CN111146865A - Intelligent monitoring system for operation and maintenance state of power equipment - Google Patents

Abstract

The invention relates to an intelligent monitoring system for the operation and maintenance state of electric power equipment, which comprises a signal sensing device, a data acquisition device and an information operation diagnosis device which are sequentially connected, wherein the signal sensing device comprises distributed optical fiber sensors which are distributed and installed on the electric power equipment to be tested, and the information operation diagnosis device stores programs for executing the following steps: sending out a data reading trigger signal at fixed time; on the basis of the acquired data, performing prediction processing by using a constructed prediction model based on a linear autoregressive method to realize data flow optimization and obtain final sampling data of the power equipment to be detected; and obtaining a monitoring result of the power equipment to be tested based on the final sampling data. Compared with the prior art, the invention has the advantages of reducing the defect occurrence rate of the optical fiber detection system, improving the working efficiency, ensuring the safe and efficient operation of the power grid and the like.

Description

Intelligent monitoring system for operation and maintenance state of power equipment

Technical Field

The invention relates to a laser signal perception, optimization processing and diagnosis decision technology of power equipment, in particular to an intelligent monitoring system for operation and maintenance states of the power equipment.

Background

The power substation is an important carrier of power transmission, and with the rapid development of a power supply network and the reduction of a power supply radius, safe operation and accidents of the power substation occur occasionally, wherein one important reason is the lack of online monitoring on operation and maintenance of the power substation. For example, in the traditional operation and maintenance of the power transformer substation, a point type monitoring device such as a vibration sensor is installed at the position of the power transformer substation for vibration monitoring, the method is more economical and practical in some occasions such as the condition that transformer substation equipment is few, but the method is not enough to reflect the vibration condition of the local position of the power transformer substation, but cannot realize the on-line monitoring of the whole power transformer substation and the vibration of the laid environment, and is complicated to install, unreliable, and high in electromagnetic interference. In order to better detect the operation and maintenance state of the power substation, a distributed optical fiber sensing technology is commonly used at present.

At present, the research and development and application of the home and abroad optical fiber sensing monitoring technology to the safety monitoring of distribution network equipment are mainly put in the transformer station for equipment and cable monitoring, for example, the optical fiber point type temperature measurement system provided by MICRON OPTICS company realizes the temperature monitoring of a switch cabinet, and the power cable monitoring system provided by LIOS company realizes the temperature monitoring of a power cable and an overhead line. However, most of the safety monitoring of the distribution network equipment by the optical fiber sensing monitoring technology of domestic and foreign research institutions is still in the initial research stage, the application basically stays in monitoring of individual equipment and certain parts, and an information networking comprehensive platform of the power cable online monitoring system covering cable anti-theft, cable key area point type temperature measurement, vibration and distribution network automation is not formed. Therefore, the application of the all-fiber sensing monitoring technology to comprehensively and safely monitor the distribution network equipment is still blank at present. The wide-coverage distribution network automation needs to rely on advanced optical fiber sensing monitoring technology to boost the construction of the distribution network automation so as to ensure the smooth implementation and operation of equipment.

The distributed optical fiber nodes are unbalanced in deployment, frequent in acquisition frequency, strong in association of perception data and the like due to the fact that the power substation equipment is wide in distribution range and far away from each other, and due to the fact that the authenticity of the operation and maintenance data of the power substation is damaged due to the influence of loss and noise of the whole system, missing reports and false reports are formed, and application of the operation and maintenance data is limited. In order to enable the optical fiber sensing monitoring technology to be better applied to practice, higher requirements are put forward on the data quality and the data management technology of the optical fiber sensing monitoring technology, so that the quality of the operation and maintenance sensing data of the power transformer substation can be more accurately improved, and effective calculation and accurate alarm are carried out. At present, the simplest method for improving the data acquisition quality of a sensor node is to collect enough sampling data for each sensor by utilizing multiple times of sampling, and then take the average value of the sampling data as final sampling data.

Disclosure of Invention

The invention aims to overcome the defects in the prior art and provide an intelligent monitoring system for the operation and maintenance state of electrical equipment.

The purpose of the invention can be realized by the following technical scheme:

the utility model provides an electrical equipment operation and maintenance state intelligent monitoring system, includes signal perception device, data acquisition device and the information operation diagnostic device that connects gradually, signal perception device includes distributed optical fiber sensor, distributes and installs on the electrical equipment that awaits measuring, information operation diagnostic device stores the procedure of carrying out following step:

sending out a data reading trigger signal at fixed time;

on the basis of the acquired data, performing prediction processing by using a constructed prediction model based on a linear autoregressive method to realize data flow optimization and obtain final sampling data of the power equipment to be detected;

and obtaining a monitoring result of the power equipment to be tested based on the final sampling data.

Further, the signal sensing device comprises a laser signal sensing unit, a current capacity monitor and a temperature monitor.

Further, the prediction model is an autoregressive (ar) (n) model, expressed as:

wherein, β_iIs an autoregressive parameter, x is a sample point, ε_tFor the purpose of a random error that is not observable,

is epsilon_tN is the model order.

Further, when the random error ε_tWhen the probability of 0 occurrence is less than a set threshold, the prediction model is adjusted to:

wherein, β'_t＝β_i+E(ε_t)/(nx_t-i)，ε'_t＝ε_t-E(ε_t)，E(ε_t) Is the average prediction error.

Further, the data flow optimization further comprises:

calculating abnormal statistics of the prediction data, wherein the expression is as follows:

wherein, ω is²Is a function of representing the N corresponding errors epsilon backwards in the time series from the current time instant_tThe average of the sum of squares;

when lambda > J_ijThen, x is determined_t+1As abnormal data, J_ijFor decisions based on generalized Jaccard coefficient definitionsAnd (4) a threshold value.

Further, the final sampling data are subjected to information fusion processing to obtain a monitoring result of the power equipment to be detected.

Further, the information fusion comprises data level fusion, feature level fusion and decision level fusion.

Further, the program further executes: sending optimization ending information to a data acquisition device;

and after receiving the optimization ending information, the data acquisition device starts the next data acquisition operation.

Compared with the prior art, the invention has the following beneficial effects:

1. the invention can utilize the optical fiber sensor to carry out omnibearing real-time intelligent monitoring on the operation condition of the power transformer in the power grid, processes the information of each sensor through comprehensive analysis, and ensures the normal operation of the power grid by controlling corresponding linkage equipment to take certain measures when abnormal conditions occur.

2. The method is characterized in that a prediction model and a prediction mechanism of sensor data flow are provided by adopting a linear autoregressive method, aiming at the defects that distributed optical fiber nodes are unbalanced in deployment, frequent in acquisition frequency, strong in correlation of perception data and the like due to the fact that the distribution range of transformer substation equipment is wide and the distance is long, and the authenticity of the measured electric transformer substation data is damaged due to the influence of the loss and noise of the whole system, so that the application is limited due to the fact that the data are not reported and are misreported.

3. The invention adopts the information fusion fault diagnosis technology, namely the fusion of a data layer, a feature layer and a decision layer, and researches such as reliability analysis or fault diagnosis and the like are carried out on a research object according to the result of the data fusion, so that the fault can be more accurately judged without forming false alarm.

4. In order to reduce the prediction error, the invention provides an automatic prediction model adjustment strategy, so that when the prediction error exceeds a preset threshold value, the prediction model is automatically adjusted, the quality of the information of the distributed optical fiber sensor is improved, the data can be effectively corrected when abnormal conditions occur, the defect occurrence rate of an optical fiber detection system is reduced to a certain extent, a large amount of field work is avoided, and the work efficiency is improved.

5. Through three levels of the information fusion fault diagnosis technology, the state monitoring data of the distributed optical fiber is more reliable, the calculation amount is improved, the anti-interference capability is enhanced, and the distributed optical fiber fault diagnosis method and the distributed optical fiber fault diagnosis system are more suitable for various application environments of the distributed optical fiber.

6. The invention can monitor the temperature, the current-carrying capacity, the theft invasion and the like of the power transformer substation, and ensure the safe and efficient operation of the power grid.

Drawings

FIG. 1 is a schematic structural diagram of example 2 of the present invention;

FIG. 2 is a signal flow diagram of the present invention;

FIG. 3 is a schematic diagram of a data level fusion process of the present invention;

FIG. 4 is a schematic diagram of a feature level fusion process of the present invention;

FIG. 5 is a schematic diagram of a decision-level fusion process of the present invention;

FIG. 6 is a schematic diagram of test data without abnormality and its predicted value;

FIG. 7 is a diagram illustrating test data with abnormalities and their predicted values.

Detailed Description

The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.

Example 1

The embodiment provides an intelligent monitoring system for operation and maintenance states of electrical equipment, which comprises a signal sensing device, a data acquisition device and an information operation diagnosis device which are sequentially connected, wherein the signal sensing device comprises a distributed optical fiber sensor, the signal sensing device is distributed and installed on the electrical equipment to be tested, and the information operation diagnosis device stores a program for executing the following steps:

sending out a data reading trigger signal at fixed time;

on the basis of the acquired data, performing prediction processing by using a constructed prediction model based on a linear autoregressive method to realize data flow optimization and obtain final sampling data of the power equipment to be detected;

and obtaining a monitoring result of the power equipment to be tested based on the final sampling data.

1. Signal sensing device

The distributed optical fiber sensor of the signal sensing device comprises a laser signal sensing unit, a current-carrying capacity monitor and a temperature monitor, and can monitor the temperature, the current-carrying capacity, the theft invasion and the like of the power transformer substation.

2. Data acquisition device

The data acquisition device is used for rapidly acquiring a large amount of operation and maintenance data of the transformer substation and some rapidly-transformed electric energy information of the transformer substation. The data acquisition device can adopt a data acquisition card, adopts a real-time sampling mode and depends on an A/D converter with high-speed transformation to carry out the processes of acquisition, quantification and storage on each sampling point. The sampling mode needs to realize continuous waveform acquisition, corresponding acquired data is put into a data acquisition card, and after information flow is optimized, next data acquisition is started, and original data can be covered so as to reduce the buffer space of intermediate data.

3. Prediction model based on linear autoregressive method

The traditional simplest method for improving the quality of data collected by sensor nodes is to collect enough sampled data for each sensor by using multiple sampling, and then take the average value of the sampled data as final sampled data, but the method faces a great challenge in the operation and maintenance detection of a transformer substation in the distributed optical fiber sensing technology. Firstly, the lines of the transformer substation are wide in distribution range and far away from each other, so that the distributed optical fiber nodes are unbalanced in deployment, frequent in acquisition frequency, strong in perceptual data correlation and the like, and the influence of system loss and noise is large; secondly, for sensor nodes with different state information, the accuracy of sensing devices, the node environment and the sensing process may be different, and it is not practical if all sensors are required to obtain the same number of samples. If the collected data is directly and simply transmitted to a computer for operation, inaccurate data detection is easy to occur, and false alarm and missing report are formed. In order to eliminate the influence and improve the quality of data acquired by the sensor node, a proper intelligent sampling method needs to be selected so as to obtain real data which can better reflect the health state of the measured transformer, such as a temperature curve, insulation and aging degree.

The system adopts an autoregressive AR (n) model as a prediction model, and performs data flow optimization processing on the transformer substation operation and maintenance data obtained by intelligent sampling. Meanwhile, in order to reduce the prediction error, a prediction model automatic adjustment strategy is adopted, so that when the prediction error exceeds a preset threshold value, the prediction model is automatically adjusted.

(a) Establishment of autoregressive AR (n) model

Obtaining N sample points { x of sensor according to intelligent sampling algorithm_t}:x₁，x₂，…，x_N. Let x be₁，x₂，…，x_t-1And x_tIs estimated value of

There is a linear relationship:

β therein₁,…,β_nIs the parameter to be calculated, ε_tIs an unobservable random error and is a random variable. Namely:

the formula (1) becomes an autoregressive AR (n) model, β_iIn order to be a parameter of the auto-regression,

is epsilon_tThe variance of (c).

Will { x_tThe sequence was directly substituted for formula (1), yielding the following linear system of equations:

expressed in matrix form as:

y＝xβ+ε (3)

wherein

According to the multiple regression theory, the least squares estimate of the parameter matrix β is:

for the autoregressive ar (n) model, the selection of the appropriate model order n is a priori taken into account. Because the order n is selected to be too large, the calculation amount is too large, the digital signal processor is not suitable for real-time detection, but the order n cannot be selected to be too small, and the data sequence cannot be effectively represented. The specific model and the order n can be selected by reference to relevant documents, and are not described in detail here.

(b) Prediction mechanism of autoregressive AR (n) model

If ε is shown in formula (1)_tIs unknown, then

Is x_tPredicted value of (2)

When parameter β_i、x_iIs known, then can be calculated

The value of (c). Now suppose x_tIs the actual value at the current time t, the data value at the future time l (l ═ 1,2, …) can be predicted by equation (1), that is:

when l is 1, data is predicted only at one future time, and when x is_t+l-i,…,x_tAre all actual observations, namely:

when l is>1 hour indicates that the data at the time of the future step l is to be predicted, since x is predicted_t+l-1Is unknown, so

To replace x_t+l-1Performing a prediction, namely:

in the actual perceived data, x_tThere may be data that does not fit the parametric model, i.e. anomalous data. At this time, if x is used_tDe-prediction of x_t+1,x_t+2…, then prediction distortion will occur. Prediction is then typically used

To replace x_tThe prediction is performed so that the error is much smaller.

(c) Auto-tuning strategy for Autoregressive (AR) (n) model

According to the nature of the autoregressive AR (n) model,

i.e. satisfying a standard normal distribution, E (epsilon)_t) To make the error value epsilon of the model 0_tAs much as 0, the parameters of the AR (n) model need to be automatically adjusted in the prediction process, i.e. when ε_tWhen the probability of 0 occurrence is less than a threshold, the model needs to be adjusted. Let the prediction error at the current time be epsilon_tThe average prediction error is E (ε)_t) And an error after adjustment is ∈'_tFrom ε'_tAs much as 0, may be_t-E(ε_t) Has a value of approximately ε'_tIs ` ε'_t＝ε_t-E(ε_t) Then model of original

Can become into

From β'_tSubstitute β_i+E(ε_t)/(nx_t-i) From ε'_tInstead of epsilon_t-E(ε_t) Then the adjusted model becomes:

(d) threshold for autoregressive AR (n) model

In the model automatic adjustment strategy, a threshold value needs to be determined. The threshold value of the invention is determined by comparing the similarity degree of the measurement sequence between two adjacent nodes by using the similarity measurement of the sensing vectors of the two sensor nodes. Firstly, a maximum-minimum normalization method is utilized to normalize the perception vector, and then a generalized Jaccard coefficient (also called Tanimoto coefficient) is selected as a similarity measurement threshold. Assuming that i and j represent two adjacent nodes, the sensing data of the two nodes after normalization processing are x respectively_iAnd x_jThe dot product of the two perceptual vectors is x_i·x_jThen the threshold defined by the generalized Jaccard coefficient can be expressed as:

(e) anomalous data detection

From the formula (1)

And is

If B is a step-back operator, i.e. x_t-1＝Bx_tThen, then

In the formula (9)

Order to

And set up

Then there are:

in the formula (10), ω²Is a function of representing the N corresponding errors epsilon backwards in the time series from the current time instant_tMean value of the sum of squares, λ representing the current error value ε_t+1Square of and ω²So that λ can be used as the monitor x_t+1Statistics of whether or not there is an anomaly. When lambda > J_ij，x_t+1Is the exception data. Here J_ijThe threshold value is defined according to the generalized Jaccard coefficient and can be set according to the actual requirements and the data characteristics, for example, J is set_ij2, this means that if the sum of squares of the current errors is twice the sum of squares of the average errors, the data is regarded as abnormal data. In general J_ijThe larger the setting, the lower the requirement for abnormal data detection.

The basic process of the abnormal data detection algorithm is as follows:

input, sensing data and threshold J_ij；

Output, judging whether the data is abnormal data;

Step 1:

s＝0；J_ijinitial value

Step 2 accepting data

Step 3: while (with perceptual data) do

Step 4:begin

Step 5:if detect_anomaly(data)then s＝s+1；

Step 6:if s>＝J_ijthen

a) Detecting anomalous data

b)s＝0；

Step 7:endif

Step 8 accepting data

Step 9:endwhile

(f) Validation of data flow optimization techniques

In order to test the prediction mechanism based on the auto-linear regression model AR (n), the effectiveness of the dynamic adjustment strategy and the abnormal data detection algorithm. In the embodiment, two indexes of temperature and vibration noise are tested, the data acquisition nodes are arranged in different position areas of the transformer substation, the sampling period of the data is 60s, the test data comprises two items of temperature and vibration, wherein the temperature unit is temperature, and the vibration noise unit is represented by voltage V. The specific implementation process is that in a certain period of time, 100 experimental data are selected as sample data, a matrix equation provided by the formula (3) is adopted to construct a prediction model of the formula (6), the order is selected to be 4, an automatic adjustment strategy is carried out according to the principle of the formula (7), the threshold is determined to be 2 according to the formula (8), and the generation of abnormal data is simulated by artificial heating and noise adding. Fig. 6 shows the prediction effect corresponding to the screenshot of a part of data when there is no abnormal data in a certain period of time, and the test data is basically the same as the prediction data. Fig. 7 shows the prediction effect corresponding to the partial data screenshot when the abnormal data is artificially simulated in a certain period of time, and it can be seen from fig. 7 that the prediction result is not greatly interfered by the abnormal data and is basically maintained in the normal range.

4. Acquisition of monitoring results

And after the final sampling data is obtained, carrying out information fusion processing on the data so as to more accurately provide basis for judgment of a decision layer and avoid forming false alarm, and obtaining a monitoring result of the power equipment to be detected based on the fused data. The information fusion of the embodiment includes data level fusion, feature level fusion and decision level fusion.

The data level hierarchical fusion is mainly to directly process the raw data collected by the distributed optical fiber sensor and to perform the research such as reliability analysis or fault diagnosis on the research object according to the result of the data fusion. As shown in fig. 3.

As shown in fig. 4, the feature level fusion is a process of processing a large amount of data obtained by a sensor first to reduce the amount of data, obtaining feature vectors of the data after optimizing the large amount of information flow, performing fusion processing on the feature vectors is called feature level fusion, obtaining a fusion result, and then performing research such as reliability analysis or failure diagnosis on a research object. The feature level fusion is undoubtedly further than the data level fusion, and therefore the feature level fusion is also called the second level fusion. The biggest advantage of feature level fusion is that the real-time performance is better than that of data level fusion, because the data is processed before the data fusion, the calculation amount of the fusion is reduced, and the time required by the fusion is shortened.

Decision-level hierarchical fusion is to further process data on the basis of feature hierarchical fusion, so that the number of data is further reduced, the fusion calculation amount is further reduced, after the decision vectors are obtained, the decision vectors are fused, a decision-level hierarchical fusion result is obtained after fusion, and research such as reliability analysis and fault diagnosis is performed on a research object according to the result, as shown in fig. 5.

Compared with the first two levels of fusion, the information fusion of the decision level is the level fusion with the minimum calculated amount, and the fusion of the decision level is the fusion of the highest level of the information fusion. The advantages of the decision hierarchy are obvious: because the data is processed continuously, the quantity of the data is small, the calculated amount is small, and the real-time performance is better. In addition, the requirement of fusion of decision level layers on the sensor is low, so that the dependence degree on the sensor is low, and the fusion anti-jamming capability of the decision level layers is good. Through three levels of the information fusion fault diagnosis technology, the state monitoring data of the distributed optical fiber is more reliable, the calculation amount is improved, the anti-interference capability is enhanced, and the distributed optical fiber fault diagnosis method and the distributed optical fiber fault diagnosis system are more suitable for various application environments of the distributed optical fiber.

The system utilizes the optical fiber sensor to carry out omnibearing real-time intelligent monitoring on the operation condition of the power transformer in the power grid, adopts the fault diagnosis technology of information fusion, and respectively fuses a data layer, a feature layer and a decision layer, and researches reliability analysis or fault diagnosis and the like on a research object according to the result after the data fusion so as to judge the fault more accurately without forming false alarm. The intelligent monitoring system can monitor the temperature, the current-carrying capacity, the theft invasion and the like of the power transformer substation, and ensure the safe and efficient operation of a power grid; and comprehensively analyzing and processing the information of each sensor, and when an abnormal condition occurs, controlling corresponding linkage equipment to take certain measures to ensure the normal operation of the power grid.

Example 2

The intelligent monitoring system for the operation and maintenance state of the power equipment provided by the embodiment is shown in fig. 1, and is used for monitoring the power equipment 1 in real time. The power equipment mainly refers to equipment needing corresponding state monitoring, and comprises a transformer of a transformer substation, a switch cabinet, a power transmission line and the like. The monitored power equipment in the embodiment is a substation. The signal flow of the monitoring system is shown in fig. 2.

In the monitoring system, a laser signal sensing unit comprises an optical fiber 2, an optical coupler 3, an optical fiber amplifier 4, an anti-stokes optical filter 5, a stokes optical filter 6, photoelectric conversion signal amplifiers 7 and 8, a laser 17, a pulse driving circuit 9, an anti-theft early warning monitoring circuit 10 and a high-voltage circuit 11.

The data acquisition device is a high-speed data acquisition circuit 12. The ampacity monitor and the temperature monitor together constitute an ampacity/temperature monitoring circuit 13.

The information operation diagnosis device consists of DSPs 14 and 15 and a central information processing monitoring platform 16, wherein the DSP14 is respectively connected with the laser 17 and the pulse driving circuit 9 and used for anti-theft monitoring, and the DSP 15 is respectively connected with the high-speed data acquisition circuit 12, the current-carrying capacity/temperature monitoring circuit 13 and the high-voltage circuit 11 and used for monitoring the temperature/load current-carrying capacity. The central information processing monitoring platform 16 is connected to the DSPs 14 and 15 and the high speed data acquisition circuit 12, respectively. The anti-theft early warning monitoring circuit, the high-voltage circuit and the current-carrying capacity/temperature monitoring circuit are the existing circuit modules. In this embodiment, the DSPs 14 and 15 employ a main processor DSPC 6000.

Laser signal perception is generally used in application scenarios where distribution points are many and direct contact is not needed for power equipment signal notification, such as the situation where the base of a transformer sinks, the vibration of the transformer, the theft of a power transmission line, and the like. The signal perception process is as follows: the pulse driving circuit → the laser → the optical fiber amplifier → the optical coupler → the optical fiber → the anti-theft early warning and monitoring circuit → the DSP and other circuits in the figure 1 are adopted to perform distributed detection on the vibration signals around the power cable based on the phi-OTDR technology, realize the noise elimination, the signal analysis and recognition and the event analysis of the induction micro-vibration signals, realize the simultaneous detection of a plurality of vibration events as long as 30 kilometers, have high spatial resolution, and the spatial resolution can reach +/-35 meters (the spatial resolution can be customized according to the monitoring distance). The distributed optical fiber anti-theft early warning monitoring subsystem of the power transformer substation of the embodiment adopts a third-generation semiconductor laser device and an optical fiber technology, and has higher power and precision than a laser manufactured by a second-generation semiconductor, so that the operation and maintenance monitoring subsystem of the power transformer substation designed by the invention has the advantages of high measurement sensitivity, small spatial resolution, high response speed, low false alarm rate, large measurement length, stable work and the like. Similarly, the distributed optical fiber current-carrying capacity/temperature safety monitoring subsystem for operation and maintenance of the transformer substation is composed of circuits such as a high-voltage isolation circuit, a photoelectric signal amplifier, a current-carrying capacity/temperature detection circuit, an optical fiber and a DSP (digital signal processor) in the figure 1, and by accumulating and analyzing temperature data of cable operation, the relation between operation temperature at the bottleneck of the cable and current-carrying capacity change is searched, so that the purpose of effectively utilizing the cable design to allow current-carrying capacity and achieve economic operation is achieved; researching the influence degree of buildings or facilities near the cable line on the operation of the transformer substation through the data of temperature monitoring; by monitoring the operation temperature of the transformer substation, a basis is provided for researching the relationship between the rising temperature of the transformer substation and the insulation aging of the transformer substation; the temperature of the cable transformer is monitored by setting the alarm temperature for the operation and maintenance temperature of the transformer, and the problem that hidden dangers exist in the operation process of the transformer substation is found.

The transformer substation vibration, anti-theft early warning monitoring and current-carrying capacity/temperature safety monitoring processes are as follows: the light wave receiving circuit collects backward scattering light, the backward scattering light is respectively generated into anti-stokes scattering light and stoke scattering light through the filter, weak optical signals form corresponding electric signals through the photoelectric conversion circuit, an APD in the photoelectric conversion circuit needs to work in a high-voltage environment, and the high-voltage biasing circuit is designed for controlling the voltage at two ends of the APD. After the optical signal is converted into the electrical signal, because the electrical signal is very weak, the signal must be enhanced by the high gain effect of the signal amplifying circuit. The reinforced signals are collected and transmitted to a main processor through a high-speed data collecting circuit for operation processing, so that temperature values can be obtained, each scattering position point is determined through an optical time domain positioning technology, the current carrying capacity/temperature collected at the moment corresponds to the position point, and finally the conditions of anti-theft early warning monitoring and current carrying capacity/temperature distribution of the tested line can be generated through multiple times of collection, accumulation and denoising.

The monitoring system simultaneously adopts a real-time sampling mode and depends on an A/D converter with high-speed transformation to carry out the processes of sampling, quantification and storage on each sampling point. The sampling mode needs to realize continuous waveform acquisition, corresponding acquired data is put into a data acquisition card, and after information flow is optimized, next data acquisition is started, and original data can be covered so as to reduce the buffer space of intermediate data. When the system works, two paths of signals to be detected are converted into 0-5V signals after pre-amplification and main amplification and respectively sent into respective A/D converters, and a group of data signal streams are obtained after A/D conversion (namely multipoint sampling of the signals in one period is realized, and the conversion rate is the sampling frequency). Under the action of control circuit, the signals are respectively sent to data memory for storage. The control circuit initiates a write cycle upon receiving a trigger signal from the trigger circuit. In the writing cycle, the control circuit automatically sends out the writing addresses with increasing sequence, so that each group of data is written into the corresponding storage unit, after the data is stored to a certain amount of requirement, the content of the storage unit is sent to the corresponding storage unit of the DTS host computer in a DMA transmission mode, and the purpose of digital average is achieved through the linear accumulation of the computer.

In this embodiment, the central information processing monitoring platform 16 is a flexibly extensible application framework based on an SOA architecture to meet the needs of adding monitoring equipment and upgrading service functions in the future, and a consistent interface adapter mode is designed to flexibly access monitoring equipment of multiple manufacturers and multiple interface types; meanwhile, a centralized data calculation and analysis module is established to provide comprehensive monitoring and analysis capability for the equipment state; and the service management mode of decentralized management and centralized monitoring is met.

The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions that can be obtained by a person skilled in the art through logic analysis, reasoning or limited experiments based on the prior art according to the concept of the present invention should be within the protection scope determined by the present invention.

Claims (8)

1. The utility model provides an electrical equipment operation and maintenance state intelligent monitoring system which characterized in that, includes signal perception device, data acquisition device and the information operation diagnostic device that connects gradually, signal perception device includes distributed optical fiber sensor, distributes and installs on the electrical equipment that awaits measuring, the information operation diagnostic device storage has the procedure of carrying out following step:

sending out a data reading trigger signal at fixed time;

on the basis of the acquired data, performing prediction processing by using a constructed prediction model based on a linear autoregressive method to realize data flow optimization and obtain final sampling data of the power equipment to be detected;

and obtaining a monitoring result of the power equipment to be tested based on the final sampling data.

2. The intelligent monitoring system for the operation and maintenance state of the electrical equipment according to claim 1, wherein the signal sensing device comprises a laser signal sensing unit, a current capacity monitor and a temperature monitor.

3. The intelligent monitoring system for the operation and maintenance state of the power equipment according to claim 1, wherein the prediction model is an autoregressive (ar) (n) model represented as:

wherein, β_iIs an autoregressive parameter, x is a sample point, ε_tFor the purpose of a random error that is not observable,

is epsilon_tN is the model order.

4. The intelligent monitoring system for operation and maintenance states of electric power equipment according to claim 3, wherein when the random error is epsilon_tWhen the probability of 0 occurrence is less than a set threshold, the prediction model is adjusted to:

wherein, β'_t＝β_i+E(ε_t)/(nx_t-i)，ε'_t＝ε_t-E(ε_t)，E(ε_t) Is the average prediction error.

5. The intelligent monitoring system for the operation and maintenance state of the power equipment according to claim 1, wherein the data flow optimization further comprises:

calculating abnormal statistics of the prediction data, wherein the expression is as follows:

wherein, ω is²Is a function of representing the N corresponding errors epsilon backwards in the time series from the current time instant_tThe average of the sum of squares;

when lambda > J_ijThen, x is determined_t+1As abnormal data, J_ijIs a decision threshold defined based on the generalized Jaccard coefficient.

6. The intelligent monitoring system for the operation and maintenance state of the electrical equipment according to claim 1, wherein the final sampling data is subjected to information fusion processing to obtain a monitoring result of the electrical equipment to be tested.

7. The intelligent monitoring system for the operation and maintenance state of the power equipment according to claim 6, wherein the information fusion comprises data-level fusion, feature-level fusion and decision-level fusion.

8. The intelligent monitoring system for the operation and maintenance state of the power equipment according to claim 1, wherein the program further executes: sending optimization ending information to a data acquisition device;

and after receiving the optimization ending information, the data acquisition device starts the next data acquisition operation.

Images