CN103701637A - Method for analyzing running trend of electric power communication transmission network - Google Patents

Abstract

The invention discloses a method for analyzing a running trend of an electric power communication transmission network. The method is characterized by comprising the following steps: (1) acquiring basic data, and providing dynamic parameters and static parameters for a trend analysis model; (2) performing mutation analysis on performance; judging whether the acquired dynamic parameters are out of limit or not based on a performance parameter early warning threshold value; directly generating a performance alarm if the acquired dynamic parameters are out of limit; (3) performing trend analysis when the performance is steadily changed; constructing an error rate and optical power analysis model according to dynamically acquired performance parameters by adopting secondary dynamic exponential smoothing; obtaining a trend analysis weight through a steepest descent iteration method by combining the dynamic parameters and historical data, and further, constructing a linear prediction equation; (4) performing early warning computation, and outputting a performance prediction result. The method disclosed by the invention is applied to a communication management system, and therefore, a powerful data support is provided for the whole-process intelligent computational analysis from data acquisition, trend prediction, early warning analysis and a maintenance strategy.

Description

A kind of electric power communication transmission network operation trend analytical method

Technical field

The present invention relates to electric power communication transmission network operation trend analytical method, particularly relate to a kind of electric power communication transmission network operation trend analytical method based on operation trend analytical model and performance index prediction algorithm, belong to electric power power communication technical field.

Background technology

Construction along with intelligent grid, when finish " 12 ", powerline network scale will be 3 times of current network scale, power communication is not only the important technology support of the safety in production of electrical network in traditional sense and enterprise operation and management, is also that intelligent grid is realized automation, informationization, interactive basis.And the operation maintenance of communication network is running status management mode and the Management for repair based on cycle planning based on equipment alarm at present substantially.In order to change the passive tupe that relies on merely equipment alarm all the time, need a set of electric power communication transmission network operation trend analytical model of research and algorithm, predict telecommunication transport network development trend.

In powerline network is safeguarded at present, the error rate and luminous power are to weigh the important indicator of the signal transmission quality of communication network, and therefore preferential selected these two performance index are as research object.Understand the variation tendency of these indexs, referring again to history data, just can estimate the running status of network, thereby deteriorate into before affecting the normal operation of network and process in advance in performance index, reduce the incidence of fault, improve network running quality.

Operation maintenance about communication network is running status management mode and the Management for repair based on cycle planning based on equipment alarm at present substantially.In research process of the present invention, find following problem:

(1) manual deal with data poor in timeliness and waste a large amount of manpowers.For this work, mainly take monthly manual collection luminous power, error rate data from professional webmaster at present, and carry out data checks, analysis and prediction by the mode of manual sorting Excel form.

(2) fail to realize the core technologies such as index analysis model, warning algorithm.Present stage is to the method for index analysis and how by achievement data, to judge, and provide effective fault pre-alarming, still lack theoretical foundation accurately and reliably, the selection unification of parameter and mainly rely on empirical value or standard value, depend on O&M personnel's ability to work, be easily subject to the interference of human factor.

Summary of the invention

Technical problem to be solved by this invention is, for the deficiency in the operation maintenance of existing communication transmission network, to provide a kind of analysis that can realize power telecom network operation trend.

For solving the problems of the technologies described above, the invention provides a kind of electric power communication transmission network operation trend analytical method, it is characterized in that, comprise the following steps:

Step 1: obtain basic data, for trend-analyzing model provides dynamic parameter and static parameter;

Step 2: performance mutation analysis, based on performance parameter early warning threshold values, whether the dynamic parameter that judgement collects is out-of-limit, if out-of-limit, directly produces performance alarm;

Step 3: trend analysis when smooth performance changes, the performance parameter arriving according to dynamic acquisition, adopts secondary based Dynamic Exponential Smoothing method to build the error rate and Analysis of optical power model; And by steepest Descent iteration method, in conjunction with dynamic parameter and historical data, obtain trend analysis weights, and then build linear prediction equation;

Step 4: early warning is calculated, the output of performance prediction result, comprehensive dynamic and static state parameters and linear prediction equation, calculate following performance index value constantly, in index trend curve mode, exports; And in conjunction with early warning threshold values, outputting alarm or report.

In described step 1, dynamic parameter is the northbound interface by transmission network management (northbound interface is the interface that carries out data interaction with upper layer NMS that facility network guard system provides) error rate directly collecting and the performance data of luminous power, system, by standard equipment webmaster northbound interface, definition data format and communication protocol, realizes the accuracy of performance data between equipment webmaster and comprehensive network management; And for guaranteeing the reliability of data, according to the significance level of on-site actual situations and equipment, collection period is set.

In described step 1, static parameter is the smoothing parameter initial value α of based Dynamic Exponential Smoothing method ₀, steepest Descent iteration method control precision and single exponential smoothing initial value, by adaptive algorithm, obtain smoothing parameter α value, solved the problem of determining smoothing parameter α value with experience, for Future Data prediction provides optimum value.

Step 2 adopts the performance sudden change alarm based on early warning threshold values, realized the real-time monitoring to super threshold values performance in communication transmission passage, more upper limit rate analysis, the analysis of upper limit rate, more lower limit rate analysis, lower limit rate analytic function are provided respectively, for different out-of-limit situations, provide concrete different alarms.

Step 3 is for not out-of-limit luminous power and the error rate, adopt based Dynamic Exponential Smoothing method to set up trend-analyzing model, the prediction of realization to performance change trend, utilize whole historical datas and relevant information, the principle that adopts " thick near thin far away " to historical data be weighted on average, smoothing data, set up secondary based Dynamic Exponential Smoothing Model.

Especially, the comprehensive dynamic parameter of step 3, static parameter and trend-analyzing model, obtain linear prediction equation, calculate following performance index constantly, further matching performance trend curve, the new data coming for each collection, calculate following performance index constantly, and in conjunction with performance trend curve of historical data matching, and then the variation tendency of reflection target itself, to may there is out-of-limit front system is anticipated.

Especially, step 4 output performance predicts the outcome, the performance alarm or performance prediction value or the performance trend curve that in output step 2 and step 3, produce.

The present invention can realize that to take History Performance Data in communications network be basis, and binding ability trend-analyzing model, reaches the intelligent exponential smoothing parameter of determining, has solved the object that relies on O&M personnel experience to determine this parameter; And by selecting the error rate and 2 performance index of luminous power as research object, in conjunction with set up Capability trend analyse model and algorithm, predict future is performance index constantly, and output performance trend curve, for the further performance early warning of system and scheduled overhaul provide data supporting.

Accompanying drawing explanation

Fig. 1 is electric power communication transmission network operation trend analysis process figure of the present invention;

Fig. 2 is trend analysis flow chart of the present invention;

Fig. 3 is performance evaluation curve chart.

Embodiment

Below in conjunction with accompanying drawing, further illustrate specific embodiment of the invention content.

In conjunction with the present invention, realize Capability trend analyse method flow as shown in Figure 2, final completely realize electric power communication transmission network operation trend analysis process as shown in Figure 1, its detailed step is as follows:

The first step: obtain basic data, for trend-analyzing model provides dynamic parameter and static parameter;

1, dynamic parameter collection

Dynamic parameter comprises: the performance data such as luminous power, the error rate, time delay, packet loss in transmission network.The present invention carries out trend analysis mainly for luminous power and the error rate, and time delay and packet loss are as the reference of performance prediction.

Dynamic parameter acquisition mode: electric power communication transmission network is under multi-vendor equipment environment, because each manufacturer's network management system exists certain difference at aspects such as architecture, Management Information Model, network management protocols.Therefore, take the performance index of different vendor to analyze one by one, and take the technological means such as format conversion, object matching, data translation, shielding difference, realizes the standardization of performance index.Simultaneously by general, standardized northbound interface adaptor model, solve the EML(Network Element Layer that existing network management system exists) with NML(network layer) the incompatible problem of interface, realize the performance collection of Liao Dui different vendor transmission network equipment.

Collection period is the important indicator that dynamic parameter gathers, can be according to the significance level of on-site actual situations and equipment, and collection period is set to 15 minutes to twenty four hours, and each collection period system will complete a data acquisition automatically.Acquisition range can be single device, multiple devices or whole network equipment.After setting completes, performance collection platform, by automatic connection device webmaster, collects device data automatically, for operation trend analytical model provides dynamic data basis.

2, static parameter input

Static parameter comprises:

The predetermined constant of trend analysis linear equation, i.e. single exponential smoothing initial value

The initial value α of smoothing parameter in based Dynamic Exponential Smoothing method ₀; In exponential smoothing, α is as Y _tweighting weights, but the present invention has adopted based Dynamic Exponential Smoothing method,

as Y _tweighting weights, so α is called smoothing parameter,

be called dynamic smoothing parameter;

For calculating smoothing parameter α and dynamic smoothing parameter

the control precision ε of the steepest Descent iteration method setting.α is the return value of steepest Descent iteration method, is input variable in exponential smoothing, here for to make based Dynamic Exponential Smoothing method and exponential smoothing keep certain correlation, has also been set to parameter;

Second step: performance mutation analysis.Based on performance parameter early warning threshold values, whether the dynamic parameter that judgement collects is out-of-limit, if out-of-limit, directly produces performance alarm.

, mainly for power transmitting device, there is the analysis of catastrophic failure in performance mutation analysis.To the performance data collecting, whether all properties that first checks transmission channel surpasses warning index, be mainly reflected in ES (SES), SES (Severely Errored Second), BBE (Background Block Error), UAS(unavailable second to the channel layer of the parameters such as OOP (light transmitted power), IOP (light receiving power) and multiplex layer, regeneration zone, each speed) etc. the whether out-of-limit inspection of performance data, if surpass user-defined performance threshold, directly produce performance alarm.

The 3rd step: trend analysis when smooth performance changes: the performance parameter arriving according to dynamic acquisition, adopt secondary based Dynamic Exponential Smoothing method, build the error rate and Analysis of optical power model; And by steepest Descent iteration method, in conjunction with dynamic parameter and historical data, obtain trend analysis weights, and then build linear prediction equation.

1, operation trend analytical model: the error rate and luminous power are constantly to change in time, can be regarded as a time series, and exponential smoothing is that a kind of algorithm is simple, result stable, widely used classical Time Series Forecasting Methods, therefore can adopt this model and algorithm prediction electric power communication transmission network operation trend.Smoothing parameter α in traditional Secondary Exponential Smoothing Method model is normalized, obtains dynamic smoothing parameter: establish { Y _tbe time series and measured value thereof, concrete secondary based Dynamic Exponential Smoothing method model is:

Wherein α is given smoothing parameter (0< α <1);

for t phase based Dynamic Exponential Smoothing parameter;

for t phase single exponential smoothing value;

for t phase double smoothing value;

predicted value (T is prediction issue) for the t+T phase; a _tand b _tfor t phase linear forecasting parameter;

Initial value is: $\left\{\begin{array}{c}{S}_0^{\left(1\right)}=\frac{\mathrm{\&alpha;}}{1-{(1-\mathrm{\&alpha;})}^t}\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}{(1-\mathrm{\&alpha;})}^{t-i}{Y}_i\\ S_0^{\left(2\right)}=\frac{\mathrm{\&alpha;}}{1-{(1-\mathrm{\&alpha;})}^t}\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}{(1-\mathrm{\&alpha;})}^{t-i}{S}_i^{\left(1\right)}\end{array}\right.$ The i.e. error rate and luminous power sequence { Y to collect _tand its smooth value weighted average separately respectively as level and smooth initial value, solved level and smooth initial value and be difficult to definite problem, wherein Y _ifor { Y _toccurrence in sequence.

2, smoothing parameter adaptive algorithm: smoothing parameter α is a key factor that affects Smoothing Prediction precision, therefore takes adaptive algorithm to obtain smoothing parameter α value, and then obtains dynamic smoothing parameter

and the exponential smoothing model of optimizing.So-called adaptive algorithm obtains smoothing parameter α value and refers to, under the condition of 0< α <1, makes the sum of square of deviations of Prediction sum squares SSE(experimental error size) the α value that reaches hour.

Set up and determine that the Optimized model of smoothing parameter α value is:

if use single exponential smoothing $S_t^{\left(1\right)}=\mathrm{\&alpha;}{Y}_t+(1-\mathrm{\&alpha;})S_{t-1}^{\left(1\right)},$ ? $S_t^{\left(1\right)}=\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}\mathrm{\&alpha;}{(1-\mathrm{\&alpha;})}^{t-i}{Y}_i+{(1-\mathrm{\&alpha;})}^t{S}_0^{\left(1\right)}$ Give predicted value

assignment, Optimized model can be expressed as: $\mathrm{Min\; SSE}=\underset{t=1}{\overset{n}{\mathrm{\&Sigma;}}}{\{Y_t-\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}\mathrm{\&alpha;}{(1-\mathrm{\&alpha;})}^{t-i}{Y}_i-{(1-\mathrm{\&alpha;})}^t{S}_0^{\left(1\right)}\}}^2,$ By using convergence rate steepest decline iterative algorithm very fast, that easily carry out, solve this nonlinear equation and can obtain best smoothing parameter α value, concrete grammar is as follows:

1) the initial value α of given smoothing parameter α ₀, control precision ε and by the single exponential smoothing initial value calculating

2) calculate (wherein the differential equation of error sum of squares SSE to smoothing parameter α) if

α ₀for approximate optimal solution; Otherwise carry out linear search, solve optimal step size λ _k-1(k>=1, K is the parameter that represents iterations in iterative process), calculates ${\mathrm{\&alpha;}}_k={\mathrm{\&alpha;}}_{k-1}-{\mathrm{\&lambda;}}_{k-1}\&dtri;\mathrm{SSE}\left({\mathrm{\&alpha;}}_{k-1}\right),$ Until meet $\left|\right|\&dtri;\mathrm{SSE}\left({\mathrm{\&alpha;}}_k\right)\left|\right|\&le;\mathrm{\&epsiv;},$ Try to achieve optimal solution α _k;

3) finally by α _ksubstitution

obtain dynamic smoothing parameter

set up new based Dynamic Exponential Smoothing Model.

The 4th step: early warning is calculated, the output of performance prediction result.Comprehensive dynamic and static state parameters and linear prediction equation, calculate following performance index value constantly, in index trend curve mode, exports; And in conjunction with early warning threshold values, outputting alarm or report.

The output of performance prediction result, out-of-limit analysis during combination property sudden change and the linear prediction equation in smooth performance change procedure, with the form of performance alarm, performance trend change curve and performance report, output transmission network Capability trend analyse result (as Fig. 3 has shown performance evaluation curve).Particular content comprises:

1, performance sudden change interpretation of result

The performance data arriving according to system acquisition, the setting based on the design's method to performance warning index, for the out-of-limit rank of performance number, forms concrete performance alarm, with the formal output of performance index value and Trouble Report.

2, performance trend mutation analysis

By performance trend-analyzing model and algorithm, calculate following performance index value constantly, output performance desired value and Capability trend analyse curve, the potential safety hazard and the weak link that for timely discovering device, exist provide data to rely on, and for further performance early warning and maintenance scheduling provide data supporting.

Other concrete technology of the method for the invention and system are described the description that need consult appropriate section in the above-mentioned explanation of the present invention in detail, are not repeated.

Those skilled in the art can change or the design of modification but do not depart from thought of the present invention and scope the present invention.Therefore, if of the present invention these are revised and modification belongs to the claims in the present invention and the technical scope that is equal within, the present invention is also intended to comprise these changes and modification interior.

Claims (5)

1. an electric power communication transmission network operation trend analytical method, is characterized in that, comprises the following steps:

Step 1: obtain basic data, for trend-analyzing model provides dynamic parameter and static parameter;

Step 2: performance mutation analysis, based on performance parameter early warning threshold values, whether the dynamic parameter that judgement collects is out-of-limit, if out-of-limit, directly produces performance alarm;

Step 3: trend analysis when smooth performance changes, the performance parameter arriving according to dynamic acquisition, adopts secondary based Dynamic Exponential Smoothing method to build the error rate and Analysis of optical power model; And by steepest Descent iteration method, in conjunction with dynamic parameter and historical data, obtain trend analysis weights, and then build linear prediction equation;

Step 4: early warning is calculated, the output of performance prediction result, comprehensive dynamic and static state parameters and linear prediction equation, calculate following performance index value constantly, in index trend curve mode, exports; And in conjunction with early warning threshold values, outputting alarm or report.

2. electric power communication transmission network operation trend analytical method according to claim 1, it is characterized in that, in described step 1, the error rate that dynamic parameter directly collects for the northbound interface by transmission network management and the performance data of luminous power, realize the accuracy of performance data between equipment webmaster and comprehensive network management by standard equipment webmaster northbound interface, definition data format and communication protocol.

3. electric power communication transmission network operation trend analytical method according to claim 1, is characterized in that, in described step 1, static parameter comprises:

The predetermined constant of trend analysis linear equation, i.e. single exponential smoothing initial value

The initial value α of smoothing parameter in based Dynamic Exponential Smoothing method ₀;

For calculating smoothing parameter α and dynamic smoothing parameter

the control precision ε of the steepest Descent iteration method setting.

4. electric power communication transmission network operation trend analytical method according to claim 1, is characterized in that, in described step 3,

Operation trend analytical model is: establish { Y _tbe time series and measured value thereof, concrete secondary based Dynamic Exponential Smoothing method model is:

Wherein α is given smoothing parameter (0< α <1);

for t phase based Dynamic Exponential Smoothing parameter;

for t phase single exponential smoothing value;

for t phase double smoothing value; for the predicted value of t+T phase, T is prediction issue; a _tand b _tfor t phase linear forecasting parameter;

Initial value is: $\left\{\begin{array}{c}{S}_0^{\left(1\right)}=\frac{\mathrm{\&alpha;}}{1-{(1-\mathrm{\&alpha;})}^t}\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}{(1-\mathrm{\&alpha;})}^{t-i}{Y}_i\\ S_0^{\left(2\right)}=\frac{\mathrm{\&alpha;}}{1-{(1-\mathrm{\&alpha;})}^t}\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}{(1-\mathrm{\&alpha;})}^{t-i}{S}_i^{\left(1\right)}\end{array}\right.$ The i.e. error rate and luminous power sequence { Y to collect _tand its smooth value weighted average separately respectively as level and smooth initial value, wherein Y _ifor { Y _toccurrence in sequence.

5. electric power communication transmission network operation trend analytical method according to claim 1, is characterized in that, in described step 3,

Smoothing parameter adaptive algorithm is: described adaptive algorithm obtains smoothing parameter α value and refers to, under the condition of 0< α <1, and the α value that Prediction sum squares SSE is reached hour;

Set up and determine that the Optimized model of smoothing parameter α value is:

if use single exponential smoothing $S_t^{\left(1\right)}=\mathrm{\&alpha;}{Y}_t+(1-\mathrm{\&alpha;})S_{t-1}^{\left(1\right)},$ ? $S_t^{\left(1\right)}=\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}\mathrm{\&alpha;}{(1-\mathrm{\&alpha;})}^{t-i}{Y}_i+{(1-\mathrm{\&alpha;})}^t{S}_0^{\left(1\right)}$ Give predicted value

assignment, Optimized model can be expressed as: $\mathrm{Min\; SSE}=\underset{t=1}{\overset{n}{\mathrm{\&Sigma;}}}{\{Y_t-\underset{i=1}{\overset{t}{\mathrm{\&Sigma;}}}\mathrm{\&alpha;}{(1-\mathrm{\&alpha;})}^{t-i}{Y}_i-{(1-\mathrm{\&alpha;})}^t{S}_0^{\left(1\right)}\}}^2,$ By using steepest decline iterative algorithm, solve this nonlinear equation and can obtain best smoothing parameter α value, specifically comprise the following steps:

1) the initial value α of given smoothing parameter α ₀, control precision ε and by the single exponential smoothing initial value calculating

2) calculate

wherein

the differential equation of error sum of squares SSE to smoothing parameter α, if

α ₀for approximate optimal solution; Otherwise carry out linear search, solve optimal step size λ _k-1, k>=1 wherein, K is the parameter that represents iterations in iterative process, calculates ${\mathrm{\&alpha;}}_k={\mathrm{\&alpha;}}_{k-1}-{\mathrm{\&lambda;}}_{k-1}\&dtri;\mathrm{SSE}\left({\mathrm{\&alpha;}}_{k-1}\right),$ Until meet $\left|\right|\&dtri;\mathrm{SSE}\left({\mathrm{\&alpha;}}_k\right)\left|\right|\&le;\mathrm{\&epsiv;},$ Try to achieve optimal solution α _k;

3) finally by α _ksubstitution

obtain dynamic smoothing parameter

set up new based Dynamic Exponential Smoothing Model.

Images