WO2018045689A1 - 光伏二极管箝位型三电平逆变器的决策树svm故障诊断方法 - Google Patents
光伏二极管箝位型三电平逆变器的决策树svm故障诊断方法 Download PDFInfo
- Publication number
- WO2018045689A1 WO2018045689A1 PCT/CN2016/113645 CN2016113645W WO2018045689A1 WO 2018045689 A1 WO2018045689 A1 WO 2018045689A1 CN 2016113645 W CN2016113645 W CN 2016113645W WO 2018045689 A1 WO2018045689 A1 WO 2018045689A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- fault
- level inverter
- decision tree
- fault diagnosis
- energy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02S—GENERATION OF ELECTRIC POWER BY CONVERSION OF INFRARED RADIATION, VISIBLE LIGHT OR ULTRAVIOLET LIGHT, e.g. USING PHOTOVOLTAIC [PV] MODULES
- H02S50/00—Monitoring or testing of PV systems, e.g. load balancing or fault identification
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/40—Testing power supplies
- G01R31/42—AC power supplies
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/54—Testing for continuity
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02S—GENERATION OF ELECTRIC POWER BY CONVERSION OF INFRARED RADIATION, VISIBLE LIGHT OR ULTRAVIOLET LIGHT, e.g. USING PHOTOVOLTAIC [PV] MODULES
- H02S50/00—Monitoring or testing of PV systems, e.g. load balancing or fault identification
- H02S50/10—Testing of PV devices, e.g. of PV modules or single PV cells
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
Definitions
- the invention relates to the field of fault diagnosis of power electronic devices, in particular to a decision tree SVM fault diagnosis method for a photovoltaic diode clamp type three-level inverter.
- photovoltaic inverters are the most important components in photovoltaic systems, which are related to the safe, stable and efficient operation of the entire system.
- the three-level inverter is widely used in photovoltaic power generation systems due to its series voltage equalization, low switching loss, low output voltage harmonic content and high working efficiency. application.
- the three-level inverter increases the number of switching devices, and the reliability of the circuit is correspondingly reduced. Failure of any one device may cause the circuit to operate abnormally, or even cause a secondary failure, resulting in huge economic loss.
- the problem of fault diagnosis of photovoltaic three-level inverter is mainly in three aspects: First, in the aspect of circuit failure mode, only the failure of a single device open circuit is considered. In recent years, the multi-fault mode diagnosis of simultaneous failure of multiple devices has been discussed, but There are still few studies in this area, and the analysis problem is not comprehensive enough.
- the existing fault diagnosis method of two switching devices is complicated. The second is that the detection signals are mostly output voltage and output current, because the output exists. Inductive load, slow current change, which often increases the fault diagnosis time;
- Third, the diagnostic algorithm, intelligent diagnosis algorithm is gradually applied to the field of inverter fault diagnosis, such as artificial neural network, support vector machine, extreme learning machine and so on. Among them, the application of neural networks is more, and the neural network itself has many defects: more parameters need to be set, slow convergence, easy to fall into local optimum, etc., which seriously hinder the application of neural networks.
- Photovoltaic three-level inverters have many switching devices, and the types of faults are complicated. At the same time, in order to meet the real-time requirements of the system, the traditional methods can not meet the actual needs.
- the fault diagnosis and identification of the photovoltaic inverter is realized by appropriate feature extraction and analysis methods.
- Wavelet analysis is a signal time-frequency domain analysis method that can describe both the time domain and the frequency domain of a signal. And the localized information of the signal can be obtained, which is a hot spot of fault feature extraction in recent years.
- the particle swarm clustering algorithm is generalized on the basis of the particle swarm optimization algorithm.
- the particle swarm optimization algorithm is an emerging evolutionary computing technology based on group intelligence.
- the group intelligence guides the optimized search through individual cooperation and competition in the group. , has a strong versatility.
- Support Vector Machine (SVM) is a machine learning algorithm based on statistical learning theory. It has unique advantages in solving pattern recognition characterized by high dimensionality, nonlinearity and small samples. It is also very practical in the field of power electronics fault diagnosis. Value and application prospects.
- a decision tree SVM fault diagnosis method for a photovoltaic diode clamp type three-level inverter comprising: establishing a model of a photovoltaic diode clamp type three-level inverter circuit; and extracting a three-level inverter main circuit Open circuit fault characteristics; construct particle swarm clustering fault diagnosis decision tree; training and test decision tree support vector machine SVM fault classification model, and finally realize fault diagnosis of photovoltaic diode clamp type three-level inverter.
- Step 1 Establish a model of a photovoltaic diode clamped three-level inverter circuit
- the main circuit of the three-level inverter is mainly composed of three-phase bridge arms of A, B and C.
- Each phase bridge arm is composed of four main switch tubes, four freewheeling diodes and two midpoint clamp diodes. Due to the structure of the three-level inverter circuit itself, the maximum voltage that each switch tube can withstand during operation is only half of that of a two-level inverter, so the three-level inverter can greatly reduce the voltage of the switching device. Stress, meeting the requirements of high voltage inverter.
- the main circuit fault of the three-level inverter is mainly open circuit fault, including IGBT open circuit, series fuse blown and trigger pulse loss fault, and the midpoint clamp diode open circuit, so the fault classification is performed according to the actual operation situation. Taking the main circuit A phase as an example, there are three major categories of thirteen subcategories.
- a single device fails, that is, one of four power transistors and two midpoint clamp diodes fails, a total of six subclasses.
- Step 2 Extracting the open circuit fault characteristics of the main circuit of the three-level inverter
- each bridge arm voltage of different faults of the main circuit of the three-level inverter is different after being decomposed, that is, when the main circuit is faulty, the energy of each frequency band is affected. Under normal circumstances, the fault will be Some frequency energy enhances its effect, some frequency energy suppresses it, and fault output and normal output may differ. Therefore, energy of different frequency bands can be used as a fault feature.
- the main circuit of the diode-clamped three-level inverter with space vector pulse width modulation (SVPWM) control is modeled. After modeling, the j-wavelet multi-scale decomposition of the bridge arm voltage at various faults is performed. Extract j+1 signal features. The wavelet multi-scale decomposition coefficients are reconstructed, the energy of each frequency band is extracted, and the energy of the frequency band signal is calculated. Let E n be the energy of the nth decomposition coefficient sequence S n , then
- T 1 ' pE 0 /E E 1 /E ... E j /E] (3)
- E is the total energy of the signal
- each element in T 1 ' corresponds to the percentage of energy of each band.
- T [T 1 ' T 2 ' T 3 '] (4)
- the bridge arm voltage in each fault condition is extracted according to the above process, and finally the data sample is constructed.
- the third step constructing a particle swarm clustering fault diagnosis decision tree
- the particle swarm clustering algorithm needs to be initialized first, randomly initialize the particle swarm, set the parameters such as the number of clusters, the number of particles, the number of iterations, etc., and then randomly classify each sample, and calculate the clustering center and adapt as the initial clustering. Parameters such as degree, set the particle initial velocity to zero. In this way, the individual particle optimal position p id and the global optimal position p gd can be obtained from the initial particle group. And use the formula
- v id (t+1) ⁇ v id (t)+c 1 r 1 (p id -x id (t))+c 2 r 2 (p gd -x id (t)) (8)
- x id (t+1) x id (t)+v id (t+1) (9)
- ⁇ is the inertia weight
- v id is the velocity of the particle
- c 1 and c 2 Is the acceleration factor
- the clustering of each sample is determined, and according to the new clustering, the new clustering center is calculated and the fitness is updated. Again fitness comparison, if it is better than the optimal location of p id, update p id; if it is better than the global optimum position p gd, update p gd. If the maximum number of iterations is reached, the algorithm ends, otherwise iteration continues.
- the generation of the decision tree structure needs to first take all the training sample sets as the initial nodes, and use the particle swarm clustering algorithm to divide them into two categories to form two sub-nodes. Determine whether the child node contains only one type of fault sample. If yes, the algorithm ends. Otherwise, the particle clustering algorithm is used to process and divide into two new child nodes. This is divided until all child nodes contain only one type of fault sample. The algorithm ends. By clustering all the fault samples in this way, the fault diagnosis decision tree can be constructed in reverse.
- Step 4 Training and test decision tree support vector machine SVM fault classification model
- the fault data samples are divided into training set and test set.
- the training set trains the support vector machine SVM classification model according to the distribution of faults on the decision tree structure.
- the support vector machine SVM classification model uses the radial basis kernel function. And optimize the parameters of the support vector machine SVM.
- the test set is used to test the decision tree support vector machine SVM fault diagnosis model, and the diagnostic accuracy and other indicators are obtained, and finally the fault diagnosis of the photovoltaic diode clamp type three-level inverter is realized.
- the decision tree SVM fault diagnosis method of the diode clamp type three-level inverter proposed by the present invention is based on the data-driven idea, and the core is the first clustering and classification, the wavelet multi-scale decomposition, and the particle group clustering. Combined with the support vector machine algorithm, real-time data fault diagnosis of the photovoltaic diode clamp type three-level inverter is realized.
- the present invention clusters the data samples of the diode clamp type three-level inverter by particle swarm clustering algorithm until each subclass contains only one type of fault information, and then constructs a decision tree in reverse, so that Make the indexability between each sub-class as strong as possible, not only improve the diagnostic accuracy, but also strengthen the anti-interference ability.
- the invention adopts the fault diagnosis model structure of the decision tree, and has fewer classification models, greatly improves the fault diagnosis efficiency, and uses the radial basis as a kernel function to optimize the parameters, thereby effectively implementing the three-level inverter. Fault diagnosis.
- Figure 1 shows the fault diagnosis process of a diode-clamped three-level inverter.
- Figure 2 shows the main circuit topology of a diode-clamped three-level inverter.
- FIG. 3 shows the A-phase topology of the inverter main circuit
- Figure 4 shows the bridge voltage for a single device failure.
- Figure 5 shows the bridge voltage when two devices are simultaneously open.
- Figure 6 is a fault eigenvector histogram of the inverter when it is normal.
- Figure 7 is a decision tree structure diagram after clustering
- the decision tree SVM fault diagnosis flowchart of the diode clamp type three-level inverter of the present invention is shown in FIG. 1.
- the specific implementation of the method of the present invention includes the following steps:
- FIG. 2 the main circuit topology diagram of the diode clamp type three-level inverter is shown. To simplify the analysis, only the working state of the A phase in the inverter inverter state is studied.
- the circuit topology is shown in Fig. 3.
- the A-phase arm has three working states:
- P state S a1 and S a2 are turned on, S a3 and S a4 are turned off, and when the current direction is positive, current flows from point P through S a1 and Sa 2 into point A, ignoring the forward voltage drop of the switching device.
- the output terminal A potential is equal to the P point potential, that is, U dc /2; when the current direction is negative, the current flows from point A through the freewheeling diodes VD a1 and VD a2 into the point P, and the output terminal A potential is still equal to P. Point potential.
- N state S a3 and S a4 are turned on, S a1 and S a2 are turned off, when the current direction is positive, the current flows from point N through VD a3 and VD a4 into point A, and the potential at point A of the output terminal is equal to the potential of point N. That is, -U dc /2; when the current direction is negative, the current flows from point A through S a3 and Sa 4 into point N, and the potential at output terminal A is still equal to the potential at point N.
- the faults are divided into three categories, thirteen subclasses, that is, the fault type of the diode clamp type three-level inverter.
- a single device fails, that is, any one of the power tubes S a1 , S a2 , S a3 , S a4 and the clamp diodes VD a5 , VD a6 fails, a total of six subclasses.
- Two devices fail. There are two subclasses in this category. First, the two switching tubes that are faulty are not in the same bridge arm. You can refer to the failure of a single device, not counting faults. Second, the two faults.
- the switching tubes are on the same bridge arm, namely the power tubes (S a1 , S a2 ), (S a1 , S a3 ), (S a1 , S a4 ), (S a2 , S a3 ), (S a2 , S a4 ) Or (S a3 , S a4 ) Any of a group of failures, a total of six sub-categories.
- a three-phase three-level SVPWM inverter model is established, and the bridge arm voltage is selected as the research object, and the bridge arm voltage model under various fault conditions can be obtained, as shown in Fig. 3 and Fig. 4.
- the sym4 wavelet basis function is selected to perform three-layer multi-scale decomposition on the middle, upper and lower arm voltages, and each bridge arm voltage is decomposed into four small signals. After reconstruction, the signal is calculated. After the energy, unified dimension, construct the fault feature vector of each bridge arm.
- the fault eigenvectors of the bridge arms are integrated, and the fault eigenvectors of the system are constructed in the order of middle, upper and lower, and data samples are constructed according to different fault types.
- the histogram of the corresponding fault feature vector is as shown in FIG. 5.
- the particle swarm clustering algorithm is used to cluster all the obtained data samples to determine the subclass. If the subclass contains only one type of fault sample, the partitioning is stopped. Otherwise, the clustering is continued until all subclasses are used. Only one type of fault information is included. After the division is completed, the decision tree can be constructed in reverse.
- the structure of the fault diagnosis decision tree constructed as shown in Figure 6 is shown in Figure 6. It can be seen from the figure that the method only needs to construct 12 classification models for the 13 fault problems, and the one-to-one structure SVM needs to construct N(N-1)/2 classification models, namely 78 classification models. Adopting a decision tree structure will undoubtedly greatly reduce the number of model constructions and improve computational efficiency.
- the data samples are divided into training sets and test sets. According to the constructed decision tree structure, SVM1 ⁇ SVM12 are trained respectively, and a total of 12 support vector machine classification models are used. Radial basis kernel functions are used and the parameters of each support vector machine are optimized.
- the original data is compared with the white noise of 10% and 15% of the signal amplitude, and the BP neural network (BPNN) and the one-to-one structure of the support vector machine (SVM) are also compared horizontally.
- BPNN BP neural network
- SVM one-to-one structure of the support vector machine
- Table 1 The diagnostic accuracy of the decision tree support vector machine (DT-SVM) is summarized in Table 1. It can be seen from the table that with the increase of noise, the diagnostic accuracy of each algorithm has a certain decline, but the decision tree support vector machine algorithm has less diagnostic accuracy and stronger anti-interference ability.
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Power Engineering (AREA)
- Inverter Devices (AREA)
Abstract
一种光伏二极管箝位型三电平逆变器的决策树SVM故障诊断方法,针对光伏微网中的三电平逆变器故障诊断问题,以逆变状态为例,首先分析逆变器主电路的运行情况并进行故障分类,然后以中、上、下三种桥臂电压为测量信号,采用小波多尺度分解法提取特征信号,进而利用粒子群聚类算法生成决策树SVM分类模型,最终实现了光伏二极管箝位型三电平逆变器的多模式故障诊断。其优点是:该算法能够明显区分光伏二极管箝位型三电平逆变器的各个故障状态,使用了较少的分类模型完成故障诊断任务,且诊断精度高,抗干扰能力强。
Description
本发明涉及电力电子装置故障诊断领域,尤其是一种光伏二极管箝位型三电平逆变器的决策树SVM故障诊断方法。
全球范围的能源危机和环境危机促使人们寻求更加清洁、绿色的新型能源,在清洁型能源中太阳能凭借其:无污染、可持续、普遍性、灵活性和可靠性等优点,受到了广泛关注。在光伏系统越来越多并入电网运行的情况下,光伏逆变器作为光伏系统中最为核心的部件,关系着整个系统的安全、稳定和高效运行。与传统的两电平逆变器相比,三电平逆变器因其开关器件串联均压、开关损耗小、输出电压谐波含量低和工作效率高等优点,在光伏发电系统中得到广泛的应用。然而三电平逆变器增加了开关器件的数量,电路的可靠性也相应地降低,任意一个器件发生故障就可能导致电路非正常运行,甚至会导致二次故障,造成巨大的经济损失。
光伏三电平逆变器故障诊断的问题主要在三个方面:一是在电路故障模式方面,仅考虑单个器件开路的故障,近些年才开始讨论多器件同时故障的多故障模式诊断,但这方面的研究还较少,分析问题还不够全面,而现有的2个开关器件同时开路的故障诊断方法算法结构都较为复杂;二是检测信号多为输出电压和输出电流,由于输出端存在感性负载,电流变化缓慢,这往往会增加故障诊断时间;三是诊断算法方面,智能诊断算法逐渐运用到逆变器故障诊断领域,如人工神经网络、支持向量机、极限学习机等等。其中,神经网络的应用较多,而神经网络自身存在较多缺陷:需要设置的参数较多、收敛速度慢、易陷入局部最优等等,这些都严重阻碍着神经网络的应用。
光伏三电平逆变器的开关器件较多,故障问题种类繁杂,同时为了满足系统的实时性要求,传统的方法已经不能满足实际需求。这里利用数据驱动的思想,利用逆变器系统运行过程中不断产生并反映系统运行机理和状态的数据,通过合适的特征提取和分析方法,实现光伏逆变器的故障诊断和识别。
小波分析是一种信号时频域分析方法,它可以同时描述信号的时域和频域,
并且可以获得信号的局部化信息,是近年来故障特征提取的热点。粒子群聚类算法是在粒子群优化算法的基础上推广得到的,粒子群优化算法是一种基于群智能的新兴进化计算技术,通过群体中个体的合作和竞争而产生的群体智能指导优化搜索,具有较强的通用性。支持向量机是一种基于统计学习理论的机器学习算法,它在解决高维、非线性及小样本为特点的模式识别中,具有独特的优势,在电力电子的故障诊断领域也有很好的实用价值和应用前景。
发明内容
本发明的目的是提供一种光伏二极管箝位型三电平逆变器的决策树SVM故障诊断方法。
一种光伏二极管箝位型三电平逆变器的决策树SVM故障诊断方法,其特征包括:建立光伏二极管箝位型三电平逆变器电路的模型;提取三电平逆变器主电路开路故障特征;构建粒子群聚类故障诊断决策树;训练和测试决策树支持向量机SVM故障分类模型,最终实现光伏二极管箝位型三电平逆变器的故障诊断。
第一步:建立光伏二极管箝位型三电平逆变器电路的模型
三电平逆变器主电路主要由A、B和C三相桥臂构成,每相桥臂由四个主开关管、四个续流二极管和两个中点钳位二极管组成。由于三电平逆变器电路本身的结构,每个开关管在工作过程中所可能承受的最高电压只有两电平逆变器的一半,因此三电平逆变器可以大大降低开关器件的电压应力,满足高压逆变的要求。三电平逆变器主电路故障主要是开路故障,包括IGBT开路、串联熔断器熔断和触发脉冲丢失故障,同时还有中点钳位二极管开路的情况,因此根据实际运行的情况进行故障分类,以主电路A相为例,共三大类十三小类。
1)所有IGBT开关管都正常运行,逆变器无故障,共一小类。
2)单个器件发生故障,即在四个功率管和两个中点钳位二极管中任意一个发生故障,共六小类。
3)两个器件发生故障,这大类存在两种情况,一是故障的两个开关管不在同一桥臂,这种情况可以归结为不同桥臂上的单个器件故障,可以参考单个器件发生故障的情况;二是故障的两个开关管在同一桥臂,即四个开关管中任意两个开关管发生故障的情况,共六小类。
第二步:提取三电平逆变器主电路开路故障特征
三电平逆变器主电路不同故障的各个桥臂电压经分解后在各频带上的投影
是不同的,即主电路故障时,会对各频带的能量发生影响,一般情况下,故障会对某些频率能量其增强作用,某些频率能量其抑制作用,故障输出和正常输出会有差异,所以可以采用不同频带的能量作为故障特征。对采用空间矢量脉宽调制(SVPWM)控制的二极管箝位型三电平逆变器主电路进行建模,建模后对各种故障发生时的桥臂电压进行j层小波多尺度分解,分别提取j+1个信号特征。再对小波多尺度分解系数进行重构,提取各个频带的能量,计算频带信号的能量。设En为第n个分解系数序列Sn的能量,则
式中,Xn,n=0,1,…,j为重构信号的Sn的离散点幅值。进而得到各个桥臂电压的能量后就可以构建特征向量,其中特征向量T1为:
T1=[E0 E1 ... Ej] (2)统一量纲,归一化处理特征向量:
T1'=pE0/E E1/E ... Ej/E] (3)其中,E为信号的总能量,T1′中各个元素对应各个频带能量的百分比。采用同样的方法再处理上、下桥臂,可以分别得到特征向量T2′和T3′,定义故障特征向量为:
T=[T1' T2' T3'] (4)将各个故障情况下的桥臂电压按照上述过程进行特征提取,最后构建数据样本。
第三步:构建粒子群聚类故障诊断决策树
三电平逆变器共有13种故障小类,若要构建决策树将故障完全区分,就需要用到粒子群聚类算法将故障不断地划分成两类。粒子群聚类算法需要先进行初始化,随机初始化粒子群,设置聚类数目、粒子数目、迭代次数等参数,再将每个样本随机分类,并作为最初的聚类划分后计算聚类中心、适应度等参数,将粒子初速度设为零。这样就可以根据初始粒子群,得到的粒子个体最优位置pid和全局最优位置pgd。并利用公式
vid(t+1)=ωvid(t)+c1r1(pid-xid(t))+c2r2(pgd-xid(t)) (8)
xid(t+1)=xid(t)+vid(t+1) (9)更新所有粒子的速度和位置;其中ω为惯性权重;vid为粒子的速度;c1和c2为
加速度因子;r1和r2是分布于[0,1]之间的随机数;R1=1;R2=0.5;R3=4;R4=2;ωmax=1.2;ωmin=0.4;iter为当前迭代次数;itermax为最大的迭代次数。依据粒子的聚类中心编码,按照最近邻法则,确定每个样本的聚类划分,并按照新的聚类划分,计算新的聚类中心,更新适应度。再一次比较适应度,若其优于个体最优位置pid,则更新pid;若其优于全局最优位置pgd,则更新pgd。如果达到最大迭代次数,则算法结束,否则继续迭代。
决策树结构的生成需要先将全部训练样本集作为初始节点,利用粒子群聚类算法,将其划分成两类,形成两个子节点。判断子节点是否只包含一类故障样本,若是则算法结束,否则继续采用粒子群聚类算法进行处理,划分成两个新的子节点,这样不断划分,直到所有子节点只包含一类故障样本,算法结束。这样将所有的故障样本进行聚类划分,就可以反向构建故障诊断决策树。
第四步:训练和测试决策树支持向量机SVM故障分类模型
按照4:1的比例将故障数据样本划分成训练集和测试集,训练集按照决策树结构上故障的分布训练支持向量机SVM分类模型,支持向量机SVM分类模型均采用径向基核函数,并各自优化支持向量机SVM的参数。训练完成后,利用测试集测试决策树支持向量机SVM故障诊断模型,得到诊断精度等指标,最终实现光伏二极管箝位型三电平逆变器的故障诊断。
本发明的有益效果是:
1)本发明所提出的二极管箝位型三电平逆变器的决策树SVM故障诊断方法,是基于数据驱动的思想,核心是先聚类后分类,将小波多尺度分解、粒子群聚类和支持向量机算法结合起来,实现光伏二极管箝位型三电平逆变器的数据实时故障诊断。
2)本发明通过粒子群聚类算法,将二极管箝位型三电平逆变器的数据样本进行聚类划分,直到每个子类只包含一类故障信息,然后反向构建决策树,这样可以使各个子类之间的可分度尽可能地强,既提高诊断精度,又加强抗干扰能力。
3)本发明采用决策树的故障诊断模型结构,构建的分类模型较少,大大提高故障诊断效率,采用径向基作为核函数,进行各个参数的寻优,从而有效地实现三电平逆变器的故障诊断。
图1为二极管箝位型三电平逆变器的故障诊断流程
图2为二极管箝位型三电平逆变器主电路拓扑结构
图3为逆变器主电路的A相拓扑
图4为单个器件故障时的桥臂电压
图5为两个器件同时开路时的桥臂电压
图6为逆变器正常时的故障特征向量直方图
图7为聚类划分后的决策树结构图
下面结合附图对本发明做进一步说明。
本发明的二极管箝位型三电平逆变器的决策树SVM故障诊断流程图如图1所示,本发明方法的具体实施包括以下步骤:
如图2所示为二极管箝位型三电平逆变器主电路拓扑结构图,为简化分析,只研究逆变器逆变状态下A相的工作状态,其电路拓扑如图3所示。A相桥臂有三种工作状态:
P状态:Sa1和Sa2导通,Sa3和Sa4关断,电流方向为正时,电流从P点经Sa1和Sa2流进A点,忽略开关器件的正向导通压降后,输出端A点电位等于P点电位,即Udc/2;当电流方向为负时,电流从A点经续流二极管VDa1和VDa2流进P点,输出端A点电位仍等于P点电位。
O状态:Sa2和Sa3导通,Sa1和Sa4关断,电流方向为正时,电流从中性点O经VDa5和Sa2流进A点,输出端A点电位等于O点电位,即中性点电位;当电流方向为负时,电流从A点经Sa3和VDa6流进O点,输出端A点电位仍等于O点电位。
N状态:Sa3和Sa4导通,Sa1和Sa2关断,电流方向为正时,电流从N点经VDa3和VDa4流进A点,输出端A点电位等于N点电位,即-Udc/2;当电流方向为负时,电流从A点经Sa3和Sa4流进N点,输出端A电位仍等于N点电位。
根据拓扑结构,将故障分为三大类十三小类,即二极管箝位型三电平逆变器的故障类型。
1)所有IGBT开关管都正常运行,逆变器无故障,共一小类。
2)单个器件发生故障,即功率管Sa1、Sa2、Sa3、Sa4和钳位二极管VDa5、VDa6中任意一个发生故障,共六小类。
3)两个器件发生故障,这大类存在两种小类,一是故障的两个开关管不在同一桥臂,可以参考单个器件发生故障的情况,不计入故障分类;二是故障的两个开关管在同一桥臂,即功率管(Sa1,Sa2)、(Sa1,Sa3)、(Sa1,Sa4)、(Sa2,Sa3)、(Sa2,
Sa4)或(Sa3,Sa4)任意一组发生故障的情况,共六小类。
建立三相三电平SVPWM逆变器模型,选取桥臂电压为研究对象,可以得到各种故障情况下的桥臂电压模型,如图3和图4所示。考虑到输出电压的特点,选取sym4小波基函数,分别对中、上和下桥臂电压进行三层多尺度分解,每个桥臂电压被分解成4个小信号,经过重构后,计算信号的能量,统一量纲后,构建每个桥臂的故障特征向量。整合桥臂的故障特征向量,按照中、上和下的顺序构建系统的故障特征向量,并按照不同的故障类型,构建数据样本。其中,当逆变器正常工作时,其对应的故障特征向量的直方图如图5所示。
采用粒子群聚类算法,将获得的全部数据样本进行聚类划分,判断得到的子类,若该子类只包含一类故障样本,则停止划分,否则继续进行聚类划分,直到所有子类只包含一类故障信息。划分结束后可以反向构建决策树,最终构建的故障诊断决策树结构如图6所示。从图中可以看出该方法对于13种故障的问题,只需要构建12个分类模型,而一对一结构的SVM需要构建N(N-1)/2个分类模型,即78个分类模型,采用决策树结构无疑将大大减少模型构建数目,提高运算效率。
将数据样本分成训练集和测试集。按照构建的决策树结构,分别训练SVM1~SVM12,共12个支持向量机分类模型,均采用径向基核函数,并优化各个支持向量机的参数。为了验证算法的抗干扰能力,对原始数据加入信号幅值10%和15%的白噪声进行对比,同时还横向比较了BP神经网络(BPNN)、一对一结构的支持向量机(SVM)和决策树支持向量机(DT-SVM)的诊断精度,最终的故障诊断结果汇总如表1所示。从表中可以看出随着噪声的增加,各个算法的诊断精度都有一定下降,但决策树支持向量机算法的诊断精度下降较小,抗干扰能力较强。
表1故障诊断结果
上述实施例仅仅是为清楚地说明本发明所做的举例,而并非是对本发明的实施方式限定,对于所属领域的普通技术人员来说,在上述说明的基础上还可以做出其他不同形式的变化或变动。
Claims (1)
- 一种光伏二极管箝位型三电平逆变器的决策树SVM故障诊断方法,其特征包括:建立光伏二极管箝位型三电平逆变器电路的模型;提取三电平逆变器主电路开路故障特征;构建粒子群聚类故障诊断决策树;训练和测试决策树支持向量机SVM故障分类模型,最终实现光伏二极管箝位型三电平逆变器的故障诊断;第一步:建立光伏二极管箝位型三电平逆变器电路的模型三电平逆变器主电路主要由A、B和C三相桥臂构成,每相桥臂由四个主开关管、四个续流二极管和两个中点钳位二极管组成,由于三电平逆变器电路本身的结构,每个开关管在工作过程中可能承受的最高电压只有两电平逆变器的一半,因此三电平逆变器可以大大降低开关器件的电压应力,满足高压逆变的要求;三电平逆变器主电路故障主要是开路故障,包括IGBT开路、串联熔断器熔断和触发脉冲丢失故障,同时还有中点钳位二极管开路的情况,因此根据实际运行的情况进行故障分类,以主电路A相为例,共三大类十三小类;1)所有IGBT开关管都正常运行,逆变器无故障,共一小类;2)单个器件发生故障,即在四个功率管和两个中点钳位二极管中任意一个发生故障,共六小类;3)两个器件发生故障,这大类存在两种情况,一是故障的两个开关管不在同一桥臂,这种情况可以归结为不同桥臂上的单个器件故障,可以参考单个器件发生故障的情况;二是故障的两个开关管在同一桥臂,即四个开关管中任意两个开关管发生故障的情况,共六小类;第二步:提取三电平逆变器主电路开路故障特征三电平逆变器主电路不同故障的各个桥臂电压经分解后在各频带上的投影是不同的,即主电路故障时,会对各频带的能量发生影响,一般情况下,故障会对某些频率能量其增强作用,某些频率能量其抑制作用,故障输出和正常输出会有差异,所以可以采用不同频带的能量作为故障特征,对采用空间矢量脉宽调制SVPWM控制的二极管箝位型三电平逆变器主电路进行建模,建模后对各种故障发生时的桥臂电压进行j层小波多尺度分解,分别提取j+1个信号特征,再对小波多尺度分解系数进行重构,提取各个频带的能量,计算频带信号的能量,设En为第n个分解系数序列Sn的能量,则式中,Xn,n=0,1,…,j为重构信号序列Sn的离散点幅值,进而得到各个桥臂电压的能量后就可以构建特征向量,其中特征向量T1为:T1=[E0 E1 ... Ej] (2)进一步统一量纲,归一化处理特征向量得到:T1'=[E0/E E1/E ... Ej/E] (3)其中,E为信号的总能量,T1′中各个元素对应各个频带能量的百分比,采用同样的方法再处理上、下桥臂,可以分别得到特征向量T2′和T3′,定义故障特征向量为:T=[T1' T2' T3'] (4)将各个故障情况下的桥臂电压按照上述过程进行特征提取,最后构建数据样本;第三步:构建粒子群聚类故障诊断决策树由于三电平逆变器共有13种故障小类,若要构建决策树将故障完全区分,就需要用到粒子群聚类算法将故障不断地划分成两类;粒子群聚类算法需要先进行初始化,随机初始化粒子群,设置聚类数目、粒子数目、迭代次数等参数,再将每个样本随机分类,并作为最初的聚类划分后计算聚类中心、适应度等参数,设置粒子初速度为零;这样就可以根据初始粒子群得到的粒子个体的最优位置pid和全局最优位置pgd,并利用公式vid(t+1)=ωvid(t)+c1r1(pid-xid(t))+c2r2(pgd-xid(t)) (8)xid(t+1)=xid(t)+vid(t+1) (9)更新所有粒子的速度和位置;其中ω为惯性权重;vid为粒子的速度;c1和c2为加速度因子;r1和r2是分布于[0,1]之间的随机数;R1=1;R2=0.5;R3=4;R4=2;ωmax=1.2;ωmin=0.4;iter为当前迭代次数;itermax为最大的迭代次数;依据粒子的聚类中心编码,按照最近邻法则,确定每个样本的聚类划分,并按照新的聚类划分,计算新的聚类中心,更新适应度,再一次比较适应度,若其优于个体最优位置pid,则更新pid;若其优于全局最优位置pgd,则更新pgd,如果达到最大迭代次数,则算法结束,否则继续迭代;决策树结构的生成需要先将全部训练样本集作为初始节点,利用粒子群聚 类算法,将其划分成两类,形成两个子节点,判断子节点是否只包含一类故障样本,若是则算法结束,否则继续采用粒子群聚类算法进行处理,划分成两个新的子节点,这样不断划分,直到所有子节点只包含一类故障样本,算法结束,这样将所有的故障样本进行聚类划分,就可以反向构建故障诊断决策树;第四步:训练和测试决策树支持向量机SVM故障分类模型按照4:1的比例将故障数据样本划分成训练集和测试集,训练集按照决策树结构上故障的分布训练支持向量机SVM分类模型;支持向量机SVM分类模型均采用径向基核函数,并各自优化支持向量机SVM的参数,训练完成后,利用测试集测试决策树支持向量机SVM故障诊断模型,得到诊断精度等指标,最终实现光伏二极管箝位型三电平逆变器的故障诊断。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US15/551,299 US10234495B2 (en) | 2016-09-07 | 2016-12-30 | Decision tree SVM fault diagnosis method of photovoltaic diode-clamped three-level inverter |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610805280.2A CN106443297B (zh) | 2016-09-07 | 2016-09-07 | 光伏二极管箝位型三电平逆变器的决策树svm故障诊断方法 |
| CN2016108052802 | 2016-09-07 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2018045689A1 true WO2018045689A1 (zh) | 2018-03-15 |
Family
ID=58164073
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2016/113645 Ceased WO2018045689A1 (zh) | 2016-09-07 | 2016-12-30 | 光伏二极管箝位型三电平逆变器的决策树svm故障诊断方法 |
Country Status (3)
| Country | Link |
|---|---|
| US (1) | US10234495B2 (zh) |
| CN (1) | CN106443297B (zh) |
| WO (1) | WO2018045689A1 (zh) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108872882A (zh) * | 2018-08-09 | 2018-11-23 | 西南交通大学 | 一种三电平级联逆变器的故障诊断装置及其诊断方法 |
| CN108875796A (zh) * | 2018-05-28 | 2018-11-23 | 福州大学 | 基于线性判别分析和支持向量机的光伏阵列故障诊断方法 |
| CN109684749A (zh) * | 2018-12-27 | 2019-04-26 | 合肥工业大学 | 一种考虑运行特性的光伏电站等值建模方法 |
| CN112183606A (zh) * | 2020-09-23 | 2021-01-05 | 内蒙古电力(集团)有限责任公司乌海电业局 | 基于c4.5算法的电力系统故障识别分类方法及系统 |
| CN112307997A (zh) * | 2020-11-06 | 2021-02-02 | 华北电力大学 | 一种利用主模式分解的功率信号重构方法和系统 |
| CN112801127A (zh) * | 2020-12-09 | 2021-05-14 | 西安华谱电力设备制造有限公司 | 基于振荡波的电缆局部放电缺陷识别方法 |
| CN113095173A (zh) * | 2021-03-29 | 2021-07-09 | 西安交通大学 | 一种Adaboost融合多分类器的光伏系统故障电弧检测方法 |
| CN113420832A (zh) * | 2021-07-16 | 2021-09-21 | 国网辽宁省电力有限公司 | 基于多维数据相似性匹配的电网故障诊断系统及方法 |
| CN113552443A (zh) * | 2021-06-08 | 2021-10-26 | 广西大学 | 基于交直流脉冲和随机森林的混连高压线路故障识别方法 |
| CN113687292A (zh) * | 2020-05-18 | 2021-11-23 | 宁夏隆基宁光仪表股份有限公司 | 大数据、云环境下电能表错误接线的检测系统及方法 |
| CN113702767A (zh) * | 2021-08-31 | 2021-11-26 | 中南大学 | 基于小波滑窗能量的孤岛直流微电网故障诊断方法 |
| CN114139635A (zh) * | 2021-12-03 | 2022-03-04 | 西安重装韩城煤矿机械有限公司 | 煤矿变频局部通风机工况监测与故障预警方法 |
| CN114676573A (zh) * | 2022-03-28 | 2022-06-28 | 合肥工业大学 | 三电平光伏逆变器开路故障统一数学建模方法 |
| CN114720834A (zh) * | 2022-02-27 | 2022-07-08 | 昆明理工大学 | 一种模块化多电平换流器igbt开路故障的识别方法 |
| CN114897103A (zh) * | 2022-06-13 | 2022-08-12 | 南京林业大学 | 一种基于近邻成分损失优化多尺度卷积神经网络的工业过程故障诊断方法 |
| CN115035336A (zh) * | 2022-06-08 | 2022-09-09 | 集美大学 | 基于gasf和cnn的三电平t型逆变器故障诊断方法 |
| CN115100296A (zh) * | 2022-07-07 | 2022-09-23 | 阳光电源股份有限公司 | 一种光伏组件故障定位方法、装置、设备及存储介质 |
| CN115267606A (zh) * | 2022-07-26 | 2022-11-01 | 合肥工业大学 | Anpc型三电平并网逆变器中igbt的开路故障诊断方法 |
| CN116167508A (zh) * | 2023-01-16 | 2023-05-26 | 南通电力设计院有限公司 | 一种基于气象因素分解的短期光伏出力快速预测方法及系统 |
| CN117272116A (zh) * | 2023-10-13 | 2023-12-22 | 西安工程大学 | 一种基于loras平衡数据集的变压器故障诊断方法 |
| CN117609871A (zh) * | 2023-11-09 | 2024-02-27 | 国网江苏省电力有限公司南京供电分公司 | 一种分布式光伏并网故障模型分析方法及系统 |
| CN118114134A (zh) * | 2024-03-19 | 2024-05-31 | 创维互联(北京)新能源科技有限公司 | 一种缺失遥信和辐照数据的逆变器停机归因分析方法 |
| CN118444123A (zh) * | 2024-07-08 | 2024-08-06 | 电子科技大学 | 基于杂散电感与集射电压的igbt故障诊断方法 |
| CN120074369A (zh) * | 2025-01-03 | 2025-05-30 | 中国长江电力股份有限公司 | 基于ALasso-IKOA-XGBoost的光伏系统故障诊断方法 |
| CN120217126A (zh) * | 2025-05-29 | 2025-06-27 | 宁波日月电线电缆制造有限公司 | 基于深度树学习的机器人电缆故障分类方法及系统 |
Families Citing this family (66)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109709411A (zh) * | 2017-10-25 | 2019-05-03 | 中国电力科学研究院有限公司 | 一种光伏逆变器故障诊断方法和系统 |
| CN107894564B (zh) * | 2017-11-09 | 2020-02-18 | 合肥工业大学 | 一种基于交叉小波特征的模拟电路故障诊断方法 |
| US11625628B2 (en) * | 2018-01-05 | 2023-04-11 | Daniel Suklja | Method of improving processing efficiency decision making within a computer system |
| CN108255783B (zh) * | 2018-01-12 | 2021-03-26 | 上海海事大学 | 一种基于小波分析的船舶电力驱动系统故障诊断方法 |
| JP6911785B2 (ja) * | 2018-02-02 | 2021-07-28 | 日本電信電話株式会社 | 判定装置、判定方法及び判定プログラム |
| CN108647716B (zh) * | 2018-05-09 | 2021-08-06 | 北京理工大学 | 一种基于复合信息的光伏阵列故障诊断方法 |
| US11415950B2 (en) * | 2018-08-24 | 2022-08-16 | The Board Of Trustees Of The University Of Alabama | Artificial neural network control system for autonomous volt/VAR control of inverter in interfaced units |
| CN109409420B (zh) * | 2018-10-08 | 2022-05-03 | 西安热工研究院有限公司 | 一种在非均匀辐照度下的光伏组串故障诊断方法 |
| CN109560770B (zh) * | 2018-10-29 | 2020-03-13 | 西安交通大学 | 基于自适应核函数和瞬时频率估计的光伏系统故障电弧检测方法 |
| CN109462231B (zh) * | 2018-11-15 | 2020-09-01 | 合肥工业大学 | 居民微电网的负荷优化调度方法、系统和存储介质 |
| CN109389104A (zh) * | 2018-11-30 | 2019-02-26 | 浙江碳银互联网科技有限公司 | 一种户用光伏电站逆变器故障预测方法 |
| CN109635880B (zh) * | 2019-01-08 | 2023-06-27 | 浙江大学 | 基于强鲁棒自适应算法的采煤机故障诊断系统 |
| CN109784279B (zh) * | 2019-01-17 | 2023-04-18 | 西安电子科技大学 | 基于多小波分析及svm的x射线高压电源故障诊断方法 |
| CN110058110A (zh) * | 2019-04-16 | 2019-07-26 | 重庆大学 | 一种有源逆变器间歇故障诊断方法 |
| CN110084316B (zh) * | 2019-05-06 | 2023-04-07 | 安徽工业大学 | 一种基于精细时移多尺度排列熵与萤火虫算法优化支持向量机的故障诊断方法 |
| CN110133538B (zh) * | 2019-05-16 | 2021-03-12 | 合肥工业大学 | 一种anpc三电平逆变器开路故障诊断方法及实验平台 |
| CN110322048B (zh) * | 2019-05-31 | 2023-09-26 | 南京航空航天大学 | 一种生产物流输送装备故障预警方法 |
| CN110135121B (zh) * | 2019-06-13 | 2023-04-18 | 中国人民解放军海军航空大学 | 一种基于拉格朗日-粒子群更新算法的故障诊断方法 |
| CN110208721B (zh) * | 2019-07-09 | 2024-06-11 | 西南交通大学 | 一种级联三电平逆变器的故障诊断方法及装置 |
| CN110333426A (zh) * | 2019-07-26 | 2019-10-15 | 沈阳工业大学 | 一种模块化多电平储能系统开路故障诊断装置及方法 |
| CN110346736B (zh) * | 2019-08-14 | 2021-07-02 | 合肥工业大学 | 基于改进小树变换的npc三电平逆变器故障诊断方法 |
| CN110716820A (zh) * | 2019-10-10 | 2020-01-21 | 厦门钛尚人工智能科技有限公司 | 一种基于决策树算法的故障诊断方法 |
| CN111798405B (zh) * | 2019-11-21 | 2024-04-19 | 南京航空航天大学 | 一种基于深度学习的套管红外图像的故障诊断方法 |
| CN112906727B (zh) * | 2019-12-04 | 2024-08-30 | 天翼云科技有限公司 | 用于实时在线检测虚拟机状态的方法和系统 |
| CN111160712B (zh) * | 2019-12-06 | 2023-09-22 | 北京国电通网络技术有限公司 | 一种用户的用电参数调节方法及装置 |
| CN111291783A (zh) * | 2020-01-15 | 2020-06-16 | 北京市燃气集团有限责任公司 | 一种燃气调压设备的智能化故障诊断方法、系统、终端及存储介质 |
| TWI708470B (zh) * | 2020-02-21 | 2020-10-21 | 龍華科技大學 | 一種電源轉換器 |
| CN111490525A (zh) * | 2020-03-19 | 2020-08-04 | 国网江苏省电力有限公司盐城供电分公司 | 一种三相光伏逆变器缺相检测及容错控制方法 |
| CN111983517B (zh) * | 2020-03-23 | 2023-07-04 | 河南理工大学 | 一种利用粗糙集贪心算法的逆变器开路故障诊断方法 |
| CN111652479B (zh) * | 2020-05-19 | 2023-06-02 | 三峡大学 | 一种针对电力系统动态安全评估的数据驱动方法 |
| CN112085060A (zh) * | 2020-08-07 | 2020-12-15 | 中国民航大学 | 基于SVT-DTSVMs的双偏振气象雷达降水粒子分类方法及装置 |
| CN114077858B (zh) | 2020-08-17 | 2025-09-09 | 浙江宇视科技有限公司 | 向量数据处理方法、装置、设备及存储介质 |
| JP7546224B2 (ja) * | 2020-08-26 | 2024-09-06 | パナソニックIpマネジメント株式会社 | 異常検出装置、異常検出方法及びプログラム |
| CN112632845B (zh) * | 2020-10-23 | 2022-10-25 | 西安交通大学 | 基于数据的小型反应堆在线故障诊断方法、介质及设备 |
| CN112327206B (zh) * | 2020-10-27 | 2023-06-27 | 河南理工大学 | 一种用于三电平逆变器的故障诊断方法 |
| CN112327999B (zh) * | 2020-11-02 | 2022-03-11 | 东南大学 | 基于最大功率点跟踪数据的光伏快速检测及精确诊断方法 |
| CN112733878A (zh) * | 2020-12-08 | 2021-04-30 | 国网辽宁省电力有限公司锦州供电公司 | 一种基于kmeans-SVM算法的变压器故障诊断方法 |
| CN112633345A (zh) * | 2020-12-17 | 2021-04-09 | 鄂尔多斯应用技术学院 | 一种基于多视图方法的采煤机数字字典的建立方法 |
| CN112731019B (zh) * | 2020-12-21 | 2022-10-14 | 合肥工业大学 | 一种anpc三电平逆变器故障诊断方法 |
| CN112560214B (zh) * | 2020-12-24 | 2024-09-06 | 国网北京市电力公司 | 变电站母线平衡故障诊断方法、系统、设备及存储介质 |
| CN112884198B (zh) * | 2021-01-13 | 2023-06-09 | 西安理工大学 | 结合门限回归和改进支持向量机面板坝坝顶沉降预测方法 |
| CN113325288B (zh) * | 2021-05-18 | 2023-08-29 | 长安大学 | 一种基于电流注入法的npc三电平逆变器故障上电自检方法 |
| US12074820B2 (en) * | 2021-09-24 | 2024-08-27 | Huawei Technologies Co., Ltd. | Protocol and frame format for coordinated beamforming |
| CN113866675B (zh) * | 2021-09-26 | 2023-05-26 | 河南理工大学 | 一种h桥逆变器短路故障推理方法 |
| CN113765336B (zh) * | 2021-11-10 | 2022-02-22 | 青岛鼎信通讯股份有限公司 | 一种应用于svg的i型三电平逆变器开关管自检方法 |
| CN114089181B (zh) * | 2021-11-19 | 2022-11-08 | 浙江大学 | 基于XG-Boost的MMC开关管开路故障检测方法 |
| CN114781179B (zh) * | 2022-05-12 | 2023-03-28 | 广东华矩检测技术有限公司 | 基于光纤通信信息采集光伏电站发电量损失核定方法 |
| CN115097752B (zh) * | 2022-06-15 | 2026-01-30 | 华东理工大学 | 一种基于fpga的多通道采集与控制系统 |
| CN115184839A (zh) * | 2022-07-26 | 2022-10-14 | 合肥工业大学 | 一种级联h桥并网逆变器系统的开路故障功率管定位方法 |
| CN114997745B (zh) * | 2022-07-28 | 2022-10-04 | 华能国际电力江苏能源开发有限公司清洁能源分公司 | 一种基于深度特征提取的光伏故障诊断溯源方法 |
| CN115270965B (zh) * | 2022-07-29 | 2023-05-12 | 国网浙江省电力有限公司经济技术研究院 | 一种配电网线路故障预测方法和装置 |
| CN115329815B (zh) * | 2022-08-11 | 2025-10-31 | 合肥工业大学 | 气泡熵和aoa优化svm模型参数的轴承故障诊断方法及系统 |
| CN115166517B (zh) * | 2022-09-07 | 2022-11-25 | 沈阳科来沃电气技术有限公司 | 一种针对电机传动系统中逆变器故障诊断方法 |
| CN115577308B (zh) * | 2022-10-27 | 2025-11-18 | 长春工业大学 | 一种故障特征选择的蛇优化支持向量机的轴承故障分类方法 |
| CN115828164A (zh) * | 2022-12-29 | 2023-03-21 | 浙江工业大学台州研究院 | 一种基于数据驱动的电动钉枪故障类型识别方法 |
| CN116050318B (zh) * | 2022-12-30 | 2025-09-19 | 中车永济电机有限公司 | 一种基于fpga的三电平anpc桥臂igbt断路故障建模方法 |
| CN116541772B (zh) * | 2023-05-05 | 2023-10-10 | 兰州理工大学 | 基于多源融合残差网络的级联h桥逆变器故障诊断方法 |
| CN117251788B (zh) * | 2023-09-19 | 2026-04-24 | 阳光智维科技股份有限公司 | 状态评估方法、装置、终端设备及存储介质 |
| CN118228006B (zh) * | 2024-05-23 | 2024-08-02 | 大连卓志创芯科技有限公司 | 一种基于fpga技术的芯片检测方法及系统 |
| CN118656955B (zh) * | 2024-06-12 | 2025-04-18 | 国家电网有限公司华东分部 | 故障电流计算方法、系统、电子设备及可读存储介质 |
| CN119093870B (zh) * | 2024-09-02 | 2025-05-02 | 华能国际电力江苏能源开发有限公司清洁能源分公司 | 一种光伏逆变器智能组串诊断方法和系统 |
| CN119147874B (zh) * | 2024-11-14 | 2025-06-06 | 南通康比电子有限公司 | 一种单相整流器的信号故障检测方法 |
| CN119167665B (zh) * | 2024-11-21 | 2025-07-18 | 南通康比电子有限公司 | 基于孪生模拟的光伏二极管制造参数优化方法 |
| CN120086702B (zh) * | 2025-05-06 | 2025-07-15 | 太原科技大学 | 一种气固流化床床层状态识别方法及系统 |
| CN120103221B (zh) * | 2025-05-12 | 2025-08-12 | 东方电子股份有限公司 | 一种三电平双向直流变换器功率器件故障诊断方法 |
| CN120763724B (zh) * | 2025-09-09 | 2025-11-11 | 华能国际电力江苏能源开发有限公司清洁能源分公司 | 一种基于svm算法的开关管故障分析构造系统及方法 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102704054A (zh) * | 2012-05-29 | 2012-10-03 | 河海大学常州校区 | 基于时间序列挖掘的梳棉机智能匀整系统 |
| CN102937688A (zh) * | 2012-10-30 | 2013-02-20 | 浙江万里学院 | 二极管中点钳位三电平逆变器的器件开路故障诊断电路 |
| CN104698397A (zh) * | 2015-03-16 | 2015-06-10 | 浙江万里学院 | 一种多电平逆变器故障诊断方法 |
| KR20150130594A (ko) * | 2014-05-13 | 2015-11-24 | 아주대학교산학협력단 | 누설전류로 인한 스위치 파괴를 방지한 3-레벨 npc 인버터 |
-
2016
- 2016-09-07 CN CN201610805280.2A patent/CN106443297B/zh not_active Expired - Fee Related
- 2016-12-30 US US15/551,299 patent/US10234495B2/en not_active Expired - Fee Related
- 2016-12-30 WO PCT/CN2016/113645 patent/WO2018045689A1/zh not_active Ceased
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102704054A (zh) * | 2012-05-29 | 2012-10-03 | 河海大学常州校区 | 基于时间序列挖掘的梳棉机智能匀整系统 |
| CN102937688A (zh) * | 2012-10-30 | 2013-02-20 | 浙江万里学院 | 二极管中点钳位三电平逆变器的器件开路故障诊断电路 |
| KR20150130594A (ko) * | 2014-05-13 | 2015-11-24 | 아주대학교산학협력단 | 누설전류로 인한 스위치 파괴를 방지한 3-레벨 npc 인버터 |
| CN104698397A (zh) * | 2015-03-16 | 2015-06-10 | 浙江万里学院 | 一种多电平逆变器故障诊断方法 |
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108875796A (zh) * | 2018-05-28 | 2018-11-23 | 福州大学 | 基于线性判别分析和支持向量机的光伏阵列故障诊断方法 |
| CN108872882B (zh) * | 2018-08-09 | 2023-09-19 | 西南交通大学 | 一种三电平级联逆变器的故障诊断装置及其诊断方法 |
| CN108872882A (zh) * | 2018-08-09 | 2018-11-23 | 西南交通大学 | 一种三电平级联逆变器的故障诊断装置及其诊断方法 |
| CN109684749A (zh) * | 2018-12-27 | 2019-04-26 | 合肥工业大学 | 一种考虑运行特性的光伏电站等值建模方法 |
| CN109684749B (zh) * | 2018-12-27 | 2023-03-24 | 合肥工业大学 | 一种考虑运行特性的光伏电站等值建模方法 |
| CN113687292A (zh) * | 2020-05-18 | 2021-11-23 | 宁夏隆基宁光仪表股份有限公司 | 大数据、云环境下电能表错误接线的检测系统及方法 |
| CN113687292B (zh) * | 2020-05-18 | 2024-02-02 | 宁夏隆基宁光仪表股份有限公司 | 大数据、云环境下电能表错误接线的检测系统及方法 |
| CN112183606A (zh) * | 2020-09-23 | 2021-01-05 | 内蒙古电力(集团)有限责任公司乌海电业局 | 基于c4.5算法的电力系统故障识别分类方法及系统 |
| CN112307997A (zh) * | 2020-11-06 | 2021-02-02 | 华北电力大学 | 一种利用主模式分解的功率信号重构方法和系统 |
| CN112801127A (zh) * | 2020-12-09 | 2021-05-14 | 西安华谱电力设备制造有限公司 | 基于振荡波的电缆局部放电缺陷识别方法 |
| CN113095173B (zh) * | 2021-03-29 | 2024-04-02 | 西安交通大学 | 一种Adaboost融合多分类器的光伏系统故障电弧检测方法 |
| CN113095173A (zh) * | 2021-03-29 | 2021-07-09 | 西安交通大学 | 一种Adaboost融合多分类器的光伏系统故障电弧检测方法 |
| CN113552443A (zh) * | 2021-06-08 | 2021-10-26 | 广西大学 | 基于交直流脉冲和随机森林的混连高压线路故障识别方法 |
| CN113552443B (zh) * | 2021-06-08 | 2022-12-02 | 广西大学 | 基于交直流脉冲和随机森林的混连高压线路故障识别方法 |
| CN113420832B (zh) * | 2021-07-16 | 2024-02-02 | 国网辽宁省电力有限公司 | 基于多维数据相似性匹配的电网故障诊断系统及方法 |
| CN113420832A (zh) * | 2021-07-16 | 2021-09-21 | 国网辽宁省电力有限公司 | 基于多维数据相似性匹配的电网故障诊断系统及方法 |
| CN113702767A (zh) * | 2021-08-31 | 2021-11-26 | 中南大学 | 基于小波滑窗能量的孤岛直流微电网故障诊断方法 |
| CN114139635A (zh) * | 2021-12-03 | 2022-03-04 | 西安重装韩城煤矿机械有限公司 | 煤矿变频局部通风机工况监测与故障预警方法 |
| CN114720834A (zh) * | 2022-02-27 | 2022-07-08 | 昆明理工大学 | 一种模块化多电平换流器igbt开路故障的识别方法 |
| CN114676573A (zh) * | 2022-03-28 | 2022-06-28 | 合肥工业大学 | 三电平光伏逆变器开路故障统一数学建模方法 |
| CN115035336A (zh) * | 2022-06-08 | 2022-09-09 | 集美大学 | 基于gasf和cnn的三电平t型逆变器故障诊断方法 |
| CN114897103A (zh) * | 2022-06-13 | 2022-08-12 | 南京林业大学 | 一种基于近邻成分损失优化多尺度卷积神经网络的工业过程故障诊断方法 |
| CN115100296A (zh) * | 2022-07-07 | 2022-09-23 | 阳光电源股份有限公司 | 一种光伏组件故障定位方法、装置、设备及存储介质 |
| CN115267606A (zh) * | 2022-07-26 | 2022-11-01 | 合肥工业大学 | Anpc型三电平并网逆变器中igbt的开路故障诊断方法 |
| CN115267606B (zh) * | 2022-07-26 | 2024-04-05 | 合肥工业大学 | Anpc型三电平并网逆变器中igbt的开路故障诊断方法 |
| CN116167508A (zh) * | 2023-01-16 | 2023-05-26 | 南通电力设计院有限公司 | 一种基于气象因素分解的短期光伏出力快速预测方法及系统 |
| CN116167508B (zh) * | 2023-01-16 | 2024-03-15 | 南通电力设计院有限公司 | 一种基于气象因素分解的短期光伏出力快速预测方法及系统 |
| CN117272116A (zh) * | 2023-10-13 | 2023-12-22 | 西安工程大学 | 一种基于loras平衡数据集的变压器故障诊断方法 |
| CN117272116B (zh) * | 2023-10-13 | 2024-05-17 | 西安工程大学 | 一种基于loras平衡数据集的变压器故障诊断方法 |
| CN117609871A (zh) * | 2023-11-09 | 2024-02-27 | 国网江苏省电力有限公司南京供电分公司 | 一种分布式光伏并网故障模型分析方法及系统 |
| CN118114134A (zh) * | 2024-03-19 | 2024-05-31 | 创维互联(北京)新能源科技有限公司 | 一种缺失遥信和辐照数据的逆变器停机归因分析方法 |
| CN118444123A (zh) * | 2024-07-08 | 2024-08-06 | 电子科技大学 | 基于杂散电感与集射电压的igbt故障诊断方法 |
| CN120074369A (zh) * | 2025-01-03 | 2025-05-30 | 中国长江电力股份有限公司 | 基于ALasso-IKOA-XGBoost的光伏系统故障诊断方法 |
| CN120217126A (zh) * | 2025-05-29 | 2025-06-27 | 宁波日月电线电缆制造有限公司 | 基于深度树学习的机器人电缆故障分类方法及系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106443297B (zh) | 2018-05-22 |
| CN106443297A (zh) | 2017-02-22 |
| US10234495B2 (en) | 2019-03-19 |
| US20180238951A1 (en) | 2018-08-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106443297B (zh) | 光伏二极管箝位型三电平逆变器的决策树svm故障诊断方法 | |
| Kou et al. | Fault diagnosis for open‐circuit faults in NPC inverter based on knowledge‐driven and data‐driven approaches | |
| WO2018120077A1 (zh) | 一种基于经验模态分解和决策树rvm的三电平逆变器故障诊断方法 | |
| CN103761372B (zh) | 一种基于主元分析与多分类相关向量机的多电平逆变器故障诊断策略 | |
| Si et al. | Fault diagnosis based on attention collaborative LSTM networks for NPC three-level inverters | |
| CN110084106A (zh) | 基于小波变换和概率神经网络的微网逆变器故障诊断方法 | |
| Shi et al. | Open switch fault diagnosis method for PWM voltage source rectifier based on deep learning approach | |
| CN110133538A (zh) | 一种anpc三电平逆变器开路故障诊断方法及实验平台 | |
| CN106483405B (zh) | 基于隐马尔可夫模型的npc光伏逆变器的故障诊断方法 | |
| Deng et al. | A high-accuracy-light-AI data-driven diagnosis method for open-circuit faults in single-phase PWM rectifiers | |
| CN105095566A (zh) | 一种基于小波分析和svm的逆变器故障诊断方法 | |
| CN110619321A (zh) | 一种基于深度神经网络的充电桩技术状况诊断方法及系统 | |
| CN116484299B (zh) | 基于梯度提升树与多层感知机融合的充电桩故障诊断方法 | |
| Cui et al. | T-type inverter fault diagnosis based on GASF and improved AlexNet | |
| CN107025341A (zh) | 一种光伏逆变器故障诊断方法 | |
| Yang et al. | Multiple IGBT open-circuit fault diagnosis strategy for MMC using feature reconstruction-based recurrent neural network | |
| Yuan et al. | A secondary classification fault diagnosis strategy based on PCA-SVM for cascaded photovoltaic grid-connected inverter | |
| CN112213665A (zh) | 一种z源逆变器功率管的开路故障诊断方法、系统及应用 | |
| CN115310604A (zh) | 一种基于神经网络的ups系统在线故障诊断方法 | |
| Long et al. | Fault diagnosis for IGBTs open-circuit faults in photovoltaic grid-connected inverters based on statistical analysis and machine learning | |
| CN110968073B (zh) | 一种hvdc系统换相失败故障原因双层溯源辨识方法 | |
| Xin et al. | A deep learning and softmax regression fault diagnosis method for multi-level converter | |
| Fu et al. | Fault diagnosis of power electronic device based on wavelet and neural network | |
| Gong et al. | Fault diagnosis of micro grid inverter based on wavelet transform and probabilistic neural network | |
| Wang et al. | A data-driven open-circuit fault diagnosis method for three-phase T-type three-level inverters |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| WWE | Wipo information: entry into national phase |
Ref document number: 15551299 Country of ref document: US |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16915623 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 16915623 Country of ref document: EP Kind code of ref document: A1 |
