阅读说明：本技术 一种轧钢工艺流程复合故障追溯方法 (Steel rolling process flow composite fault tracing method ) 是由 马亮 杨萍萍 彭开香 董洁 于 2021-09-07 设计创作，主要内容包括：本发明提供一种轧钢工艺流程复合故障追溯方法,属于生产过程的控制和监测技术领域。所述方法包括：构建轧钢工艺流程复合故障模式分类器；基于构建的轧钢工艺流程复合故障模式分类器实现层次化的复合故障追溯。采用本发明,能够在发生复合故障时及时诊断和推理辨识出故障的根本原因。(The invention provides a steel rolling process flow composite fault tracing method, and belongs to the technical field of control and monitoring of a production process. The method comprises the following steps: constructing a steel rolling process flow composite fault mode classifier; and the hierarchical composite fault tracing is realized based on the constructed steel rolling process flow composite fault mode classifier. By adopting the invention, the root cause of the fault can be diagnosed and reasoned in time when the composite fault occurs.)

1. A steel rolling process flow composite fault tracing method is characterized by comprising the following steps:

constructing a steel rolling process flow composite fault mode classifier;

and the hierarchical composite fault tracing is realized based on the constructed steel rolling process flow composite fault mode classifier.

2. The method for tracing the composite fault of the steel rolling process flow according to claim 1, wherein the composite fault mode comprises: propagation type, coupling type, multiple concurrent type, and composite type.

3. The method for tracing the composite fault of the steel rolling technological process according to claim 1, wherein the step of constructing the composite fault mode classifier of the steel rolling technological process comprises the following steps:

analyzing the correlation between propagation type, coupling type, multiple parallel type and compound type fault data and normal data by using a correlation analysis method, and determining a characteristic subspace of a compound fault mode of a steel rolling process flow according to a correlation analysis result;

and identifying and classifying the obtained feature subspaces by using the steel rolling process expert experience and process knowledge, marking a composite fault mode of the fault data and the normal data according to the classification result, and constructing a steel rolling process flow composite fault mode classifier capable of reflecting the composite fault mode based on the marking result and the corresponding feature subspaces, wherein the composite fault mode of the normal data is marked as none.

4. The method for tracing the complex fault of the steel rolling process flow according to claim 3, wherein the step of analyzing the correlation between the propagation type, the coupling type, the multiple concurrent type and the complex fault data and the normal data by using the correlation analysis method comprises the following steps of:

analyzing the correlation between propagation type, coupling type, multiple concurrent type and compound fault data and normal data, and extracting data characteristics capable of reflecting the fault characteristics of the steel rolling process flow according to the correlation analysis result to obtain a characteristic vector of a compound fault mode of the steel rolling process flow;

and screening the obtained feature vectors by using a similarity analysis method to obtain a feature subspace of a composite fault mode, wherein the change amplitude of the similarity index is in a preset interval and has the fault characteristic of the steel rolling process flow.

5. The steel rolling process flow composite fault tracing method according to claim 1, wherein said layering refers to: from the full flow to the subsystem.

6. The method for tracing composite faults of steel rolling technological processes according to claim 1, wherein the step of implementing hierarchical composite fault tracing based on the constructed composite fault mode classifier of the steel rolling technological processes comprises the following steps:

determining a composite fault mode to which steel rolling process flow data belongs by using the constructed steel rolling process flow composite fault mode classifier, and extracting an explicit optimal feature projection matrix and a implicit optimal feature projection matrix according to a feature subspace corresponding to each composite fault mode and explicit and implicit states of composite fault symptoms of the steel rolling process flow in time and space;

and weighting the extracted dominant and recessive optimal feature projection matrixes to obtain a comprehensive feature projection matrix, constructing a composite fault initial tracing model according to the obtained comprehensive feature projection matrix, and realizing the layered steel rolling process flow composite fault tracing according to the constructed composite fault initial tracing model and the composite fault full-flow and subsystem detection results.

7. The method for tracing the composite fault of the steel rolling process flow according to claim 6, wherein the method for extracting the dominant optimal feature projection matrix and the recessive optimal feature projection matrix comprises the steps of determining the composite fault mode to which the steel rolling process flow data belongs by using the constructed steel rolling process flow composite fault mode classifier, and extracting the dominant optimal feature projection matrix and the recessive optimal feature projection matrix according to the feature subspace corresponding to each composite fault mode and the dominant and recessive states of the composite fault symptom of the steel rolling process flow in time and space:

determining a composite fault mode to which steel rolling process flow data to be analyzed belong by using a constructed steel rolling process flow composite fault mode classifier, projecting a feature subspace corresponding to each composite fault mode to a dominant correlation mode of a fault sign aiming at a dominant state of the steel rolling process flow composite fault sign presented in time and space, and extracting a dominant optimal feature projection matrix by using an index discrimination analysis method according to the obtained dominant correlation mode;

aiming at the hidden states of the composite fault symptoms of the steel rolling process flow in time and space, projecting the feature subspace corresponding to each composite fault mode to the hidden correlation mode of the fault symptoms, and extracting a hidden optimal feature projection matrix by using a multitask feature selection and causal relationship analysis method according to the obtained hidden correlation mode.

8. The method for tracing composite faults of steel rolling process flows according to claim 1, wherein the method for tracing composite faults of the steel rolling process flows based on the built composite fault mode classifier further comprises the following steps:

when a new fault occurs, a dominant correlation mode and a recessive correlation mode between the composite fault and the fault symptom are established by using new fault data, and a feature subspace corresponding to the new fault is projected onto the dominant correlation mode and the recessive correlation mode, so that the self-adaptive updating of the initial tracing model of the composite fault is realized.

Technical Field

The invention relates to the technical field of control and monitoring of a production process, in particular to a composite fault tracing method for a steel rolling process flow.

Background

In recent years, the abnormal operation state of the hot rolling process flow is generally a bottom-up development mode of the abnormal state of subsystems caused by the composite fault of a bottom layer loop, so that the abnormal state of other subsystems and the whole flow is further caused. Correspondingly, the compound fault tracing is a 'top-down' process for tracing the fault reason of the bottom layer loop from the abnormal state of the whole flow, and the compound fault has a wide potential distribution range and hysteresis of fault symptom expression, so that the research of the compound fault tracing problem is challenging. Therefore, the composite fault tracing technology is researched based on the composite fault autonomous detection result, the root cause of the fault is timely diagnosed and inferred and identified when the operation state is abnormal, and the method has important theoretical and engineering significance for guaranteeing the safety and stability of the operation of the hot rolling process flow.

The steel rolling process flow mainly comprises a plurality of production processes such as heating, rough rolling, flying shears, finish rolling and the like, and a long product processing flow taking a series structure as a main body is formed from a raw material to a final product; meanwhile, the corresponding integrated automation system has obvious hierarchy and mainly comprises an equipment layer, a real-time control layer, a process control layer, a manufacturing execution layer and the like, as shown in fig. 1. The work of each layer is definite and is in mutual cooperation and association, and in addition, the raw material components, equipment states, process parameters and the like of the layers are difficult to perceive in real time or comprehensively, so that the analysis of the safety and the stability of the layers is complex and changeable, and the fault propagation and even evolution can be caused by the abnormality of any one or more links, thereby causing the production halt and the maintenance of enterprises due to the return of goods of quality objection users and influencing the economic benefit of the enterprises. However, in the prior art, the root cause of the fault cannot be diagnosed and identified in time and in an inference mode when the compound fault occurs.

Disclosure of Invention

The embodiment of the invention provides a steel rolling process flow composite fault tracing method, which can timely diagnose and reason and identify the root cause of a fault when the composite fault occurs. The technical scheme is as follows:

the embodiment of the invention provides a steel rolling process flow composite fault tracing method, which comprises the following steps:

constructing a steel rolling process flow composite fault mode classifier;

and the hierarchical composite fault tracing is realized based on the constructed steel rolling process flow composite fault mode classifier.

Further, the composite failure mode includes: propagation type, coupling type, multiple concurrent type, and composite type.

Further, the constructing of the steel rolling process flow composite fault mode classifier comprises:

analyzing the correlation between propagation type, coupling type, multiple parallel type and compound type fault data and normal data by using a correlation analysis method, and determining a characteristic subspace of a compound fault mode of a steel rolling process flow according to a correlation analysis result;

and identifying and classifying the obtained feature subspaces by using the steel rolling process expert experience and process knowledge, marking a composite fault mode of the fault data and the normal data according to the classification result, and constructing a steel rolling process flow composite fault mode classifier capable of reflecting the composite fault mode based on the marking result and the corresponding feature subspaces, wherein the composite fault mode of the normal data is marked as none.

Further, the analyzing the correlation between the propagation type, coupling type, multiple concurrent type and composite type fault data and the normal data by using the correlation analysis method, and the determining the characteristic subspace of the composite fault mode of the steel rolling process flow according to the correlation analysis result comprises the following steps:

analyzing the correlation between propagation type, coupling type, multiple concurrent type and compound fault data and normal data, and extracting data characteristics capable of reflecting the fault characteristics of the steel rolling process flow according to the correlation analysis result to obtain a characteristic vector of a compound fault mode of the steel rolling process flow;

and screening the obtained feature vectors by using a similarity analysis method to obtain a feature subspace of a composite fault mode, wherein the change amplitude of the similarity index is in a preset interval and has the fault characteristic of the steel rolling process flow.

Further, the hierarchy refers to: from the full flow to the subsystem.

Further, the composite fault tracing for realizing layering based on the constructed steel rolling process flow composite fault mode classifier comprises the following steps:

determining a composite fault mode to which steel rolling process flow data belongs by using the constructed steel rolling process flow composite fault mode classifier, and extracting an explicit optimal feature projection matrix and a implicit optimal feature projection matrix according to a feature subspace corresponding to each composite fault mode and explicit and implicit states of composite fault symptoms of the steel rolling process flow in time and space;

and weighting the extracted dominant and recessive optimal feature projection matrixes to obtain a comprehensive feature projection matrix, constructing a composite fault initial tracing model according to the obtained comprehensive feature projection matrix, and realizing the layered steel rolling process flow composite fault tracing according to the constructed composite fault initial tracing model and the composite fault full-flow and subsystem detection results.

Further, the determining a composite fault mode to which the steel rolling process flow data belongs by using the constructed steel rolling process flow composite fault mode classifier, and extracting an explicit optimal feature projection matrix and a implicit optimal feature projection matrix according to the feature subspace corresponding to each composite fault mode and explicit and implicit states of the steel rolling process flow composite fault symptom presented in time and space comprises:

determining a composite fault mode to which steel rolling process flow data to be analyzed belong by using a constructed steel rolling process flow composite fault mode classifier, projecting a feature subspace corresponding to each composite fault mode to a dominant correlation mode of a fault sign aiming at a dominant state of the steel rolling process flow composite fault sign presented in time and space, and extracting a dominant optimal feature projection matrix by using an index discrimination analysis method according to the obtained dominant correlation mode;

aiming at the hidden states of the composite fault symptoms of the steel rolling process flow in time and space, projecting the feature subspace corresponding to each composite fault mode to the hidden correlation mode of the fault symptoms, and extracting a hidden optimal feature projection matrix by using a multitask feature selection and causal relationship analysis method according to the obtained hidden correlation mode.

Further, the composite fault tracing for realizing layering based on the constructed steel rolling process flow composite fault mode classifier further comprises:

when a new fault occurs, a dominant correlation mode and a recessive correlation mode between the composite fault and the fault symptom are established by using new fault data, and a feature subspace corresponding to the new fault is projected onto the dominant correlation mode and the recessive correlation mode, so that the self-adaptive updating of the initial tracing model of the composite fault is realized.

The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:

in the embodiment of the invention, a steel rolling process flow composite fault mode classifier is constructed; the structured steel rolling process flow composite fault mode classifier is used for realizing layered composite fault tracing, so that the root cause of the fault can be diagnosed and reasoned in time when the composite fault occurs, and the method has important engineering significance for ensuring the safety of the steel rolling process flow and the stability of the product quality.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a schematic diagram of a comprehensive automation system for a hot rolling process according to an embodiment of the present invention;

FIG. 2 is a schematic flow chart of a complex fault tracing method for a steel rolling process flow according to an embodiment of the present invention;

fig. 3 is a detailed flow diagram of a steel rolling process flow composite fault tracing method provided by an embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The present embodiment takes a steel rolling process flow as an example. It should be noted that the method for tracing the composite fault of the steel rolling process flow provided by the embodiment of the present invention is not limited to the steel rolling process flow, and is also applicable to other production processes, such as a chemical production process.

FIG. 1 is a schematic diagram of a comprehensive automation system of a steel rolling process flow. As shown in fig. 1, the steel rolling process flow of the present embodiment mainly includes a plurality of production processes such as heating, rough rolling, flying shear, finish rolling, laminar cooling, and coiling, and a long product processing flow mainly including a series structure is formed from a raw material to a final product. Meanwhile, the corresponding integrated automation system has obvious levels, mainly comprises an equipment layer, a real-time control layer, a process control layer, a manufacturing execution layer, an enterprise management layer, an enterprise strategic layer and the like, and all levels are in definite work and are in mutual cooperative association; wherein the content of the first and second substances,

the device layer includes: the system comprises a heating furnace, a rough rolling unit, a flying shear, a finishing rolling unit, a laminar cooling and coiling machine, and mainly achieves the functions of main and auxiliary transmission, electric, hydraulic and pneumatic operation execution, instrument data acquisition and the like;

the real-time control layer mainly completes the sequence and logic control of the whole-line equipment according to an operation instruction issued by the process control layer, and undertakes the task of controlling the overall length and quality of the strip steel, thereby realizing the basic automation;

the process control layer is mainly used for tracking each production procedure of the hot continuous rolling whole line in real time, acquiring data and optimally setting process parameters according to an operation plan issued by the manufacturing execution layer, obtaining the optimally set parameters of various production equipment by calculating through a series of mathematical models according to actual working conditions at a proper moment, and determining the quality of a final product, particularly the quality of a head part, and greatly influencing the production sequence;

the manufacturing execution layer mainly completes the functions of production planning, production scheduling, quality management, inventory management, logistics tracking and the like, fully considers the production constraints of all the working procedures, gives consideration to different quality requirements and contract delivery date, and adopts an integrated scheduling strategy and scheduling strategy to realize material flow matching and energy flow matching;

the enterprise strategic layer and the enterprise management layer take decision management and production and general management as cores respectively, emphasize the planning of enterprises, and simultaneously take customer orders and market demands as planning sources to carry out macroscopic planning and grasp, so that various resources in the enterprises are fully utilized, and the enterprise benefit is improved.

The combined action of the multi-level and full-flow manufacturing modes brings challenges to the accurate tracing of the composite faults of the steel rolling process flow.

As shown in fig. 2, an embodiment of the present invention provides a method for tracing a composite fault of a steel rolling process flow, including:

s101, constructing a steel rolling process flow composite fault mode classifier; wherein the compound failure mode comprises: the method comprises the following steps of:

a1, analyzing the correlation between propagation type, coupling type, multiple parallel types and compound type fault data and normal data by using a correlation analysis method, and determining a characteristic subspace of a compound fault mode of a steel rolling process flow according to a correlation analysis result;

as shown in fig. 3, firstly, a correlation analysis method such as information entropy is used to analyze the correlation between propagation type, coupling type, multiple concurrent type and composite fault data and normal data (data under normal operation of the system), and data features capable of reflecting the fault characteristics of the steel rolling process flow are extracted according to the correlation analysis result to obtain a feature vector of a composite fault mode of the steel rolling process flow;

and then, screening the obtained feature vectors by using similarity analysis methods such as neighbor similarity, cosine similarity and the like to obtain a feature subspace of a composite fault mode, wherein the index variation amplitude of the neighbor similarity and the cosine similarity is in a preset interval (for example, between 0 and 1) and the feature subspace has the fault characteristic of the steel rolling process flow.

A2, identifying and classifying the obtained feature subspaces by using the steel rolling process expert experience and process knowledge, marking a composite fault mode of fault data and normal data according to the classification result, and constructing a steel rolling process flow composite fault mode classifier capable of reflecting the composite fault mode based on the marking result and the corresponding feature subspaces, wherein the composite fault mode of the normal data is marked as none.

S102, realizing hierarchical composite fault tracing based on the constructed steel rolling process flow composite fault mode classifier, wherein the hierarchical means: from the whole process to the subsystem, the method specifically comprises the following steps:

b1, determining a composite fault mode to which steel rolling process flow data belongs by using the constructed steel rolling process flow composite fault mode classifier, and extracting an explicit optimal feature projection matrix and a implicit optimal feature projection matrix according to a feature subspace corresponding to each composite fault mode and explicit and implicit states of composite fault symptoms of the steel rolling process flow in time and space;

in this embodiment, the dominant state refers to: the fault symptom is obvious, the position and the device of trouble are easy to be located, the recessive condition indicates: the fault signs are hidden, and the fault parts and devices are not easy to locate.

As shown in fig. 3, firstly, a composite fault mode to which steel rolling process flow data to be analyzed (including fault data and normal data) belong is determined by using a constructed steel rolling process flow composite fault mode classifier, a feature subspace corresponding to each composite fault mode is projected onto a dominant correlation mode of a fault symptom aiming at a dominant state of the steel rolling process flow composite fault symptom presented in time and space, the problems of few composite fault data samples, separability among data and the like are fully considered, and according to the obtained dominant correlation mode, an exponential discriminant analysis method is used for extracting a dominant optimal feature projection matrix by constructing an optimal discriminant direction, reasonably designing and optimizing a target function and solving the optimal target function;

then, aiming at the recessive state of the composite fault symptom of the steel rolling process flow in time and space, fully considering the commonalities and characteristic relations between faults, faults and symptoms, symptoms and symptoms of the composite fault, and the like, projecting the characteristic subspace corresponding to each composite fault mode onto the recessive association mode of the fault symptom, and extracting a recessive optimal characteristic projection matrix by using a multi-task characteristic selection and causal relation analysis method according to the obtained recessive association mode.

And B2, weighting the extracted dominant and recessive optimal feature projection matrixes by using algorithms such as particle swarm optimization, information entropy and the like to obtain a comprehensive feature projection matrix, constructing a composite fault initial tracing model according to the obtained comprehensive feature projection matrix, and realizing the hierarchical accurate tracing of the composite fault of the steel rolling process flow according to the constructed composite fault initial tracing model, the whole flow of the composite fault and the detection result of a subsystem.

In the embodiment, firstly, the characteristics of propagation, coupling, multiple concurrency and the like of the composite fault are comprehensively considered, the independent detection method of the composite fault of the subsystem is researched, the independent detection of the composite fault of the subsystem is realized, and the detection result of the subsystem is obtained; wherein, the subsystem detection result includes: and the fault or normal operation state of the subsystem level of the steel rolling process flow such as full-flow heating, rough rolling, flying shear, finish rolling, laminar cooling, coiling and the like.

Then, on the basis of fully considering the static association and dynamic cooperation relation among the subsystems, the information fusion and machine learning methods such as variational Bayesian inference, ensemble learning and the like are utilized to fuse the composite fault autonomous detection information of different subsystems, so as to realize the full-process composite fault autonomous detection and obtain the full-system detection result; wherein, the whole process detection result includes: the fault (namely, abnormity) or normal operation state of the whole process level of the steel rolling process flow.

In this embodiment, the abnormal operation state of the hot rolling process flow is usually a "bottom-up" development mode in which the abnormal state of the subsystems is caused by a composite fault occurring in the bottom loop, and further causes the abnormal state of other subsystems and the whole flow. Correspondingly, the composite fault tracing process is a 'top-down' process for tracing the fault reason of the bottom layer loop from the abnormal state of the whole process.

In this embodiment, the composite fault tracing for implementing layering based on the constructed steel rolling process flow composite fault mode classifier further includes:

when a new fault occurs, a dominant correlation mode and a recessive correlation mode between a composite fault and a fault symptom are established by using new fault data, and a feature subspace corresponding to the new fault is projected onto the dominant correlation mode and the recessive correlation mode, so that the self-adaptive updating of the initial tracing model of the composite fault is realized, the generalization capability of the model is enhanced, and information support is provided for quickly and accurately searching the cause of the abnormal operation of the steel rolling process flow and maintenance decision

According to the steel rolling process flow composite fault tracing method, a steel rolling process flow composite fault mode classifier is constructed; the structured steel rolling process flow composite fault mode classifier is used for realizing layered composite fault tracing, so that the root cause of the fault can be diagnosed and reasoned in time when the composite fault occurs, and the method has important engineering significance for ensuring the safety of the steel rolling process flow and the stability of the product quality.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.