，即：In the n+1th road section, if the predicted remaining energy consumption value meets the minimum energy consumption requirement value, the energy consumption prediction is performed on the n+2 road section, and after the weighted energy consumption prediction value is obtained, the n+th The predicted remaining energy consumption value of one road segment minus the predicted remaining energy consumption value of the n+2th road segment, and then compared with the minimum energy consumption requirement value again until the predicted remaining energy consumption value of the road segment is less than or equal to the minimum energy consumption The required value ends the algorithm. In this process, the mileage can be accumulated according to the predicted remaining energy consumption value meeting the minimum energy consumption requirement value until the algorithm ends, and the sum of the accumulated mileage is the maximum operating mileage of pure electric passenger vehicles. Right now:

$\mathrm{<span><span>L</span>L</span>}<span><span>=</span>=</span>\underset{\mathrm{<span><span>n</span>no</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}{\overset{\mathrm{<span><span>k</span>k</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}}{\mathrm{<span><span>w</span>w</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>{\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>,</span>,</span>{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>-</span>-</span>{\mathrm{<span><span>E</span>E.</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>-</span>-</span>\mathrm{<span><span>1</span>1</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>\&GreaterEqual;</span>\&Greater\; Equal;</span>\mathrm{<span><span>0</span>0</span>}$

需要说明的是，续驶里程是模糊聚类中心对应行驶路段的里程长度累加结果。由于纯电动乘用车的目的地和行驶路段具有不确定性，因此在累加续驶里程时候不同线路对应的能耗和续驶里程有所不同。如果已经完成相同里程的行驶，则其差别体现在不同的剩余能耗值上。若在实际行驶过程中能够获取实际路段能耗值及实际行驶里程，可用实际值替代预测剩余能耗值进行预测运算。该方式既可以实际行驶中提升预测的准确性，又可以在不对续驶里程运算构成任何影响的情况对路段能耗权值的进行动态调整。It should be noted that the driving mileage is the cumulative result of the mileage length of the corresponding driving section of the fuzzy clustering center. Due to the uncertainty of the destination and driving route of pure electric passenger vehicles, the energy consumption and driving range of different lines are different when accumulating driving range. If the same mileage has been traveled, the difference is reflected in different remaining energy consumption values. If the actual road segment energy consumption value and the actual mileage can be obtained during the actual driving process, the actual value can be used instead of the predicted remaining energy consumption value for prediction calculation. This method can not only improve the accuracy of prediction during actual driving, but also dynamically adjust the energy consumption weight of the road section without any impact on the mileage calculation.

上述实施例为本发明较佳的实施方式，但本发明的实施方式并不受上述实施例的限制，其他的任何未背离本发明的精神实质与原理下所作的改变、修饰、替代、组合、简化，均应为等效的置换方式，都包含在本发明的保护范围之内。The above-mentioned embodiment is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited by the above-mentioned embodiment, and any other changes, modifications, substitutions, combinations, Simplifications should be equivalent replacement methods, and all are included in the protection scope of the present invention.

Claims (7)

1. The method for predicting the driving range of the pure electric passenger vehicle is characterized by comprising the following steps of:

s1, on the basis of road section energy consumption, taking all influence factors in the operation process of the pure electric passenger car as clustering dimensions, and modeling in a fuzzy clustering mode;

s2, describing and identifying the energy consumption among road sections and the energy consumption condition of the influence factors through a fuzzy clustering center, and modeling the energy consumption condition of the pure electric passenger car through a fuzzy time series analysis algorithm;

s3, introducing a membership weight matrix, taking the value of the membership of the observed value on each fuzzy set as a weight predicted by using the fuzzy matrix, establishing a fuzzy time series model, reflecting the relation between the observed value and each fuzzy subinterval, and avoiding adding a prediction rule when predicting, thereby realizing the improvement of the accuracy of prediction;

s4, dividing energy consumption into five energy consumption state descriptions of ultra-high energy consumption, normal energy consumption, low energy consumption and ultra-low energy consumption according to the actual observation value and the actual condition of the clustering center and a normal distribution algorithm for the energy consumption of the road section;

s5, obtaining an energy consumption clustering center and the actual driving length of a corresponding road section after fuzzy clustering of road section energy consumption and influence factors thereof, obtaining an energy consumption matching value after weighted similarity matching by combining the influence factor predicted value calculated by a fuzzy time sequence, and finally obtaining a weighted energy consumption predicted value after the energy consumption matching value is operated with a road section energy consumption weight;

s6, the method for predicting the running range of the pure electric passenger vehicle comprises the steps of subtracting the weighted energy consumption predicted value from the residual energy consumption value after the weighted energy consumption predicted value is obtained, obtaining a predicted residual energy consumption value, and converting the predicted residual energy consumption value into the running range;

and S7, in the (n + 1) th road section, if the predicted residual energy consumption value meets the minimum energy consumption requirement value, performing energy consumption prediction on the (n + 2) th road section to obtain a weighted energy consumption predicted value, subtracting the predicted residual energy consumption value of the (n + 2) th road section from the predicted residual energy consumption value of the (n + 1) th road section, and then comparing with the minimum energy consumption requirement value again until the predicted residual energy consumption value of the road section is less than or equal to the minimum energy consumption requirement value, thereby finishing the algorithm.

2. The pure electric passenger vehicle driving range prediction method according to claim 1, further comprising, before step S1, step S0: establishing an energy consumption model and grade division, and dividing the route according to road sections on the basis of the planned driving route, wherein the route length, the driver driving model, the current, the voltage and the road condition of each road section are factors influencing the energy consumption of the road sections; according to the traffic operation rule, the energy consumption condition of the pure electric passenger vehicle is divided on the basis of the driving condition and the factors influencing the energy consumption at the same time every week and the average energy consumption value of the road section as the energy consumption reference.

3. The method for predicting the driving range of the pure electric passenger vehicle according to claim 1, wherein the fuzzy clustering algorithm with the energy consumption data of the pure electric passenger vehicle as an object comprises the following specific steps:

selecting a random value epsilon>0, selecting and initializing a cluster center V⁽⁰⁾Make it have energy consumption label, let s = 0;

the first step is as follows: determining parameters b and initializing fuzzy classification matrix U⁽⁰⁾；

The second step is that: updating U^(s)Is U^(s+1)(ii) a i =1, …, c; j =1, …, N, and is as follows

Iterate and calculate U^(s)Of the hour

： <math> <mrow> <msubsup> <mi>v</mi> <mi>i</mi> <mrow> <mo>(</mo> <mi>s</mi> <mo>)</mo> </mrow> </msubsup> <mo>=</mo> <mfrac> <mrow> <munderover> <mi>&Sigma;</mi> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msup> <mrow> <mo>[</mo> <mover> <mi>P</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <msub> <mi>&omega;</mi> <mi>i</mi> </msub> <mo>|</mo> <msub> <mi>x</mi> <mi>j</mi> </msub> <mo>,</mo> <mover> <mi>&theta;</mi> <mo>^</mo> </mover> <mo>)</mo> </mrow> <mo>]</mo> </mrow> <mi>b</mi> </msup> <msub> <mi>x</mi> <mi>j</mi> </msub> </mrow> <mrow> <munderover> <mi>&Sigma;</mi> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msup> <mrow> <mo>[</mo> <mover> <mi>P</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <msub> <mi>&omega;</mi> <mi>i</mi> </msub> <mo>|</mo> <msub> <mi>x</mi> <mi>j</mi> </msub> <mo>,</mo> <mover> <mi>&theta;</mi> <mo>^</mo> </mover> <mo>)</mo> </mrow> <mo>]</mo> </mrow> <mi>b</mi> </msup> </mrow> </mfrac> <mo>,</mo> </mrow></math>i=1,2,…,c；

The third step: comparing U with matrix norm^(s)And U^(s+1)If | | | U^(s)-U^(s+1)||<Epsilon, the iteration stops; otherwise, s = s +1, return to the second step.

4. The pure electric passenger vehicle driving range prediction method according to claim 1, wherein in step S2, in the road section energy consumption samples, the attribute-identifying energy consumption data, the influence factors of the road network operation indexes, and the characteristics of the clustering center samples having time series therebetween are identified, the time series characteristics of the influence factors directly influence the road section energy consumption value, and further directly influence the energy consumption distribution of the planned route, and for this purpose, a time series model based on the road section samples is established, which is suitable for establishing the driving range prediction model.

5. The pure electric passenger vehicle driving range prediction method according to claim 1, wherein step S4 specifically includes:

let energy universe U = { U =₁,u₂,…,u_n}, n =5, fuzzy set A of discourse domain U_iCan be expressed as <math> <mrow> <msub> <mi>A</mi> <mi>i</mi> </msub> <mo>=</mo> <mfrac> <mrow> <msub> <mi>f</mi> <msub> <mi>A</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> </mrow> <msub> <mi>u</mi> <mn>1</mn> </msub> </mfrac> <mo>+</mo> <mfrac> <mrow> <msub> <mi>f</mi> <msub> <mi>A</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <msub> <mi>u</mi> <mn>2</mn> </msub> <mo>)</mo> </mrow> </mrow> <msub> <mi>u</mi> <mn>2</mn> </msub> </mfrac> <mo>+</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>+</mo> <mfrac> <mrow> <msub> <mi>f</mi> <msub> <mi>A</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <msub> <mi>u</mi> <mi>n</mi> </msub> <mo>)</mo> </mrow> </mrow> <msub> <mi>u</mi> <mi>n</mi> </msub> </mfrac> <mo>=</mo> <munder> <mo>&Integral;</mo> <mi>U</mi> </munder> <mfrac> <mrow> <msub> <mi>f</mi> <msub> <mi>A</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <msub> <mi>u</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mrow> <msub> <mi>u</mi> <mi>k</mi> </msub> </mfrac> <mo>,</mo> <mo>&ForAll;</mo> <msub> <mi>u</mi> <mi>k</mi> </msub> <mo>&Element;</mo> <mi>U</mi> <mo>,</mo> <mi>n</mi> <mo>=</mo> <mn>5</mn> <mo>,</mo> <mi>k</mi> <mo>=</mo> <mn>1,2</mn> <mo>,</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>,</mo> <mn>5</mn> <mo>,</mo> </mrow></math>Wherein

Is a fuzzy set A_iIs used (triangle, trapezoid, etc. can be selected, according to the characteristics of the pure electric passenger car, isosceles trapezoid is selected as fuzzy function in this text), the symbol "+" represents the connector

And is

u_kAs a fuzzy set A_iIs a function of one of the elements of (1),is an element u_kBelong to fuzzy set A_iDegree of membership of, and membership function

Satisfy the requirement of <math> <mrow> <msub> <mi>f</mi> <msub> <mi>A</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <msub> <mi>u</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>&Element;</mo> <mo>[</mo> <mn>0,1</mn> <mo>]</mo> <mo>,</mo> </mrow></math>1≤k≤n。

6. The pure electric passenger vehicle driving range prediction method according to claim 1, wherein a fuzzy relation set R (t-1, t) exists and is satisfied

F (t) can be derived from F (t-1); wherein,

is a relational operator, if F (t-1) = A_i，F(t)=A_iThe Fuzzy Logic Relationship (FLR) between two consecutive data F (t-1) and F (t) can be represented as F (t-1) → F (t), F (t-1) being called the left-hand relationship of the fuzzy logic relationship, and F (t) being the right-hand relationship of the fuzzy logic relationship; according to the actual situation of the pure electric passenger vehicle, the fuzzy logic relations of all the road section energy consumptions are unique correspondences, a Fuzzy Logic Group (FLG) can be defined and expressed as: f (t-1) → F (t), t =1,2, …, 17; for a uniquely ordered fuzzy time series, the energy consumption weight value for each road segment is expressed as follows:

<math> <mrow> <mi>W</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mo>[</mo> <msub> <mi>w</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>,</mo> <msub> <mi>w</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>,</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <msub> <mi>w</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>]</mo> </mrow></math>

<math> <mrow> <mo>=</mo> <mo>[</mo> <mfrac> <mrow> <msub> <mi>c</mi> <mn>1</mn> </msub> <msub> <mi>E</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <msub> <mi>c</mi> <mn>1</mn> </msub> <msub> <mi>E</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>c</mi> <mn>2</mn> </msub> <msub> <mi>E</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>+</mo> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>,</mo> <mfrac> <mrow> <msub> <mi>c</mi> <mn>2</mn> </msub> <msub> <mi>E</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <msub> <mi>c</mi> <mn>1</mn> </msub> <msub> <mi>E</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>c</mi> <mn>2</mn> </msub> <msub> <mi>E</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>+</mo> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>,</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>,</mo> <mfrac> <mrow> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <msub> <mi>c</mi> <mn>1</mn> </msub> <msub> <mi>E</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>c</mi> <mn>2</mn> </msub> <msub> <mi>E</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>+</mo> <mmultiscripts> <mrow> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mmultiscripts> </mrow> </mfrac> <mo>]</mo> </mrow></math>

<math> <mrow> <mo>=</mo> <mo>[</mo> <mfrac> <mrow> <msub> <mi>c</mi> <mn>1</mn> </msub> <msub> <mi>E</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mi>&Sigma;</mi> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>,</mo> <mfrac> <mrow> <msub> <mi>c</mi> <mn>2</mn> </msub> <msub> <mi>E</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mi>&Sigma;</mi> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>,</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mo>&CenterDot;</mo> <mfrac> <mrow> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mi>&Sigma;</mi> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>c</mi> <mi>n</mi> </msub> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>]</mo> </mrow></math>

wherein w_n(t) is the weight of energy consumption of one of the road sections; c. C_nFuzzy membership degree of energy consumption grade corresponding to the road section

The corresponding defuzzification normal distribution value; e_nAnd (t) is the energy consumption value corresponding to the clustering center corresponding to the road section.

7. The pure electric passenger vehicle driving range prediction method according to claim 1, wherein in the step S7, the driving range is accumulated according to a condition that the predicted remaining energy consumption value meets the minimum energy consumption requirement value until the algorithm is finished, and a total accumulated driving range is a maximum operating driving range of the pure electric passenger vehicle, that is:

<math> <mrow> <mi>L</mi> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <msub> <mi>w</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <msub> <mi>L</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>,</mo> <msub> <mi>E</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>E</mi> <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>&GreaterEqual;</mo> <mn>0</mn> <mo>.</mo> </mrow></math>

Images