Power operation and maintenance alarm data fusion method based on evidence theory

Abstract

The present invention relates to the technical field of power operation and maintenance alarm and more particularly relates to a power operation and maintenance alarm data fusion method based on the evidence theory. The method comprises the steps of (S1) obtaining basic probability allocation BPA according to an evidence confidence level, and (S2) fusing data to obtain a fusion result according toan improved fusion rule. Through introducing the evidence confidence level and using a Bhattacharyya distance, the BPA is adjusted, at the same time, a D-S combination rule is improved, and a new fusion rule is proposed to fuse the data. According to the method, a problem of evidence conflict in the D-S evidence theory is solved well, information contained in conflict evidence is not ignored, theaccuracy of the fusion result is improved, and a fault diagnosis speed and fault handling efficiency of a power communication network are improved.

Description

technical field

[0001] The present invention relates to the technical field of electric power operation and maintenance alarms, and more specifically, relates to a data fusion method for electric power operation and maintenance alarms based on evidence theory. Background technique

[0002] At present, the power system communication network has become an indispensable part of the intelligent dispatching and modern management of the power system. The technologies of alarm collection, warehousing and classification in the alarm mechanism are relatively mature, but there are still many new methods for alarm data fusion technology. Alarm data fusion is an important task for fault diagnosis of power communication network, and it is of great significance to the management and maintenance of power communication network. The power communication network has the characteristics of a general communication network. The data collected by the system is delayed, incomplete and inaccurate. ...

Examples

Embodiment 1

[0038] Example 1: Complete Conflict Evidence:

[0039] Assuming that the identification framework is Θ={A, B, C}, there are four evidence source data, and BPA is expressed as follows:

[0040]

[0041] Among them, the evidence m 1 and m 2 for complete conflict evidence. In the traditional DS combination rules, this situation cannot be calculated. However, according to the calculation method in this paper, the fusion result can be obtained as follows:

[0042] m(A)=0.912, m(B)=0.044, m(C)=0.044.

Embodiment 2

[0043] Example 2: Zero Trust Paradox:

[0044] Assuming that the identification framework is Θ={A, B, C}, there are four evidence source data, and BPA is expressed as follows:

[0045]

[0046] Through calculation, the total conflict coefficient of the four evidences is k=0.99, and the DS combination rule is used for fusion, and the fusion result is:

[0047] m(A)=0, m(B)=0.727, m(C)=0.273,

[0048] Evidence m 2 Completely negate proposition A, so no matter how much other pieces of evidence support proposition A, the support degree of proposition A in the composite result obtained is always 0. However, through the method of this paper for fusion, the obtained fusion result is:

[0049] m(A)=0.5153, m(B)=0.2363, m(C)=0.2484.

Embodiment 3

[0050] Embodiment 3: A paradox of trust:

[0051] Assuming that the identification framework is Θ={A, B, C}, there are four evidence source data, and BPA is expressed as follows:

[0052]

[0053] By calculation, according to The total conflict coefficient of the four pieces of evidence is k=0.9998, and the DS combination rule is used for fusion, and the fusion result is:

[0054] m(A)=0, m(B)=1, m(C)=0,

[0055] It can be seen that the BPA assigned to Proposition B by all the evidence is very small, but the result obtained is that Proposition B is fully supported, which is contrary to common sense in practical applications. Through the method of this paper, the fusion results are as follows:

[0056] m(A)=0.070, m(B)=0.061, m(C)=0.869.

technical field

[0001] The present invention relates to the technical field of electric power operation and maintenance alarms, and more specifically, relates to a data fusion method for electric power operation and maintenance alarms based on evidence theory. Background technique

[0002] At present, the power system communication network has become an indispensable part of the intelligent dispatching and modern management of the power system. The technologies of alarm collection, warehousing and classification in the alarm mechanism are relatively mature, but there are still many new methods for alarm data fusion technology. Alarm data fusion is an important task for fault diagnosis of power communication network, and it is of great significance to the management and maintenance of power communication network. The power communication network has the characteristics of a general communication network. The data collected by the system is delayed, incomplete and inaccurate. ...

Examples

Embodiment 1

[0038] Example 1: Complete Conflict Evidence:

[0039] Assuming that the identification framework is Θ={A, B, C}, there are four evidence source data, and BPA is expressed as follows:

[0040]

[0041] Among them, the evidence m 1 and m 2 for complete conflict evidence. In the traditional DS combination rules, this situation cannot be calculated. However, according to the calculation method in this paper, the fusion result can be obtained as follows:

[0042] m(A)=0.912, m(B)=0.044, m(C)=0.044.

Embodiment 2

[0043] Example 2: Zero Trust Paradox:

[0044] Assuming that the identification framework is Θ={A, B, C}, there are four evidence source data, and BPA is expressed as follows:

[0045]

[0046] Through calculation, the total conflict coefficient of the four evidences is k=0.99, and the DS combination rule is used for fusion, and the fusion result is:

[0047] m(A)=0, m(B)=0.727, m(C)=0.273,

[0048] Evidence m 2 Completely negate proposition A, so no matter how much other pieces of evidence support proposition A, the support degree of proposition A in the composite result obtained is always 0. However, through the method of this paper for fusion, the obtained fusion result is:

[0049] m(A)=0.5153, m(B)=0.2363, m(C)=0.2484.

Embodiment 3

[0050] Embodiment 3: A paradox of trust:

[0051] Assuming that the identification framework is Θ={A, B, C}, there are four evidence source data, and BPA is expressed as follows:

[0052]

[0053] By calculation, according to The total conflict coefficient of the four pieces of evidence is k=0.9998, and the DS combination rule is used for fusion, and the fusion result is:

[0054] m(A)=0, m(B)=1, m(C)=0,

[0055] It can be seen that the BPA assigned to Proposition B by all the evidence is very small, but the result obtained is that Proposition B is fully supported, which is contrary to common sense in practical applications. Through the method of this paper, the fusion results are as follows:

[0056] m(A)=0.070, m(B)=0.061, m(C)=0.869.

Images