CN-122001090-A - Remote intelligent operation and maintenance monitoring method and system for new energy power station

CN122001090ACN 122001090 ACN122001090 ACN 122001090ACN-122001090-A

Abstract

The application provides a remote intelligent operation and maintenance monitoring method and system for a new energy power station, which relate to the technical field of operation and maintenance monitoring and comprise the steps of constructing photoelectric hysteresis loop area density and topological thermoelectric resonance deviation according to a body response time-space alignment data set; the method comprises the steps of constructing a synchronous matching degree aging phase space, taking the area density of a photoelectric hysteresis loop and the topological thermoelectric resonance deviation degree as coordinate axes, dividing the synchronous matching degree aging phase space into a plurality of mutual exclusion state subareas and determining the mutual exclusion state subareas to which the state points belong according to judging conditions by calculating track curvature and divergence coefficients of the state points in the synchronous matching degree aging phase space so as to generate an aging degradation index, generating a risk conduction score based on the aging degradation index so as to identify a key risk conduction link, and backtracking a fault source according to the risk conduction score and generating a remote monitoring diagnosis report. Thoroughly solves the defects of fuzzy fault location and single operation and maintenance strategy in the prior art.

Inventors

CHEN WEIQI
YI KANGYU
CHEN ZHENYU
SONG ZIYI
ZHANG YUHONG
Zeng Penggao
SUN PEIPEI
Tan Enlai
TAN FUQIANG
XIAO ZHONGHUI

Assignees

湘能楚天电力装备股份有限公司

Dates

Publication Date: 20260508
Application Date: 20260407

Claims (9)

1. The remote intelligent operation and maintenance monitoring method for the new energy power station is characterized by comprising the following steps of: step 1, collecting multi-source heterogeneous data of each power generation unit in a continuous time window; step 2, preprocessing based on the multi-source heterogeneous data to obtain a body response space-time alignment data set; Step 3, constructing photoelectric hysteresis loop area density and topological thermoelectric resonance deviation according to the body response time-space alignment data set; step 4, constructing a synchronous matching degree aging phase space, taking the area density of the photoelectric hysteresis loop and the topological thermoelectric resonance deviation degree as coordinate axes, and dividing the synchronous matching degree aging phase space into a plurality of mutual exclusion state subareas and determining the mutual exclusion state subareas to which the state points belong according to judgment conditions by calculating the track curvature and the divergence coefficient of the state points in the synchronous matching degree aging phase space so as to generate an aging degradation index; step 5, generating a risk conduction score based on the aging degradation index to identify a key risk conduction link; and 6, backtracking the fault source according to the risk conduction score and generating a remote monitoring diagnosis report.
2. The method for monitoring the remote intelligent operation and maintenance of the new energy power station according to claim 1, wherein the multi-source heterogeneous data comprises an electroluminescence image sequence, an infrared thermal imaging temperature field sequence and direct current voltage and current time sequence data.
3. The remote intelligent operation and maintenance monitoring method for new energy power stations according to claim 2, wherein the construction of the photoelectric hysteresis loop area density according to the body response time-space alignment data set comprises the following steps: Step 31, extracting a body damage evolution intensity sequence and a loop response efficiency sequence of each power generation unit in a time window based on the body response time-space alignment data set, wherein the body damage evolution intensity sequence comprises a plurality of body damage evolution intensity values, the loop response efficiency sequence comprises a plurality of loop response efficiency values, and the body damage evolution intensity values and the loop response efficiency values form data points; Step 32, constructing a first plane rectangular coordinate system, wherein a first abscissa of the first plane rectangular coordinate system is a body damage evolution intensity sequence, and a first ordinate of the first plane rectangular coordinate system is the loop response efficiency sequence; step 33, based on each sampling time in the time window, taking the evolution intensity value of the body damage corresponding to the current sampling time as a first abscissa value, taking the response efficiency value of the loop corresponding to the current sampling time as a first ordinate value, and constructing a state point according to the first abscissa value and the first ordinate value; And step 34, sequentially connecting the state points at all sampling moments in a time window according to time sequence to form a track loop, calculating the area of a geometric area surrounded by the track loop in the first plane rectangular coordinate system, and obtaining the area density of the photoelectric hysteresis loop according to the area of the geometric area.
4. The method for remote intelligent operation and maintenance monitoring for new energy power stations according to claim 3, wherein the topology thermoelectric resonance deviation is constructed according to the body response time-space alignment data set, comprising the following steps: Step 35, extracting a direct neighbor unit based on the current power generation unit; Step 36, performing trending processing based on the infrared thermal imaging temperature field sequence to obtain a zero-mean temperature fluctuation sequence, performing fast Fourier transform on the zero-mean temperature fluctuation sequence to obtain a frequency domain complex sequence, calculating power spectrum density according to the frequency domain complex sequence, and defining the frequency corresponding to the maximum value of the power spectrum density as a dominant frequency component; step 37, obtaining a phase angle at the dominant frequency component, calculating the absolute value of the phase difference of the current power generation unit and each direct neighbor unit on the dominant frequency component according to the phase angle, and obtaining an average phase mismatch angle according to the absolute value of the phase difference; Step 38, calculating the ratio of the characteristic temperature fluctuation energy of the current power generation unit to the characteristic temperature fluctuation energy of the direct neighbor unit to obtain an energy ratio; And step 39, obtaining the topological thermoelectric resonance deviation degree according to the average phase mismatch angle and the energy ratio.
5. The remote intelligent operation and maintenance monitoring method for new energy power stations according to claim 4, wherein a synchronous matching degree aging phase space is constructed, the photoelectric hysteresis loop area density and the topological thermoelectric resonance deviation degree are taken as coordinate axes, the synchronous matching degree aging phase space is divided into a plurality of mutually exclusive state subregions according to judgment conditions by calculating track curvature and divergence coefficients of state points in the synchronous matching degree aging phase space, and the mutually exclusive state subregions to which the state points belong are determined to generate an aging degradation index, and the method comprises the following steps: Step 41, constructing a second plane rectangular coordinate system as a synchronous matching degree aging phase space, wherein a second abscissa of the second plane rectangular coordinate system is photoelectric hysteresis loop area density, and a second ordinate of the second plane rectangular coordinate system is topological thermoelectric resonance deviation degree; step 42, mapping a second abscissa value and a second ordinate value of the current power generation unit into phase space state points, backtracking the phase space state points at a history preset moment to form a track sequence, and calculating track curvature of each phase space state point in the track sequence; Step 43, determining an instantaneous step length according to the adjacent phase space state points of the track sequence, determining a step length growth factor according to the instantaneous step length, and generating a divergence coefficient according to the step length growth factor; Step 44, presetting a target threshold value, namely a first threshold value, a second threshold value, a curvature critical value and a divergence critical value, and dividing the synchronous matching degree aging phase space into three mutually exclusive state sub-areas, namely a first state sub-area, a second state sub-area and a third state sub-area according to the target threshold value, a second abscissa value of a second plane rectangular coordinate system, a second ordinate value of the second plane rectangular coordinate system, a divergence coefficient and track curvature; And 45, generating an aging degradation index according to the second abscissa value and the second ordinate value of the current phase space state point when the phase space state point is in the first state subarea, generating the aging degradation index according to the second abscissa value, the second ordinate value and the track curvature of the current phase space state point when the phase space state point is in the second state subarea, and generating the aging degradation index according to the second abscissa value, the second ordinate value and the divergence coefficient of the current phase space state point when the phase space state point is in the third state subarea.
6. The new energy power station oriented remote intelligent operation and maintenance monitoring method according to claim 5, wherein when the second abscissa value of the phase space state point is smaller than the first threshold value and the second ordinate value is smaller than the second threshold value, it is determined that the phase space state point is in the first state sub-area; When the second abscissa value of the phase space state point is larger than or equal to the first threshold value or the second ordinate value is larger than or equal to the second threshold value, and the curvature of the track is smaller than the curvature critical value and the divergence coefficient is smaller than the divergence critical value, judging that the phase space state point is in the second state subarea; And when the instantaneous curvature of the phase space state point is greater than or equal to a curvature critical value or a divergence coefficient is greater than or equal to a divergence critical value, judging that the phase space state point is in a third state subarea.
7. The new energy power station oriented remote intelligent operation and maintenance monitoring method of claim 6, wherein generating risk conduction scores based on the aging degradation index to identify critical risk conduction links comprises the steps of: Step 51, a topological graph is constructed by taking power generation units as nodes and electric connection between power generation nodes as edges, the aging degradation index of each node is taken as a node attribute, and the weight of the edges is the absolute value of the difference value of the aging degradation indexes between adjacent nodes; Step 52, constructing a communication path according to the nodes and edges, obtaining an instantaneous aging increment based on the aging degradation indexes of the nodes of the communication path at the current moment and the last moment, and obtaining an aging increment accumulated value according to the instantaneous aging increment; And step 53, obtaining a weight accumulated value according to the weight of the edge of the communication path, generating a risk conduction score according to the weight accumulated value and the aging increment accumulated value, and defining the communication path corresponding to the largest risk conduction score as a key risk conduction link.
8. The new energy power station oriented remote intelligent operation and maintenance monitoring method according to claim 7, wherein tracing back the fault source and generating a remote monitoring diagnosis report according to the risk conduction score comprises the following steps: Step 61, identifying a target abnormal node based on the aging degradation index; Step 62, based on the communication paths, screening all first paths taking the target abnormal node as an end point, and identifying a key conduction path set in the first paths according to the risk conduction score; Step 63, traversing all nodes in turn according to the direction from the end point to the start point based on the key conduction path, and obtaining the initial growth moment of the aging degradation index of each node; step 64, if each node of the key conduction path meets the first condition and the second condition, taking the node as a candidate root node; Step 65, counting the selected accumulated times of each candidate root node, if the accumulated times are larger than a preset frequency threshold, taking the candidate root node with the largest accumulated times as a fault root node, otherwise, selecting the candidate root node with the largest current aging degradation index value from all the candidate root nodes as the fault root node; Step 66, acquiring a first contribution rate, a second contribution rate, a third contribution rate and a fourth contribution rate according to the fault source node and the target abnormal node; Step 67, if the first contribution rate is greater than the zero point five and the third contribution rate is greater than the zero point five, determining that the mismatch type of the target abnormal node is homologous time lag type mismatch, and generating a first maintenance suggestion instruction; If the second contribution rate is greater than the zero point five and the fourth contribution rate is greater than the zero point five, determining that the mismatch type of the target abnormal node is homologous space collaborative type mismatch, and generating a second maintenance suggestion instruction; Otherwise, judging the mismatch type of the target abnormal node as evolution compound mismatch, and generating a third maintenance suggestion instruction.
9. A remote intelligent operation and maintenance monitoring system for a new energy power station, configured to execute the remote intelligent operation and maintenance monitoring method for a new energy power station according to any one of claims 1 to 7, and the remote intelligent operation and maintenance monitoring system is characterized by comprising the following modules: The acquisition module is used for acquiring multi-source heterogeneous data of each power generation unit in a continuous time window; the preprocessing module is used for preprocessing based on the multi-source heterogeneous data to obtain a body response space-time alignment data set; The data processing module is used for constructing photoelectric hysteresis loop area density and topological thermoelectric resonance deviation according to the body response time-space alignment data set; The data analysis module is used for constructing a synchronous matching degree aging phase space, taking the area density of the photoelectric hysteresis loop and the topological thermoelectric resonance deviation degree as coordinate axes, dividing the synchronous matching degree aging phase space into a plurality of mutual exclusion state subregions and determining the mutual exclusion state subregions to which the state points belong according to judgment conditions by calculating the track curvature and the divergence coefficient of the state points in the synchronous matching degree aging phase space so as to generate an aging degradation index; an identification module for generating a risk conduction score based on the aging degradation index to identify a critical risk conduction link; and the diagnosis module is used for backtracking the fault source according to the risk conduction score and generating a remote monitoring diagnosis report.

Description

Remote intelligent operation and maintenance monitoring method and system for new energy power station Technical Field The application relates to the technical field of operation and maintenance monitoring, in particular to a remote intelligent operation and maintenance monitoring method and system for a new energy power station. Background Along with the continuous expansion of the installation scale of the new energy power station, the demands for the refinement and the intellectualization of the operation and maintenance management of the new energy power station are increasingly urgent. The new energy power generation unit (such as a photovoltaic power generation unit) refers to a minimum functional module which has independent grid connection capability and can be independently metered and regulated by a monitoring system, and is generally formed by collecting, inverting and boosting hundreds to thousands of photovoltaic modules (namely solar panels) in series-parallel connection, wherein the photovoltaic modules are the underlying basic hardware for converting light energy into electric energy, and the power generation unit is a basic operation object for power station operation management, power control and fault isolation. At present, fault diagnosis and health state evaluation of the power generation unit mainly depend on two main technical means of electric parameter monitoring and image detection. The electrical parameter monitoring (such as direct current voltage, current, power and the like) can reflect macroscopic electrical output performance of the power generation unit in real time, but often has hysteresis, namely when the electrical parameter is remarkably abnormal, internal physical damage is usually developed to a serious stage, while the image detection technology (such as Electroluminescence (EL) imaging and infrared thermal Imaging (IR)) can visually present microscopic physical defects (such as hidden cracks, hot spots, bypass diode faults and the like) in the assembly, has early discovery capability, is usually off-line or low-frequency sampling, and is difficult to realize accurate synchronization with high-frequency electrical data on a time axis. In the prior art, when fusion analysis is carried out on the multi-source data, the defects of space-time splitting and associated shallow layer formation are common. On one hand, the traditional method is used for analyzing microscopic physical defects visible to images and macroscopic electrical output cutting and cracking visible to electrical parameters, or is used for carrying out qualitative judgment based on image characteristics or trend prediction based on electrical time sequence data, and lacks of coupling mechanism research for placing the two under a unified space-time frame, on the other hand, even if some of the prior art tries to carry out data fusion, a simple linear regression, weighted average or threshold value comparison equivalent static statistical method is adopted, and an attempt is made to establish a direct mapping relation between a physical damage area and electrical performance loss. However, the aging failure of the new energy power generation unit is a complex nonlinear dynamic process, and the evolution of microscopic physical damage (such as crack propagation and slow increase of contact resistance) and the change of macroscopic electrical response are not simply linear correspondence, but have obvious time-space inconsistency. The non-uniformity is particularly expressed as two typical working conditions, namely, a latency phenomenon, namely, obvious microscopic physical damage (visible images) occurs in the interior of the component, but due to redundant design or environmental factor compensation, no statistically significant decline of macroscopic electrical output occurs yet, so that a monitoring means based on pure electrical data is not reported, an optimal maintenance window is missed, and false positive fluctuation, namely, severe fluctuation occurs in electrical performance due to the influence of cloud shielding, temperature transient and other environmental noise, but the physical image of the component is normal, so that the monitoring means based on a pure electrical threshold value is misreported, and unnecessary operation and maintenance cost is caused. The root cause of the defects is that the prior art fails to construct a dynamic analysis method capable of simultaneously accommodating the evolution rate of microscopic damage and the response efficiency of a macroscopic loop, and cannot capture the hysteresis effect and the resonance deviation characteristics of a physical field and an electric field in the space-time dimension, so that the nonlinear conduction path from microscopic damage accumulation to macroscopic performance degradation is difficult to quantitatively describe. Due to the lack of deep description of the dynamic evolution process, when facing complex and changeable e