CN103647677A - Power communication network state detection method - Google Patents

Abstract

The invention discloses a power communication network state detection method. The method comprises the following steps: S1 a state parameter system of a power communication network state maintenance system is established; S2 a judgment matrix is constructed and then the judgment matrix is transformed through the compatibility matrix method to make the judgment matrix satisfy the consistency condition; S3 a score is calculated and a suggestion is given; S4 if the final score is not in an actual experience value interval, then the step returns to the step S2, and the importance contrast value of two parameters in a same module is adjusted according to the actual conditions until the final score satisfy the experience value interval. According to the invention, the time of the next state maintenance can be deduced through the dynamic analysis of device state evaluation, so the multiple maintenance and insufficient maintenance caused by fixed period type maintenance can be prevented, and the device maintenance cost can be effectively reduced while ensuring the normal operation of devices.

Description

A kind of detection method of power communication net state

Technical field

The present invention relates to a kind of detection method of communication network state, especially relate to a kind of detection method of power communication net state.

Background technology

Power telecom network requires to have high reliability, yet fault is inevitable, and this just requires system to carry out real-time monitoring.According to the current operation conditions arrangement maintenance of equipment, can not only find in time the defect of power telecom network, improve equipment dependability, also can save human, financial, and material resources power, to improving equipment performance, guarantee that power communication reliability service is significant.Meanwhile, can infer the time of repair based on condition of component next time by dynamic analysis historical state data, thereby prevent many degree maintenance and maintenance deficiency that fixed cycle inspection and repair shop causes, can, in the normal work of support equipment, effectively reduce overhaul of the equipments cost.But, in traditional detection method, the foundation of state parameter system and forecast model always is a difficult point, unsuitable state parameter system directly has influence on the correctness of condition grading, equally, unsuitable forecast model can cause power telecom network to-be to do the estimation making mistake.

Summary of the invention

Technical problem to be solved by this invention, just be to provide a kind of detection method of power communication net state, in the method, the foundation of state parameter system is based on analytic hierarchy process (AHP), can to state parameter, carry out the analysis of refinement, finally obtain relatively definite state final score; Meanwhile, the forecast model that method proposes can be described the power telecom network state trend in future better.

Solve the problems of the technologies described above, the technical solution used in the present invention is:

A detection method for power communication net state, is characterized in that: comprise the following steps:

S1 builds the state parameter system of power telecom network Condition-based maintenance system

The classification of S1-1 state parameter: according to the correlation of parameter, all parameters are divided in several modules, have corresponding state parameter under each module;

S1-2 state parameter scoring: according to the standards of grading of each state parameter, give each state parameter scoring;

S1-3 sets up the state parameter system of stratification

By step S1-1, state parameter system is divided into two-layer: module layer and parameter layer; Corresponding module proportion between setting module and module, sets the ratio of importance degree between the parameter under equal modules;

S2 Judgement Matricies

S2-1 analytic hierarchy process (AHP) Judgement Matricies, judgment matrix A=(a _ij) meet:

$\left\{\begin{array}{cc}{a}_{\mathrm{ij}}>0& (i,j=\mathrm{1,2},\·\·\·,n)\\ a_{\mathrm{ii}}=0& (i=\mathrm{1,2},\·\·\·,n)\\ a_{\mathrm{ij}}=\frac{1}{a_{\mathrm{ji}}}& (i,j=\mathrm{1,2},\·\·\·,n)\end{array}\right.$

Wherein, a _ijfor the index X of lower floor _iand X _jratio to the relative significance level of its upper strata index, and this ratio is quantized to (these values all obtain by expert consulting and field experiment, as shown in table 4);

S2-2 utilizes consistent matrix method conversion judgment matrix

Conventionally, between index, there is transitivity, if i.e. X ₁with X ₂relative Link Importance be a ₁₂, X ₁with X ₃relative Link Importance be a ₁₃, so can be according to a ₁₂and a ₁₃obtain X ₂and X ₃between relative Link Importance be

be generalized to general situation, condition for consistence is or a _ija _jk=a _ik, i wherein, j, k=1,2 ..., n; But the element in the judgment matrix of taking is herein determined according to the relative importance between 2 indexs, can not guarantee the consistency of judgment matrix, and we use consistent matrix, can meet condition for consistence, avoid above problem;

Consistent matrix B=(b _ij) by following formula, obtain:

$b_{\mathrm{ij}}=\sqrt[n]{\underset{k=1}{\overset{n}{\mathrm{\Π}}}{a}_{\mathrm{ik}}\·a_{\mathrm{kj}}}$

It meets condition for consistence: b _ij=b _jkb _kj; Therefore, consistent matrix method is the judgment matrix A=(a to weight calculation _ij) in each element revise, make it be converted to the judgment matrix B=(b processing through consistency _ij), this matrix meets condition for consistence, and

can find out that matrix B is the revised matrix of matrix A, it meets the consistency of matrix;

An example:

Table 4 is the judgment matrix of user's input, $A=\left(\begin{array}{ccc}1& 0.5& 4\\ 2& 1& 1\\ 0.25& 1& 1\end{array}\right),$ Be completing steps S2-1, then carry out S2-2, use consistent matrix to carry out consistency processing, obtain: $B=\left(\begin{array}{ccc}\sqrt[3]{(1\×1)\×(0.5\×2)\×(4\×0.25)}& \sqrt[3]{(1\×0.5)\×(0.5\×1)\×(4\×1)}& \sqrt[3]{(1\×4)\×(0.5\×1)\×(4\×1)}\\ \sqrt[3]{(2\×1)\×(1\×2)\×(1\×0.25)}& \sqrt[3]{(2\×0.5)\×(1\×1)\×(1\×1)}& \sqrt[3]{(2\×4)\×(1\×1)\×(1\×1)}\\ \sqrt[3]{(0.25\×1)\×(1\×2)\×(1\×0.25)}& \sqrt[3]{(0.25\×0.5)\×(1\×1)\×(1\×1)}& \sqrt[3]{(0.25\×4)\×(1\×1)\×(1\×1)}\end{array}\right)$ Calculate the result of matrix B, obtain the judgment matrix of finally processing through consistency.

S3 counts the score and provides suggestion

Determine each evaluation index weight, according to each evaluation index score value, calculate the final score of system, thereby provide the quantized values of evaluation scheme system quality, provide maintenance suggestion, and predict the repair time next time, reduce maintenance of equipment cost;

The computational methods of S3-1 weight:

$W_j=\frac{c_j}{{\mathrm{\Σ}}_{k=1}^n{c}_k}(j=\mathrm{1,2},\·\·\·,n)$

Wherein $c_j=\sqrt[n]{{\mathrm{\Π}}_{k=1}^n{b}_{\mathrm{jk}}}(j=\mathrm{1,2},\·\·\·,n);$

S3-2 final score can calculate by following formula:

S=∑A _j∑B _iC _i

The state score that wherein S is equipment, A _jfor the weight of parts, B _ifor the weight of state parameter, C _iscore for state parameter;

Examination and repair system is according to the state score of equipment, and whether judgment device needs maintenance, and the parts that need maintenance, finally provides maintenance suggestion, generates service bulletin; Service bulletin grade is listed in Table 1;

S3-3 status predication: the rate of change q of define equipment state evaluation result, wherein R _iand R _i-1the double repair based on condition of component evaluation of estimate of the equipment that is respectively, the time interval that t is double repair based on condition of component;

$q_i=\frac{R_{i-1}-R_i}{t_i}$

There iing under the prerequisite of abundant historical data the rate of change of state evaluation while predicting that according to formula below i+1 carries out repair based on condition of component; Wherein, a is a constant, and while being used for identifying the last repair based on condition of component, state evaluation rate of change is to the influence degree of state evaluation next time;

$q_{i+1}={\mathrm{aq}}_i+{\mathrm{\Σ}}_{j=1}^n{(1-a)}^j{q}_{i-j}$

If establish a=1/2, can obtain:

$q_{i+1}=q_i/2+{\mathrm{\Σ}}_{j=1}^n{(1/2)}^j{q}_{i-j}$

If establishing the threshold value of warning of this equipment state is V, require the time of i+1 next state maintenance to meet:

v≤R _i-q _i+1t _i+1

Change the formula that can draw below, prediction should be at t _i+1in time, arrange the repair based on condition of component next time to this equipment;

$t_{i+1}\≤\frac{R_i-V}{q_{i+1}}$

S4 adjusts judgment matrix, again scoring

If final score not in actual empirical value interval, is adjusted judgment matrix (adjusting the value of the importance degree contrast of two parameters under same module according to actual conditions), make state estimation more reasonable.

Beneficial effect: the present invention is by the dynamic analysis that equipment state is evaluated, infer the time of repair based on condition of component next time, can prevent many degree maintenance and maintenance deficiency that fixed cycle inspection and repair shop causes, can, in the normal work of support equipment, effectively reduce overhaul of the equipments cost.

Accompanying drawing explanation

Fig. 1: detection method flow chart of the present invention;

Fig. 2: evaluation model figure of the present invention;

Fig. 3: scoring operational flowchart of the present invention;

Fig. 4: state parameter system figure of the present invention, a system has several modules, lower minute a plurality of parameters of each module;

Fig. 5: the prognostic chart of equipment state.

Table 1: service bulletin grade, for overhauling the feedback of result.

Table 2: state parameter methods of marking, this table is an example of state parameter methods of marking, for using the method to carry out the system of equipment state overhauling.

Table 3: score value table, the state parameter numerical value in corresponding diagram 2 gathers;

Table 4: weight table, the Primary Judgement Matrix of the state parameter in corresponding diagram 2;

Table 5: module specific density table, is associated with the state parameter weight calculation in Fig. 2;

(state evaluation that table 4,5,6 finally obtains in Fig. 2 calculates)

Embodiment

Referring to Fig. 1 and Fig. 2, the detection method of power communication net state of the present invention, comprises the following steps:

S1 builds the state parameter system of power telecom network Condition-based maintenance system

The classification of S1-1 state parameter: according to the correlation of parameter, all parameters are divided in several modules, have corresponding state parameter under each module;

Referring to Fig. 4, for example machine room site inspection is divided into equipment machine room environment, power supply and ground connection, Routing Protocol configuration and four modules of configuration reasonability, under each module, there is corresponding state parameter, as there being following state parameter under equipment machine room environment module: machine room cleannes, computer room temperature and computer room temperature;

S1-2 state parameter scoring: according to the standards of grading of each state parameter, give each state parameter scoring;

As shown in table 2, be the standards of grading of three state parameter machine room cleannes, computer room temperature and computer room temperatures of above-mentioned example; According to equipment actual conditions and standards of grading, just can make suitable scoring, as table 3.

Table 1

Table 2

Table 3

?	Machine room cleannes	Computer room temperature	Machine room humidity
				Score value	8	9	8

S1-3 sets up the state parameter system of stratification

By step S1-1, state parameter system is divided into two-layer: module layer and parameter layer; Between setting module and module, corresponding module proportion (as shown in table 5), sets the ratio of setting importance degree between the parameter under equal modules, i.e. judgment matrix (as shown in table 4);

Table 5

Table 4

?	Machine room cleannes	Computer room temperature	Machine room humidity
				Machine room cleannes	1	0.5	4
Computer room temperature	2	1	1
				Machine room humidity	0.25	1	1

S2 Judgement Matricies

S2-1 analytic hierarchy process (AHP) Judgement Matricies, judgment matrix A=(a _ij) meet:

$\left\{\begin{array}{cc}{a}_{\mathrm{ij}}>0& (i,j=\mathrm{1,2},\·\·\·,n)\\ a_{\mathrm{ii}}=0& (i=\mathrm{1,2},\·\·\·,n)\\ a_{\mathrm{ij}}=\frac{1}{a_{\mathrm{ji}}}& (i,j=\mathrm{1,2},\·\·\·,n)\end{array}\right.$

Wherein, a _ijfor the index X of lower floor _iand X _jratio to the relative significance level of its upper strata index, and this ratio is quantized to (these values all obtain by expert consulting and field experiment, as shown in table 4);

S2-2 utilizes consistent matrix method conversion judgment matrix

Conventionally, between index, there is transitivity, if i.e. X ₁with X ₂relative Link Importance be a ₁₂, X ₁with X ₃relative Link Importance be a ₁₃, so can be according to a ₁₂and a ₁₃obtain X ₂and X ₃between relative Link Importance be

be generalized to general situation, condition for consistence is

or a _ija _jk=a _ik, i wherein, j, k=1,2 ... n; But the element in the judgment matrix of taking is herein determined according to the relative importance between 2 indexs, can not guarantee the consistency of judgment matrix, and we use consistent matrix, can meet condition for consistence, avoid above problem;

Consistent matrix B=(b _ij) by following formula, obtain:

$b_{\mathrm{ij}}=\sqrt[n]{\underset{k=1}{\overset{n}{\mathrm{\Π}}}{a}_{\mathrm{ik}}\·a_{\mathrm{kj}}}$

It meets condition for consistence: b _ij=b _jkb _kj; Therefore, consistent matrix method is the judgment matrix A=(a to weight calculation _ij) in each element revise, make it be converted to the judgment matrix B=(b processing through consistency _ij), this matrix meets condition for consistence, and can find out that matrix B is the revised matrix of matrix A, it meets the consistency of matrix;

An example:

Table 4 is the judgment matrix of user's input, $A=\left(\begin{array}{ccc}1& 0.5& 4\\ 2& 1& 1\\ 0.25& 1& 1\end{array}\right),$ Be completing steps S2-1, then carry out S2-2, use consistent matrix to carry out consistency processing, obtain: $B=\left(\begin{array}{ccc}\sqrt[3]{(1\×1)\×(0.5\×2)\×(4\×0.25)}& \sqrt[3]{(1\×0.5)\×(0.5\×1)\×(4\×1)}& \sqrt[3]{(1\×4)\×(0.5\×1)\×(4\×1)}\\ \sqrt[3]{(2\×1)\×(1\×2)\×(1\×0.25)}& \sqrt[3]{(2\×0.5)\×(1\×1)\×(1\×1)}& \sqrt[3]{(2\×4)\×(1\×1)\×(1\×1)}\\ \sqrt[3]{(0.25\×1)\×(1\×2)\×(1\×0.25)}& \sqrt[3]{(0.25\×0.5)\×(1\×1)\×(1\×1)}& \sqrt[3]{(0.25\×4)\×(1\×1)\×(1\×1)}\end{array}\right)$ Calculate the result of matrix B, obtain the judgment matrix of finally processing through consistency.

S3 counts the score and provides suggestion

Determine each evaluation index weight, according to each evaluation index score value, calculate the final score of system, thereby provide the quantized values of evaluation scheme system quality, provide maintenance suggestion, and predict the repair time next time, reduce maintenance of equipment cost;

The computational methods of S3-1 weight:

$W_j=\frac{c_j}{{\mathrm{\Σ}}_{k=1}^n{c}_k}(j=\mathrm{1,2},\·\·\·,n)$

Wherein $c_j=\sqrt[n]{{\mathrm{\Π}}_{k=1}^n{b}_{\mathrm{jk}}}(j=\mathrm{1,2},\·\·\·,n);$

S3-2 final score can calculate by following formula:

S=∑A _j∑B _iC _i

The state score that wherein S is equipment, A _jfor the weight of parts, B _ifor the weight of state parameter, C _iscore for state parameter;

Examination and repair system is according to the state score of equipment, and whether judgment device needs maintenance, and the parts that need maintenance, finally provides maintenance suggestion, generates service bulletin; Service bulletin grade is listed in Table 1;

S3-3 status predication: the rate of change q of define equipment state evaluation result, wherein R _iand R _i-1the double repair based on condition of component evaluation of estimate of the equipment that is respectively, the time interval that t is double repair based on condition of component;

$q_i=\frac{R_{i-1}-R_i}{t_i}$

There iing under the prerequisite of abundant historical data the rate of change of state evaluation while predicting that according to formula below i+1 carries out repair based on condition of component; Wherein, a is a constant, and while being used for identifying the last repair based on condition of component, state evaluation rate of change is to the influence degree of state evaluation next time;

$q_{i+1}={\mathrm{aq}}_i+{\mathrm{\Σ}}_{j=1}^n{(1-a)}^j{q}_{i-j}$

If establish a=1/2, can obtain:

$q_{i+1}=q_i/2+{\mathrm{\Σ}}_{j=1}^n{(1/2)}^j{q}_{i-j}$

If establishing the threshold value of warning of this equipment state is V, require the time of i+1 next state maintenance to meet:

v≤R _i-q _i+1t _i+1

Change the formula that can draw below, prediction should be at t _i+1in time, arrange the repair based on condition of component next time to this equipment;

$t_{i+1}\≤\frac{R_i-V}{q_{i+1}}$

S4 adjusts judgment matrix, again scoring

If the error of existence (being that final score does not conform to actual conditions), adjusts judgment matrix (adjusting the value of the importance degree contrast of two parameters under same module according to actual conditions), make state estimation more reasonable.

State parameter classification is good by program setting in advance, need to arrange the proportion of each module of parameter.(table 5)

The weight table of parameter under each module of equipment, arranges weight between every two parameters, and this weight is exactly the ratio of the importance degree of these two parameters.(table 4)

The corresponding score value of parameter under each module is set, the high indication equipment of score value in good condition.(table 3) calculates score value and corresponding detection feedback opinion.

According to the grade trend of the historical data predict future of equipment.(Fig. 5).

Claims (1)

1. a detection method for power communication net state, is characterized in that: comprise the following steps:

S1 builds the state parameter system of power telecom network Condition-based maintenance system

The classification of S1-1 state parameter: according to the correlation of parameter, all parameters are divided in several modules, have corresponding state parameter under each module;

S1-2 state parameter scoring: according to the standards of grading of each state parameter, give each state parameter scoring;

S1-3 sets up the state parameter system of stratification: by step S1-1, state parameter system is divided into two-layer: module layer and parameter layer; Corresponding module proportion between setting module and module, sets the ratio of importance degree between the parameter under equal modules;

S2 Judgement Matricies

S2-1 analytic hierarchy process (AHP) Judgement Matricies, judgment matrix A=(a _ij) meet:

$\left\{\begin{array}{cc}{a}_{\mathrm{ij}}>0& (i,j=\mathrm{1,2},\·\·\·,n)\\ a_{\mathrm{ii}}=0& (i=\mathrm{1,2},\·\·\·,n)\\ a_{\mathrm{ij}}=\frac{1}{a_{\mathrm{ji}}}& (i,j=\mathrm{1,2},\·\·\·,n)\end{array}\right.$

Wherein, a _ijfor the index X of lower floor _iand X _jratio to the relative significance level of its upper strata index, and this ratio is quantized;

S2-2 utilizes consistent matrix method conversion judgment matrix

Consistent matrix B=(b _ij) by following formula, obtain:

$b_{\mathrm{ij}}=\sqrt[n]{\underset{k=1}{\overset{n}{\mathrm{\Π}}}{a}_{\mathrm{ik}}\·a_{\mathrm{kj}}}$

It meets condition for consistence: b _ij=b _jkb _kj; And

Calculate the result of matrix B, obtain the judgment matrix of finally processing through consistency;

S3 counts the score and provides suggestion

Determine each evaluation index weight, according to each evaluation index score value, calculate the final score of system, thereby provide the quantized values of evaluation scheme system quality, provide maintenance suggestion, and predict the repair time next time, reduce maintenance of equipment cost;

The computational methods of S3-1 weight:

$W_j=\frac{c_j}{{\mathrm{\Σ}}_{k=1}^n{c}_k}(j=\mathrm{1,2},\·\·\·,n)$

Wherein $c_j=\sqrt[n]{{\mathrm{\Π}}_{k=1}^n{b}_{\mathrm{jk}}}(j=\mathrm{1,2},\·\·\·,n);$

S3-2 final score can calculate by following formula:

S=∑A _j∑B _iC _i;

The state score that wherein S is equipment, A _jfor the weight of parts, B _ifor the weight of state parameter, C _iscore for state parameter;

Examination and repair system is according to the state score of equipment, and whether judgment device needs maintenance, and the parts that need maintenance, finally provides maintenance suggestion, generates service bulletin;

S3-3 status predication: the rate of change q of define equipment state evaluation result, wherein R _iand R _i-1the double repair based on condition of component evaluation of estimate of the equipment that is respectively, the time interval that t is double repair based on condition of component;

$q_i=\frac{R_{i-1}-R_i}{t_i};$

There iing under the prerequisite of abundant historical data the rate of change of state evaluation while predicting that according to formula below i+1 carries out repair based on condition of component; Wherein, a is a constant, and while being used for identifying the last repair based on condition of component, state evaluation rate of change is to the influence degree of state evaluation next time;

$q_{i+1}={\mathrm{aq}}_i+{\mathrm{\Σ}}_{j=1}^n{(1-a)}^j{q}_{i-j}$

If establish a=1/2, can obtain:

$q_{i+1}=q_i/2+{\mathrm{\Σ}}_{j=1}^n{(1/2)}^j{q}_{i-j}$

If establishing the threshold value of warning of this equipment state is V, require the time of i+1 next state maintenance to meet:

v≤R _i-q _i+1t _i+1

Change the formula that can draw below, prediction should be at t _i+1in time, arrange the repair based on condition of component next time to this equipment;

$t_{i+1}\≤\frac{R_i-V}{q_{i+1}};$

S4 adjusts judgment matrix, again scoring

If final score, in actual empirical value interval, does not return to S2, according to actual conditions, adjust the value of the importance degree contrast of two parameters under same module, until final score meets the interval of empirical value.

Images