JP5549607B2 - Manufacturing process operation support apparatus, method and program - Google Patents

Description

  The present invention relates to an operation support apparatus, method, and program suitable for supporting the operation of a manufacturing process in which a continuous product continuously flows, such as a continuous casting process.

  In order to realize a stable operation of continuous casting, it is necessary for an operator to always judge and select an appropriate operation. For that purpose, it is important to accurately predict the state of equipment and products.

  For the above prediction, operators, particularly experts, have constructed a cognitive model of the continuous casting process by integrating piecewise information about the state of steel in the continuous casting process through causal relationships. And since operators have knowledge of the cognitive model of the continuous casting process in their heads, they have avoided trouble by estimating the state of the steel, predicting the future state, and performing appropriate operations.

  On the other hand, in recent years, productivity decline due to a decrease in skilled workers, operation trouble occurrence, and the like are increasing, and it is demanded to support appropriate operations even if they are not skilled workers.

  As a countermeasure to the above-mentioned decrease in skilled persons, there is a method of predicting an operation state by performing a simulation using a numerical model expressing a physical phenomenon by a statistical analysis method. For example, Patent Document 1 predicts an operation state using a numerical model created by applying any one of principal component analysis, independent component analysis, and wavelet analysis. Since various operating factors can be incorporated into the model, a numerical model can be created even for complex processes.

  Furthermore, there is a method for searching for similar cases of past operation states from a time series database and predicting the future of operation states. For example, Patent Document 2 discloses a complicated process in which a large number of operation factors having a short period with respect to a time scale are included by reducing noise in time series data as preprocessing for searching for similar cases. It is said that a practically sufficient prediction accuracy can be obtained.

JP 2010-214417 A JP 2009-76037 A

Tadao Murata: Petri net analysis and application, Modern Science (1992)

  However, in general, in a continuous process, the state of a product takes various forms depending on temperature and components, and the state changes depending on the time and operation of the equipment. The product state changes the state of the equipment. , It also affects the condition of the product, etc. There are complex mutual interferences. The techniques disclosed in Patent Document 1 and Patent Document 2 are aimed at predicting the operation state by a simple method, but the above-described model of the phenomenon in which the state of the equipment or product interferes in a complicated manner is accurate. It is difficult to build high.

  Therefore, paying attention to the fact that the cognitive model in the expert's head takes into account the complex mutual interference of the equipment and product states as described above, There is a need to make the cognitive model of the continuous casting process objective, that is, to model causal relationships in the state transition of steel so that it can provide guidance on avoiding operational troubles and actions when operational troubles occur.

  The present invention has been made in view of the above points, and discretely models a manufacturing process in which a continuous product continuously flows, and uses the discrete model to support the operation of the manufacturing process. For the purpose.

The operation support apparatus of the manufacturing process of the present invention is an operation support apparatus for supporting the operation of the manufacturing process in which a continuous product continuously flows,
About how to divide the stages of the manufacturing process , the attribute and state of the product for each stage, the transition path of the state of the attribute of the product, the equipment and the state for each stage, and the transition path of the state of the equipment The manufacturing process is divided into a plurality of stages in the product flow direction, and a state transition model of the attribute of the product and a state transition model of the equipment are created for each stage, A discrete modeling means for receiving an input by an operator of defining an influence relationship between the state of the attribute of the product and the state of the equipment in each stage, and creating a discrete model by incorporating the influence relationship;
First state transition diagram creation that sets an initial state for each stage using the discrete model created by the discrete modeling means and creates a state transition diagram of each stage indicating a state that can be reached from the initial state Means,
And a second state transition diagram creating means for creating a state transition diagram of the entire manufacturing process using the state transition diagrams of all stages created by the first state transition diagram creating means.
Further, another feature of the manufacturing process operation support device of the present invention is that the state transition of the entire manufacturing process is performed using the state transition diagram of the entire manufacturing process created by the second state transition diagram creating means. In the figure, there is provided first risk matrix creating means for creating a risk matrix representing the relationship between the occurrence probability and the risk of each state in the figure.
Another feature of the manufacturing process operation support device according to the present invention is that, in the discrete model, between the attributes of the products between the stages, which is the influence received from the products existing in different stages at the same time. This is in the point of further defining the influence relationship.
Another feature of the manufacturing process operation support device according to the present invention is that the second state transition diagram creating means is configured to perform the manufacturing based on an influence relationship between attributes of the product between the stages. The point is to create a state transition diagram for the entire process.
Further, another feature of the manufacturing process operation support device according to the present invention is that, in the discrete model, the state of the attribute of the product in the upstream stage is on the downstream side as the product moves. It is in the point that further defines the relationship passed to the stage.
Further, another feature of the manufacturing process operation support device of the present invention is that, from the state transition diagram of the entire manufacturing process created by the second state transition diagram creating means, all states other than the initial state and Extract the transitions that lead to it,
The first state transition diagram creating means creates a new state transition diagram of each stage as a new initial state that reflects each of the extracted states and the transitions leading to the discrete model,
The second state transition diagram creating means creates a new state transition diagram of the entire manufacturing process using the new state transition diagrams of all the stages created by the first state transition diagram creating means. There is a point to repeat.
Further, another feature of the operation support apparatus of the manufacturing process of the present invention is that each of the extracted states and the transitions leading to the extracted time, the elapsed time from the start of manufacturing the product, and each of the extracted There is a second risk matrix creating means for creating a risk matrix representing the relationship between the occurrence probability of the state and the risk index.
The operation support method of the manufacturing process of the present invention is an operation support method for supporting the operation of the manufacturing process in which a continuous product continuously flows,
Discrete modeling means, the division of the stage of the manufacturing process , the attribute and state of the product for each stage, the transition path of the state of the attribute of the product, the equipment and the state for each stage, the state of the equipment The manufacturing process is divided into a plurality of stages in the flow direction of the product, and a state transition model of the attribute of the product and a state transition model of the equipment are provided for each stage. A discrete model is created by receiving the input of the definition of the influence relation between the state of the attribute of the product and the state of the equipment in each stage by the operator, and creating a discrete model incorporating the influence relation Modeling steps,
A state transition diagram of each stage showing a state that can be reached from the initial state by setting an initial state for each stage using the discrete model created by the discrete modeling step by the first state transition diagram creating means A first state transition diagram creation step for creating
A second state transition diagram creating step in which the second state transition diagram creating means creates a state transition diagram of the entire manufacturing process using the state transition diagrams of all the stages created in the first state transition diagram creating step; It is characterized by having.
The program of the present invention is a program for supporting the operation of a manufacturing process in which a continuous product continuously flows,
About how to divide the stages of the manufacturing process , the attribute and state of the product for each stage, the transition path of the state of the attribute of the product, the equipment and the state for each stage, and the transition path of the state of the equipment The manufacturing process is divided into a plurality of stages in the product flow direction, and a state transition model of the attribute of the product and a state transition model of the equipment are created for each stage, A discrete modeling process that receives an input by an operator of definition of an influence relationship between the state of the attribute of the product and the state of the equipment in each stage, and creates a discrete model by incorporating the influence relationship;
First state transition diagram creation that sets an initial state for each stage using the discrete model created by the discrete modeling process and creates a state transition diagram of each stage indicating a state that can be reached from the initial state Processing,
The computer is caused to execute a second state transition diagram creating process for creating a state transition diagram of the entire manufacturing process using the state transition diagrams of all stages created by the first state transition diagram creating process.

  According to the present invention, a manufacturing process in which a continuous product flows continuously is divided into a plurality of stages in a direction in which the product flows, and is discretely modeled. Using the discrete model, a state that can be reached from the initial state Create a state transition diagram for each stage. Then, by creating a state transition diagram for the entire manufacturing process using the state transition diagram for all stages, it is possible to support the operation of the manufacturing process, such as guidance on actions when an operation trouble is avoided or when an operation trouble occurs. .

It is a figure which shows the outline | summary of a continuous casting process. It is a flowchart which shows the procedure of the discrete modeling of a continuous casting process. It is a figure explaining the attribute of the product of each stage. It is a figure explaining a Petri net. It is a figure which shows the state transition model of the attribute of the product in a TD stage, and the state transition model of an installation. It is a figure which shows the state transition model of the temperature and hot_water | molten_metal surface level of the product in a casting_mold | template stage, and the state transition model of bulging of the product in a vertical area | region stage. It is a figure which shows the state transition model of the flow volume of the product in a nozzle stage, and the state transition model of the hot_water | molten_metal surface level of the product in a casting_mold | template stage. It is a figure which shows the structural example of the operation support apparatus of a continuous casting process. It is a flowchart which shows the procedure of the operation support by the operation support apparatus. It is a figure which shows the reachable tree created from the state transition model of the TD stage. It is a figure which shows the reachable tree produced from the state transition model of the temperature of a product in a casting_mold | template stage and a hot_water | molten_metal surface level, and the reachable tree created from the state transition model of the bulging of the product in a vertical area | region stage. It is a figure which shows the reachable tree of the casting_mold | template stage and vertical area | region stage corresponding to FIG. It is a figure which shows the state transition model of the temperature and hot_water | molten_metal surface level of the product in a casting_mold | template stage, and the state transition model of bulging of the product in a vertical area | region stage. It is a figure which shows the reachable tree of the casting_mold | template stage and vertical area | region stage corresponding to FIG. It is a figure which shows the example of a risk matrix. It is a figure which shows the example of a risk matrix. It is a flowchart which shows the procedure of the operation support by the operation support apparatus. It is a figure which shows the example of a risk matrix.

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.
(Outline of continuous casting process)
In the continuous casting process, as shown in FIG. 1, molten steel made in a converter (not shown) is put into a ladle 100 through secondary refining and is transported to the top of a continuous casting facility. Molten steel is poured from the bottom of the ladle 100 to the lower tundish 200. The molten steel poured into the tundish 200 is poured from the bottom of the tundish 200 into the mold 400 via the nozzle 300. The molten steel that has contacted the mold 400 is cooled while being precisely adjusted and solidified to form a steel slab, which is cut to an appropriate length by a gas cutting machine (not shown) at the end of the roll row while being transported by a roll (not shown). The

(Discrete modeling of continuous casting process)
First, discrete modeling of the continuous casting process as shown in FIG. 1 will be described. In the continuous casting process, a continuous (physically linked) product (molten steel, billet) flows continuously (moves). Discrete modeling of a continuous casting process having the feature of “continuous” in this way is equivalent to extracting the causal relationship recognized in “continuous” and defining it as a discrete model. Therefore, the state of the product is defined for each stage and discretized.

  FIG. 2 shows a procedure for discrete modeling of the continuous casting process. First, the continuous casting process is divided into a plurality of stages in the direction in which the product flows (step S101). In this embodiment, as shown in FIG. 3, the tundish stage (TD stage), the nozzle stage, the mold stage, and the vertical region stage are divided from the upstream side.

  Next, the attribute of the product and its state are defined for each stage (step S102). In the present embodiment, as shown in FIG. 3, temperature (high / common / low) and viscosity (high / common) are defined as the attributes and states of the product in the TD stage. Moreover, the temperature (high / normal / low), the flow rate (appropriate / low), and the flow (rectification / diffusion) are defined as attributes and states of the product in the nozzle stage. In addition, the attributes of the product at the mold stage and its state include temperature (high, ordinary, low), temperature uniformity (uniform / non-uniform), shell thickness (appropriate / thin), shell uniformity (uniform / non-uniform). ), Inclusions (with / without), surface tensile stress (strong / weak), and level (appropriate / low). In addition, as attributes and states of the product in the vertical region stage, temperature (high / normal / low), shell thickness (appropriate / thin), and presence / absence of bulging (presence / absence) are defined. In addition, the attribute of the product mentioned here and its state are examples, and are not limited to this.

  Next, a transition path of the attribute state of the product is defined for each stage (step S103). For example, in each stage, for the temperature of the product, a transition path is defined such that the movement between the high temperature and low temperature always moves after passing through the temperature.

  Next, a state transition model M1 of product attributes is created for each stage (step S104). In the present embodiment, the state transition model M1 of the product attribute is created by a Petri net. A Petri net is a graphical and executable mathematical model that expresses a discrete event system consisting of multiple processes that run in parallel and asynchronously simultaneously. As shown in FIG. 4, a state is described as a place P and a transition as a transition T. By firing T, the token moves from the input-side places P1 and P2 to the output-side place P3, thereby expressing the target behavior. The Petri net is described in detail in Non-Patent Document 1 and the like.

  FIG. 5 shows a state transition model TD-M1 of product attributes at the TD stage. Regarding the temperature, a place where the temperature is high, a place where the temperature is normal, and a place where the temperature is low are connected via transitions T9 to T12. If there is a token in the place where the temperature is low, the transition T11 can be ignited, and when it is ignited, the token moves to a place where the temperature is normal. In addition, when there is a token at a place where the temperature is normal, the transition T9 can be ignited, and when it is ignited, the token moves to a place where the temperature is high. In addition, when there is a token in a place where the temperature is high, the transition T10 can be ignited, and when it is ignited, the token moves to a place where the temperature is normal. In addition, when there is a token in a place where the temperature is normal, the transition T12 can be ignited, and when ignited, the token moves to a place where the temperature is low.

  In addition, with regard to viscosity, a place with high viscosity and a place with high viscosity are connected via transitions T13 and T14. When there is a token in a place with high viscosity, the transition T14 can be ignited, and when ignited, the token moves to a place where the viscosity is normal. In addition, when there is a token in the viscous place, the transition T13 can be ignited, and when ignited, the token moves to a viscous (high) place.

  Here, the state transition model TD-M1 of the product attribute at the TD stage has been described in detail. However, the state transition model NOZ-M1 of the product attribute is also respectively applied to the nozzle stage, the mold stage, and the vertical region stage. MD-M1 and V-M1 are created.

  If the state transition model M1 of the product attribute is created for all stages (step S105), the equipment and its state are defined for each stage (step S106). In this embodiment, a plasma heating device (ON / OFF / failure) and bubbling (ON / OFF / failure) are defined as equipment and its state on the TD stage. In addition, nozzles (stationary (non-clogging), non-stationary, failure (uneven clogging), failure (clogging)), slide (stationary / non-stationary (wide), failure) ) Is defined. In addition, the equipment and conditions at the mold stage include powder feeders (steady / unsteady (separate pattern) / failure), sprinklers (steady / unsteady (strong) / failure), rollers (steady / Non-stationary use (slow withdrawal speed / failure) is defined. In addition, the facilities mentioned here and the state are examples, and are not limited to this.

  Next, a transition path of equipment state is defined for each stage (step S107). For example, in the TD stage, for the plasma heating device (ON / OFF / failure), a transition path is defined such that the plasma heating device (failure) always moves from the plasma heating device (OFF).

  Next, an equipment state transition model M2 is created for each stage (step S108). In the present embodiment, the equipment state transition model M2 is also created by a Petri net.

  FIG. 5 shows a state transition model TD-M2 of the equipment at the TD stage. In the plasma heating apparatus, an ON place, an OFF place, and a failure place are connected via transitions T1 to T3. When there is a token in the OFF place, the transition T1 can be ignited, and when ignited, the token moves to the place of failure. Further, when there is a token in the OFF place, the transition T3 can be fired, and when fired, the token moves to the ON place. In addition, when there is a token in the ON place, the transition T2 can be ignited, and when ignited, the token moves to the OFF place.

  For bubbling, an ON place, an OFF place, and a failure place are connected via transitions T5 to T7. If there is a token in the OFF place, the transition T5 can be ignited, and when ignited, the token moves to the place of failure. Further, when there is a token in the OFF place, the transition T7 can be fired, and when fired, the token moves to the ON place. Further, when there is a token in the ON place, the transition T6 can be fired, and when fired, the token moves to the OFF place.

  Here, the equipment state transition model TD-M2 at the TD stage has been described in detail. However, the equipment state transition models NOZ-M2, MD-M2, and V for the nozzle stage, the mold stage, and the vertical region stage, respectively. -Create M2.

  If the equipment state transition model M2 is created for all stages (step S109), the influence relationship within the stage is defined for each stage (step S110). Table 1 shows the contents defining the influence relationship between the state of the equipment in the TD stage and the state of the attribute of the product. As shown in Table 1, when the plasma heating device is ON, the temperature of the product attribute changes to a normal temperature if the temperature is low, and to a high temperature if the temperature is normal. Defined. Further, when bubbling is ON, it is defined that, among the attributes of the product, the temperature transitions to a normal temperature when the temperature is high and transitions to a low temperature when the temperature is normal.

  Next, the influence relationship between the state of the equipment and the state of the product attribute defined in step S110 is incorporated into the state transition model (the state transition model M1 of the product attribute and the state transition model M2 of the equipment) (step S111). ). As shown in FIG. 5, when there is a token at the ON place in the state transition model of the plasma heating apparatus, the transition T4 is ignited via the conditional arc. As a result, when there is a token at a place where the temperature is low in the temperature state transition model, the transition T11 connected with the transition T4 by the permission synchronization arc is ignited, and the token moves to the place where the temperature is normal. In addition, when there is a token at a place where the temperature is normal, the transition T9 connected to the transition T4 by the permission synchronization arc is ignited, and the token moves to the place where the temperature is high. Further, when there is a token in the ON place in the bubbling state transition model, the transition T8 is ignited via the conditional arc. As a result, when there is a token at a place where the temperature is high in the temperature state transition model, the transition T10 connected with the transition T8 by the permission synchronization arc is ignited, and the token moves to the place where the temperature is normal. In addition, when there is a token in a place where the temperature is normal, the transition T12 connected with the transition T8 by the permission synchronization arc is ignited, and the token moves to the place where the temperature is high. The permitted synchronization arc indicates that the end point transition always fires when the start point transition fires. The conditional arc indicates that the token can be ignited when a token is present at the starting place.

  Here, the state transition model TD-M1 of the attribute of the product in the TD stage and the equipment state transition model TD-M2 are described in detail. However, the nozzle stage, the mold stage, and the vertical region stage are also staged respectively. The influence relationship between the state of the equipment and the state of the product attribute is incorporated into the state transition model (the state transition model M1 of the product attribute and the state transition model M2 of the equipment).

  Further, here, the influence relationship between the state of the equipment in the stage and the state of the product attribute has been described. However, for example, the influence relationship between the product attributes in the stage may be defined. For example, the temperature uniformity affects the shell uniformity in the mold stage (if the temperature is uniform, the shell is uniform, and if the temperature is not uniform, the shell is not uniform), the state transition model (state of the product attribute) It may be incorporated in the transition model M1).

  If the influence relationship within the stage is taken in for all stages (step S112), the influence relationship between the attributes of the product between the stages is defined (step S113). The influence relationship between the attributes of the product between the stages is an effect received from the product existing in another stage at substantially the same time. For example, it is known that when bulging occurs in the vertical region, the molten metal surface vibration is generated in the upstream mold at approximately the same time, affecting the molten metal surface level. Therefore, as indicated by the broken line arrow 31 in FIG. 3, the presence or absence of bulging in the vertical region stage is defined as affecting the level of the molten metal surface in the upstream mold stage. That is, when there is bulging in the vertical region stage, it is defined that the level shifts to low if the molten metal level at the mold stage is appropriate. Further, it is defined that when there is no bulging on the vertical region stage, an appropriate transition is made if the molten metal level at the mold stage is low.

  Next, the influence relationship between the product attributes between the stages (arrow 31 in FIG. 3) defined in step S113 is expressed as a state transition model (the product attribute state transition model M1 and the equipment state transition model M2). (Step S114). FIG. 6 shows part of the state transition model MD-M1 of the product attribute at the mold stage (temperature and hot water level) and one of the state transition model V-M1 of the product attribute at the vertical region stage. Part (bulging). These state transition models M1 are created in step S104. Regarding the temperature at the mold stage, if there is a token in a place where the temperature is low, the transition T301 can be ignited, and when ignited, the token moves to a place where the temperature is high. In addition, when there is a token in a place where the temperature is high, the transition T302 can be ignited, and when ignited, the token moves to a place where the temperature is low. Although it has been described that (high / normal / low) is defined as the temperature state, only the (high / low) state is illustrated in FIG. 6 in order to simplify the description. With respect to the level of the molten metal surface at the mold stage, when there is a token in an appropriate place, the transition T303 can be ignited, and when ignited, the token moves to a place where the token is low. In addition, when there is a token in a low place, the transition T304 can be fired, and when fired, the token moves to an appropriate place. As for bulging in the vertical region stage, when there is a token in a place without bulging, the transition T401 can be fired, and when fired, the token moves to a place with bulging. In addition, when there is a token in a place where there is a token, the transition T402 can be ignited, and when it is ignited, it moves to a place where there is no token.

  As shown in FIG. 6, when there is a token at a place in the state transition model for bulging in the vertical region stage, a transition T501 is ignited via a conditional arc. As a result, when there is a token at an appropriate place in the state transition model at the molten metal level at the mold stage, the transition T303 connected with the transition T501 by the permission synchronization arc is ignited, and the token moves to a place with a low token. Further, when there is a token in a place where there is no bulging state transition model in the vertical region stage, the transition T502 is ignited via a conditional arc. As a result, when there is a token in a low place in the state transition model of the molten metal level at the mold stage, the transition T304 connected to the transition T501 by the permission synchronization arc is ignited, and the token moves to an appropriate place.

  In the present embodiment, the example in which the attribute of the product at the downstream stage affects the attribute of the product at the upstream stage, but the opposite, that is, the product at the upstream stage. Can affect the attributes of the product at the downstream stage.

  Next, a relationship is defined in which the attribute state of the product in the upstream stage is transferred to the downstream stage as the product moves (step S115). As shown by the solid line arrow 32 in FIG. 3, for example, the temperature at the TD stage is defined as being transferred to the temperature at the nozzle stage on the downstream side. That is, if the temperature at the TD stage increases, the temperature at the downstream nozzle stage also increases, and if the temperature at the TD stage becomes normal, the temperature at the downstream nozzle stage also becomes normal, If the temperature at the TD stage decreases, the temperature at the nozzle stage on the downstream side also decreases. Similarly, it is defined that the temperature at the nozzle stage is transferred to the temperature at the mold stage, and the temperature at the mold stage is transferred to the temperature at the vertical region stage.

  Further, it is defined that the flow rate at the nozzle stage is transferred to the molten metal level at the downstream mold stage. That is, if the flow rate at the nozzle stage is appropriate, the level of the molten metal level at the downstream mold stage is also appropriate, and if the flow rate at the nozzle stage is reduced, the molten metal level at the downstream mold stage. Also lower.

  Further, it is defined that the flow at the nozzle stage is transferred to the temperature uniformity at the downstream mold stage. That is, if the flow at the nozzle stage is rectified, the temperature uniformity at the downstream mold stage is also uniform, and if the flow at the nozzle stage is uneven, the temperature is uniform at the downstream mold stage. The characteristics are also uneven.

  Next, the state transition model defines the relationship (arrow 32 in FIG. 3) in which the state of the product attribute at the upstream stage defined in step S115 is transferred to the downstream stage as the product moves. They are incorporated into (attribute state transition model M1 and facility state transition model M2) (step S116). FIG. 7 shows a part (flow rate) of the state transition model NOZ-M1 of the product attribute at the nozzle stage and a part (water surface level) of the state transition model MD-M1 of the product attribute at the mold stage. ). These state transition models M1 are created in step S104. As for the flow rate at the nozzle stage, when there is a token in an appropriate place, the transition T201 can be ignited, and when ignited, the token moves to a place with few tokens. In addition, when there are tokens in a small number of places, the transition T202 can be fired, and when fired, the tokens move to an appropriate place. The hot water level on the mold stage is as described with reference to FIG.

  Then, as shown in FIG. 7, an appropriate place of the state transition model of the flow rate at the nozzle stage and an appropriate place of the state transition model of the molten metal level at the downstream mold stage are transferred via the transition T601. Further, a small number of places in the state transition model of the flow rate at the nozzle stage and a low place in the state transition model at the molten metal level at the downstream mold stage are connected via the transition T602. Hereinafter, the transition defined in step S116 is referred to as a transfer transition. If there is a token at an appropriate place in the state transition model of the flow rate at the nozzle stage, the transfer transition 601 can be ignited, and when ignited, the state transition model at the molten metal level at the mold stage on the downstream side is ignited. The token moves. In addition, when there are tokens in a small number of places in the state transition model of the flow rate at the nozzle stage, the transfer transition 602 can be ignited, and when ignited, the low state transition model at the mold surface level on the downstream side of the mold stage is low. The token moves to the place.

  In addition, when connecting to the downstream stage, the corresponding place is connected via a transfer transition. In the example of FIG. 7, since the hot water level does not connect to the downstream stage, delivery transitions T603 and T604 in which no place is connected to the outlet side are defined.

  The transfer transitions defined in step S116 (the transfer transitions T601 to T604 in the example of FIG. 7) are not ignited alone, and all of them are ignited simultaneously when the state of the product attribute changes in any stage. . For example, in the state shown in FIG. 7 (a token is in an appropriate place in the state transition model of the flow rate at the nozzle stage and a token is in an appropriate place in the state transition model of the mold level at the mold stage) When the flow rate is low (when a token moves to a place with a low flow rate), all transfer transitions fire simultaneously. As a result, the token of the place with a small flow rate moves to a place with a low hot water level due to the firing of the transfer transition T602, and the token of the appropriate place at the hot water level disappears due to the firing of the passing transition T603. Again, when the product attributes change at any stage, all delivery transitions fire simultaneously.

  Further, for example, in the state shown in FIG. 7, if the temperature of the product of any stage changes, all the transfer transitions are ignited simultaneously. As a result, the token of the place with the appropriate flow rate moves to the appropriate place of the hot water level due to the ignition of the transfer transition T601, and the token of the appropriate place of the hot water level disappears due to the ignition of the transfer transition T603. In this arrangement, the flow rate of the nozzle stage and the state of the mold surface level of the mold stage are not affected by the change in temperature and do not change from the previous state.

  The discrete modeling described above can be performed using a computer device that can execute an algorithm of discrete modeling. That is, when an operator inputs through a stage dividing method, a product attribute and its state for each stage, and a transition path of a product attribute state (steps S101 to S103), the computer apparatus automatically performs the stage. A state transition model for each product attribute is created (step S104). Similarly, when the operator inputs the equipment for each stage and its state, and the transition path of the equipment state (steps S106 to S107), the computer apparatus automatically creates the equipment state transition model for each stage (step S106). S108). When the operator inputs an influence relationship within the stage (step S110), the computer apparatus automatically incorporates the influence relationship into the state transition model (step S111). When the operator inputs the influence relationship between the attributes of the product between the stages (step S113), the computer apparatus automatically incorporates the influence relationship into the state transition model (step S114). Further, when the operator inputs a relationship in which the attribute state of the product at the upstream stage is transferred to the downstream stage as the product moves (step S115), the computer device automatically The relationship is incorporated into the state transition model (step S116).

  Although the Petri net has been described in this embodiment, the present invention is not limited to this as long as it models a discrete event system, and the present invention is applied to other graph models such as a directed graph and an undirected graph. Is also possible. For example, in a directed graph, a place in a Petri net model is represented by a point, and a transition is represented by a line with an arrow. A line with an arrow is a moving operation end for moving a product, that is, a token from point to point, and serves as a moving path. The arrow indicates the direction in which the token moves from line to point or from point to line. Even when the present invention is applied to a graph model having such characteristics, a series of operations are the same as the operations in the above-described Petri net model, and detailed description thereof is omitted.

(Operation support using discrete models created by discrete modeling 1)
FIG. 8 is a diagram illustrating a configuration example of an operation support apparatus for a continuous casting process. 1 is a discrete modeling unit, and as described above, the continuous casting process is divided into a plurality of stages in the product flow direction, and a state transition model of product attributes and a state transition model of equipment are created for each stage. A discrete model created by defining the influence relationship in each stage is created. In addition, if necessary, the influence relationship between the attributes of the product between the stages (arrow 31 in FIG. 3), which is an influence received from the product existing in another stage at the same time, the above-mentioned relationship in the upstream stage. The state of the attribute of the product defines a relationship (arrow 32 in FIG. 3) that is transferred to the downstream stage as the product moves.

  Reference numeral 2 denotes a reachable tree creation unit which is a first state transition diagram creation unit, which uses the discrete model created by the discrete modeling unit 1 to set an initial state for each stage, and is reachable from the initial state A reachable tree that is a state transition diagram of each stage showing is generated. 3 is an overall reachable tree creation unit which is a second state transition diagram creation unit, and is a state transition diagram of the entire continuous casting process using reachable trees of all stages created by the reachable tree creation unit 2. Create a reachable tree.

  4 is a risk matrix creation unit, which uses the reachable tree of the entire continuous casting process created by the overall reachable tree creation unit 3 to represent a risk assessment matrix that represents the relationship between the occurrence accuracy and the risk level (in this application, simply the risk matrix). To create). Reference numeral 5 denotes an input device. For example, the discrete modeling unit 1 inputs and sets various definitions when creating a discrete model, and the initial state when the reachable tree creating unit 2 creates a reachable tree is input and set. To do. An output device 6 corresponds to, for example, a display device that displays a risk matrix created by the risk matrix creation unit 4 on the screen.

  Although illustrated as one device in FIG. 8, the functional configuration shown in FIG. 8 may be realized by cooperation of a plurality of devices, for example, discrete modeling is performed by another device. .

  FIG. 9 shows a procedure for operation support by the operation support apparatus. First, a discrete model is created by the above-described discrete modeling (step S201). In the case of this example, the purpose is to present to the operator in detail the risks that may occur in the near future, including the risks that are unlikely to occur. I don't think. That is, a discrete model is created in steps S101 to S114 in FIG. 2, and steps S115 and S116 in FIG. 2 are not performed. That is, the state of the attribute of the product in the upstream stage changes with the movement of the product. A discrete model is created without incorporating the relationship (arrow 32 in FIG. 3) passed to the downstream stage.

  Next, using the discrete model created in step S201, an initial state is set for each stage, for example, the current state of each stage, and a reachable tree that can be transitioned (reachable) from the initial state is created (step) S202). As shown in FIG. 10, the reachable tree is obtained by connecting “states” (ellipses describing the states) via “transitions (corresponding to the contents of the operation)” (ellipses with dots). . FIG. 10 shows a reachable tree created from the state transition model of the TD stage (see FIG. 5). Here, the initial state is “low temperature, high viscosity, plasma heating device OFF, bubbling failure”, and the next “state” that can be reached is “low temperature, high viscosity, plasma heating when transition T1 ignites. “Device failure, bubbling failure” or “Temperature normal, high viscosity, plasma heating device ON, bubbling failure” when transition T3 is ignited (accordingly, transitions T4, T11 are also ignited). In this way, the “state” that can be reached next is sequentially obtained to create a reachable tree.

  Although the reachable tree in the TD stage has been described in detail here, the initial state is set for each of the nozzle stage, the mold stage, and the vertical region stage, and a reachable tree is created. FIG. 11 shows a reachable tree created from a state transition model (see FIG. 6) of the product temperature and the molten metal level at the mold stage, and a state transition model (FIG. 11) of the product bulging at the vertical region stage. 6)). In the mold stage, when the initial state is “low temperature, appropriate level of molten metal level”, the “state” that can be reached next is “high temperature, appropriate level of molten metal level” when transition T301 ignites, or transition T303. "Low temperature, low water level" when the fire ignites. In the vertical region stage, when the initial state is “no bulging”, the next “state” that can be reached is “with bulging” when the transition T401 is ignited. In this way, the “state” that can be reached next is sequentially obtained to create a reachable tree.

  In the meantime, the reachable tree of the TD stage in FIG. 10 is a state transition diagram in which the state transitions only in one direction (cannot return to the original state), whereas the mold stage and the vertical region stage in FIG. The reachable tree is a state transition diagram that can be freely returned to the original state. This is a state transition diagram in which the transition state T12 cannot be fired even if the state transitions due to the firing of the transition T11 because the “bubble failure” is the initial state in the TD stage, and the transition is only in one direction. . Thus, the reachable tree differs depending on how the initial state is given.

  Next, a reachable tree (one reachable tree) of the entire continuous casting process is created using the reachable trees of all stages created in step S202 (step S203).

  Here, depending on the stage, there may be a case where the influence relationship (the arrow 31 in FIG. 3) between the attributes of the product is defined between the stages, and a case where it is not defined. First, with reference to FIG. 11, a description will be given of how to create reachable trees for both stages when an influence relationship between product attributes is defined between stages. FIG. 11 uses a reachable tree created from the state transition model of the product temperature and the molten metal level at the mold stage and a reachable tree created from the state transition model of the product bulging at the vertical region stage. This shows how to create reachable trees for both stages.

  As described above with reference to FIG. 6, when there is a token at a certain place in the bulging state transition model in the vertical region stage, the transition T501 is ignited via the conditional arc. As a result, when there is a token at an appropriate place in the state transition model at the molten metal level at the mold stage, the transition T303 connected with the transition T501 by the permission synchronization arc is ignited, and the token moves to a place with a low token. Further, when there is a token in a place where there is no bulging state transition model in the vertical region stage, the transition T502 is ignited via a conditional arc. As a result, when there is a token in a low place in the state transition model of the molten metal level at the mold stage, the transition T304 connected to the transition T501 by the permission synchronization arc is ignited, and the token moves to an appropriate place.

  In this case, as shown in FIG. 11, the “state” of the mold stage and the vertical region stage is regarded as a place, “transition” is regarded as a transition, and transitions T501 and T502 that connect the stages are regarded as transitions T700 to T703. Transition ”and places P701 and P702. Modeling the “switching” where the hot water level transitions from “appropriate” to “low” only when bulging is “present” and the hot water level transitions from “low” to “appropriate” only when bulging is not present Yes. Although it does not exist in this example, the hot water level changes from “low” to “appropriate” only when the bulging is changed from “present” to “other” (for example, “somewhat present”). do not do. In order to make the transition to the “appropriate” level, the bulging state must always transit to “none”. In order to express this, modeling as shown in FIG. 11 is performed. The transition T501 is a transition that always fires when a token exists with bulging, and the transition T502 is a transition that always fires when a token exists without bulging. The place P701 is a place that becomes a condition for firing the transition T700, and the place P702 is a place that becomes a condition for firing the transition T701. The transition T700 is a transition that always fires when a token exists in the place P701, and the transition T701 is a transition that always fires when a token exists in the place P702. Further, when there is a token in the place P701, the transition T702 is ignited via the conditional arc, and as a result, the transition T304 connected to the transition T702 by the permission synchronization arc is ignited. Further, when there is a token in the place P702, the transition T703 is ignited via the condition arc, and as a result, the transition T303 connected to the transition T703 by the permission synchronization arc is ignited.

  Thereby, as shown in FIG. 12, reachable trees of both stages having the initial state “low temperature, appropriate level of molten metal level, no bulging” are created.

  Next, with reference to FIG. 13 and FIG. 14, a description will be given of how to create reachable trees of both stages when the influence relationship between the attributes of the product between the stages is not defined. Here, including the meaning of the comparison, as shown in FIG. 13, it is assumed that the influence relationship between the attributes of the product is not defined between the mold stage and the vertical region stage, and the reachable trees of both stages are defined. The case of creating will be described. In this case, the reachable tree is created with all combinations of the “state” of the mold stage and the “state” of the vertical area stage as “states”, respectively. That is, in the case of the example of FIG. 14, a reachable tree having eight “states” that are all combinations of the four “states” of the mold stage and the two “states” of the vertical region stage is created. Thereby, as shown in FIG. 14, reachable trees of both stages having the initial state “low temperature, appropriate level of molten metal level, no bulging” are created.

  As shown in FIG. 12, when the influence relationship between product attributes is defined between stages, transition is possible among all combinations of “state” of the mold stage and “state” of the vertical area stage. Only “state” will be selected. That is, the combination of “states” is limited, and the number of “states” is reduced as compared with the case where the influence relationship between the attributes of the product between the stages is not defined as shown in FIG.

  According to the above algorithm, the reachable tree of the entire continuous casting process is created using the reachable trees of all stages. This single reachable tree means a set of states in which the products and equipment of the entire continuous casting process can transition from an initial state.

  Next, using the reachable tree of the entire continuous casting process created in step S203, a risk matrix representing the relationship between the occurrence probability and the risk is created (step S204). FIG. 15 shows an example of the risk matrix. In the illustrated example, for simplicity of explanation, an example is shown in which a risk matrix is created using reachable trees of the mold stage and the vertical region stage. Create a matrix.

  The horizontal axis of the risk matrix represents the probability of occurrence, and the vertical axis represents the degree of risk. The “state” of the reachable tree of the entire continuous casting process (the reachable tree of the mold stage and the vertical area stage in the example of FIG. 15) is rearranged. It is a thing. In the horizontal axis direction, an initial state is first positioned, and then a “state” that can be reached from the initial state is positioned. The minimum number of transitions required to reach each “state” from the initial state is defined as the “distance” from the “initial state” to each “state”. The smaller the “distance”, the lower the probability of occurrence. Position to be. Here, the number of executions of transitions from the initial state is defined as “distance”. However, considering the probability of each “transition” being executed (probability of firing), etc., each “state” from the initial state is You may make it prescribe | regulate a distance.

  In the transitions connecting the initial state and each “state” in FIG. 12, transitions other than the transition of the minimum distance are removed, and the “state” is generated with a lower probability of occurrence as the “distance” from the initial state is smaller. To do.

  Further, a risk level is determined in advance for each attribute of the product, and the risk level is determined according to the combination. Here, the “state” of “high temperature, appropriate level of hot water level, no bulging” is the safest state of operation, whereas “state” of “low temperature, low hot water level, with bulging” is It becomes the most dangerous state in operation.

  By displaying and presenting such a risk matrix on the screen, it is possible to present information that is effective in operator decision making, such as estimation of trouble occurrence. In the reachable tree mapped to the risk matrix, “state of high occurrence accuracy and low risk”, “state of low occurrence accuracy and low risk”, etc. are shown, and any “state” Is comprehensively visualized, it is easy to look down on the cause of the trouble. For example, it can be recognized that the “state” located at the lower left of the risk matrix is “a state where the occurrence probability is high and the degree of danger is high”, and that attention should be paid together with the route to that state. In other words, “transition” in which the degree of danger is reduced from the initial state “low temperature, appropriate level of molten metal level, no bulging” means an operation to be performed.

  When the risk matrix is displayed on the screen, for example, the “highest occurrence probability and high risk state” may be displayed prominently. Further, as indicated by a box 1501 in FIG. 15, a route (scenario) leading to a dangerous state (or a route leading to a safe state (scenario) conversely) may be surrounded and displayed.

  Although the number of “states” is large as described above, the reachable tree of the entire continuous casting process can be obtained without defining the influence relationship between the attributes of the product between the stages (arrow 31 in FIG. 3). Can be created. That is, based on the discrete model created only in steps S101 to S112 in FIG. 2, a reachable tree for each stage is created, and a reachable tree for the entire continuous casting process is created using them to create a risk matrix. You may make it do. For reference, FIG. 16 shows an example of creating a risk matrix using reachable trees of the mold stage and the vertical region stage when the influence relationship between product attributes between stages is not defined.

(Operation support using a discrete model created by discrete modeling 2)
In the operation support 1 described above, an example in which a reachable tree indicating a state transition that can be reached from the initial state is created and presented as a risk matrix has been described. In other words, Operation Support 1 is used when it is desired to grasp the danger in the near future. A risk matrix is created under the condition that the steel remains at the same stage, and there is a possibility that a risk in the near future can occur. The purpose was to present it to the operator in detail, including the low risk. On the other hand, in the operation support 2, an example in which state transitions accompanying the progress of the continuous casting process are presented as a risk matrix will be described. This operation support 2 is used when it is desired to grasp the danger in the far future, and a risk matrix including the far future can be created in consideration of moving to the stage where the steel follows. That is, the operation support 2 does not present risks that are unlikely to occur in the near future, but can present risks to the operator in the distant future, that is, when the product moves to the subsequent stage.

  The configuration example of the operation support device for the continuous casting process is the same as that of FIG. 8, but in the case of this example, the discrete modeling unit 1 shows that the state of the product attribute at the upstream stage is accompanied by the movement of the product. Therefore, the relationship (arrow 32 in FIG. 3) to be transferred to the downstream stage is always defined.

  Then, after giving the initial state and creating the reachable tree of the entire continuous casting process by the overall reachable tree creating unit 3, the state is obtained from the reachable tree of the entire continuous casting process created by the overall reachable tree creating unit 3 ”And the“ transition ”leading to it, and the reachable tree creation unit 2 reaches the reach of each stage with the initial state as the state in which the extracted“ state ”and the“ transition ”leading to it are reflected in the discrete model The step of creating a tree and the step of creating the reachable tree of the entire continuous casting process by using the reachable trees of all stages created by the reachable tree creating unit 2 by the overall reachable tree creating unit 3 .

  FIG. 17 shows a procedure for operation support by the operation support apparatus. First, a discrete model is created by the above-described discrete modeling (step S301). In the case of this example, steps S115 and S116 in the figure are also performed, and the state of the product attribute at the upstream stage is transferred to the downstream stage as the product moves (arrow 32 in FIG. 3). ) To create a discrete model. In the operation support 2 as well, as in the operation support 1, it is preferable to define the influence relationship between the attributes of the product between the stages (arrow 31 in FIG. 3), but it is not always necessary to define it.

  Next, the initial state, for example, the current state of each stage, is set for each stage using the discrete model created in step S301, and a reachable tree for each stage is created (step S302). Next, a reachable tree (one reachable tree) for the entire continuous casting process is created using the reachable trees of all stages created in step S302 (step S303). Then, a risk matrix is created using the reachable tree of the entire continuous casting process created in step S303 (step S304). In addition, the process of step S302-S304 is the same as that of step S202-S204 of the operation support 1, The detailed description is abbreviate | omitted here.

  Next, it is determined whether or not the necessary number of times (necessary operation time) has been completed (step S305). If the required number of times has not been completed, the “state” on the risk matrix created in step S304 based on the occurrence accuracy and the risk (for example, the occurrence accuracy and the risk are quantified and the product thereof is the highest), Alternatively, the “state” artificially selected by the operator is extracted, and “transition” leading to the extracted “state” is also extracted (step S306). The “transition” and “state” not extracted here are not presented to the operator. The “state” extracted here and the “transition” leading to it need not be one, and a plurality of “states” may be extracted.

  Then, the state of the discrete model created in step S301 is changed based on the “state” extracted in step S306 and the “transition” leading to it. When the state changes, the state of the product attribute in the upstream stage is transferred to the downstream stage as the product moves (arrow 32 in FIG. 3), and the state at this time is initialized. A reachable tree for each stage is created as a state (step S307).

  Thereafter, returning to step S303, the reachable tree (one reachable tree) of the entire continuous casting process is created using the reachable trees of all the stages created in step S307 (step S303). Then, a risk matrix is created using the reachable tree of the entire continuous casting process created in step S303 (step S304). If a plurality of “states” and “transitions” leading to them are extracted in step S306, the processes of steps S307, S303, and S304 are executed for each. Although it has been described that the risk matrix is created in step S304, it is not always necessary to create the risk matrix, and “state” and “transition” leading to it are extracted from the reachable tree of the entire continuous casting process created in step S303. You may do it.

  Until the required number of times is completed in this manner, the creation of reachable trees for each stage, the creation of reachable trees for the entire continuous casting process, and the creation of the risk matrix are repeated. A risk matrix representing the relationship with the index is created (step S308). FIG. 18 shows an example of the risk matrix. The horizontal axis of the risk matrix represents the casting time, and the vertical axis represents the occurrence accuracy and the risk. On the vertical axis, A means high risk, B means high risk, C means low risk, a means high occurrence accuracy, b means high occurrence accuracy, and c means low occurrence accuracy. In FIG. 18, a state 1801 is the initial state set in step S302, and a state 1802 is the “state” extracted in step S306. States 1803 and 1804 are “states” extracted in step S306 through a single loop of steps S307, S303, and S304.

  As described above, it is possible to predict the action to be taken for each cause and the phenomenon that will occur as a result, or to predict the phenomenon that will occur if no action is taken. It is possible to support the operation of the continuous casting process, for example, by guiding the actions.

  In the present invention, software (program) for realizing the functions of the present invention is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, etc.) of the system or apparatus reads the program. It can also be realized by executing.

  In the above embodiment, the continuous casting process has been described as an example. However, the present invention is applicable to any manufacturing process in which a continuous product flows continuously. For example, hot rolling (a steel plate flows continuously) or petroleum It can also be applied to plants (liquid flows continuously).

1: discrete modeling unit 2: reachable tree creation unit 3: overall reachable tree creation unit 4: risk matrix creation unit 5: input device 6: output device

Claims (9)

An operation support device for supporting operation of a manufacturing process in which continuous products continuously flow,
About how to divide the stages of the manufacturing process , the attribute and state of the product for each stage, the transition path of the state of the attribute of the product, the equipment and the state for each stage, and the transition path of the state of the equipment The manufacturing process is divided into a plurality of stages in the product flow direction, and a state transition model of the attribute of the product and a state transition model of the equipment are created for each stage, A discrete modeling means for receiving an input by an operator of defining an influence relationship between the state of the attribute of the product and the state of the equipment in each stage, and creating a discrete model by incorporating the influence relationship;
First state transition diagram creation that sets an initial state for each stage using the discrete model created by the discrete modeling means and creates a state transition diagram of each stage indicating a state that can be reached from the initial state Means,
Manufacturing comprising: second state transition diagram creating means for creating a state transition diagram of the entire manufacturing process using the state transition diagrams of all stages created by the first state transition diagram creating means Process operation support device.   Using the state transition diagram of the entire manufacturing process created by the second state transition diagram creating means, a risk matrix that represents the relationship between the occurrence probability of each state in the state transition diagram of the entire manufacturing process and the risk level is created The manufacturing process operation support apparatus according to claim 1, further comprising: a first risk matrix creating unit configured to perform the manufacturing process.   3. The discrete model further defines an influence relationship between product attributes between stages, which is an influence received from a product existing in another stage at the same time. Operation support device for the described manufacturing process.   The said 2nd state transition diagram preparation means creates the state transition diagram of the said whole manufacturing process based on the influence relationship between the attributes of the product between the said stages. Manufacturing process operation support equipment.   The discrete model further defines a relationship in which an attribute state of the product at an upstream stage is transferred to a downstream stage as the product moves. The operation support apparatus of the manufacturing process of any one of 1-4. From the state transition diagram of the entire manufacturing process created by the second state transition diagram creating means, extract all states other than the initial state and transitions leading to them,
The first state transition diagram creating means creates a new state transition diagram of each stage as a new initial state that reflects each of the extracted states and the transitions leading to the discrete model,
The second state transition diagram creating means creates a new state transition diagram of the entire manufacturing process using the new state transition diagrams of all the stages created by the first state transition diagram creating means. 6. The operation support device for a manufacturing process according to claim 5, wherein the operation support device is repeated.   Using each of the extracted states and transitions leading thereto, a risk matrix is created that represents the relationship between the elapsed time from the start of manufacture of the product and the occurrence accuracy and risk index of each of the extracted states. The manufacturing process operation support apparatus according to claim 6, further comprising second risk matrix creation means. An operation support method for supporting operation of a manufacturing process in which a continuous product continuously flows,
Discrete modeling means, the division of the stage of the manufacturing process , the attribute and state of the product for each stage, the transition path of the state of the attribute of the product, the equipment and the state for each stage, the state of the equipment The manufacturing process is divided into a plurality of stages in the flow direction of the product, and a state transition model of the attribute of the product and a state transition model of the equipment are provided for each stage. A discrete model is created by receiving the input of the definition of the influence relation between the state of the attribute of the product and the state of the equipment in each stage by the operator, and creating a discrete model incorporating the influence relation Modeling steps,
A state transition diagram of each stage showing a state that can be reached from the initial state by setting an initial state for each stage using the discrete model created by the discrete modeling step by the first state transition diagram creating means A first state transition diagram creation step for creating
A second state transition diagram creating step in which the second state transition diagram creating means creates a state transition diagram of the entire manufacturing process using the state transition diagrams of all the stages created in the first state transition diagram creating step; A method for supporting the operation of a manufacturing process, comprising: A program for supporting the operation of a manufacturing process in which a continuous product flows continuously,
About how to divide the stages of the manufacturing process , the attribute and state of the product for each stage, the transition path of the state of the attribute of the product, the equipment and the state for each stage, and the transition path of the state of the equipment The manufacturing process is divided into a plurality of stages in the product flow direction, and a state transition model of the attribute of the product and a state transition model of the equipment are created for each stage, A discrete modeling process that receives an input by an operator of definition of an influence relationship between the state of the attribute of the product and the state of the equipment in each stage, and creates a discrete model by incorporating the influence relationship;
First state transition diagram creation that sets an initial state for each stage using the discrete model created by the discrete modeling process and creates a state transition diagram of each stage indicating a state that can be reached from the initial state Processing,
A program for causing a computer to execute a second state transition diagram creating process for creating a state transition diagram of the entire manufacturing process using the state transition diagrams of all stages created by the first state transition diagram creating process.

Images