JP6918681B2 - Design support device and design support method - Google Patents

Description

The present invention relates to a design support device and a design support method for supporting the design of a product in the manufacture of a chemical product, a food product, or the like.

In the manufacture of chemicals, foods, etc., it is required to accurately and quickly design a recipe for manufacturing a product having one or more characteristics required by the manufacturer or its customer. However, when the characteristics of the product are composed of a plurality of raw materials, the combination pattern becomes enormous, and it takes a lot of time to design.

On the other hand, in recent years, design support has been considered by utilizing simulations utilizing first-principles calculation and machine learning. However, it may take a lot of time to calculate this time when trying to perform a precise simulation.

In response to this problem, Patent Document 1 (Japanese Unexamined Patent Publication No. 2016-200903) states that "a method for creating an approximate model of a structure includes a structure and a plurality of design variables that define the materials constituting the structure. Targets a structure and a plurality of characteristic values that define the materials that make up the structure. Obtain a first approximate model with the characteristic values as the objective function, which is created using a non-linear response relationship. The first Pareto solution is extracted using the model. In the first Pareto solution, at least one of the upper limit value and the lower limit value is extracted, and at least one of the extracted upper limit value and the lower limit value is fixed and extracted. Set a new value of the design variable by changing the design variable that has not been created. Create a second approximate model using the new design variable, perform multi-objective optimization calculation, and extract the second Pareto solution. ”Is disclosed (see Solution).

Japanese Unexamined Patent Publication No. 2016-200903

However, in the technique described in Patent Document 1, since the search is executed after inputting the conditions related to all the purposes and restrictions in advance, there is a solution that does not satisfy the desired characteristics if the cost other than the characteristics and the restrictions on the raw materials are excessively applied. There is a possibility that a large number will be output. That is, when searching for a predetermined solution, it is not considered that a solution may not be obtained if the constraint conditions are not set appropriately.

An object of the present invention has been made in view of the above, and an object of the present invention is to provide a design support device and a method for facilitating input of constraint conditions of a manufacturing recipe of a product having a predetermined characteristic.

In order to solve the above problems, to give an example of the design support device of the present invention, a design support device that supports the design of a design target by targeting a product or an intermediate product made by using two or more raw materials. Therefore, the target is the control means for managing at least the raw material composition information regarding the composition or composition ratio of the raw material as the management target, the characteristic requirement information required as the characteristic of the design target, and the composition or composition ratio of the design target. The input means for inputting the composition parameter that defines the permissible range of the permissible value as the value, and the raw material composition information is processed based on the characteristic requirement information input by the input means, and the conditions specified in the characteristic requirement information are satisfied. A processing means and a process for specifying the composition or composition ratio of the design object to be satisfied, and specifying the combination of raw materials belonging to the range of the conditions specified by the composition parameter with respect to the composition or composition ratio of the specified design object. It has a display means for displaying information by the output of the means, and the processing means specifies a combination of conditions different from the conditions used when specifying the combination of raw materials, and displays the specified combination of different conditions. It is characterized by displaying in.

According to the present invention, it is possible to provide a design support device and a method that facilitates input of constraint conditions in a manufacturing recipe of a product having predetermined characteristics.

It is a block diagram which shows the system structure and the function structure in this embodiment. It is a figure which shows the hardware configuration in this embodiment. It is a figure of the experimental recipe data in this embodiment. It is a figure of the cost composition data in this embodiment. It is a figure of the experimental physical characteristic data in this embodiment. It is a figure of the query data in this embodiment. It is a figure of the raw material filter data in this embodiment. It is a figure of the composition discrepancy parameter in this embodiment. It is a figure of the search result physical property data in this embodiment. It is a figure of the search result composition ratio data in this embodiment. It is a figure of the search result recipe data in this embodiment. It is a figure of the search result discrepancy data in this embodiment. It is a flowchart which shows the process flow of experiment data registration in this embodiment. It is a flowchart which shows the processing flow of the prediction model learning in this embodiment. It is a flowchart which shows the process flow of the raw material combination search in this embodiment. It is a figure of the query registration screen in this embodiment. It is a figure of the physical characteristic display screen in this embodiment. It is a figure of the raw material combination display screen in this embodiment. This is an example of a composition data display screen corresponding to the raw material type and amount in the present embodiment.

Hereinafter, typical embodiments for carrying out the present invention will be described with reference to the drawings as appropriate.

As a general rule, the explanation of the same or similar parts will not be repeated unless it is particularly necessary. Even when the reference numerals described in the predetermined figures are described in other figures, the description of the same or similar parts will be omitted.

Further, in the following embodiments, when it is necessary for convenience, the description will be divided into a plurality of sections or embodiments, but unless otherwise specified, they are not unrelated to each other, and one of them is not related to each other. There is a relationship between some or all of the other modifications, details, supplementary explanations, and so on.

In addition, in the following embodiments, when the number of elements (including the number, numerical value, quantity, range, etc.) is referred to, when it is specified in particular, or when it is clearly limited to a specific number in principle, etc. Except, the number is not limited to the specific number, and may be more than or less than the specific number.

Further, in the following embodiments, the components (including element steps and the like) are not necessarily essential unless otherwise specified or clearly considered to be essential in principle. Needless to say.

In the present embodiment, in order to support the design of a product, a plastic product, or the like having desired characteristics, for example, characteristics that meet the requirements (characteristics that meet the required specifications) in manufacturing the product. The outline of the design support device is described. Not limited to plastics, any chemicals such as rubber, foods, chemicals or pharmaceuticals can be used. Collectively, these are referred to as chemicals or simply products.

In addition, the product is not limited to the finished product, and it is possible to design an intermediate product that supports the production of the finished product that is used, mixed, or combined as a part of the finished product.

The present invention is not limited to the above-described embodiment, and when a product having characteristics that meet the requirements means a finished product as a final product, for example, a finished product is produced from the raw material plastic. It can also be applied to intermediate products (intermediate products) such as pellets and compounds obtained in the process.

In this case, the product or intermediate product is the design target. It has been described as a product including such an intermediate product. Further, it is possible to add / delete / replace a part of the configuration of the embodiment with another configuration.

In the case of foods, the above chemicals are texture, taste, hardness, texture, shape, degree of baking, etc., and in the case of chemicals or pharmaceuticals, the amount of multiple raw materials to obtain predetermined characteristics such as shape and efficacy. Because it is necessary to combine, a very complicated calculation is required.

Therefore, in the conventional method of searching for a solution after inputting conditions related to all purposes and constraints in advance, selection is made in consideration of cost and restrictions on raw materials among recipes and manufacturing methods that can achieve desired characteristics. , There is a problem that the designer cannot cope with the actual process. In addition, if too many constraints are added at the beginning, a solution that satisfies the conditions may not be found.

In actual operation, there is a problem that it is often necessary to recreate the model that has been used so far when the raw material itself becomes difficult to obtain or a new kind is obtained. No considered design support devices and methods have been provided.

The processing of the design support device in this embodiment is divided into a basic data registration phase, a prediction model construction phase, and a search phase, and are processed in order. The flow of a series of processes will be described below for each phase.

In the basic data registration phase, the basic data is the data obtained when an experiment was conducted in the past to produce a predetermined product (chemical product) by combining raw materials, and corresponds to a product having a predetermined characteristic (1) Product. Various basic data on the physical properties of products (chemicals) such as raw material combinations of (chemicals), (2) raw material costs (unit costs), composition of raw materials such as molecular weight, and (3) gloss and fluidity. Register in the database.

Next, in the prediction model construction phase, a prediction model (simulator) that predicts physical properties from the composition ratio of raw materials is constructed based on the basic data registered in the database in the basic data registration phase (experimental data registration phase).

The composition ratio includes chemical components such as elements and compounds constituting the substance, nutritional components such as vitamins, calcium, proteins and sugars, and effects, effects and the like. Further, when a chemical reaction occurs due to heat treatment or the like, information before and after the reaction may be included.

In the search phase, first, query data representing the physical property information desired by the user who is the designer of the product, and parameters indicating how much the request is allowed to be inconsistent are accepted. Next, the search for the composition ratio of the product that satisfies the physical properties of the product (chemical product) is executed using the prediction model.

Further, a recipe (raw material combination) in which the composition ratio of the product falls within a predetermined discrepancy category and the cost (unit cost) of the product or manufacturing is reduced is searched for. Then, the search result data (physical properties, composition ratio, raw material combination) is registered in the database. Not limited to the above unit cost, it is also possible to search on the condition of product manufacturing time, availability of raw materials, and the like. In this specification, a search method using a unit cost will be described as a representative.

A product that falls into the category of discrepancy is a product that does not exactly match the predetermined characteristics but has characteristics similar to the predetermined characteristics as a product. That is, it has characteristics different from the predetermined characteristics but satisfies the characteristics as a product.

That is, the first condition used to identify the ingredient combination of the product recipe in which the product satisfies the predetermined characteristics is input, and the manufactured product satisfies the characteristics similar to the predetermined characteristics, and the first condition is satisfied. Identify a second condition that is different from.

At this time, a combination of conditions different from the conditions used when specifying the raw material combination is specified, and the specified contents are registered in the database. The search result (data registered in the database) is presented to the user (displayed on the user's display terminal).

The processing of the above search phase may actually be repeated in the middle depending on the interaction with the user.

<System configuration>
Next, the configuration of the design support device 1 in the present embodiment will be described with reference to FIG. FIG. 1 is a configuration diagram showing a system configuration and a functional configuration in the present embodiment.
In FIG. 1, the design support device 1 according to the present embodiment includes at least a calculation server 11 that manages data, predicts physical properties, and searches for a combination of raw materials, and an operation terminal 12 that allows a user to input and output data. The calculation server 11 has a data registration unit 111, a prediction model construction unit 112, a search unit 113, and a data management unit 114.

The operation terminal 12 has an operation unit 121. Each component of the design support device 1 is connected to each other by a network 13 such as a LAN (Local Area Network). In the present embodiment, each component is connected by LAN, but it may be connected via WWW (World Wide Web). Further, the number of elements of the calculation server 11 and the operation terminal 12 may be 2 or more.

<Functions and hardware>
Next, the functions and hardware will be described with reference to FIGS. 1 and 2. FIG. 2 is a configuration diagram showing a hardware configuration according to the present embodiment.

In the data registration unit 111, the prediction model construction unit 112, the search unit 113, and the data management unit 114 included in the calculation server 11 of the design support device 1, the CPU (Central Processing Unit) 1H101 is the ROM (Read Only Memory) 1H102 or the external storage. The program stored in the device 1H104 is read into the RAM (Read Access Memory) 1H103, and the communication I / F (Interface) 1H105, the external input device 1H106 represented by a mouse or keyboard, the external output device 1H107 represented by a display, etc. It is realized by controlling.

The data management unit 114 functions as a management means for managing at least the raw material composition information (cost composition data 1D2 in FIG. 4) regarding the composition or composition ratio of the raw material as a management target. The search unit 113 processes the raw material composition information (cost composition data 1D2 in FIG. 4) based on the characteristic request information (query data 1D4 in FIG. 6) input by the operation unit 121 (input means) of the operation terminal 12. The composition or composition ratio of the design object satisfying the conditions specified in the characteristic requirement information (query data 1D4 of FIG. 6) is specified (step 1F303 of FIG. 15), and the composition or composition ratio of the specified design object is relative to the specified composition or composition ratio. It functions as a processing means for specifying a combination of raw materials belonging to the permissible range of the conditions specified by the composition parameter (composition discrepancy parameter 1D6 in FIG. 8) (step 1F304 in FIG. 15).

In the operation unit 121 included in the operation terminal 12 of the design support device 1, the CPU (Central Processing Unit) 1H101 converts the program stored in the ROM (Read Only Memory) 1H102 or the external storage device 1H104 into the RAM (Read Access Memory) 1H103. It is realized by controlling the reading and communication I / F (Interface) 1H105, the external input device 1H106 represented by a mouse or keyboard, and the external output device 1H107 represented by a display or the like.

The operation unit 121 (external input device 1H106) allows the characteristic requirement information (query data 1D4 in FIG. 6) required as the characteristics of the design object (product) and the composition or composition ratio of the design object as the target values. It functions as an input means for inputting at least a composition parameter (composition discrepancy parameter 1D6 in FIG. 8) that defines an allowable range of the values to be obtained. The operation unit 121 (external output device 1H107) functions as a display means for displaying information output from the processing means (search unit 113).

<Data structure>
Next, the experimental recipe data 1D1 used for constructing the prediction model, which is managed by the data management unit 114, will be described with reference to FIG. FIG. 3 is a block diagram of the experimental recipe data in the present embodiment. The experimental recipe data 1D1 is data representing a combination of raw materials for each experiment performed in the past, and includes an experiment number 1D101 and raw materials "1" to "20" (1D102 to 1D121), and is provided in the data management unit 114. Stored.

Experiment number 1D101 is a number that uniquely identifies past experiments. Raw material "1" to raw material "20" (1D102 to 1D12) represent the amount of each raw material used in each experiment. In the present embodiment, the raw materials are 20 types from the raw material "1" to the raw material "20", but this may be increased or decreased depending on the embodiment. Further, in the following, except when the raw material of any one of the raw materials "1" to "20" is specified, it may be simply referred to as a raw material.

Next, the cost composition data 1D2 managed by the data management unit 114 will be described with reference to FIG. FIG. 4 is a block diagram of the cost composition data in the present embodiment.

The cost composition data 1D2 is data representing the cost and composition for each raw material, and includes the raw material name 1D201, the unit cost 1D202, the composition "1" to the composition "10" (1D203 to 1D222), and the data management unit 114. Stored in. The raw material name 1D201 is a name that uniquely identifies the raw material.

The unit cost 1D202 is a cost per unit amount (1 kg or the like). It may be another unit amount. The compositions "1" to "10" (1D203 to 1D222) are one or more feature quantities representing the composition of the raw material, such as the molecular weight and the chemical structure of the raw material.

In the present embodiment, the composition of the raw material is 10 kinds of composition "1" to composition "10", but this may be increased or decreased depending on the embodiment.

Hereinafter, the composition may be simply referred to as a composition except when the composition of any one of the compositions "1" to "10" is specified. Further, the raw material name 1D201, the unit cost 1D202, and the information belonging to the composition "1" to the composition "10" (1D203 to 1D222) refer to the cost composition information regarding the cost (unit cost) for the raw material and the composition or composition ratio of the raw material. The raw material name 1D201 and the information belonging to the composition "1" to the composition "10" (1D203 to 1D222) constitute the raw material composition information regarding the composition or composition ratio of the raw material.

As described above, according to the concept of the unit cost 1D202, it is possible to provide a recipe for a product in which a combination of raw materials having predetermined characteristics and a low manufacturing cost is specified.

Next, the experimental physical property data 1D3 managed by the data management unit 114 will be described with reference to FIG. FIG. 5 is a block diagram of experimental physical property data corresponding to the experiment in the present embodiment. The experimental physical property data 1D3 is the data of the physical properties measured in the experimental recipe, and includes experiment number 1D301, gloss 1D302, fluidity 1D303, flexural rigidity 1D304, fracture elongation 1D305, and softening temperature 1D306. It is stored in the management unit 114.

The experimental physical property data is information indicating the physical properties of a product manufactured in the past or a product manufactured experimentally. As for the physical properties, glossy 1D302, fluidity 1D303, and the like will be described as representatives, but the characteristics of the other products described above can be used.

Physical characteristics are the characteristics of measured or analyzed manufactured products. Further, not only the physical characteristics of the product actually manufactured in the past but also the physical characteristics corresponding to the recipe of the product designed by the design support device can be used.

Experiment number 1D301 corresponds to the above-mentioned experiment number 1D101 and is a number that uniquely identifies a past experiment. The gloss 1D302, the fluidity 1D303, the flexural rigidity 1D304, the breaking elongation 1D305, and the softening temperature 1D306 are indexes representing the physical properties of the product (chemical product), respectively. In this embodiment, five types of physical property indexes are used, but this may be increased or decreased depending on the embodiment.

By displaying the experimental physical property data 1D3 in this way, it is possible to refer to the characteristics and constraints of the product after confirming the past results, so that the user can easily input the characteristics and constraints. be able to.

Further, the product is not limited to the finished product, and the data of the intermediate product that supports the production of the finished product that is used, mixed or combined as a part of the finished product can be registered in the experimental physical property data 1D3. In this case as well, the user can easily input the characteristics and constraints. It should be noted that the data of the intermediate product can be registered or designed not only in the experimental physical property data 1D3 but also in other data.

Next, the query data 1D4 managed by the data management unit 114 will be described with reference to FIG. FIG. 6 is a block diagram of query data in the present embodiment. The query data 1D4 is data for indicating to the design support device 1 what kind of physical properties or characteristics the user wants to design when designing a new product (characteristics required as characteristics of the product to be designed). Request information), which includes a query number 1D401, a physical property name 1D402, a search type 1D403, and parameters "1" and "2" (1D404, 1D405), and is stored in the data management unit 114.

The query number 1D401 is a number for uniquely identifying the query. The physical characteristic name 1D402 represents the required physical characteristic name described in the query, for example, "gloss", "fluidity", "flexural rigidity", "break elongation", and "softening temperature".

The search type 1D403 is a label indicating what the physical properties should be, and in the present embodiment, it is maximized, minimized, greater than or equal to (greater than or equal to a certain value), greater than or equal to a certain value, or less than or equal to less than or equal to (a certain value). , Less than a certain value, within a range (a range of a certain value 1 and a certain value 2), and the like.

The parameters "1" and "2" (1D404, 1D405) have parameters according to the search type 1D403, and None is set for maximization and minimization. If (less than or equal to a certain value) or less than a certain value, a value corresponding to "a certain value" is set in the parameter "1" (1D404).

For example, "7.5" is set below, "90" is set above, and "None" is set for each of the parameters "2" (1D405), and the range (a certain value 1 and a certain value 2) is set. Within the range), the parameter "1" (1D404) is set as "a certain value 1", for example, "170", and the parameter "2" (1D405) is set as "a certain value 2", for example, "220". can do.

In the present embodiment, the above-mentioned seven types of search type 1D403 are selected, but the number may be increased or decreased depending on the embodiment. Further, when the parameter to be set increases or decreases according to the increase or decrease, the parameters "3", "4" and the like may be added, and the parameters "1" and "2" (1D404, 1D405) may be added. Semi-structured data may be held inside.

Further, the search type 1D403 and the parameters "1" and "2" (1D404, 1D405) are conditions specified (specified) in the characteristic requirement information corresponding to each physical property, and are "gloss" and "flexural elongation". , "Maximization" are used in the objective function, and "fluidity", "flexural rigidity", "softening temperature", "less than or equal to", "range", and "greater than or equal to" are used as constraints.

Next, the raw material filter data 1D5 managed by the data management unit 114 shown in FIG. 1 will be described with reference to FIG. 7. FIG. 7 is a block diagram of raw material filter data in the present embodiment.

The raw material filter data 1D5 is data for limiting the raw materials used in the recipe (raw material combination) output in the search phase (raw material filter information indicating whether or not the raw material is used), and the query number 1D501, the raw material "1". -The raw material "20" (1D502 to 1D521) is provided and stored in the data management unit 114.

The query number 1D501 is a number corresponding to the query number 1D401 of the query data 1D4, and is a number for uniquely identifying the query. The raw material "1" to the raw material "20" (1D502 to 1D521) are values indicating whether or not the corresponding raw material is used (not used).

If it is used, it takes a value of "use", and if it is not used, it takes a value of "not used". These values are determined by the user. This makes it possible to identify product recipes using ingredients designated for use.

In addition, it is possible to specify a product recipe using a raw material that does not use a raw material designated as "non-used", that is, a raw material that is different from the non-used raw material by the raw material filter. When there are unused raw materials, it becomes difficult to specify and input constraints. However, since the present invention can specify a recipe for a product having characteristics similar to a predetermined characteristic, a recipe satisfying the characteristics as a product. Can be provided.

Using the above-mentioned past experimental data and product information, the user can also select a product or parameter having characteristics similar to the desired product.

Next, the composition discrepancy parameter 1D6 managed by the data management unit 114 will be described with reference to FIG.

FIG. 8 is a configuration diagram of composition discrepancy parameters in the present embodiment. The composition discrepancy parameter 1D6 allows a discrepancy in the composition ratio when searching for a recipe (ingredient combination) with respect to the composition ratio of the product at the end of the search phase (for example, the target value of the composition and the search). If the absolute value of the difference from the result is a specific example of the discrepancy and the value is within the range that is practically satisfactory), and if the corresponding recipe is not found under the specified conditions, the condition is automatically set. It is data (composition discrepancy information) indicating whether or not mitigation may be performed, and includes query number 1D601, composition name 1D602, automatic mitigation 1D603, and permissible discrepancy 1D604, and is stored in the data management unit 114.

By performing automatic relaxation, when a recipe cannot be found, it becomes possible to find a recipe close to the desired recipe with relaxed conditions. The permissible discrepancy indicates a range of characteristics that are different from the predetermined characteristics but are satisfactory as a product. The permissible discrepancy is, for example, a value representing the upper limit of the practically satisfactory range. Also called a similar specific range.

The query number 1D601 is a number corresponding to the query number 1D401 of the query data 1D4, and is a number for uniquely identifying the query. The composition name 1D602 indicates a composition name that specifies an allowable range of discrepancies, for example, "composition 1" and "composition 2". The automatic relaxation 1D603 is a value indicating whether or not a certain composition name (for example, query number = "1", composition name = composition "1") in a certain query may be relaxed until an executable solution is found at the time of searching.

If it can be relaxed, the value is "OK", and if not, the value is "NG". Allowable discrepancy 1D604 is a discrepancy in the composition ratio allowed when searching for a recipe. For example, if the permissible discrepancy 1D604 is 3% and the automatic relaxation 1D603 is "NG", the composition ratio when the recipe (executable solution) is found is guaranteed to be within plus or minus 3% of the target value. Will be done.

By this automatic relaxation 1D603, it is possible to identify a recipe of a product having characteristics that satisfy the product even when there is no exact solution that falls within the target value of the predetermined characteristics. In addition, a characteristic that falls within the range that can be relaxed is called a characteristic that is similar to the target value.

In the present embodiment, the composition name for which nothing is specified is set to the default value of "NG" for the automatic relaxation 1D603 and "3%" for the allowable discrepancy 1D604. On the other hand, when "OK" is set, it may be relaxed beyond that range.

In the present embodiment, the upside and downside of the discrepancy are not distinguished, but they may be distinguished if necessary. Further, the default value, the method of specifying the discrepancy and the mitigation may be changed as appropriate according to the intended use.

Further, the automatic relaxation 1D603 and the permissible discrepancy 1D604 are conditions specified (specified) by the composition discrepancy parameter 1D6 corresponding to the composition name 1D602, and the permissible discrepancy 1D604 is the target for the composition or composition ratio of the product. Although the difference has been described as an example of the discrepancy with the value (difference between the target value and the permissible limit value), it is a parameter (information) that defines the permissible range having characteristics that satisfy the product. Not limited to the difference, other examples of the concept of discrepancy include the square of the difference, log, approximation, etc. (values allowed as target values), which can be appropriately selected depending on the parameters.

Next, the search result physical property data 1D7 managed by the data management unit 114 will be described with reference to FIG. FIG. 9 is a block diagram of the search result physical property data in the present embodiment.

The search result physical property data 1D7 is data representing the physical property values searched in the early stage of the search phase, and includes query number 1D701, candidate number 1D702, gloss 1D703, fluidity 1D704, flexural rigidity 1D705, and fracture elongation. It is provided with 1D706 and a softening temperature of 1D707, and is stored in the data management unit 114.

The query number 1D701 is a number corresponding to the query number 1D401 of the query data 1D4, and is a number for uniquely identifying the query. The candidate number 1D702 is a number for uniquely identifying the candidates obtained as a result of searching for the query for the same query.

Next, the gloss 1D703, the fluidity 1D704, the flexural rigidity 1D705, the fracture elongation 1D706, and the softening temperature 1D707 are prediction results (search results of the search unit 113) regarding their respective physical properties. In the present embodiment, the prediction result is a pair of data such as [86.4, 1.2] in which the query number 1D701 corresponds to "1", and represents the average value (expected value) and the standard deviation, respectively. The method of expressing the prediction result may be a simpler average value or a parameter of the estimated probability density function.

Next, the search result composition ratio data 1D8 managed by the data management unit 114 will be described with reference to FIG. FIG. 10 is a block diagram of the search result composition ratio data in the present embodiment.

The search result composition ratio data 1D8 is data representing the composition values corresponding to the physical property values searched in the early stage of the search phase, and includes query numbers 1D801 and candidate numbers 1D802, and compositions "1" to "10" (1D802). ~ 1D812), and is stored in the data management unit 114.

The query number 1D801 is a number corresponding to the query number 1D401 of the query data 1D4, and is a number for uniquely identifying the query. The candidate number 1D802 is a number for uniquely identifying the candidates obtained as a result of searching for the query corresponding to the candidate number 1D702 for the same query.

The compositions "1" to "10" (1D802 to 1D812) are the values of the searched composition ratios (composition ratios of products), respectively. It should be noted that, usually, a plurality of candidates are output for a certain single unit.

Next, the search result recipe data 1D9 managed by the data management unit 114 will be described with reference to FIG.

FIG. 11 is a block diagram of the search result recipe data in the present embodiment. The search result recipe data 1D9 is recipe (raw material combination) data output at the end of the search phase, and includes query number 1D901, search number 1D902, cost 1D903, and raw material "1" to raw material "20" (1D904). ~ 1D923), and is stored in the data management unit 114.

The query number 1D901 is a number corresponding to the query number 1D401 of the query data 1D4, and is a number for uniquely identifying the query. The search number 1D902 is a number assigned to identify a feasible solution when a feasible solution is obtained as a result of searching for a raw material for a certain candidate of a certain query, and the search number 1D902 is a number assigned to identify a certain query number 1D901. A number that uniquely identifies the search result.

The cost 1D903 is the cost (raw material cost) per unit of the searched recipe. The raw material "1" to the raw material "20" (1D904 to 1D923) each represent a blending amount of each raw material, and represent a combination of blending as a whole. In this embodiment, the manufacturing costs are not explicitly treated as being about the same, but a variable manufacturing cost or the like may be considered as necessary according to the recipe.

Next, the search result discrepancy data 1D10 managed by the data management unit 114 will be described with reference to FIG. FIG. 12 is a configuration diagram of search result discrepancy data in the present embodiment.

The search result recipe data 1D10 is data indicating how much the target composition ratio (target value of the composition ratio) is different from the recipe (raw material combination) output at the end of the search phase. The query number 1D1001, the search number 1D1002, the number of types 1D1003, and the compositions "1" to "10" (1D1003 to 1D1013) are provided and stored in the data management unit 114.

The query number 1D1001 is a number corresponding to the query number 1D401 of the query data 1D4, and is a number for uniquely identifying the query. The search number 1D1002 is a number corresponding to the search number 1D902 of the search result recipe data 1D9, and is a number that uniquely identifies the search result with respect to a certain query number 1D01. The number of types 1D1003 is the number of raw materials used.

The compositions "1" to "10" (1D1003 to 1D1013) are information indicating the degree of relaxation indicating how much each composition is inconsistent with respect to the target value. It takes a positive value in the case of an upside and a negative value in the case of a downside. In addition, "0" indicates that the degree of relaxation is 0 and the value of the composition ratio is equal to the target value.

By displaying the cost 1D903 corresponding to the search number 1D902 shown in FIG. 11 and the number of types 1D1003 shown in FIG. 12, the user can know the correspondence between the manufacturing cost of the product in the product recipe and the number of types, which is more efficient. Recipes can be selected. In addition, these can also display the number of types 1D103 corresponding to products with similar characteristics within the range of discrepancy, making it possible to compare recipes having characteristics that satisfy the product and improve user convenience. do.

<Processing flow>
Next, the processing flow of the basic data registration phase of the design support device 1 in the present embodiment will be described with reference to FIG. FIG. 13 is a flowchart showing a processing flow of the basic data registration phase (experimental data registration) in the present embodiment. This process is started by the CPU 1H101 starting the data registration unit 111 that functions as a data registration program.

First, the data registration unit 111 registers the experimental recipe data 1D1 input by the user via the operation unit 121 of the operation terminal 12 in the data management unit 114 as a database (1F101).

Next, the data registration unit 111 registers the cost composition data 1D2 input by the user via the operation unit 121 of the operation terminal 12 in the data management unit 114 (step 1F102).

Finally, the data registration unit 111 registers the experimental physical characteristic data 1D3 input by the user via the operation unit 121 of the operation terminal 12 in the data management unit 114, and ends this process (step 1F103). At this time, the data management unit 114 functions as a management means for managing various registered data (information) as management targets.

Next, the processing flow of the prediction model construction phase of the design support device 1 in the present embodiment will be described with reference to FIG. FIG. 14 is a flowchart showing a processing flow of predictive model learning in this embodiment. This process is started by the CPU 1H101 starting the prediction model construction unit 112 that functions as a prediction model construction program.

First, the prediction model construction unit 112 calculates the composition ratio (composition ratio of raw materials) in each experiment using the experimental recipe data 1D1 registered in the data management unit 114 in the basic data registration phase and the cost composition data 1D2. (Step 1F201).

Next, the prediction model construction unit 112 learns the prediction model (simulator) using the composition ratio in each of the above experiments and the experimental physical property data 1D3 registered in the data management unit 114 in the basic data registration phase (step 1F202). ).

In this embodiment, CGAN (Conditional Generative Adversarial Nets) is used as a prediction model. That is, a neural network is used to construct a generative model of physical properties under a certain composition ratio condition. In the search phase described below, for example, 100 samples can be generated from CGAN and their averages and variances can be used.

Although CGAN is used in this embodiment, other statistical and machine learning models can be used. For example, a linear regression model, a support vector regression, a decision tree, another neural network, or the like can be used alone or in combination as needed. When searching for the composition ratio, it may be effective that the prediction model is lightweight (processing time is fast).

In such a case, a plurality of prediction models may be appropriately used, such as using only the final stage of the composition ratio search, or using a linear regression model or the like except for the evaluation of the final variation (variance, etc.).

In addition, this embodiment deals with a method of constructing a prediction model from various experimental data by learning, but if the prediction result can be calculated by the first-principles method, the calculation formula is adopted as the prediction model. May be good.

However, in the present embodiment, a prediction model is constructed by learning from experimental data, and for the sake of simplicity, a case where only one prediction model is handled will be described. Further, in the present embodiment, the manufacturing conditions and the like are not included in the prediction model, but they may be input as conditions if necessary.

Next, the processing flow of the search phase of the design support device 1 in the present embodiment will be described with reference to FIG. FIG. 15 is a flowchart showing a processing flow of the raw material combination search in the present embodiment. This process is started when the CPU 1H101 activates the search unit 113 that functions as a search program.

First, the search unit 113 receives the query data 1D4 (characteristic request information including physical property information) and the raw material filter data 1D5 input by the user via the operation unit 121 of the operation terminal 12, and registers them in the data management unit 114. (Step 1F301).

Next, the search unit 113 receives the composition discrepancy parameter 1D6 input from the user via the operation unit 121 of the operation terminal 12, and registers it in the data management unit 114 (step 1F302).

Next, the search unit 113 searches for a composition ratio (composition ratio of the product) that satisfies the physical properties specified in the query data 1D4 (step 1F303). NSGA-II (Non-dominated Sorting Genetic Algorithm II), which is a standard method of multi-objective optimization, is used for the search.

By adopting the multi-objective optimization method, it is possible to obtain solution candidates that maintain diversity as compared with the case of single-objective optimization of a specific physical property or a composite value in which two or more physical properties are weighted. This is ultimately effective in maintaining the diversity of recipes (ingredient combinations).

Further, in the query data 1D4, regarding the conditions specified by the search type 1D403 below, above, the range, etc., the solution candidates that do not match the specified conditions are not generated at the time of creating the solution candidates.

For example, if the flexural rigidity is restricted to be in the range of 170 to 220 and the flexural rigidity of the solution candidate generated by differential evolution is 230, the solution candidate is not adopted. , Repeat the solution candidate generation until a solution candidate that satisfies the constraint is found. In this embodiment, NSGA-II, which is a multi-objective optimization method, is adopted, but other optimization methods may be used.

Further, in the above prediction model, since samples of each physical property are obtained, not only the average value (expected value) indicating the predicted value but also the degree of blurring / variation (dispersion, etc.) can be evaluated. In the form, only the average value (expected value) is used, but if necessary, for example, it is possible to apply a penalty when the degree of blurring / variation in the prediction is large.

Next, the search unit 113 specifies that the composition ratio (composition ratio of the product) falls within the discrepancy specified by the composition discrepancy parameter 1D6 (within the allowable range specified by the composition discrepancy parameter 1D6) and is specified by the raw material filter data 1D5. A recipe (ingredient combination) that reduces the cost is searched for using only the raw materials (raw materials used) (step 1F304). Linear programming is used for the search.

That is, the range of the composition ratio of the product (for example, 12.1-5 <composition "1" <12.1 + 5) is set as a constraint condition, a problem that minimizes the cost is set, and the best solution is searched for. In addition, "12.1" shows the composition ratio of composition "1", and "5" shows the permissible discrepancy (%).

Since there are many solvers that can be calculated at high speed in the linear programming method, both commercial and non-commercial, it is possible to quickly find a low-cost recipe that satisfies the query by formulating in this way. It should be noted that the unused raw materials should not be included in the coefficient of determination. In addition, it is generally preferable that the number of types of raw materials used is small. Therefore, the result of narrowing down the raw materials to be used in the form of 2 types, 3 types, 4 types, etc. is created.

If no feasible solution is found as a result of the search, the restrictions on the range of the composition ratio of the product are gradually relaxed. For example, when the target value of the composition ratio of the composition "1" is 10% and the allowable discrepancy is 1%, the composition ratio of the composition "1" is initially smaller than 11%, and the composition ratio of the composition "1" is 9%. There are two restrictions that are greater than, but if automatic relaxation is OK, the composition ratio of composition "1" is less than 11.5%, and the composition ratio of composition "1" is greater than 8.5%. , And so on, the restrictions will be relaxed by 0.5%. In this embodiment, it is assumed that the cost is linear with respect to the amount of each raw material used.

If the manufacturing cost breaks that assumption, for example, if the cost becomes constant when the amount is less than a certain amount, it is necessary to consider a search method other than linear programming. It can also be used.

Further, as for the method of relaxing the constraint, the method of relaxing by 0.5% is only an example, and may be relaxed by another method. That is, when a solution that is a recipe corresponding to a predetermined constraint condition cannot be obtained, the condition may be set to be relaxed step by step.

Next, the search unit 113 sends the search result physical property data 1D7, the search result composition ratio data 1D8, the search result recipe data 1D9, and the search result discrepancy data 1D10, which are the search results obtained in step 1F304, to the data management unit 114. Register (step 1F305).

Step 1 In the process of F305, the method of narrowing down the raw materials to be used in the form of 2 types, 3 types, 4 types, etc. is simply to select all combinations of 2 types of raw materials and 3 types of raw materials. Set the problem in all combinations and solve the set problem. However, if the calculation load of the recipe search becomes a problem due to a large total number of raw materials, the combination of raw materials to be used may also be searched by using the exchange Monte Carlo method or the like.

Finally, the operation unit 121 presents various search results and related past results (experimental recipe data 1D1, cost composition data 1D2, experimental physical property data 1D3) having similar physical properties to the user, and ends the process (step 1F306). .. That is, these pieces of information are displayed on the screen of the operation unit 121.

As described above, a prediction model of physical properties with respect to the composition ratio of raw materials is constructed, and in the first stage search, an objective function and constraints related to the physical properties of the product are constructed, and the composition ratio of the product is used as a determining variable. do.

Next, in the second stage search, the objective function and constraints are constructed using the raw material cost (unit cost) and the composition ratio of the product, and the raw material combination (combination of raw materials) is searched by using the raw material combination as a determinant. ) Can be obtained.

By searching for raw materials by a two-step search through such a composition ratio, it is possible to flexibly set complicated objective functions and constraints according to the user's thoughts.

The recipe can be identified as a solution without the user having to enter the constraints exactly. By performing a two-step search, it is possible not only to identify recipes that match the characteristics of the input product, but also to identify recipes that are similar to the characteristics of the input product in consideration of cost and past experimental information. can. In addition, the prediction model can be reused.

In addition, when the search space becomes very large, the effect of reducing the search space can be obtained by dividing the problem step by step. In general, even if it becomes difficult to handle complicated constraints by multi-objective optimization or it often requires prior knowledge of the problem, in this embodiment, only a part (specifically, physical characteristics) is required. Perform multi-objective optimization.

After that, regarding the number of types of raw materials, the problem itself is divided into multiple linear programming problems to constrain and search, so while maintaining the diversity of candidates, the search strictly imposes important constraints of interest. Results can be obtained efficiently.

Further, the search result of the first stage and the search result of the second stage can be displayed on the screen of the operation unit 121 of the operation terminal 12. At this time, when the operation unit 121 executes an operation of changing or adding any of the conditions (information) specified in the query data 1D4, the raw material filter data 1D5, and the composition discrepancy parameter 1D6, the search unit 113 Based on the changed or added conditions, the first-stage search and the second-stage search can be executed again.

<User interface>
The query registration screen 1G1 for the operation unit 121 of the operation terminal 12 to receive the query data 1D4 and the raw material filter data 1D5 from the user will be described with reference to FIG. FIG. 16 is a configuration diagram of a query registration screen according to the present embodiment.

The query registration screen 1G1 includes a physical property input pane 1G101, a raw material filter input pane 1G102, an OK button 1G103, and a reset button 1G104. At this time, the external output device 1H107 of the operation terminal 12 functions as a display unit (query registration screen display unit) or display means for displaying the query registration screen 1G1 for receiving data from the external input device 1H106. These external output devices 1H107 and external input devices 1H106 can be implemented either integrally or separately.

The physical property input 1G101 has a function of registering search types and parameters of gloss, fluidity, flexural rigidity, fracture elongation, and softening temperature corresponding to query data 1D4. The initial state is a state in which nothing is set. The user sets the physical characteristics of the request, including one or more maximization or minimization.

The raw material filter input pane 1G102 has a function of registering the use / non-use of each raw material corresponding to the raw material filter data 1D5. "○" corresponds to use, "x" corresponds to non-use, and the initial values are all "○". Each time this symbol part is pressed, it changes to "x" in the case of "○" and "○" in the case of "x".

The user can register various data in the data management unit 114 of the calculation server 11 by inputting the physical property input 1G101 and the input pane 1G102 and pressing the OK button 1G103. When the data registration process is completed, the search process will be performed. On the other hand, when it is desired to input new physical properties from the beginning, the initial state of various input contents is reset by pressing the reset button 1G104.

A physical characteristic display screen 1G2 for the operation unit 121 of the operation terminal 12 to present the search result of the physical characteristics to the user will be described with reference to FIG. FIG. 17 is a configuration diagram of a physical characteristic display screen according to the present embodiment.

The physical property display screen 1G2 has a physical property display pane 1G201 that displays the searched physical properties, a detailed display pane 1G202 that displays the detailed physical properties and composition ratio of the candidates selected in the physical property display pane 1G201, and a raw material combination display that displays the raw material combination. It includes a button 1G203 and a back button 1G204. At this time, the external output device 1H107 of the operation terminal 12 functions as a display unit (physical property display unit) that displays the searched physical properties of the search unit 113.

In the physical property display pane 1G201, a set of candidate (Pareto solution) 1G201a obtained as a result of executing multi-objective optimization related to physical properties for which one or more maximization or minimization is set, and a past experimental result 1G201b are displayed. Here, the past experimental results are various experimental data (1D1, 1D2, 1D3).

In the example of FIG. 17, the past experimental results having similar characteristics to the candidates (Pareto solution) related to gloss and fracture elongation and the solutions in the vicinity thereof are displayed. Further, when the user selects a point, the physical property information or the outline thereof regarding the candidate or the past experiment result is displayed in the pop-up 1G201c.

Note that FIG. 17 shows a case where maximization is specified for two physical properties (“gloss” and “break elongation”), so that the graph is a two-dimensional plot.

If only one maximization or minimization is specified, the solution near the best candidate is displayed as a one-dimensional graph.

Further, when three or more maximization or minimization are specified, the three-dimensional display may be performed, or the two-dimensional graph may be displayed side by side only in the specified combination of physical properties. Further, in the present embodiment, only the physical properties for which maximization or minimization is specified are displayed on the graph, but even if the physical characteristics for which the following, greater than or equal to, larger, smaller, range, etc. are specified are displayed on the graph. good.

As described above, the detailed display pane 1G202 has detailed physical properties (“gloss”, “fluidity”, “flexural rigidity”, “break elongation”, “softening temperature”) of the candidates selected by the user in the physical property display pane 1G201. ), The predicted value (expected value) and the composition ratio are displayed in the subpanes (1G202a, 1G202b) based on the search result physical property data 1D7 and the search result composition ratio data 1D8.

At that time, by displaying the variation in the prediction result of the physical properties (the degree of blurring indicating the variance), it is possible to confirm not only the goodness of the candidate as a simple expected value but also the certainty of the prediction. Although the variance will be described as an example of the variation, other means such as mean, deviation, and correlation may be used.

By pressing the raw material combination display button 1G203, it is possible to transition to a screen (raw material combination display screen 1G3 (described later)) for displaying a recipe (raw material combination) related to the candidate selected by the user in the physical property display pane 1G201. On the other hand, by pressing the back button 1G204, it is possible to return to the setting of the query (required physical characteristics).

By providing such a user interface, the search result and the past experiment result can be compared. In addition, the physical properties corresponding to the search results and the past experimental results, and their predicted values and variations can be displayed, which enhances convenience.

Further, since the combination of raw materials corresponding to the displayed search result and the past experimental result can be displayed, it becomes easy to select the characteristic to be emphasized among the characteristics of the product.

The raw material combination display screen 1G3 for the operation unit 121 of the operation terminal 12 to present a recipe (raw material combination) to the user will be described with reference to FIG. FIG. 18 is a configuration diagram of a raw material combination display screen according to the present embodiment.

The raw material combination display screen 1G3 is a display screen for displaying the search result of the search unit 113, and includes a cost display pane 1G301, a raw material combination display pane 1G302, a relaxation condition display pane 1G303, a report output button 1G304, and a back button 1G305. At this time, the external output device 1H107 of the operation terminal 12 in FIG. 2 functions as a display unit (search result display unit) that displays the search result of the search unit 113 shown in FIG.

The degree of relaxation is the range or value of automatic relaxation, or the range or value of discrepancy. This relaxation degree is output or automatically set in the relaxation degree display pane 1G303, and the range or value is displayed.

The cost display pane 1G301 displays the cost (raw material cost) for each number of raw materials used for the candidates selected on the physical property display screen 1G2 based on the search result recipe data 1D9. For example, the cost display pane 1G301 displays information on a combination of low-cost raw materials (use of 3 types to 6 types). In this case, the use of 6 types is the minimum cost.

The raw material combination display pane 1G302 displays the combination of raw materials for each number of raw materials used for the candidates selected on the physical property display screen 1G2 based on the search result recipe data 1D9. For example, the raw material combination display pane 1G302 displays that the raw material "2" whose blending amount of the raw material is "24" and the other two kinds of raw materials (not shown) are combined in the case of "use of 3 types". become. In the case of using four types, the recipe is a combination of the raw material 2 and the raw material 3 and two other raw materials (not shown).

At this time, when the raw material filter data 1D5 indicating whether or not the raw material shown in FIG. 7 is used is input from the operation unit 121, the search unit 113 shown in FIG. 1 does not use the raw material filter data 1D5 among the raw materials. Identify combinations that do not use the specified ingredients.

Next, the combinations that do not use the specified raw materials are excluded from the combinations of raw materials that belong to the permissible range of the conditions specified by the composition discrepancy parameter 1D6. Among the combinations of raw materials that belong to the permissible range of the conditions specified by the composition discrepancy parameter 1D6, the combinations of raw materials excluding the combinations that do not use raw materials are specified. The specified raw material combination is displayed on the raw material combination display pane 1G302.

The relaxation condition display pane 1G303 displays the relaxation condition of the composition ratio for each number of raw materials used for the candidates selected on the physical property display screen 1G2 based on the search result discrepancy data 1D10. For example, the relaxation degree display pane 1G303 displays that the relaxation degree of the composition ratio of "composition 1" is "1.3%" in the case of "use of three types".

By pressing the report output button 1G304, a report for the user to request an experiment can be output. On the other hand, if you want to confirm another candidate, you can return to the physical characteristic display screen 1G2 by pressing the back button 1G305.

In this embodiment, screen examples relating to the use of 3 to 6 types of raw materials are shown, but the number can be increased or decreased according to the search results and the user's request. It is also possible to display search results based on the characteristics of the product input and the information input by the user in the past, even if it is not the user's request.

In this way, by displaying each raw material number using the search result, the user can simply compare the number of raw materials including the number of raw materials that are difficult to convert into costs.

In the example of FIG. 18, the composition ratio for each number of raw materials used is not displayed, but it may be displayed together if necessary.

Further, the contents of the search result physical property data 1D7, the search result composition ratio data 1D8, the search result recipe data 1D9, and the search result discrepancy data 1D10 are displayed on the screen of the operation unit 121 of FIG. 1 according to the search result of the search unit 113. You can also do it. It is not necessary to display all items, and the display items can be changed according to the user's settings.

Each user interface such as the physical characteristic display pane 1G201, the detailed display pane 1G202, the cost display pane 1G301, the raw material combination display pane 1G302, and the relaxation condition display pane 1G303 described with reference to FIGS. 17 and 18 can be displayed in any combination. can. In addition, there may be a pane that is not displayed if necessary.

As described above, according to the present embodiment, after specifying the composition or composition ratio of the product having the required characteristics, different conditions (for example, the composition and composition discrepancy parameter defined in the cost composition data 1D2) are specified. The conditions specified in 1D6, including the conditions that can be changed by reusing the conditions specified in the query data 1D4, etc.) are considered (added) to search for the combination of raw materials. It is possible to identify recipes for products with matching properties.

In addition, it is possible to support product design in consideration of different constraints. Furthermore, as long as there is no change in the composition or composition ratio of the raw material to be considered, the prediction model for the composition or composition ratio of the raw material can be reused even if the types of raw materials increase or decrease.

As an example of the characteristics that meet the requirements, those that satisfy the physical properties required as a product and satisfy the conditions specified in the query data 1D4 (characteristic requirement information) of FIG.

In addition, by registering the past experiment results of multiple users, it becomes possible to share the past experiment results among users from the viewpoint of the physical characteristics described in the query data, and design with a similar recipe. The situation can be avoided.

In addition, since the recipe is evaluated for the degree of relaxation indicating the discrepancy between the composition ratio and the target value, the user can select a recipe with less degree of relaxation. It is possible to search for the possibility of a new formulation by selecting a place with a large degree of relaxation.

These enter the first constraint corresponding to the product for which the present invention satisfies the desired property, identify the combination of raw materials corresponding to the first constraint, and make the property similar to the specified combination of raw materials. By searching for the combination of raw materials specified by the corresponding second constraint condition, it can be realized by the idea of searching for a plurality of conditions different from the originally input constraint condition.

A modified example of the cost composition data 1D2 described with reference to FIG. 4 will be described with reference to FIG. Unlike FIG. 4, FIG. 19 is a configuration diagram of composition data corresponding to the type and amount of the raw material in the present embodiment.

The composition data 1D2a corresponding to the type and amount of the raw material is data showing the composition that changes depending on the type and amount of the raw material. The type 1D201a of the raw material to be mixed or combined, the amount 1D202a of the raw material 1, and the amount 1D203a of the raw material 2 are shown.

When the raw materials are mixed or combined, the amount of the raw material may not increase linearly with the increase amount but may increase non-linearly depending on the composition. An equation showing this non-linear and increasing relationship is shown in Equation 1D204a of composition "1" and the like. Similarly, from the formula 1D205a of the composition "2", the formula 1D222a of the predetermined composition "n" is shown.

When the user selects the formula 1D204a of the composition "1", the graph displaying the raw material "1" and the raw material "2" as parameters can be displayed, and the user can know the amount of the composition for a predetermined combination of raw materials. ..

When the composition is non-linear with respect to the type and amount of raw materials as in this example, the search in the second stage does not become a linear programming problem, and is therefore achieved by using, for example, a genetic algorithm.

Further, in this example, two types of raw materials are used, but a combination of three or more types can be input and displayed in a three-dimensional graph or the like. The display method is not limited to these, and may be another expression method.

By showing the composition when a plurality of types of raw materials are mixed or combined in this way, the user can know the combination of raw materials having predetermined characteristics, and it is easy to input the constraint conditions required for the recipe of the product. It becomes.

To explain an example of the present embodiment, the data management unit 114 of FIG. 1 functions as a management means for managing at least the raw material composition information (cost composition data 1D2) regarding the composition or composition ratio of the raw material as a management target.

The operation unit 121 (external input device 1H106) of FIG. 1 allows the characteristic requirement information (query data 1D4 of FIG. 6) required as the characteristics of the design object and the composition or composition ratio of the design object as the target values. It functions as an input means for inputting at least a composition parameter (composition discrepancy parameter 1D6 in FIG. 8) that defines an allowable range of the values to be obtained.

The search unit 113 of FIG. 1 processes the raw material composition information (cost composition data 1D2 of FIG. 4) based on the characteristic request information (query data 1D4 of FIG. 6) input by the input means, and processes the characteristic request information (query data). The composition or composition ratio of the design target satisfying the conditions specified in 1D4) is specified (step 1F303 in FIG. 15), and the composition parameter (composition discrepancy parameter in FIG. 8) is compared with the composition or composition ratio of the specified design target. It functions as a processing means for specifying a combination of raw materials belonging to the permissible range of the conditions specified in 1D6) (step 1F304).

The operation unit 121 (external output device 1H107) of FIG. 1 can function as a display means for displaying information output from the processing means (search unit 113). The processing means (search unit 113) specifies a combination of conditions (type of raw material) different from the condition used when specifying the combination of raw materials, and displays the specified combination of different conditions by the display means (raw material combination display screen 1G3). ) Is displayed. As a result, it is possible to easily input the constraint conditions of the manufacturing recipe of the product having the predetermined characteristics.

Further, the processing means (search unit 113) indicates that the cost composition information (cost composition data 1D2) relating to the cost for the raw material and the composition or composition ratio of the raw material exists as a management target of the management means (data management unit 114). As a condition, a combination of low-cost raw materials is specified based on the cost composition information, and the specified low-cost raw material combination is displayed on the display means (cost display pane 1G301). Since the combination of low-cost raw materials is displayed, it is possible to determine the combination of low-cost raw materials as a product.

When the raw material filter information (raw material filter data 1D5 in FIG. 7) indicating whether or not the raw material is used is input to the input means (operation unit 121), the processing means (search unit 113) has the raw material filter information among the raw materials. The combination that does not use the specified raw material is specified in, and the combination that does not use the specified raw material is excluded from the combinations of raw materials that belong to the permissible range of the conditions specified by the composition parameter, and the composition is composed. Among the combinations of raw materials that belong to the permissible range of the conditions specified by the parameters, the combinations of raw materials excluding the combinations that do not use raw materials are specified, and the specified combinations of raw materials are displayed by the display means (raw material combinations in FIG. 18). It is displayed in the display pane 1G302). Since the combination of raw materials used for the product having predetermined characteristics is displayed, the combination of the raw materials used for the product can be confirmed.

The processing means (search unit 113) is information managed by the management means (data management unit 114), and is index information related to an index indicating the physical properties of a product or intermediate product different from the design target (experimental physical property data in FIG. 3). Based on 1D3), index information (past experimental results) relating to an index indicating the physical properties of a product or intermediate product manufactured before the design target is displayed on the display device (physical characteristic display pane 1G201 in FIG. 17). Since the past achievements are displayed, this information can be used as a reference when inputting the constraints of the manufacturing recipe of the product having a predetermined characteristic.

In the process of specifying the composition or composition ratio of the design target, the processing means (search unit 113) identifies the predicted value of the physical property of the design target and its variation, and determines the predicted value of the specified physical property of the design target and its variation. It is displayed on the display device (sub-pane 1G202a in FIG. 17). The predicted values of the physical properties of the design target and their variations can be visualized, and the visualized information can be used as a reference when inputting the constraint conditions of the manufacturing recipe of the product having a predetermined characteristic.

The processing means (search unit 113) further specifies the number of types of raw materials belonging to the combination of low-cost raw materials based on the specified combination of low-cost raw materials, and the number of specified types of raw materials (FIG. 12 search result discrepancy data (number of types of 1D10) is displayed on the display device. Display and visualize the number of types of raw materials that belong to low-cost raw material combinations.

When the processing means (search unit 113) cannot specify the combination of raw materials belonging to the allowable range of the conditions specified by the composition discrepancy parameter, the processing means (search unit 113) gradually relaxes the permissible range of the conditions specified by the composition discrepancy parameter. It is possible to prevent the combination of raw materials from becoming unspecified and speed up the process.

The processing means (search unit 113) is based on the specified combination of low-cost raw materials, and shows a degree of relaxation indicating the difference between the composition of the design target or the target value of the composition ratio with respect to the combination of low-cost raw materials. (Compositions "1" to "10" of the search result discrepancy data 1D10 in FIG. 12) are further specified, and the specified relaxation degree is displayed on the display device (relaxation condition display pane 1G303 in FIG. 18). Since the information on the degree of relaxation is displayed and visualized, this information can be used as a reference when inputting the constraint conditions of the manufacturing recipe of the product having a predetermined characteristic.

When the condition specified by the composition discrepancy parameter is changed or added by the input means (external input device 1H106), the processing means (search unit 113) specifies a combination of raw materials satisfying the changed or added condition. It is possible to identify a combination of raw materials that satisfy the changed or added conditions, in which the conditions for inputting the constraints of the manufacturing recipe are changed or added.

1 Design support device, 11 Calculation server, 12 Operation terminal, 111 Data registration unit, 112 Prediction model construction unit, 113 Search unit, 114 Data management unit, 121 Operation unit

Claims (15)

A design support device that supports the design of the design target by designing a product or intermediate product made from two or more raw materials.
Managing means for managing information and composition or composition ratio the material has said raw material associated with the material composition information,
An input means for inputting at least the characteristic requirement information required as the characteristic of the design object and the composition parameter defining the permissible range of the allowable value as the target value for the composition or composition ratio of the design object.
The raw material composition information is processed based on the characteristic requirement information input by the input means to specify the composition or composition ratio of the design target satisfying the conditions specified in the characteristic requirement information, and the specification is specified. A processing means for specifying a combination of the raw materials belonging to the allowable range of the conditions specified by the composition parameter with respect to the composition or composition ratio of the design target.
It has a display means for displaying information by the output of the processing means, and has.
The processing means
A design support device characterized in that a combination of conditions different from the conditions used when specifying the combination of raw materials is specified, and the specified combination of different conditions is displayed on the display means. The design support device according to claim 1.
The processing means
As the information managed by the control means, it is low based on the cost composition information on condition that there is cost composition information in which the cost for the raw material and the composition or composition ratio of the raw material are associated with the raw material. A design support device characterized in that a combination of raw materials to be cost is specified, and the specified combination of raw materials to be low cost is displayed on the display means. The design support device according to claim 1.
The processing means
When raw material filter information indicating whether or not the raw material is used is input to the input means,
Among the raw materials, a combination that does not use the raw material specified not to be used in the raw material filter information is specified, and the combination that does not use the specified raw material is within the permissible range of the conditions specified by the composition parameter. The combination of the raw materials excluded from the combination of the raw materials to which the raw material belongs, and the combination of the raw materials excluding the combination not using the raw material among the combinations of the raw materials belonging to the permissible range of the conditions specified by the composition parameter is specified. A design support device characterized in that the specified combination of raw materials is displayed on the display means. The design support device according to claim 1.
The processing means
The physical properties of a product or intermediate product manufactured in the past than the design target based on the index information regarding the index indicating the physical properties of the product or intermediate product different from the design target, which is the information managed by the control means. A design support device characterized in that index information relating to an index indicating the above is displayed on the display means. The design support device according to claim 1.
The processing means
In the process of specifying the composition or composition ratio of the design target, the predicted value of the physical property of the design target and its variation are specified, and the predicted value of the specified physical property of the design target and the variation thereof are displayed on the display means. A design support device characterized by being made to. The design support device according to claim 2.
The processing means
Based on the specified combination of low-cost raw materials, the number of types of raw materials belonging to the combination of low-cost raw materials is further specified, and the number of types of the specified raw materials is displayed on the display means. A design support device characterized by being made to work. The design support device according to claim 1.
The processing means
A design support device characterized in that when the combination of the raw materials belonging to the permissible range of the conditions specified by the composition parameters cannot be specified, the permissible range of the conditions specified by the composition parameters is gradually relaxed. The design support device according to claim 2.
The processing means
Based on the specified combination of low-cost raw materials, the degree of relaxation indicating the difference between the combination of low-cost raw materials and the target value of the composition or composition ratio of the design target is further specified. A design support device comprising displaying the specified relaxation degree on the display means. The design support device according to claim 1.
The processing means
A design support device characterized in that when the conditions specified by the composition parameters are changed or added by the input means, a combination of the raw materials satisfying the changed or added conditions is specified. A design support method that supports the design of a product or intermediate product made from two or more raw materials and uses a computer to support the design of the design target.
The computer, and a management step of a composition or composition ratio the material has said feedstock to manage information associated with the material composition information,
The computer inputs at least the characteristic requirement information required as the characteristic of the design object and the composition parameter that defines the permissible range of the allowable value as the target value for the composition or composition ratio of the design object. Input steps and
The computer processes the raw material composition information based on the characteristic requirement information input by the input step to specify the composition or composition ratio of the design target satisfying the conditions specified in the characteristic requirement information. In addition, a processing step of specifying a combination of the raw materials belonging to the allowable range of the conditions specified by the composition parameter with respect to the specified composition or composition ratio of the design target, and
The computer has a display step that displays information from the output of the processing step.
The calculator
In the processing step,
Identify a combination of conditions that is different from the conditions used when specifying the combination of raw materials.
In the display step,
A design support method comprising displaying a combination of the different conditions specified in the processing step. The design support method according to claim 10.
The calculator
In the processing step,
Wherein a information managed by the management step, and specifying a combination of raw materials composition or composition ratio the raw material cost for the raw material has become a low cost based on the cost composition information corresponding to the raw material,
In the display step,
A design support method comprising displaying a combination of the low-cost raw materials specified in the processing step. The design support method according to claim 10.
The calculator
In the processing step,
When raw material filter information indicating whether or not the raw material is used is input in the input step, a combination of the raw materials that does not use the raw material designated as non-use in the raw material filter information is specified and specified. The combination that does not use the raw material is excluded from the combination of the raw materials that belongs to the allowable range of the conditions specified by the composition parameter, and the combination of the raw materials that belongs to the allowable range of the conditions specified by the composition parameter is excluded. Among the combinations, the combinations of the raw materials excluding the combinations that do not use the raw materials are specified.
In the display step,
A design support method comprising displaying a combination of the raw materials specified in the processing step. The design support method according to claim 10.
The calculator
In the processing step,
The physical properties of the product or intermediate product manufactured in the past than the design target based on the index information regarding the index indicating the physical properties of the product or intermediate product different from the design target, which is the information managed in the control step. Identify the indicator information about the indicator that indicates
In the display step,
A design support method characterized by displaying index information relating to an index indicating the physical properties of a product or an intermediate product manufactured in the past than the design target, which is index information specified in the processing step. The design support method according to claim 10.
The calculator
In the processing step,
In the process of specifying the composition or composition ratio of the design target, the predicted values of the physical properties of the design target and their variations are specified.
In the display step,
A design support method characterized by displaying predicted values of physical properties of the design target specified in the processing step and variations thereof. The design support method according to claim 11.
The calculator
In the processing step,
Based on the specified combination of low-cost raw materials, the number of types of raw materials belonging to the low-cost combination of raw materials is further specified.
In the display step,
A design support method comprising displaying the number of types of the raw material specified in the processing step.

Images