阅读说明：本技术 基于群体智能的液压支架性能退化定量评估方法及系统 (Group intelligence-based hydraulic support performance degradation quantitative evaluation method and system ) 是由 王金鑫 王恩元 李忠辉 沈荣喜 于 2021-06-08 设计创作，主要内容包括：本发明提出了一种基于群体智能的液压支架性能退化定量评估方法及系统,该方法首先实时采集液压支架的运行数据,之后利用标准化公式对运行数据进行标准化处理以获取运行数据的标准化特征值,然后利用群体智能聚类算法分析聚类对象,输出聚类结果聚类完成后,性能状态相近的液压支架将被分入同一簇中。在此基础上,计算与其它支架性能差异较大的液压支架的离群因子,最大离群因子对应的液压支架即为故障液压支架。本发明能够在无样本数据情况下,准确查找到性能异常的液压支架,避免了现有方法对故障样本的依赖。(The invention provides a group intelligence-based hydraulic support performance degradation quantitative evaluation method and system. On the basis, the outlier factor of the hydraulic support with larger performance difference with other supports is calculated, and the hydraulic support corresponding to the largest outlier factor is the fault hydraulic support. The method can accurately find the hydraulic support with abnormal performance under the condition of no sample data, and avoids the dependence of the existing method on the fault sample.)

1. A group intelligence-based hydraulic support performance degradation quantitative evaluation method is characterized by comprising the following steps:

s1, acquiring the operation data of the hydraulic support in real time: the type of operational data collected includes front strut emulsion pressure X₁Front pillar emulsion temperature X₂Pressure X of emulsion on rear support₃Pressure X of emulsion liquid of front extension beam₄Pressure of advancing emulsion X₅Specific pressure to the base plate X₆；

S2, standardizing the operation data by using a standardization formula to obtain a standardized characteristic value of the operation data;

s3, analyzing the clustering objects by using a group intelligent clustering algorithm, and outputting a clustering result:

taking the standardized characteristic value of each operation data in the S2 as the attribute of the hydraulic support, taking the operation state of a single hydraulic support in the hydraulic support group as a clustering object, and carrying out clustering analysis on a plurality of hydraulic supports in the hydraulic support group by using a group intelligent clustering algorithm: first using the formulaComputing any two clustering objects O_iAnd O_jThe Manhattan distance between the hydraulic supports, p is the number of the hydraulic supports,as a clustering object O_iK-th operation data X of_k(k＝1,2,3,4,5,6)，As a clustering object O_jK-th clustering object X of_k(k＝1,2,3,4,5,6)；

Then using the formulaComputing a clustering object O_iAverage similarity to neighborhood-wide clustered objects f (O)_i) R is polyClass object O_iN is the number of other clustering objects in the neighborhood, and neighbor (r) is the set of other clustering objects in the neighborhood;

then according to the formulaComputing ant picking up clustering object O_iProbability P of_p(O_i)，k₁Is a threshold constant; according to the formulaComputing ant drop clustering object O_iProbability of (k)₂The clustering result is a threshold constant, and the clustering result is finally output to realize the automatic separation of the fault hydraulic support;

s4, on the basis of the clustering result, calculating the outlier factor of each hydraulic support by using an outlier factor calculation formula, and finally outputting the hydraulic support corresponding to the largest outlier factor as a fault hydraulic support;

the outlier factor calculation refers to calculating the deviation degree of the hydraulic support and other hydraulic supports, and an outlier factor calculation formula is adoptedt is an outlier cluster object, O_iAs non-outlier cluster object, d (t, O)_i) Is t and O_iM is the number of non-outlier clustering objects; CBGOF is an outlier of t.

2. The swarm intelligence-based hydraulic support performance degradation quantitative evaluation method of claim 1, wherein the operation data normalization process is to scale the operation data of the hydraulic support to fall into a specific interval, remove unit limitation of the data, and convert it into a dimensionless pure value: adopting a data 0-1 standardization method; wherein the standardized formula is defined as

X_i(i is 1,2,3,4,5,6) is collected operation data of the hydraulic support,for the standardized hydraulic support operation data,for hydraulic support operating data X_iAverage value of (1), X_i(max)For hydraulic support operating data X_iMaximum value of (A), X_i(min)For hydraulic support operating data X_iIs measured.

3. A group intelligence-based quantitative evaluation system for performance degradation of a hydraulic support is characterized by comprising:

the monitoring data acquisition module is used for acquiring the operating data of the hydraulic support;

the standardization processing module is used for standardizing the acquired operation data to acquire a standardized characteristic value of the operation data;

the clustering module is used for taking the standardized characteristic value of each operating data as the attribute of the hydraulic support, taking the operating state of a single hydraulic support in the hydraulic support group as a clustering object, carrying out clustering analysis on the hydraulic support group based on a group intelligent clustering algorithm, and finally outputting a clustering result to the outlier factor acquisition module;

the outlier factor acquisition module is used for calculating outlier factors of all the hydraulic supports according to the clustering result and outputting the outlier factors to the fault result output module;

and the fault result output module selects the maximum outlier factor and outputs the related information of the hydraulic support corresponding to the maximum outlier factor.

Technical Field

The invention belongs to a fault prediction method of a hydraulic support, and particularly relates to a group intelligence-based hydraulic support performance degradation quantitative evaluation method and system.

Background

In recent years, with the transition of national economy from a high-speed growth stage to a high-quality development stage, the requirements of the society on the safe, green and intelligent development of the coal industry are continuously increased. To meet this demand, researchers have proposed the concept of "smart mine". The intelligent mine aims to realize self-management and self-operation of links such as mine tunneling, mining, transportation and lifting through the deep fusion of artificial intelligence and information technology and the mine production process. The construction of smart mines involves a number of issues. Intelligent health management of mine full-station equipment facilities is an essential ring. In all equipment facilities of mine production, the hydraulic support has important functions of supporting a coal rock roof, forming a coal mining working space and the like, and is core equipment for comprehensive mechanized coal mining. The performance perception and evaluation of the hydraulic support are important contents for building an intelligent mine.

Currently, some methods have been proposed by researchers in the health assessment and management of hydraulic stents. The method is characterized in that the method comprises the steps of analyzing actual measurement fault samples or simulation data of the hydraulic support to obtain an operation data threshold value, and then comparing the current operation data of the hydraulic support with a preset threshold value to evaluate the working state and the performance degradation condition of the hydraulic support. The method has a good application effect on certain faults of the hydraulic support. However, for faults such as liquid crossing in the emulsion, deformation of the upright column, abrasion of the inner wall of the hydraulic cylinder and the like, actually measured fault sample data is very difficult to implement in engineering, the reliability of simulation data is often difficult to guarantee, and meanwhile, due to the fact that the working conditions of the hydraulic support are varied, the universality of mastered fault samples is limited. Therefore, the lack of fault sample data becomes a bottleneck limiting the development of the state evaluation technology of the hydraulic bracket.

The document retrieval of the prior art finds that a publication of 'an early warning method for the working condition of a hydraulic support' (Chinese invention patent, publication number: CN112145231A, published Japanese 2020.12.29) provides a hydraulic support fault detection early warning method, and the publication comprises the following steps: the application discloses a hydraulic support working condition early warning method, and relates to the technical field of coal mine safety mining, wherein the method comprises the following steps: monitoring the working resistance data of all hydraulic supports on the coal mine working face in real time; determining index values of a plurality of evaluation indexes for evaluating the working condition of the hydraulic support according to the working resistance change data in each coal mining cycle; and generating early warning information according to the index value of the evaluation index. Through the technical scheme, the working resistance data of all hydraulic supports of the coal mine working face are monitored in real time, a plurality of evaluation indexes for comprehensively evaluating the working condition of the hydraulic supports can be determined through the analysis of the working resistance data of the hydraulic supports in the stoping process of the coal mine fully mechanized mining working face, so that the real-time monitoring, analysis and early warning of the working condition of the hydraulic supports are realized, the problems existing in the middle of support supporting and roof management are timely positioned, and the important function is realized on the safety of the roof of the fully mechanized mining working face. The disadvantages are as follows: according to the method, the working state of the hydraulic support is judged by comparing the collected hydraulic support operation data with a preset standard value, and the setting of the standard value of the operation data usually needs to comprehensively analyze the operation data change of the hydraulic support in a normal state and a fault state, so that the implementation of the method is limited by a fault sample, the application condition is high, and the range is limited.

Disclosure of Invention

The invention aims to provide a group-intelligence-based hydraulic support performance degradation quantitative evaluation method and system, and provides a hydraulic support performance evaluation method independent of fault samples. In order to achieve the purpose, the invention adopts the following technical scheme:

a group intelligence-based hydraulic support performance degradation quantitative evaluation method comprises the following steps:

s1, acquiring the operation data of the hydraulic support in real time: the type of operational data collected includes front strut emulsion pressure X₁Front pillar emulsion temperature X₂Pressure X of emulsion on rear support₃Pressure X of emulsion liquid of front extension beam₄Pressure of advancing emulsion X₅Specific pressure to the base plate X₆；

S2, standardizing the operation data by using a standardization formula to obtain a standardized characteristic value of the operation data;

s3, analyzing the clustering objects by using a group intelligent clustering algorithm, and outputting a clustering result:

taking the standardized characteristic value of each operation data in the S2 as the attribute of the hydraulic support, taking the operation state of a single hydraulic support in the hydraulic support group as a clustering object, and carrying out clustering analysis on a plurality of hydraulic supports in the hydraulic support group by using a group intelligent clustering algorithm: first using the formulaComputing any two clustering objects O_iAnd O_jThe Manhattan distance between the hydraulic supports, p is the number of the hydraulic supports,as a clustering object O_iK-th operation data X of_k(k＝1,2,3,4,5,6)，As a clustering object O_jK-th clustering object X of_k(k＝1,2,3,4,5,6)；

Then using the formulaComputing a clustering object O_iAverage similarity to neighborhood-wide clustered objects f (O)_i) R is a clustering object O_iN is the number of other clustering objects in the neighborhood, and neighbor (r) is the set of other clustering objects in the neighborhood;

then according to the formulaComputing ant picking up clustering object O_iProbability P of_p(O_i)， k₁Is a threshold constant; according to the formulaComputing ant drop clustering object O_iProbability of (k)₂The clustering result is a threshold constant, and the clustering result is finally output to realize the automatic separation of the fault hydraulic support;

s4, on the basis of the clustering result, calculating the outlier factor of each hydraulic support by using an outlier factor calculation formula, and finally outputting the hydraulic support corresponding to the largest outlier factor as a fault hydraulic support;

the outlier factor calculation refers to calculating the deviation degree of the hydraulic support and other hydraulic supports, and an outlier factor calculation formula is adoptedt is an outlier cluster object, O_iAs non-outlier cluster object, d (t, O)_i) Is t and O_iM is the number of non-outlier clustering objects; CBGOF is an outlier of t.

Preferably, the operation data standardization process is to scale the operation data of the hydraulic support to fall into a specific interval, remove unit limitation of the data, and convert the operation data into dimensionless pure values: adopting a data 0-1 standardization method; wherein the standardized formula is defined as

X_i(i is 1,2,3,4,5,6) is collected hydraulic support operation data,for the standardized hydraulic support operation data,for hydraulic support operating data X_iAverage value of (1), X_i(max)For hydraulic support operating data X_iMaximum value of (A), X_i(min)For hydraulic support operating data X_iIs measured.

A group intelligence-based hydraulic support performance degradation quantitative evaluation system comprises:

the monitoring data acquisition module is used for acquiring the operating data of the hydraulic support;

the standardization processing module is used for standardizing the acquired operation data to acquire a standardized characteristic value of the operation data;

the clustering module is used for taking the standardized characteristic value of each operating data as the attribute of the hydraulic support, taking the operating state of a single hydraulic support in the hydraulic support group as a clustering object, carrying out clustering analysis on the hydraulic support group based on a group intelligent clustering algorithm, and finally outputting a clustering result to the outlier factor acquisition module;

the outlier factor acquisition module is used for calculating outlier factors of all the hydraulic supports according to the clustering result and outputting the outlier factors to the fault result output module;

and the fault result output module selects the maximum outlier factor and outputs the related information of the hydraulic support corresponding to the maximum outlier factor.

Compared with the prior art, the invention has the advantages that:

(1) according to the invention, the hydraulic supports with abnormal performance are separated through the self-organizing transverse comparison of the performance states of the hydraulic support groups, historical operation data of the hydraulic supports are not needed in the whole process, and the dependence of the prior art on fault samples is overcome.

(2) By applying the clustering analysis method, the key problem of missing of the fault sample is effectively solved without depending on the fault sample data, and the problem of poor universality of the fault samples of hydraulic supports of different models and structures is avoided.

(3) The concept of the hydraulic support 'outlier factor' is provided, and the defect that the fault degree of the hydraulic support cannot be quantitatively described by the traditional diagnosis method is overcome.

(4) Through parameter dimensionless, firstly, transverse comparison and calculation of different types of data are realized, the condition that the fault diagnosis result of the hydraulic support is influenced by the overlarge numerical value of a certain type of parameter is avoided, and secondly, the defect that the accuracy of the fault diagnosis method of the hydraulic support is difficult to guarantee when a large amount of data are dealt with is overcome.

(5) According to the invention, quantitative description of fault degree is realized by comparing and analyzing the outlier factors in the normal state and the abnormal state, so that targeted maintenance of the hydraulic support can be realized, the occurrence of faults is avoided, and the reliability and the economical efficiency of equipment operation are improved.

Drawings

FIG. 1 is a flow chart of the group intelligence-based hydraulic support performance degradation quantitative evaluation method.

Fig. 2 is a schematic diagram of the clustering result output in S3.

Detailed Description

The present invention will now be described in more detail with reference to the accompanying schematic drawings, in which preferred embodiments of the invention are shown, it being understood that one skilled in the art may modify the invention herein described while still achieving the advantageous effects of the invention. Accordingly, the following description should be construed as broadly as possible to those skilled in the art and not as limiting the invention.

As shown in FIG. 1, the quantitative evaluation method for hydraulic support performance degradation based on group intelligence comprises steps S1-S4.

And S1, acquiring the operation data of the hydraulic support in real time.

The operation data of the hydraulic support is acquired through the cooperative work of the sensor and the data acquisition card (the data acquisition card acquires the operation data acquired by the sensor). Wherein the type of operational data collected comprises front pillar emulsificationHydraulic pressure X₁Front pillar emulsion temperature X₂Pressure X of emulsion on rear support₃Pressure X of emulsion liquid of front extension beam₄Pressure of advancing emulsion X₅Specific pressure to the base plate X₆；

And S2, normalizing the operation data by using a normalization formula to obtain a normalized characteristic value of the operation data.

The steps have the following functions: the collected hydraulic support operation data are different in dimension, so that the numerical value of the collected hydraulic support operation data is often greatly different, and further the clustering result is influenced. The standardization process can convert the operation data with dimensions into dimensionless pure numerical values, thereby avoiding the influence of different dimensions on the clustering.

Specifically, the operation data standardization processing is to scale the operation data of the hydraulic support to make the operation data fall into a specific interval, remove the unit limitation of the data, and convert the operation data into a dimensionless pure numerical value: and converting the operation data into data in a closed interval [0,1] by adopting a data 0-1 standardization method.

Wherein, the standardized formula is shown in formula (1):

X_i(i is 1,2,3,4,5,6) is collected hydraulic support operation data,for the standardized hydraulic support operation data,for hydraulic support operating data X_iAverage value of (1), X_i(max)For hydraulic support operating data X_iMaximum value of (A), X_i(min)For hydraulic support operating data X_iIs measured.

And S3, analyzing the clustering objects by using a swarm intelligence clustering algorithm, and outputting a clustering result, wherein the clustering result is shown in figure 2.

And taking the standardized characteristic value of each operation data in the S2 as the attribute of the hydraulic support, taking the operation state of a single hydraulic support in the hydraulic support group as a clustering object, and carrying out clustering analysis on a plurality of hydraulic supports in the hydraulic support group by using an ant colony clustering algorithm.

Specifically, first, a formula is usedComputing any two clustering objects O_iAnd O_jThe Manhattan distance between the hydraulic supports, p is the number of the hydraulic supports,as a clustering object O_iK-th operation data X of_k(k＝1,2,3,4,5,6)，As a clustering object O_jK-th clustering object X of_k(k＝1,2,3,4,5,6)；

Then using the formulaComputing a clustering object O_iAverage similarity to neighborhood-wide clustered objects f (O)_i) R is a clustering object O_iN is the number of other clustering objects in the neighborhood, and neighbor (r) is the set of other clustering objects in the neighborhood;

then according to the formulaComputing ant picking up clustering object O_iProbability P of_p(O_i)， k₁Is a threshold constant; according to the formulaComputing ant drop clustering object O_iProbability of (k)₂And finally outputting a clustering result to realize the automatic separation of the fault hydraulic support as a threshold constant.

The performance states of the hydraulic support are transversely compared by adopting an ant colony clustering algorithm, namely clustering objects are randomly distributed in a two-dimensional grid, a certain number of ants are randomly arranged in the two-dimensional grid, and then the ants move one grid around randomly in each cycle, which is divided into the following conditions:

A. when the moved grid contains the clustering object and the ant does not carry other clustering objects currently, the probability P of the ant picking up the clustering object is calculated by adopting the formula (2)_p(O_i)。

The meaning of formula (2) is: clustering object O_iThe lower the average similarity with the clustered objects in the neighborhood range, the less the ants pick up the clustered objects O_iThe greater the probability of (c).

B. And when the moved grid does not contain the clustering object and the ant carries other clustering objects currently, calculating the probability that the ant puts down the clustering object to the current grid by adopting the formula (3).

Thus, the meaning of formula (3) is: clustering object O_iThe higher the average similarity with the clustered objects in the neighborhood range, the lower the ant puts down the clustered object O_iThe greater the probability of (c).

C. In other cases, no operation is performed.

After ants move repeatedly, the hydraulic supports are self-organized and clustered according to the performance states of the hydraulic supports, and the hydraulic supports with similar performance states are clustered in the same cluster.

And S4, separating the fault hydraulic support.

Specifically, on the basis of a clustering result, an outlier factor calculation formula is utilized to calculate the outlier factor of each hydraulic support, and finally, the hydraulic support corresponding to the largest outlier factor is output as a fault hydraulic support to describe the performance deviation degree of the corresponding hydraulic support.

The calculation of the outlier factor refers to calculating the deviation degree of the hydraulic support and other hydraulic supports, and an outlier factor calculation formula is adoptedt is an outlier cluster object, O_iAs non-outlier cluster object, d (t, O)_i) Is t and O_iM is the number of non-outlier clustering objects; CBGOF is an outlier of t.

The judgment method of the outlier clustering object and the non-outlier clustering object is that if no other clustering object exists in the neighborhood range, the clustering object is the outlier clustering object, otherwise, the clustering object is the non-outlier clustering object. And the hydraulic support with the maximum output outlier factor is a fault hydraulic support.

In fig. 2, (a) is a case where the clustering objects are randomly distributed in the two-dimensional grid before clustering, the black solid squares represent the faulty hydraulic support, and the hollow squares represent the normal hydraulic support. In fig. 2, (b) shows the effect after clustering. Therefore, the ant colony algorithm can effectively realize clustering of the hydraulic supports according to the performance of the hydraulic supports, and the fault hydraulic supports are separated out independently.

Based on the foregoing concept, the present embodiment further provides a group intelligence-based quantitative evaluation system for performance degradation of a hydraulic support, which includes:

the monitoring data acquisition module is used for acquiring the operating data of the hydraulic support;

the standardization processing module is used for standardizing the acquired operation data to acquire a standardized characteristic value of the operation data;

the clustering module is used for taking the standardized characteristic value of each operating data as the attribute of the hydraulic support, taking the operating state of a single hydraulic support in the hydraulic support group as a clustering object, carrying out clustering analysis on the hydraulic support group based on a group intelligent clustering algorithm, and finally outputting a clustering result to the outlier factor acquisition module;

the outlier factor acquisition module is used for calculating outlier factors of all the hydraulic supports according to the clustering result and outputting the outlier factors to the fault result output module;

and the fault result output module selects the maximum outlier factor and outputs the related information of the hydraulic support corresponding to the maximum outlier factor.

In summary, the working principle of the evaluation method and system is as follows:

in the prior art, dozens or even hundreds of hydraulic supports are arranged on one coal face, and all the hydraulic supports have consistent functions and same structures and work in cooperation with each other to jointly form a reliable coal mining space. In a normal state, the difference of the operation data of different hydraulic supports is not obvious; when a hydraulic support fails, the performance state of the hydraulic support is degraded, and the operation data of the hydraulic support is greatly deviated from other hydraulic supports. Therefore, the hydraulic support with the fault can be effectively separated by 'transversely comparing' the operation data of the hydraulic support group. The swarm intelligence technology is an effective algorithm for seeking a distributed problem and a combined optimization problem in the fields of engineering technology and the like by simulating complex intelligent behaviors expressed by ants, bees and other individuals in daily interaction and cooperation. According to the structural function and application characteristics of the hydraulic support group, the hydraulic support group is abstracted to be a group, a single hydraulic support is regarded as an independent individual, the operation data of the hydraulic support is regarded as individual attributes, and the performance states of the hydraulic supports are transversely compared based on an ant colony clustering algorithm, so that the hydraulic support group is self-organized and clustered according to the performance states. After clustering is finished, the hydraulic support with the largest outlier degree is the support with the fault, and therefore intelligent identification of the fault hydraulic support under a fault-free sample is achieved.

The above description is only a preferred embodiment of the present invention, and does not limit the present invention in any way. It will be understood by those skilled in the art that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.