CN105184385B - A kind of distributed control method of automated manufacturing system - Google Patents

Abstract

An automatic manufacturing system for embedding assembly operations into flexible processing paths, firstly, according to the current state of the system, obtain the set T _en of the transitions that can be enabled; and then judge whether each element in T _en can ensure that the system is not dead lock operation. The basis for judgment is: simulated emission element t∈T _en , according to the position of t, if the current resource can support the corresponding token to reach a key place/sub-key place of the key place continuum, for the key place continuum In the case of , it is necessary to ensure that the tokens of other parallel processes of the assembly operation can reach their corresponding sub-key locations, then t∈T _df . When all transitions are detected, T _df is obtained, t∈T _df is emitted, and the next state is entered for re-judgment. The invention can ensure the deadlock-free property of the system, obtain greater permissibility as much as possible, and realize real-time online control by adopting the mode of controlling while predicting.

Description

A Distributed Control Method for Automatic Manufacturing System

technical field

The invention belongs to the technical field of automatic manufacturing systems, and relates to a distributed control method for automatic manufacturing systems.

Background technique

In the past few decades, with the wide application of information technology, automation technology and computing technology, the traditional manufacturing system has gradually transformed into an automatic manufacturing system. The goal is to greatly reduce manufacturing costs, improve product quality and ensure production safety. And can respond quickly to market changes and customization requirements. The deadlock-free nature of an automatic manufacturing system can be compared to the stability of a continuous system. No matter how good a manufacturing system is, once a deadlock occurs, it is unacceptable, because it means that the system may stagnate production at any time. This can lead to serious or even catastrophic consequences. To solve the deadlock problem in the automatic manufacturing system, people mainly use three kinds of mathematical tools: directed graph, automaton and Petri net. Petri nets can more properly simulate, analyze and control automatic manufacturing systems. The automatic manufacturing system with flexible processing path can adapt to the automatic mechanical manufacturing system of the processing object change, can automatically adjust and realize the efficient batch production of various workpieces within a certain range, and can change products in time to meet market demand. An automated manufacturing system with assembly operations can produce products that require parallel machining processes and assembly operations. These two systems each have irreplaceable advantages of the other. The research on these two systems has been quite extensive and in-depth, but the research on the combination of the two systems is still shallow.

The basic distributed predictive control method has been well applied in the automatic manufacturing system with only flexible processing paths, such as the automatic manufacturing system of S ⁴ R structure. Specifically, it is first assumed that when any process is performing a specific operation, other processes are temporarily stopped until the operation is completed. For this specific process, it can be predicted whether any token can advance from the current position to the nearest target position or the position with the largest resource occupancy. If the prediction is true, the token can move forward; otherwise, it must pause at its current position. Therefore, these target locations or locations with the largest resource occupancy are counted as key places. The specific implementation process of this method can refer to the paper "Distributed Supervisor Synthesis for Automated Manufacturing Systems Using Petri Nets". But for systems with assembly operations, this method has limitations. Therefore, the method is good in theory, but there are flaws in the process of widespread promotion.

Especially for automatic manufacturing systems with flexible processing paths and assembly operations, the basic distributed predictive control methods are no longer applicable. It is specifically reflected in: 1. The assembly operation needs all parallel processes to be carried out synchronously, and the assembly operation is completed in the last processing stage of each parallel process. Therefore, if the key places are simply defined in a parallel process, even if Token can reach these key places, it does not guarantee that the system can complete the assembly operation, and the completion of the assembly operation requires the cooperation of all parallel processes. 2. Since the assembly operations are embedded in the flexible processing path, the complexity of the system is greatly enhanced. According to the basic distributed predictive control method, Token must check whether the existing resources can support it to reach the key place closest to it before each step forward. But in such a complex system, if the Token arrives at its nearest key storehouse as the standard, the deadlock-free nature of the system can also be guaranteed, but the permission of the system will be greatly reduced, thereby affecting the performance of the system.

Contents of the invention

In order to overcome the problems in the prior art, the purpose of the present invention is to provide a distributed control method for automatic manufacturing systems, which is suitable for automatic manufacturing systems with flexible processing paths and assembly operations, and can enhance the control of complex manufacturing systems. control capabilities, and at the same time adopting this method can obtain higher permissibility.

To achieve the above object, the present invention adopts the following technical solutions:

A distributed control method for an automatic manufacturing system, comprising the following steps:

1) Initialize, make Among them, T _en is the set of enabled transitions, and T _df is the set of transitions that make the automatic manufacturing system run without deadlock;

2) Collect the current state M of the automatic manufacturing system;

3) According to the current state M, check whether each transition in the automatic manufacturing system is enabled, if t _i is enabled, then T _en =T _en +{t _i }, where t _i is any transition in the automatic manufacturing system ;

4) Judging whether any element t in T _en is located in the sign diagram module representing the assembly operation;

5) When any element t in T _en is located in the sign diagram module representing the assembly operation, simulate the emission of element t;

6) When t ∈ T _en , and t is not located in the sign diagram module representing the assembly operation, simulate the emission element t, if the current resource can support the arrival of the moving Token corresponding to the emission element t A key place in , then T _df ＝T _df +{t}, otherwise, Carry out step 4); Wherein, It is the collection of key places of the system structure after the sign diagram module in the system is replaced by a place and its pre-set and post-set as a whole;

7) When all the transitions in T _en are detected, get the transition set T _df that makes the system run without deadlock, output any t∈T _df , the automatic manufacturing system emits t, enters the next state, and then returns step 1).

Described step 5) specifically comprises the following steps:

5.1) If the current resource can support the launch of the mobile token corresponding to t to reach a sub-key place in the key place continuum, and in all other parallel processes of the sign diagram module representing the assembly operation, there is a token that can When arriving at the corresponding sub-key place in the key place continuum, T _df =T _df +{t};

5.2) If the current resource can support the launch of the mobile token corresponding to t to reach a subkey place in the key place continuum, and there is no Token in all other parallel processes of the sign diagram module representing the assembly operation When the sub-key places in the key place continuum can be reached, then Go to step 4);

5.3) If the current resource cannot support the launch of the mobile Token corresponding to t to reach a sub-key place in the key place continuum, then Proceed to step 4).

The set C _MG of the key place continuum of the sign graph module and the set of key places of the system structure after the sign map module is replaced by a store in the system The union of is the set C of the key warehouses of the entire automatic manufacturing system.

The key storehouse unity of the sign diagram module is a sequence pair composed of several subkey storehouses; the subkey storehouses are respectively distributed in each parallel process, and the subkey storehouses have the same type; The number of sub-key stores is equal to the number of parallel processes of the logogram module.

Compared with the prior art, the present invention has the beneficial effects: the present invention collects the state M of the original system first, and according to the current state, by running the method, obtains the set T of transmissible transitions that ensure the system is free of deadlock _df , finally, in order to ensure the permissibility of the system, randomly output instructions, that is, randomly output a transition t in T _df , and then the system receives the instruction, emits element t, enters the next state, and re-acquires the current state for a new round of prediction. Steps 1) to 7) of the present invention constitute the activity supervisory controller of the automatic manufacturing system proposed in the present invention, and activity means that the system has no deadlock. The invention aims to generate an active supervisory controller of an automatic manufacturing system through a distributed control method, so as to ensure the deadlock-free property of the system and obtain greater permission as much as possible. This control method adopts the method of predicting and controlling to realize real-time online control.

The key points of the control method of the present invention are as follows:

First, places are divided into key places and non-key places (that is, all places in the system except key places). The former represents the position with the smallest or largest resource occupancy; the latter represents other positions. From the perspective of resource acquisition, after the position with the smallest resource occupation, there must be sufficient resources; after the position with the largest resource occupation, no more resources are bound to be needed. The existence of key places is unquestionable, because at the initial and target positions, no resources are occupied, and they are themselves key places. The role of the key library in the process is like a "safety island". For AESM (Automated Manufacturing System with Flexible Process Path and Assembly Operations), the invention extends the original concept of the key store. The essential meaning of the key repository is that it is the repository that uses the most or the least resources in a process. Therefore, these target locations or locations with the largest resource occupancy or the least resource occupancy are counted as key places. But in AESM, the sign diagram module represents the assembly operation, its start and end, marking that all its parallel processes must start and end at the same time. Therefore, in the symbol diagram module, the key place is no longer a simple place, but a continuum of key places, and each element in the continuum is called a sub-key place. The key-place continuum is actually an ordinal pair of sub-key-places. All the sub-key places are respectively distributed in each parallel process, representing the places that use the most or the least resources in the parallel process. In a key place continuum, the number of sub-key places is equal to the number of parallel processes in the flag map module, and the sub-key places must have the same type, that is, all of them are places that use the most resources, or all use The place with the fewest resources. For other parts of the system that do not contain the logogram module, the key place is still the place that uses the most or the least resources, because if the logogram module is replaced by a library, its structure is the S ⁴ R structure, the definition of the key place Same definition as in the basic predictive control method.

Secondly, the present invention operates in an online real-time mode while predicting and controlling. Even if the resources are sufficient to support a specific Token to move forward after taking a step, the control method only allows it to move forward. When other processes are stopped, the existing resources can fully support the Token to advance to the next key position.

Finally, the control method of the invention operates without detecting global information. The execution of each step depends only on the adequacy of existing resources. Apart from the current process, there is no need to know the status of other processes. Since the operation of the system with flexible processing paths and assembly operations depends entirely on local information, the control method cleverly implements a distributed operation mode, which greatly reduces the communication flow between the controller and each process.

Therefore, the control method of the present invention is applicable to more complex automatic manufacturing systems in practice, and enhances the control ability of complex manufacturing systems. Finally, the definition of key places is extended so that the new definition covers the original definition and can accurately describe the position of key places in AESM. Higher permissibility can be obtained by using the method of the present invention.

Description of drawings

Figure 1 is a feedback system diagram composed of a controller and a Petri net;

Fig. 2 is a flow chart of the present invention.

Fig. 3 is an explanatory diagram of three logogram modules, in which Fig. 3(a) and Fig. 3(b) are both logogram modules, and Fig. 3(c) is not a logogram module.

Fig. 4 is a schematic diagram of the process from SSM to ESSM, wherein Fig. 4(a) is a simple state machine (SSM), and Fig. 4(b) is an extended state machine (ESSM).

Figure 5 is a structural diagram of an AESM instance.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings.

The system in the present invention refers to a system that has both flexible processing paths and assembly operations, more specifically, embeds assembly operations into flexible processing paths, and the invention is an online, real-time distributed control method.

The core idea of the present invention is to treat the sign map module and other parts in the AESM differently, more specifically, to treat the key storehouse entity and the key storehouse differently. From the point of view of understanding, the essence of the method can be described as follows: First, the sign map module is virtualized as a place, because the sign map module in AESM itself can be regarded as a place replaced by a place. At this time, the network structure is a typical ^S4R , and the key place is the place that uses the most or the least resources in each process. Then, for each map module, break through each to find all the sub-key places, which are places that use the most or the least resources in each parallel process in the map module; the key store of the map module The continuum is the sequence pair combined by the sub-key stores in each parallel process. Therefore, the number of sub-key stores in the key store continuum is equal to the number of parallel paths in the symbol graph module. In addition, it should be noted that the sub-key places in the key place continuum must be of the same type, that is, they are all sub-key places that use the most resources, or all use the least resources. Finally, when running the system, when the Token has not advanced to the logo map module, the logo map module is regarded as a warehouse. When the existing resources can support the Token to advance to any key warehouse, the Token can go forward Further; when a Token enters the sign map module, each Token in each parallel process must ensure that the Token can reach its sub-key store, and the Tokens in all other parallel processes must To be able to reach the corresponding sub-key storehouse, that is, to ensure that every sub-key storehouse in a key storehouse continuum has a Token can reach it, so that this Token can move forward.

Referring to Fig. 2, the control method of the present invention is aimed at an automatic manufacturing system with flexible processing paths and assembly operations, and specifically includes the following steps:

1) Initialize, make Among them, T _en is the set of enabled transitions, and T _df is the set of transitions that make the automatic manufacturing system run without deadlock;

2) Collect the current state M of the automatic manufacturing system;

3) According to the current state M, check whether each transition in the automatic manufacturing system is enabled, if t _i is enabled, then T _en =T _en +{t _i }, where t _i is any transition in the automatic manufacturing system ;

4) Judging whether any element t in T _en is located in the sign diagram module representing the assembly operation;

5) When any element t in T _en is located in the sign diagram module representing the assembly operation, simulate the emission of element t;

5.1) If the current resource can support the launch of the movable Token corresponding to t to a sub-key place in the key place continuum, and there is a Token When the corresponding sub-key place in the key place continuum can be reached, T _df =T _df +{t};

5.2) If the current resource can support the launch of the movable Token corresponding to t to reach a sub-key place in the key place continuum, and there is no Token When Ken can reach the subkey place in the key place continuum, then Go to step 4);

5.3) If the current resource cannot support the launch of the movable Token corresponding to t to reach a sub-key place in the key place continuum, then Go to step 4);

6) When t∈T _en , and t is not located in the sign diagram module representing the assembly operation, simulate the launch of element t, if the current resource can support the launch of the movable Token corresponding to t A key place in , then T _df ＝T _df +{t}, otherwise, Carry out step 4); Wherein, It is the collection of key places of the system structure after the sign diagram module in the system is replaced by a place and its pre-set and post-set as a whole;

7) After all the transitions in T _en are detected, the set T _df of transitions that make the system run without deadlock is obtained. Output any t∈T _df , the automatic manufacturing system emits transition t, enters the next state, and then returns to step 1).

The above steps 1) to 7) constitute the active supervisory controller of the automatic manufacturing system. Referring to Figure 1, the controller gives the control result t, the controlled system emits t, and turns to a new state M, which is input to the controller, and the controller analyzes it, and then gives a new output result t, The controlled system emits t and enters another state. Repeatedly.

A collection of key warehouses for the entire automated manufacturing system Among them, C _GM is the collection of key place continuities of the sign graph module; It is a collection of key places of the system structure after replacing the sign diagram module in the system with a place. That is to say, the key place of the whole automatic manufacturing system is composed of two parts, one part is the collection of key place continuum of the sign diagram module, and the other part is the sign map module in the system is composed of a place and its pre-set and post-set Set the collection of key places of the system structure after the overall replacement. The key storehouse unity of the sign diagram module is a sequence pair composed of several subkey storehouses; the subkey storehouses are respectively distributed in each parallel process, and the subkey storehouses must have the same type, that is, all of them are used The places with the most resources are all the places that use the least resources; the number of sub-key places in the key place continuum is equal to the number of parallel processes of the logogram module.

The key places are also deduced step by step from simple to complex. The specific process is as follows.

Definition 1 Given a cycle or path that does not contain _PR , their key places are defined as C={C ₁ ,C ₂ ,C ₃ }, where:

Note: The key storage places defined by C ₁ are of type I, which means that the least resources are used, and generally _, no resources are _used ; The place with the most resources.

Definition 1 gives the definition of key places for a loop or path in a system, and all subsequent definitions are based on this definition, because the flexible processing path is composed of loops, and the sign map module is composed of composed of parallel paths.

Definition 2 For the sign graph module B(t _s ,t _e ), its key places are actually a collection of key place continuities, denoted as Among them, m∈N ⁺ needs to meet the following conditions:

1. The number of parallel paths in B(t _s ,t _e ) is n, ;

2. is a key place continuum, In the key place continuum, are the same type, specifically type I or type II.

Definition 3 For the network of S ⁴ R structure, its key places are defined as The meanings of C ₁ , C ₂ , and C ₃ are the same as those in Definition 1.

Note: The system of S ⁴ R structure does not include assembly operation, it is composed of some sequential processes, and each process is connected to each other through shared resources to affect each other, so its key places do not consider the key place unity, only need to consider each The key repository in the process. Among them, the logo module is to represent the assembly operation.

Definition 4 For AESM, the set of its key places is defined as Among them, C _MG is the collection of key place continuities of the sign graph module; is the key library of S ⁴ R ^* corresponding to AESM after replacing the sign map module with place p ^* ; after replacing the sign map module with place p ^* , AESM is a system with S ⁴ R ^* characteristics, and its key place recorded as

Note: The key repository of AESM should pay attention to distinguishing the sign map module from other parts, so it should be considered separately. The key places of the logo map module are determined according to Definition 2, and the key places of other parts need to be temporarily regarded as a place of the logo map module and determined according to Definition 3.

Referring to Fig. 1, the original Petri net and the controller constitute a feedback system. Note that the controller here is the distributed control method proposed by the present invention. The controller first collects the state of the original system, and according to the state of the original network, by running this method, it obtains the transition T _df that can be transmitted without deadlock, and finally in order to ensure the permissibility of the system, it randomly outputs instructions, that is, outputs T _df A transition t that can be transmitted, and then the system receives an instruction, transmits t, and enters the next state, and the controller re-acquires the current state for predictive simulation. This process can be represented by Figure 1, and M represents the output state of the original Petri net.

Since there is no need to consider global information, the control method of the present invention can generate event occurrence sequences dynamically and in real time. It is not difficult to see that the sequence is not unique, and each sequence can independently guide the entire system from the initial state to the target state. In extreme cases, when all processes except itself are in the initial state, a specific token in any process must be able to reach the target location; then, all resources are released, and the above steps are repeated for other tokens in the process , until all reach the target position.

A detailed description will be given below through a specific embodiment.

First of all, the present invention proposes a new Petri net structure, that is, Augmented Extended State Machine, which is translated into Chinese as Enhanced Extended State Machine, abbreviated as AESM. It embeds the assembly operation into the flexible processing path, which represents a more complex processing process, which means that in the flexible processing path, there can also be shunt operations and assembly operations, as well as parallel processing processes between them. Assembly operations are represented by sign diagram modules. AESM is given step by step by the definition below.

Definition 5 A simple state machine, denoted as SSM, is a strongly connected state machine, expressed as N=(P,T,F), where P={p ₀ }∪PA, _{p 0 is} an idle place , _p∈PA , _PA is the set of active places, and each cycle in N contains p ₀ .

Definition 6 A sign graph module B(T _s ,t _e ) can be represented by a Petri net that satisfies the following requirements: 1) 2) 3) The sub-network between t _s and t _e is a sign graph. Referring to Fig. 3, in Fig. 3, there are 3 explanatory diagrams of sign map modules, wherein, Fig. 3 (a) and Fig. 3 (b) are all sign map modules, and Fig. 3 (c) is not a sign map module, because Fig. 3 ( c) contains internal loops, and the logo module cannot contain internal loops.

Definition 7 In SSM, the extended state machine can be obtained by gradually replacing the active places and the pre-sets and post-sets of the active places with the sign diagram module, which is denoted as ESM.

Figure 4(a) is a simple state machine, where {t ₃ ,p ₅ ,t ₅ } in Figure 4(a) is replaced by the sign diagram module in Figure 3(a), where t in the sign diagram module ₁₁ and t ₁₅ respectively replace t ₃ and t ₅ in Fig. 4(a). From this, an ESM is obtained.

Definition 8 An enhanced extended state machine N=(P, T, F, W) is denoted as AESM, and it must meet the following conditions:

2. For each i∈N _K , For any i,j∈N _K , i≠j,

3. For each i∈N _K , (supplement the letter meaning) is an ESM;

4. For each r∈P _R , there exists a unique minimal P-semi-flow X _r ∈ N ^|P| such that {r}=||X _r ||∩P _R , And X _r (r)=1.

Referring to Fig. 5, Fig. 5 is an example structure diagram of an AESM. Among them, there are three processes, J ₁ ={p ₁ ,p ₂ ,p ₃ ,p ₄ ,p ₆ ,p ₇ ,p ₉ }, J ₂ ={p ₁ ,p ₂ ,p ₅ ,p ₈ , p ₉ }, J ₃ = {p ₁₀ , p ₁₁ , p ₁₂ , p ₁₃ , p ₁₄ }. Among them, {p ₃ ,p ₄ ,p ₆ ,p ₇ } is a sign map module, and t ₂ and t ₇ represent the shunt operation and assembly operation respectively. J ₁ contains a sign map module, so its key place is a unity, namely <p ₆ ,p ₇ >. The key places of J ₂ and J ₃ are p ₈ and p ₁₃ , respectively. σ ₁ =<t ₁ ,t ₁₀ ,t ₃ ,t ₁₁ ,t ₆ ,t ₁₂ ,t ₈ ,t ₁₃ ,t ₉ ,t ₁₄ >,σ ₂ =<t ₁ ,t ₁₀ ,t ₂ ,t ₁₁ , t ₄ , t ₅ , t ₁₂ , t ₇ , t ₁₃ , t ₉ , t ₁₄ >, σ ₁ , σ ₂ are two emission sequences, which can ensure the system to run without deadlock.

Claims (3)

1. A distributed control method for an automated manufacturing system, comprising the steps of:

1) is initialized toWherein, T_enIs a collection of enabled transitions, T_dfIs a set of transitions that cause the automated manufacturing system to run without deadlocks;

2) acquiring the current state M of the automatic manufacturing system;

3) according to the current stateM, checking whether each transition in the automatic manufacturing system is enabled, if t_iIs enabled, then T_en＝T_en+{t_iWhere t is_iIs any one transition in an automated manufacturing system;

4) judgment of T_enWhether any one element t in (a) is located in a flag map module representing an assembly operation;

5) when T is_enSimulating a transmitting element t when any one element t in the map module representing the assembly operation;

6) when T ∈ T_enAnd when t is not located in the signpost module representing the assembly operation, simulating the transmission transition t if the current resource can support the mobile Token arrival corresponding to the transmission transition tA key library of (1), then T_df＝T_df+ { t }, otherwise,carrying out step 4); wherein,the system is a set of key libraries of the system, wherein the marker graph module in the system is replaced by a library and the whole of a front set and a rear set of the library;

7) when T is_enAfter all transitions are detected, obtaining a set T of transitions which enables the system to run without deadlock_dfAnd outputting any T e to T_dfThe automatic manufacturing system transmits t, enters the next state and then returns to the step 1);

the step 5) specifically comprises the following steps:

5.1) if the current resource can support the transmission of a mobile token corresponding to T to reach one sub-key library location in the key library universe, and there is a token that can reach the corresponding sub-key library location in the key library universe in all other parallel processes of the marker map module representing the assembly operation, respectively, T_df＝T_df+{t}；

5.2) if the current resource can support the transmission of the mobile token corresponding to t to reach one sub-key library in the key library unified body, and in all other parallel processes of the mark map module representing the assembly operation, no token can reach the sub-key library in the key library unified body, then the method comprises the steps ofCarrying out step 4);

5.3) if the current resource can not support the mobile Token corresponding to the transmission t to reach one sub-key library in the key library unified body, thenCarrying out step 4);

the system comprises a key library, a mark graph module, a front-end set and a back-end set, wherein the key library consists of two parts, one part is a set of key libraries of the mark graph module which are integrated, and the other part is a set of key libraries of the system structure after the mark graph module is integrally replaced by one library and the front-end set and the back-end set of the mark graph module in the system; the key libraries of the marker graph module are integrated into a sequence pair consisting of a plurality of sub-key libraries.

2. The distributed control method for an automated manufacturing system according to claim 1, wherein the set C of key library unifications of the signature modules_MGSet of key libraries of the system after the whole replacement of the signature graph module in the system by one library and its pre-set and post-setIs the set C of key libraries for the entire automated manufacturing system.

3. The distributed control method of an automatic manufacturing system according to claim 1, wherein the key libraries of the signature module are unified into a sequential couple consisting of a plurality of sub-key libraries; the sub-key libraries are respectively distributed in each parallel process and have the same type; the number of the sub-key libraries in the key library integration is equal to the number of the parallel processes of the marker map module.