Reliability evaluation method for hot rolling production line considering fuzzy maintenance quality

Abstract

A hot rolling production line reliability evaluation method considering fuzzy maintenance quality, determines the task requirements of the hot rolling production line with a series structure and the key machines in the series structure, identifies the key processes of the hot rolling production line and collects process data, estimates the performance state distribution function of each machine in the series structure and evaluates the fuzzy maintenance quality parameters of each machine, so as to build a random fuzzy flow manufacturing network model of the hot rolling production line; the proportional set in the model is analyzed to determine the sub-task requirement set in the random fuzzy flow network model, and then the reliability model of each machine considering the fuzzy maintenance quality is established, and finally the reliability model of the hot rolling production line is established based on the structure relationship among the machines in the hot rolling production line, so that the dynamic evolution trend of the performance in the production process of the hot rolling production line is analyzed and the production process of the hot rolling production line is quantified under the condition that the production requirement is known. The production process of the hot rolling production line is simplified through the random fuzzy flow manufacturing network, the performance state change law is analyzed, and the reliability is characterized through the opportunity measure index based on the Markov model and fuzzy mathematics, so that the accuracy of the result is significantly improved.

Description

Technical Field

[0001] This invention relates to a technique in the field of fuzzy theory, specifically a reliability assessment method for hot rolling production lines that considers fuzzy maintenance quality. Background Technology

[0002] Existing reliability assessment methods for hot rolling production lines cannot characterize the degradation features of machines under the combined effects of fuzziness and randomness, nor can they describe the intrinsic relationships between production tasks, machines, products, and maintenance. Furthermore, due to the high complexity of modern hot rolling production lines, it is even more difficult to understand their operating mechanisms and performance degradation patterns. Summary of the Invention

[0003] This invention addresses the problems of existing hot-rolled production line reliability assessment methods, which fail to accurately assess operational status and cannot resolve reliability assessments under the fuzziness of maintenance quality due to neglecting fuzzy maintenance quality. It proposes a new hot-rolled production line reliability assessment method that considers fuzzy maintenance quality. Based on multi-state system reliability theory and the operational characteristics of hot-rolled production lines, this method examines the reliability of the hot-rolled production line system. It simplifies the production process of the hot-rolled production line through a stochastic fuzzy flow manufacturing network, analyzes the performance state change patterns, and characterizes reliability using a chance measure index based on Markov models and fuzzy mathematics, significantly improving the accuracy of the results.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to a reliability assessment method for hot-rolled production lines considering fuzzy maintenance quality. The method involves determining the task requirements of a tandem hot-rolled production line and the key machines within the tandem structure, identifying key processes and collecting process data, estimating the performance state distribution function of each machine in the tandem structure, and evaluating the fuzzy maintenance quality parameters of each machine. This constructs a stochastic fuzzy flow manufacturing network model for the hot-rolled production line. By analyzing the proportion set in the model, the method determines the sub-task requirement set in the stochastic fuzzy flow network model, thereby establishing a reliability model for each machine considering fuzzy maintenance quality. Finally, based on the structural relationships between machines in the hot-rolled production line, a reliability model for the entire hot-rolled production line is established. This enables the analysis of the dynamic evolution trend of performance during the production process of the hot-rolled production line and the quantification of the production process, given known production requirements.

[0006] In the aforementioned random fuzzy flow manufacturing network model, the capacity of different arcs is statistically independent and each node is a reliable node, i.e., only arcs are considered.

[0007] The performance status of the machine refers to the maximum working load it can withstand.

[0008] The degradation process of the hot rolling production line follows a homogeneous Markov process, meaning that the current state is only related to the state at the previous moment and is independent of previous states; and the transition strength between states is known. Specifically, the output of machine number l has two quality states: qualified and unqualified, and only the qualified output can be transferred to the next machine; the relationship between the degradation processes before and after machine maintenance follows a quasi-renewal process.

[0009] The membership functions of the fuzzy maintenance quality parameters are all triangular membership functions.

[0010] The aforementioned determination of the task requirements and related key machines for the hot rolling production line refers to: determining the key machines required for reliability assessment based on task requirements (including product quantity and quality requirements).

[0011] The aforementioned identification of key processes in hot rolling production lines and collection of process data refers to: identifying key processes that affect the reliability of hot rolling production lines for the identified key machines, and then collecting relevant machine performance, process quality, and maintenance record data.

[0012] The performance state distribution functions of each machine, based on the assumption that the performance state degradation of machines in a cascaded structure follows a homogeneous Markov process, are expressed by the Kolmogorov differential equation system dp. l (t) / dt=p l (t)X i The performance state probability function of each machine with respect to time t is obtained, where: S l Let X be the set of states of machine l. l Let p be the transition strength matrix of machine l. l (t) is the state probability vector of machine l at time t.

[0013] The fuzzy maintenance quality parameters of each machine are based on the fact that the degradation process of the machines in the series structure before and after maintenance follows a quasi-updating process, i.e. By combining the collected performance degradation data and maintenance record data, the maintenance quality parameters of machine l are evaluated. in: and Let be the probability functions of machine l being in state j at time t before the first and kth maintenance, respectively.

[0014] The aforementioned stochastic fuzzy flow manufacturing network model refers to the process of converting a hot rolling production line into a stochastic flow manufacturing network model based on key machines, the structural relationships between machines, and the machine performance degradation function under fuzzy maintenance quality considerations. This model is used to abstractly represent the production process of the hot rolling production line for subsequent reliability assessment.

[0015] The aforementioned random fuzzy flow manufacturing network model includes: a set of directed arcs and a set of proportions. The set of directed arcs is obtained by abstracting and representing the key processing machines. The set of proportions is obtained by calculating the pass rate, fail rate and material transfer ratio based on the product production pass rate information and the process flow information between the key processing machines.

[0016] The construction is as follows: the processing machine is abstractly represented as a directed arc, the machine's input / output nodes are abstractly represented as circles, and the machine's output material flow is abstractly represented as a double circle. The machines are connected according to the production and processing relationship between them to construct a random fuzzy flow manufacturing network model. The directed dashed arcs used for connection represent the transfer of material flow between machines, and the direction of the dashed arcs represents the direction of material flow transfer.

[0017] The set of proportions refers to: the set of proportions is P = {q} l,1 ,q l,2 ,pl ,j}, where: q l,1 and q l,2 Let q represent the probabilities that the material flow output by machine l is in a qualified state and an unqualified state, respectively. l,1 and q l,2 Determined based on process quality data; p l,j The proportion of qualified material flow output from machine l is transmitted to machine j.

[0018] The aforementioned set of subtask requirements refers to the following: given the production task requirement D, the input requirement for machine l is... Output requirements are Where: L represents the total number of machines in the hot rolling production line.

[0019] The aforementioned reliability function Where: the input material flow f of machine l at time t after the kth maintenance l,1(k) (t), which is less than the input requirement of machine l. chance measure Where: Cr(g) is the confidence function, Ψ is the possibility space, and the failure probability function is defined at a confidence level of δ. 1(g) is a judgment function, where 1(true) = 1 and 1(false) = 0.

[0020] The overall reliability function is as follows: The hot rolling production line consists of L machines connected in series.

[0021] This invention relates to a system for implementing the above-mentioned method, comprising: a machine state transition strength matrix estimation module, a fuzzy maintenance quality parameter estimation module, a stochastic fuzzy flow manufacturing network model parameter evaluation module, and a hot rolling production line reliability evaluation module. Specifically: the machine state transition strength matrix estimation module uses performance degradation data of the machine under unmaintained conditions to estimate the transition strength parameters between any two states, obtaining the machine's transition strength matrix; the fuzzy maintenance quality parameter estimation module uses maintenance record information and performance degradation data of the machine after maintenance to statistically compare the numerical changes of the transition strength matrix before and after maintenance, and performs fuzzification processing on the statistical results to obtain fuzzy maintenance quality parameters; the stochastic fuzzy flow manufacturing network model parameter evaluation module statistically obtains the product qualification rate, product non-qualification rate, and material transfer ratio based on process quality information and process flow information, obtaining the parameters in the stochastic fuzzy flow manufacturing network model; the hot rolling production line reliability evaluation module uses the calculation results obtained from the machine state transition strength matrix estimation module, the fuzzy maintenance quality parameter estimation module, and the stochastic fuzzy flow manufacturing network model parameter evaluation module to perform the above-mentioned reliability evaluation technology processing, obtaining the hot rolling production line reliability evaluation result.

[0022] Technical effect

[0023] This invention establishes a stochastic fuzzy flow manufacturing network model as an abstract representation of the hot rolling production line process, accurately explaining the operating mechanism of the hot rolling production line. Based on stochastic fuzzy theory, this invention evaluates the reliability of the hot rolling production line under the influence of fuzzy maintenance quality and inherent randomness. Compared with existing technologies, this invention significantly improves the accuracy of hot rolling production line reliability assessment results. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the random fuzzy flow manufacturing network model for the hot rolling production line of the present invention;

[0025] In the figure: (a) is a 3-machine series-parallel system; (b) is a schematic diagram of a random fuzzy flow manufacturing network model of a 3-machine series-parallel system.

[0026] Figure 2 This is a flowchart of the present invention;

[0027] Figure 3 Schematic diagram of key machines in the 1580 hot rolling production line;

[0028] Figure 4 A schematic diagram of a stochastic fuzzy flow manufacturing network model for a 1580 hot rolling production line;

[0029] Figure 5 A schematic diagram of a machine reliability model that takes fuzzy repair quality into account;

[0030] In the figure: (a) is the reliability model of machine 1 after the 1st, 2nd and 3rd maintenance; (b) is the reliability model of machine 2 after the 1st, 2nd and 3rd maintenance; (c) is the reliability model of machine 3 after the 1st, 2nd and 3rd maintenance.

[0031] Figure 6 A schematic diagram of the reliability model of the 1580 hot rolling production line considering fuzzy maintenance quality. Detailed Implementation

[0032] like Figure 2 As shown in the figure, this embodiment relates to a reliability assessment method for hot rolling production lines that considers fuzzy maintenance quality, including:

[0033] Step 1: Determine the task requirements and key machines for a company's 1580 hot rolling production line. After determining the task requirements (including product quantity and quality requirements), identify the key machines needed for reliability assessment, including the roughing mill (machine 1), finishing mill (machine 2), and coiling mill (machine 3). The key machines for the 1580 hot rolling production line are as follows: Figure 3 As shown.

[0034] Step 2: Identify the key processes of a company's 1580 hot rolling production line and collect relevant data. There are three key processes in the 1580 hot rolling production line: strip surface flatness, strip surface roughness, and the number of surface defects. Based on this, collect relevant data.

[0035] Step 3 estimates the performance state distribution function of each machine. The state sets of each machine are shown in Table 1, and their corresponding transition intensities are shown in Table 2. Based on the Kolmogorov differential equations, the performance state distribution function of each machine can be obtained.

[0036] Table 1. Set of Machine States.

[0037]

[0038] Table 2. Transfer intensity of each machine.

[0039]

[0040]

[0041] Step 4: Evaluate the fuzzy repair quality parameters for each machine. Machines 1, 2, and 3... The values ​​are (0.82, 0.85, 0.95), (0.82, 0.85, 0.95), and (0.82, 0.85, 0.95), respectively.

[0042] Step 5: Construct a stochastic fuzzy flow manufacturing network model for the hot rolling production line. Combining steps 1 to 4, construct a model as follows: Figure 4 The random fuzzy flow manufacturing network model shown in the figure, a l For machine l in the hot rolling production line, For a network model that generates a random fuzzy flow, input nodes with directed edge arc l are generated. For the output node of directed edge arc l in the network model of random fuzzy flow, p l,j q represents the proportion of material passed from machine l to machine j. l,1 Let q be the probability that the product output by machine l is of acceptable quality. l,2 f is the probability that the output product of machine l is defective. l The flow rate of material flowing into machine l.

[0043] Step 6 analyzes the proportion set in the random fuzzy flow network model. The flow transfer proportion between machines is p. 1,2 =p 2,3 =1. Because for the same machine, q l,1 and q l,2 The sum is 1. The q of each machine... l,1 The values ​​are 0.98, 0.97, and 0.99, respectively.

[0044] Step 7 determines the set of subtask requirements in the stochastic fuzzy flow network model. Taking the task requirement D = 60 tons of thin wide steel per day as an example, the input requirements for each machine to complete the production task are as follows:

[0045] Step 8: Establish reliability models for each machine considering fuzzy maintenance quality. The reliability models for each machine under different maintenance cycles k are as follows: Figure 5 As shown.

[0046] Step 9 establishes a reliability model of the hot rolling production line based on the structural relationships between machines within the line. Finally, as... Figure 6 As shown, a reliability model for the 1580 hot rolling production line is established.

[0047] Compared with existing technologies, this method considers the impact of fuzzy maintenance quality, thereby improving the accuracy of hot rolling production line reliability assessment results.

[0048] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A hot rolling line reliability evaluation method considering ambiguous repair quality, characterized by, The random fuzzy flow manufacturing network model is a model for abstractly representing the production process of the hot rolling production line, which is used for subsequent reliability evaluation, and is obtained by converting the hot rolling production line into a random flow manufacturing network model according to the key machines, the structural relationship between the machines and the machine performance degradation function considering the fuzzy maintenance quality. The sub-task demand set refers to: in the case of known production task demand D, the input demand of machine l is , and the output demand is , wherein: L is the total number of machines in the hot rolling production line; Reliability function where: input material flow of machine l at time t after kth repair which is less than input demand of machine l The opportunity measure is where: is the belief function, is the plausibility space, the confidence level is The failure probability function is the judgment function, 1(true)=1, 1(false)=0; The overall reliability function is where the hot rolling line is made of L machines in series. The performance state distribution function of each machine is obtained by solving the Kolmogorov differential equation set based on the assumption that the performance state degradation of the machine in series structure obeys the homogeneous Markov process , wherein the performance state probability function of each machine with respect to time t is obtained, and wherein: is a state set of the machine l, is a transition intensity matrix of the machine l, is a state probability vector of the machine l at time t. The fuzzy maintenance quality parameters of each machine are based on the condition that the degradation process of the machine in series structure before and after maintenance obeys quasi-update process, that is , combined with the collected performance degradation data and maintenance record data, the maintenance quality parameters of machine l are evaluated , wherein: and are the probability functions of machine l in state j at time t before the first and kth maintenance, respectively. The random fuzzy flow manufacturing network model is a model for abstractly representing the production process of the hot rolling production line, which is used for subsequent reliability evaluation, and is obtained by converting the hot rolling production line into a random flow manufacturing network model according to the key machines, the structural relationship between the machines and the machine performance degradation function considering the fuzzy maintenance quality. The random fuzzy flow manufacturing network model comprises a directed arc set and a proportion set, wherein: the directed arc set is obtained by abstractly representing the key processing machines; and the proportion set is obtained by calculating the pass rate, the unqualified rate and the material transfer proportion according to the product production pass rate information and the process flow information between the key processing machines. The hot rolling production line reliability evaluation specifically comprises: Step 1, determining the task demand of the hot rolling production line and the related key machines: after determining the task demand, the key machines required for reliability evaluation are determined, including rough rolling machines, finishing rolling machines and coiling machines; Step 2, identifying the key processes of the hot rolling production line and collecting related data: the hot rolling production line has three key processes, which are the flatness of the strip surface, the roughness of the strip surface and the number of defects on the strip surface, and related data are collected on this basis; Step 3, estimating the performance state distribution function of each machine: the performance state distribution function of each machine can be obtained based on the Kolmogorov differential equation set; Step 4, evaluating the fuzzy maintenance quality parameters of each machine; Step 5, constructing the random fuzzy flow manufacturing network model of the hot rolling production line: the random fuzzy flow manufacturing network model is constructed in combination with steps 1 to 4; Step 6, analysis of the proportion set in the random fuzzy flow network model: the traffic transfer proportion between machines is Because for the same machine, and The sum of the sum is 1, and the proportion of each machine is 0.98, 0.97, 0.99, respectively. Step 7, determining the sub-task demand set in the random fuzzy flow network model: taking the task demand D = 60 tons of thin wide strip steel per day as an example, in order to complete the production task demand, the input demands of each machine are respectively , , ; Step 8, establishing the reliability model of each machine considering the fuzzy maintenance quality; Step 9, establishing the reliability model of the hot rolling production line based on the structural relationship between the machines in the hot rolling production line.

2. The hot rolling line reliability evaluation method considering ambiguous repair quality according to claim 1, characterized by, The capacities of different arcs in the random fuzzy flow manufacturing network model are statistically independent, and each node is a reliable node, i.e., only the arcs are considered. The performance state of the machine refers to the maximum working load that can be borne by the machine. The degradation process of the hot rolling production line obeys a homogeneous Markov process, that is, the state at a current time is only related to the state at a previous time and is not related to previous states, and transition intensities between states are known, specifically: output of the lth machine has two quality states of qualified and unqualified, and only qualified output can be transmitted to the next machine; and the relationship between the degradation processes before and after machine maintenance obeys a quasi renewal process. The membership functions of the fuzzy maintenance quality parameters are all triangular membership functions. The task demand of the hot rolling production line and the related key machines are determined based on the task demand.

3. The hot strip mill line reliability assessment method considering ambiguous repair quality according to claim 1 or 2, characterized by, The key process of the hot rolling production line is identified, and process data are collected, that is, for the obtained key machines, a key process flow affecting the reliability of the hot rolling production line is identified, and then machine performance, process quality and maintenance record data are collected.

4. The hot strip mill line reliability assessment method considering ambiguous repair quality according to claim 1, wherein The construction is specifically: the processing machines are abstractly represented as directed arcs, the input / output nodes of the machines are abstractly represented as circles, the output material flow of the machines is abstractly represented as double circles, the machines are connected according to the production and processing relationship between the machines to construct a random fuzzy flow manufacturing network model, the directed dashed arcs used for connection represent the transmission of the material flow between the machines, and the direction of the dashed arcs represents the transmission direction of the material flow. The ratio set is a set of ratios wherein: and are the probabilities that the material stream output by machine i is in a pass and a fail state, respectively, and are determined from process quality data; is the proportion of pass material stream output by machine i that is passed to machine j.

5. A system for implementing the method for evaluating the reliability of a hot rolling line taking into account the quality of the repairs according to any one of claims 1 to 4, characterized in that, The method comprises the following steps: The machine transition intensity matrix estimation module, the fuzzy maintenance quality parameter estimation module, the random fuzzy flow manufacturing network model parameter evaluation module and the hot rolling production line reliability evaluation module, wherein: the machine transition intensity matrix estimation module uses the performance degradation data of the machine without maintenance to perform transition intensity parameter estimation processing between any two states to obtain the transition intensity matrix of the machine; the fuzzy maintenance quality parameter estimation module uses the maintenance record information and the performance degradation data of the machine after maintenance to statistically compare the numerical changes of the transition intensity matrices before and after maintenance, performs fuzzy processing on the statistical results to obtain the fuzzy maintenance quality parameters; the random fuzzy flow manufacturing network model parameter evaluation module statistically obtains the product qualified rate, the product unqualified rate and the material transmission proportion according to the process quality information and the process flow information to obtain the parameters in the random fuzzy flow manufacturing network model; and the hot rolling production line reliability evaluation module uses the calculation results obtained by the machine transition intensity matrix estimation module, the fuzzy maintenance quality parameter estimation module and the random fuzzy flow manufacturing network model parameter evaluation module to perform the reliability evaluation technology processing to obtain the hot rolling production line reliability evaluation result.

Images