METHOD FOR CLASSIFYING PREDICTED LONGITUDINAL FALLS OF MATERIAL IN AN INDUSTRIAL MANUFACTURING PROCESS, ASSOCIATED SYSTEM

Abstract

METHOD FOR CLASSIFYING FORECAST LONGITUDINAL MATERIAL WASTE IN AN INDUSTRIAL MANUFACTURING PROCESS, ASSOCIATED SYSTEM Method for processing material waste in an industrial manufacturing process comprising: Acquisition (ACQ1) of a first planning (PL1); Acquisition (ACQ2) of a first optimization plan (POA); Characterization (CAR1) of at least one forecast waste (CHk) defining at least one zone of the optimization plan (POA); Querying a first article database (BD1) comprising a set of parts (Pn) to be produced; First comparison (COMP1) between the dimensions (Kc) of the at least one forecast waste (CHk) and the dimensions of the parts to be produced (Pn) extracted from the first article database (BD1); Determination of a list (LIST1) of compatible parts to be produced; Classification (CLASS1) of each waste forecast (CHk);Update of a falls database (BD2). Figure for the abstract: Fig.1.

Description

Title of the invention: METHOD FOR CLASSIFYING FORECAST LONGITUDINAL FALLS OF MATERIAL IN AN INDUSTRIAL MANUFACTURING PROCESS, ASSOCIATED SYSTEM Field of the invention

[0001] The field of the invention relates to that of computer-implemented methods and systems for classifying material scraps produced during a manufacturing process for industrial parts. More particularly, the field of the invention relates to that of methods aimed at reducing the production of scraps in material cutting operations. The field of the invention relates more particularly to the cutting of material profiles such as metallic materials, but may also find applications in the processing of glass, plastic such as polymers, wood, composite materials or even textiles. State of the art

[0002] There are many solutions today aimed at planning and optimizing the production of a set of industrial parts cut from material profiles. Optimizing this operation is generally achieved by producing an optimization plan, i.e. optimizing the arrangement of parts within a longitudinal profile to optimize the reduction of material waste. There are algorithms aimed at optimizing this arrangement based on constraints dependent on the dimensions of the parts, their occurrence, the necessary margins, the cutting technique, etc.

[0003] However, during a planning operation for the production of industrial parts, there is still a production of significant material waste. The production of this material waste is sometimes unavoidable because a plan includes a final profile of non-optimized material or because part orders may vary over time. In addition, the dimensions of the material profiles may vary during a plan depending on the supply, and during the life cycle of an industrial production line there is a heterogeneity of the different products and parts to be produced which involves the production of material waste. Summary of the invention

[0004] The invention makes it possible to overcome the aforementioned drawbacks of the prior art.

[0005] According to a first aspect, the invention relates to a method for classifying predicted material falls in an industrial manufacturing process comprising: • Acquisition of a first planning of a set of parts to be cut, said parts to be cut being intended to be cut within a set of material profiles, each material profile each defining a material element extending along a main dimension; • Acquisition of at least one first optimization plan of parts to be cut for each profile, each part to be cut within a profile comprising geometric descriptors defining at least one given section, a given length, said given length being less than the length of the profile; • Characterization of at least one forecast drop defining at least one zone of the optimization plan not occupied by the parts to be cut, said characterization comprising at least: • an identifier for each expected fall; • dimensions of each forecast fall including at least one fall length, at least one piece of data characterizing the section of the fall; • Querying a first database of articles comprising a set of parts to extract the geometric descriptors of said parts; • First comparison between the dimensions of the at least one forecast drop and the dimensions of at least one part extracted from the first article database, said first comparison producing a correspondence indicator; • Determination of an initial list of compatible parts from the value of the correspondence indicator; • Selection of at least one part from the first list according to at least one given criterion; • Classification of each forecast fall according to a plurality of labels based on a calculation of a score characteristic of said forecast fall, said score being calculated in particular based on at least the first criterion; • Generation of a new first planning comprising the generation of at least one new optimization plan for at least one material profile comprising at least one new part selected from the first list within zones corresponding to a forecast fall having a reuse label; • Updating a falls database in which at least one labeled predicted fall is recorded.

[0006] One advantage is that it allows for the reduction of material losses from unused offcuts.

[0007] According to one embodiment, the method comprises calculating a waste rate for a given material profile category, the waste rate being calculated on a set of optimization plans of one or more schedules, said waste rate being expressed as a percentage of residual material length, said waste rate being used to calculate a desired length of a material profile category.

[0008] An advantage is to be able to quantify the optimization of a material profile optimization plan of an arrangement of parts to be cut in order to either modify the dimensions of the material profiles, or reallocate the available spaces to the production of other parts, or both.

[0009] According to one embodiment, a given profile category comprises a set of material profiles having: • length differences of less than 5% compared to an average length of this set and; • at least one characteristic dimension of the section of each profile of the set is less than a characteristic dimension of the average of a corresponding characteristic dimension of the section of the profiles of the set, the difference between the chosen characteristic dimension of the section of a profile of the set with the average characteristic dimension of the section being less than 10%.

[0010] An advantage is to take into account tolerance margins between a material profile and a part to be produced which can reasonably be produced despite a difference with an initial production specification.

[0011] According to one embodiment, the method comprises a simulation of at least one planning according to a set of profile categories, said simulation comprising an optimization of the determination of a material profile category so as to minimize the waste rate of said planning.

[0012] One advantage is to reduce waste rates by considering new material profiles at the input.

[0013] According to one embodiment, the optimization of a simulation comprises taking into account the number of parts in a schedule, the number of identical parts in a schedule, a heterogeneity factor, a homogeneity factor and / or a variability factor from one part to another, an indicator of statistical distribution of parts characterized by a length or another characteristic.

[0014] One advantage is that it allows different optimization plans to be simulated according to one or more criteria to be optimized. One advantage is that it reduces material losses while taking into account a production criterion.

[0015] According to one embodiment, the step of generating a new first planning comprises the generation of at least one new optimization plan associated with a new material profile comprising a dimension different from the first material profile, said new optimization plan comprising a forecast drop whose surface area is smaller than the forecast drop of the first optimization plan.

[0016] One advantage is that it allows the dimensions of the ordered profiles to be adjusted in order to reduce material losses during cutting operations.

[0017] According to one embodiment, the step of generating a new first planning comprises the generation of a plurality of new optimization plans each associated with a new material profile comprising a different dimension from the first material profile, said new optimization plans together comprising a surface area of forecasted offcuts whose surface area is smaller than the entire surface area of the forecasted offcuts of the first optimization plans. An advantage is to optimize a complete planning.

[0018] According to one embodiment, the method comprises the calculation of a statistical indicator corresponding to a quantification of a type of forecast fall produced on a set of schedules, said statistical indicator making it possible to generate a criterion taken into account when labeling a fall to be produced.

[0019] According to one embodiment, the classification comprises: • an allocation of a first label to the forecast falls for their reuse when the first criterion includes: • an occurrence or subdivision of parts selected in the first list in the first planning plan or another planning plan greater than a predefined threshold and; • a first or second time interval less than a predefined duration; • an allocation of a second label to the forecast falls for their conservation when the first criterion includes: • an occurrence or subdivision of parts selected in the first list in the first planning plan or another planning plan greater than a predefined threshold and • a first or second time interval greater than a predefined duration and / or; • an allocation of a third label to the forecast falls for their conservation when the first criterion corresponds to structural, dimensional or quality properties of the forecast fall corresponding to a predefined criterion and / or; • an allocation of a fourth label to the forecast waste for disposal when the first criterion includes: • an occurrence of selected parts lower than a predefined threshold and; • a first or second time interval greater than a duration predefined.

[0020] An advantage is to allow different uses of a forecasted fall and in particular to allow this fall to be reused by re-entering a new part to be produced in the location of the forecasted fall.

[0021] According to one embodiment, the first planning comprises at least one milestone corresponding to a cutting forecast of at least one material profile forming a step of an industrial production process of a plurality of parts to be produced. An advantage is to optimize a production line comprising several milestones within which different material profile optimization plans are provided.

[0022] According to one embodiment, the first planning comprises at least one material profile identifier, at least one cutting date of each material profile of said first planning and a quantity of parts to be produced. An advantage is to allow the data to be extracted from the optimization plan and not from an article database.

[0023] According to one embodiment, at least one first schedule further comprises at least one product identifier to be produced.

[0024] According to one embodiment, an initial schedule comprises at least one product identifier and an association with a plurality of first schedules, each first schedule being associated with a material profile.

[0025] One advantage is to reason according to the production of a product and not of a set of parts. For example, if a new optimization of a schedule involves an additional delay to produce a product, this can be taken into account in the labeling of the waste.

[0026] According to one embodiment, the optimization plan comprises a set of identifiers of parts to be produced, and for each identifier geometric descriptors of a part to be manufactured which is arranged within a characteristic length representing the length of a material profile represented within the optimization plan.

[0027] According to one embodiment, the characteristic length of the optimization plane is equal to the length of at least one material profile.

[0028] According to one embodiment, the dimensions of each fall include in particular a given length of the fall and characteristic dimensions of a section of the fall.

[0029] According to one embodiment, each forecast fall comprises a set of geometric descriptors including in particular a minimum length and dimensions of a section.

[0030] According to one embodiment, the characteristic dimensions of a section of each fall comprise at least two dimensions characterizing a section of the fall.

[0031] According to one embodiment, the characteristic dimensions of a section of each fall comprise at least one section profile. One advantage is to enrich the database of falls in order to facilitate their reuse.

[0032] According to one embodiment, a set of offcuts is generated from a processing of all the zones of the optimization plan not occupied by the parts to be manufactured, said processing comprising a step of segmenting the lengths of each forecast offcut from the cutting lines of the parts to be manufactured in the optimization plan and the edges of the optimization plan.

[0033] According to one embodiment, the step of segmenting the lengths of each forecasted fall comprises taking into account a set of exclusion criteria making it possible to assign an area of the optimization plan not occupied by parts to be manufactured to a scrap area which is not taken into account for the generation of a forecasted fall.

[0034] According to one embodiment, the step of segmenting the lengths of each forecast fall comprises a first characterization of the fall taking into account the dimensions of the forecast fall, the position of the forecast fall, a surface condition of the forecast fall and / or the material of the forecast fall. An advantage is to collect historical data of the origin of the fall produced. This data allows for better reassignment of the fall or better subsequent exploitation.

[0035] According to one embodiment, each forecast fall comprises a set of complementary descriptors including in particular a material profile identifier. One advantage is to allow the produced fall to inherit the structural characteristics of the material profile which are collected.

[0036] According to one embodiment, the first comparison comprises comparing the length of a predicted drop with the length of a part identified in the article database. An advantage is to quickly discriminate compatible parts from a comparison of a characteristic value.

[0037] According to one embodiment, the comparison step comprises comparing the dimensions of a section of a forecasted drop with the dimensions of a section of a part selected from the article database. An advantage is to quickly discriminate compatible parts from a comparison of a characteristic value.

[0038] According to one embodiment, the first comparison comprises the selection of a set of articles from the article database having at least one characteristic dimension of the section of said part less than a characteristic dimension of the section of the scrap, the difference between the characteristic dimension of the section of each selected part and the characteristic dimension being less than 10% of the value of the characteristic dimension. An advantage is to broaden the scope of exploitation of the predicted scraps.

[0039] According to one embodiment, the first criterion comprises: • taking into account a first occurrence of a part extracted from the article database and associated with a first planning or another planning and / or; • taking into account a batching of identical parts extracted from the article database and associated with at least one material profile from the first planning and / or; • at least a first time interval between two milestones of the same planning in which the part extracted from the article database is associated with each of the two associated milestones, each milestone being associated with a different material profile and / or; • at least one second time interval between two schedules in which the part extracted from the article database is associated with at least one milestone of each of the schedules.

[0040] According to one embodiment, assigning a retention label comprises associating a retention duration with a forecast drop having a retention label.

[0041] According to one embodiment, the generation of the new optimization plan comprises the superposition of the first optimization plan and an intermediate optimization plan, said intermediate optimization plan comprising at least one new part to be produced selected from the first list and positioned and oriented within zones corresponding to a forecast fall having a reuse label. An advantage is that the invention can be applied to numerous nesting software programs without modifying the optimization of the nesting made by these software programs.

[0042] According to one embodiment, the method comprises an update of the article database in which an occurrence of the new selected part is removed. One advantage is to allow better exploitation of the scraps produced.

[0043] According to one embodiment, the first article database comprises a set of parts pre-sectioned to be produced in at least one schedule.

[0044] According to one embodiment, the database of falls comprises the falls having the second label and / or the third label so as to provide access to said falls thus labeled to a user via a communication interface accessible from a data network.

[0045] According to another aspect, the invention relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the latter to implement the steps of the method of the invention.

[0046] According to another aspect, the invention relates to a system for generating a labeling of a forecast fall characterized in that it comprises: • at least one communication interface for acquiring planning data and optimization plans and parts from the first database, • a memory for recording said data and • a calculator for generating a labeling of said forecast falls according to the method of the invention, and in that it further comprises a communication interface for outputting the updated data of the labeled falls to computer equipment.

[0047] According to one embodiment, a system comprising a state sensor of a material profile in order to calculate a state criterion taken into account in the development of the first criterion. Brief description of the figures

[0048] Other characteristics and advantages of the invention will emerge on reading the detailed description which follows, with reference to the appended figures, which illustrate: • [Fig.l]: the different stages of the method of the invention according to one embodiment; • [Fig.2]: an example of representation of a parts database used to feed the production of optimization plans for different milestones of different industrial parts production schedules; • [Fig.3]: a first example of an optimization plan in which a calculation of forecasted waste is produced from spaces not allocated to parts to be produced; • [Fig.4]: an embodiment relating to the first example plan optimization in which a calculation of forecast falls is carried out taking into account the identification of non-reusable areas; • [Fig.5]: an embodiment relating to the first example of optimization plan in which a calculation of forecast falls is carried out taking into account a dimensional criterion of the falls to be produced; • [Fig.6]: an embodiment relating to the first example of optimization plan in which a calculation of forecast waste is carried out taking into account a criterion of dimensions of the parts to be produced subsequently and identified in a database, • [Fig.7]: an embodiment of a system of the invention. Definitions

[0049] In the remainder of the description, we call "an optimization plan" a surface of a longitudinal profile whose dimensions are known and within which a plurality of part profiles according to a main dimension {x} is arranged in an a priori optimized manner. The arrangement generally comprises the positioning of each part within the surface and the definition of a margin between each part profile arranged according to the industrial tool used and possibly additional cutting constraints.

[0050] A "material profile" is a representation of an element extending along a longitudinal main axis having a generally constant section. A material profile is characterized by its length, its section and the material. In certain embodiments, a material profile may be characterized by other characteristics such as a surface condition, a thickness or other parameters.

[0051] The term "surface of a material profile" refers to the total surface area of a two-dimensional representation of the profile. This surface area can be summarized as the length of the material profile when the section is otherwise characterized.

[0052] In the remainder of the description, the term “an anticipated fall” or “a fall forecast” is used to designate the estimated fall of material during a production step of industrial parts within a material profile.

[0053] In the remainder of the description, an “effective scrap” is a scrap actually produced by a cutting device and which can be stored or used to manufacture another part in a subsequent milestone or subsequent planning.

[0054] A part is defined as a part identified in a database or file and to which access is available to possibly extract its identifier and its dimensional characteristics, in particular to carry out length, section or surface comparison operations.

[0055] A part to be cut is a part selected or identified within a milestone of a schedule or a part which is known to be manufactured and which is in a list of parts ready to be selected within a schedule. and which is not necessarily yet integrated into an optimization plan. A part to be cut can also refer to a part already selected and integrated into an optimization plan and not yet produced. When the term "part to be cut" is used in the context of planning or an optimization plan, it refers to the 2D representation of said part within an optimization plan.

[0056] In the remainder of the description, we refer to “a schedule” as an industrial schedule comprising a set of milestones associated with forecast dates, each milestone comprising the association of a material profile and an optimization plan for the production of a set of parts to be cut. A milestone can therefore be defined by a date and a quantity of parts to be produced present in an optimization plan and possibly in an available space defined by a forecast waste.

[0057] According to an exemplary embodiment, a milestone is associated with an optimization plan and a material profile. According to another exemplary embodiment, a milestone is defined solely by association with at least one part, a production quantity of said part and a manufacturing date.

[0058] In the remainder of the description, the term "final product" refers to the product produced from a set of cut parts. The final product generally includes parts other than those cut from a material profile, however, generally a part cut from a material profile is manufactured to be included in a manufacturing process of a final product. There is therefore generally an association between the part present in an optimization plan and a final product to be produced.

[0059] The term "section" refers to the section of a profile in a cutting plane intersecting with the longitudinal axis of the largest dimension of the profile. Generally, this cutting plane is perpendicular to the main longitudinal axis of the profile, however an inclined plane can also be considered.

[0060] The term “profile” or “material profile” refers to a longitudinal piece intended to be cut to produce smaller pieces, at least one piece.

[0061] A "technical waste" is a waste induced by the cutting process. Reception of the planning

[0062] [Fig. 1] represents an embodiment of the method in which a set of steps are carried out.

[0063] A first ACQi step comprises the acquisition of a first PLi schedule. The first PLi schedule is for example defined from a software component making it possible to plan milestones of a production line of industrial parts. According to one example, the first PLi schedule comprises the choice of a profile of material or a plurality of material profiles and their dimensions. The first planning PLi may further comprise an association with a final product to be produced, such as a frame, a supporting structure, a structure forming the framework of a vehicle or other final device or system to be designed in particular by means of an assembly of parts cut from the method of the invention. According to one embodiment, the first planning PLi is coordinated with another planning comprising parts other than profiles and which cooperate with said profiles to form a final product.

[0064] The method of the invention may comprise a step aimed at acquiring or receiving a plurality of schedules. The first schedule PLi in the context of this description is called the schedule used to illustrate an embodiment. The other schedules PL; are also designated the schedules likely to be taken into account in the method of the invention when a reuse of the scraps is envisaged over a broader framework than the current schedule.

[0065] The method of the invention comprises receiving or acquiring an optimization plan denoted POA. This optimization plan POA corresponds to an arrangement of parts to be cut, i.e. a representation of part profiles arranged within a material profile defining a surface extending along a main length.

[0066] A distinction is therefore made between a profile of a part to be cut and a profile of material insofar as the profile of the part to be cut is inscribed in the surface of the profile of material.

[0067] The surface of the optimization plan is preferably identical to that of a material profile that will be used for cutting. However, the optimization plan can be defined in a smaller surface or at least included in the surface of the material profile that will be used to cut the parts arranged in the optimization plan. This ensures a certain cutting security during planning. This type of security is found in stationery publishing when it is desired to anticipate backup margins.

[0068] According to one example, a plurality of POA optimization plan receptions are carried out according to the method of the invention, each optimization plan being associated for example with a milestone of the first PLi planning.

[0069] According to an exemplary embodiment, the steps of receiving the first planning PLi and a set of optimization plans POA can be implemented together.

[0070] [Fig.2] represents an example of planning PLi comprising a plurality of milestones, each milestone being associated with a date tb t2, t3, t4. Each of the dates tb t2, t3, t4 is associated with an operation of cutting a material profile PROk aimed at producing a set of parts. Each milestone is associated with a POA optimization plan to guide the machine tool used to cut the parts on the PROk material profile.

[0071] [Fig.2] also represents a time scale representing the different milestones of the first PLi planning. According to one embodiment, the part profiles of each optimization plan are extracted from a BDi database of parts. According to another example, a digital file can contain all the data describing the parts to be produced. According to another embodiment, the geometry of the parts is directly deduced from a reading of the POA optimization plan, for example by means of a software component allowing the extraction of a contour of a part drawn within a POA optimization plan.

[0072] The method of the invention comprises a step of characterizing the free zones of the POA optimization plan. This step is denoted CARi in [Fig.l]. The free zones correspond to the zones not occupied by part profiles defining an occupation surface of the material profile. The free zones correspond in a first approximation to all the forecasted offcuts. These forecasted offcuts may comprise several distinct zones or a single zone comprising portions of continuous surfaces. It is recalled that the surfaces in question correspond to “linear surfaces”, that is to say a surface whose width is considered as a small dimension with respect to the length of the material profile considered.

[0073] Figures 3, 4, 5, 6 represent different embodiments of exploitation of free zone(s) of a material profile PROk in the form of an optimization plan POA for dimensioning forecast scraps CHk in order to envisage exploitations of these forecast scraps by different choices of arrangement and selection of parts PN according to given selection criteria.

[0074] These figures make it possible to represent cases: • in which the only sections of the part profiles Pi arranged on the optimization plane POA are planned according to a single direction transverse to the main axis of the material profile PROk, such as a perpendicular axis, this is for example the case of figures 3 and 5 and; • in which the cuts are planned both along a direction transverse to the main axis of the material profile PROk and along at least one other axis such as an axis parallel to the main axis of the material profile PROk as shown in figures 4 and 6.

[0075] [Fig. 3] represents an exemplary embodiment representing a POA optimization plan comprising a plurality of profiles of parts to be cut P; arranged within a given surface or within a given length. In this example, the offcuts CHk forecasts correspond to all the surfaces not occupied by the profiles of parts to be cut P;.

[0076] [Fig.4] represents an exemplary embodiment representing the same optimization plan POA as that of [Fig.3] and therefore the same representations of parts to be cut P; represented at the same locations as in the optimization plan POA of [Fig.3], with the exception that two additional longitudinal cuts are made in this example on either side of each part Pi to be cut so as to remove the edges.

[0077] A CHk' scrap is shown smaller than that shown in [Fig.3]. In the case of [Fig.4], the generated CHk' scrap comprises a surface trimmed with respect to the surface of the CHk scrap. Margins correspond to the edges of the CHk scrap or the material profile have been removed. According to one embodiment, the margins may be provided transversely to the longitudinal axis. According to another embodiment, other margins may be provided such as longitudinal margins as shown in [Fig.4] and [Fig.6].

[0078] Zones ZR are represented as being non-exploitable or non-usable zones. They can correspond to margins between the edge of the material profile and representations of parts to be cut P; or margins between representations of parts to be cut P; or forecast margins CHk, CHk' between parts to be cut P; and forecast waste CHk, CHk' or even a margin between a forecast waste CHk, CHk' and an edge of the material profile.

[0079] An advantage of the invention is to treat the portions of each profile of unexploited material forming potential waste CHk, CHk' of material which can be exploited or used in different ways.

[0080] Advantageously, the configuration of the margins provided around the forecasted offcuts CHk, CHk' can be inherited from the margins provided around the parts to be cut P;. According to one example, if the profiles of the parts to be cut P; do not have a longitudinal margin around them represented on the optimization plan POA, then the longitudinal margins around the forecasted offcuts CHk, CHk' are not generated. This is the case in [Fig.5]. A first interest is to apply the same cutting conditions between the parts to be cut P; and the forecasted offcuts CHk, CHk'. Indeed, these latter conditions can be imposed by the cutting machine. A second interest is to provision specific dimensions of a forecasted offcut CHk, CHk' for example to reinsert a part PN or to revalue the forecasted offcut with optimized dimensions.

[0081] In the opposite case, for example that of [Fig.6], if longitudinal margins are present on either side of the pieces to be cut P; on the POA optimization plan, the The method of the invention makes it possible to generate an identical configuration for the registration of forecasted offcuts CHk, CHk'. One advantage is to inherit a criterion already chosen for the cutting of parts to be cut P; for the cutting of future parts PN falling within the forecasted offcuts CHk, CHk'.

[0082] An objective of the invention is to classify and label each forecast material waste so as to reuse these forecast wastes or to keep possible wastes for other uses or for uses in the same use, even at a date later than that of a planning milestone. However, there generally remains a waste which is difficult to envisage exploitation. One of the possible labels for forecast wastes is therefore “reject” or “unusable waste”.

[0083] In the example of [Fig.4], the forecast waste CHk' corresponds to the entire surface area not occupied by the representation of the parts to be cut Pi from which have been subtracted non-exploitable zones ZR defined between the delimitations of the flanks FLi or between a delimitation of a flank FLi and an edge of the optimization plan POA.

[0084] In the latter case, it is considered that the ZR zones do not have dimensions large enough to allow these forecast waste to be reassigned to a use other than their disposal or that it is necessary to consider a margin to ensure exploitation of a useful zone.

[0085] A forecast fall CHk, CHk' may be a zone arranged "next to" an edge of a representation of a part to be cut P; or an edge of the optimization plan POA, however, a forecast fall CHk, CHk' may also be included in the general shape of a representation of a part to be cut P; when this part to be cut has an unretained interior zone. This may for example be the interior of a rectangular shape extending longitudinally having a sufficient width to allow exploitation of the interior zone.

[0086] Generally speaking, we note a forecast fall CHk, the notation CHk' only allowing to illustrate the different ways of considering a fall in the processing of an optimization plan in the example of [Fig.4].

[0087] In this example, it is understood that rules can be predefined to select a forecast waste CHk that can be used or not. For example, a spacing between two distinct representations of parts to be cut P; succeeding each other that is less than a threshold distance can be defined in order to label a waste waste or not to consider it for subsequent processing. This threshold distance can be defined according to the material and the thickness of the profile or according to of the cutting tool. Other criteria can be taken into account such as the surface condition of the material profile or the type of part to be cut.

[0088] Thus in this step, the zones ZR are either removed from the usable forecast waste CHk and therefore are not treated as material waste, or labeled as forecast waste which will be discarded. The method of the invention comprises the two embodiments, consisting of labeling or not treating these wastes.

[0089] According to one embodiment, the dimensions of the margins ZR between two forecast offcuts arranged side by side or the dimensions of the margins between a representation of a part to be cut Pi adjacent to a forecast offcut CHk may vary according to certain parameters, including for example: the profile of the part or the thickness of the material profile. Indeed, a curved profile of a part to be cut P; may require movements of a cutting tool covering greater amplitudes than when cutting a straight profile. Characterization of predicted falls

[0090] According to one embodiment, a characterization CARi of a forecast fall CHk comprises a forecast fall identifier IDc which is generated when determining a forecast fall within the optimization plan.

[0091] According to one example, as soon as an area not occupied by a part to be cut within the material profile PROk is sufficiently large, it is then characterized as a predicted fall. According to one embodiment, the “magnitude” threshold corresponds to a minimum of a dimension in the direction in which the material profile PROk extends.

[0092] According to another embodiment, all the identified forecast offcuts are associated with an identifier independently of their dimension(s) including when they define a margin between two representations of parts to be cut, such as the ZR zones.

[0093] Another characterization is a descriptor of the geometry Kc of the predicted fall CHk. This descriptor of the geometry Kc, also called geometric descriptor, can comprise one or more of these characteristics: - The position of each point of the contour of the predicted fall in the plane of the surface of the optimization plan; - The maximum length of the predicted fall CHk along the main axis of the material profile PROk; - The maximum width of the predicted fall CHk along an axis transverse to the main axis of the material profile PROk; - The minimum width of the predicted fall CHk along an axis transverse to the main axis of the material profile PROk; - At least one characteristic data of the section of the profile, such as a section width, a section height, a specific shape in circle, cone, square, rectangle, “L”, “T”, “H”, etc.; - At least one type of profile such as solid or hollow type; - At least one cut defined by the cutting angle with respect to the axis main section of the profile, generally the cutting angle is perpendicular to the longitudinal axis, however, some cuts may have a different cutting angle; - Two sections given when we want to characterize the two ends of a predicted fall. In one case, one of the ends of the section is the section of the material profile when it is arranged at the end of the profile; - The diagonal of a rectangle inscribed in the free zone of the optimization plan, that is to say an area not containing a representation of a part to be cut Pi, as well as the position of this diagonal. This descriptor can be interesting for profiles of a certain width; - The largest convex polygon inscribed in a free zone of the optimization plane, i.e. an area not containing any representation of parts to be cut, as well as the position of each edge of the polygon in the plane; - Etc.

[0094] The method of the invention comprises a step REQi aimed at querying a data resource which may be a database BDi or any other data file recorded in a memory in order to compare at least one descriptor of a CHk fall with dimensions of a PN part and / or a descriptor of said PN part. The PN parts are preferably parts described in a nomenclature in which all the dimensional and structural characteristics of the parts are recorded. In the remainder of the description, a PN part is a part referenced in a data system such as a database in which said parts have a nomenclature. The characteristic data of the PN parts are accessible and possibly extracted from a complete database of parts, or possibly from a list of parts extracted from a complete database according to one or more selection criteria.

[0095] Other descriptors may be taken into account according to the different embodiments, in particular the thickness of the material profile and the composition of the material. In another example, the finish of the material profile, a possible surface treatment or even its condition can be taken into account. By inheriting characteristics from the material profile to the elements obtained after cutting, these characteristics can be associated with those of the predicted CHk waste.

[0096] [Fig.2] represents such a database which can be queried to supply different planning with the integration of new parts in different POA optimization plans following one another at times tb t2, t3, t4, each being associated with a PROk material profile.

[0097] The invention makes it possible to benefit from the characterization Kc of the forecast fall CHK in order to verify to what extent a part PN of the base BDi can be selected to be produced in the free zone corresponding to the zone of the forecast fall CHk of the material profile in a milestone of a given planning PLi or even of another planning PL;.

[0098] According to one embodiment, the position along the main axis of the material profile PROk of the predicted fall CHk is a characteristic of the predicted fall. It is of interest to know the part of the material profile from which the predicted or actual fall comes. It can be arranged at the edge of the profile or inside the profile, the inside designating the central zone along the main axis of the material profile PROk. This case can occur when two ends of a material profile PROk are used to cut parts and a central part is not used. This central part can form a predicted fall CHk. Comparison of extracted parts with scraps

[0099] According to an example, data characterizing the thickness of the part PN, the type of material of the part PN or any other parameter characterizing the part PN can be taken into account in order to select a subset of parts PN that can be cut on the material profile PROk. This or these characteristic data of a part can be compared to one or more homogeneous data of a material profile or of a predicted offcut having characteristics inherited from a material profile PROk from which it originates.

[0100] According to one embodiment, at least one section characteristic is taken into account to filter the parts Pn from the database BD1. The section characteristic may be, for example, one or more of the following characteristics: - A section width, a section height, a specific shape in circle, cone, square, rectangle, “L”, “T”, “H”, of said section, etc.; - At least one type of profile such as solid or hollow type;

[0101] According to one embodiment, a tolerance or a margin between the characteristic data of the part PN and that of the material profile PROk can be defined according to a predefined criterion. For example, a thickness of a part PN not strictly corresponding to the thickness of the material profile PROk while remaining less than a given difference can be selected. For example, if the material profile has a thickness of 20mm and a part has a thickness of 19.5mm, according to a given tolerance criterion, the part PN can still be considered for selection in order to be produced within a forecasted waste CHk. The tolerance can be set to accept only a negative margin, a positive margin or both between the characteristics compared between those of the forecasted waste CHk and that of the part PN.

[0102] According to one embodiment, when first geometric constraints between the predicted fall CHk and a part PN of the database correspond, second geometric constraints can be compared by integrating a deviation tolerance between said second geometric constraints.

[0103] For example, a first constraint may correspond to a minimum length and a second constraint may correspond to a characteristic dimension of the section of said part. According to one example, a tolerance may be accepted so as to select parts whose diameter of the section is less than a characteristic dimension of the section of the predicted fall CHk. This applies to another characteristic dimension of a section.

[0104] According to one example, the difference between the characteristic dimension of the section of a selected PN part and the characteristic dimension of the section of the profile is less than 10% of the value of the characteristic dimension.

[0105] This tolerance can also be applied to a type of material of a PN part different from that of the PROk material profile but whose desired characteristics are compatible with those of the material profile. Typically, a part supposed to be made of steel can also be made of titanium if only the mechanical strength characteristics must be respected.

[0106] According to one embodiment, a tolerance is expressed by a percentage deviation of a value characterizing a physical or dimensional property with respect to a reference value of said characterization.

[0107] The tolerance may apply to a strength value, a density value, an elasticity value such as Young's modulus value, a torsional strength value, a tensile strength value, a flexural strength value, a compressive strength value, or a thermal property value.

[0108] In these latter cases, the calculation of the tolerance is obtained by comparing a characteristic value of the material profile with a characteristic value of a part from the BDb database. If the values are within a certain given range of values, then the tolerance is acceptable for retaining the part.

[0109] According to one embodiment, the association between one or more PN parts and one or more forecasted scraps CHk is carried out from at least one table of correspondences or equivalence of characteristics defining the PN part(s) and respectively the forecasted scrap(s) CHk.

[0110] A comparison step COMPi between the surface of one or more forecasted falls CHk and one or more parts PN makes it possible to select a subset of forecasted falls CHk corresponding to a subset of parts Pn-

[0111] The COMPi comparison step can be implemented according to different embodiments. A maximum length comparison algorithm can be used to automate and optimize this calculation step on a large number of PN parts. In the case of a predicted fall coming from a material profile, the length of the fall is a parameter of primary interest for carrying out the comparison with the lengths of existing parts.

[0112] Other criteria may be taken into account, including the shape of the section, the thickness or the width of the predicted fall CHk. The material or the surface treatment are also criteria that can be compared.

[0113] According to one embodiment, a first simple comparison step makes it possible to discard or select a large number of PN parts so as to pre-filter a subset of PN parts and a second comparison step is carried out to refine the comparison among a pre-selection carried out during the first comparison step.

[0114] According to an exemplary embodiment, the first comparison step comprises comparing the largest dimension of a part PN with the largest dimension of the predicted fall CHk. When the largest dimension of a part is greater than the largest dimension of the fall, this part is not selected.

[0115] According to an exemplary embodiment, the first comparison corresponds to a filtering of a part type such as a “profile” type part. According to an example, this first comparison makes it possible to filter both the part type, such as a “profile” filter, and the material of the PN part.

[0116] According to an example, another dimension is tested between the dimensions of the part PN and the dimension of the predicted fall CHk, for example a dimension perpendicular to the largest dimension of the PN part. An interest is to select PN parts quickly having at least two dimensions smaller respectively than the same two dimensions of the predicted fall CHk.

[0117] According to an exemplary embodiment, the comparison of certain dimensions of a forecasted scrap CHk with those of a part PN takes into account an orientation constraint to be respected. The orientation constraint can be defined by data specific to the material profile PROk. The orientation constraint can be for example imposed by a rolling direction, a fiber direction of a composite material. When an orientation constraint is imposed by the part PN or the material profile PROk, the dimensions compared between the part PN and the forecasted scrap CHk are chosen to respect this orientation constraint.

[0118] According to this embodiment comprising at least two steps, after this first comparison, a second comparison implementing a more complex algorithm makes it possible to process the selected PN parts.

[0119] However, according to one embodiment, the method of the invention comprises only one comparison step. The latter can be simple according to a technique previously described or complex, for example by implementing an algorithm for comparing dimensions and other characteristics.

[0120] When the material profile has a width greater than a predefined threshold, according to one example, the second comparison comprises the comparison of the surface area of each predicted fall CHk on the one hand and of each part PN on the other hand.

[0121] According to one example, the database comprising the identification of each PN part comprises a geometric descriptor corresponding to the largest dimension of the PN part. When the PN part is of the “profile” type, the largest dimension is the length of the part. List of selected parts

[0122] The comparison operation results in defining a first list LISTi of compatible PN parts of one or a plurality of forecast drops CHk of a POA optimization plan. This step of determining a list of compatible PN parts is noted DETi in [Fig.l].

[0123] A list LISTi can be generated for each milestone of a planning PLi, i.e. for each material profile PROk. According to one embodiment, the list of parts PN is generated for a set of milestones of a planning PLb, i.e. a set of material profiles PROk. One advantage is to obtain better optimization of the reuse of the forecast scraps CHk by promoting a diversity of scenarios for integrating or nesting parts PN within a plurality of forecast scraps CHk.

[0124] In its broadest interpretation, the list LISTi includes all compatible PN parts of all forecast CHk drops identified in a milestone or planning.

[0125] The list can be used in different ways depending on the embodiments of the invention.

[0126] According to a first embodiment, a first filtering on the PN parts to be produced in one or more PL schedules is carried out then a second filtering on the geometry of the profiles of the parts is carried out while taking into account criteria characterizing the parts such as their material, thickness, an orientation constraint, etc.

[0127] According to a second embodiment, the order of the filtering is reversed. First, a filtering is carried out on the geometry of a set of PN parts and, in a second step, a filtering on the PN parts to be produced in one or more PL plans; is carried out from the parts having a filtered geometry. Other criteria characterizing the parts, such as their material, thickness, an orientation constraint, etc. can be taken into account in one or other of the filtering operations to select the parts to be kept to be re-nested in a PO optimization plan;.

[0128] One advantage is to prioritize the manufacturing of parts planned in later schedules by re-nesting them in an optimization plan of a current or imminent schedule. One advantage is to gain in manufacturing speed of the different schedules. Selection criteria

[0129] The criterion for selecting a PN part from the base to be nested in a POA optimization plan at the location of at least one forecasted fall CHk is noted Ci. This criterion Ci can comprise one or more sub-criteria, also called criteria. The criteria Ci can comprise dimensional criteria inherited from the characteristics of the cutting of the forecasted fall CHk, structural criteria inherited from the material profile PROk or even temporal criteria inherited from the characteristics of the existing or identified PLb PL schedules. According to another case, the criterion Ci responds to an optimization indicator of the PN parts to be produced, that is to say the factor which best optimizes the arrangement of PN parts within forecasted falls CHk. According to another case, a criterion Ci can correspond to the taking into account of a batching of PN parts which must be manufactured together or in the same milestone or the same schedule.

[0130] In the remainder of the description, the first criterion Ci is a criterion which is taken into account as a priority, i.e. before other criteria. However, taking into account a set of criteria is sometimes necessary in order to produce the PN part, for example a criterion relating to the type of material or a dimension is often necessary to produce the PN part. It is therefore possible to dissociate the necessary criteria which correspond to the minimum or maximum value ranges chosen so that at least one PN part can be inscribed in a sufficient space, from the criteria for selecting a PN part among other parts.

[0131] According to a first example, the production planning of a part PN is taken into account as criterion Ci to select the said part(s).

[0132] A first exploitation aims to verify whether one of the PN parts of the list LISTi is part of the first planning PLi. When this is the case, the number of occurrences of the PN part planned to be produced and the milestones in which this production is carried out are then verified.

[0133] According to one embodiment, when the occurrence is less than a threshold value, for example two parts, the part is not selected to be reintegrated into a predicted drop location CHk. An advantage is to prioritize the manufacturing of parts whose occurrence is significant. An advantage of this embodiment is that the storage of these parts is a priori already well known and there is less risk of not losing the part, for example in the event of cancellation of an order.

[0134] According to another embodiment, when the occurrence is greater than a threshold value, for example four parts, the part PN is not selected to be reintegrated into at least one predicted drop location CHk. An advantage of this embodiment is to manufacture “rare” parts having few occurrences of which there is a priori little stock.

[0135] According to one embodiment, when the PN part is part of a batch of PN parts, that is to say that they are only manufactured in batches of a set of the same parts, then it is then verified that a certain length of forecasted fall is available or that the length of a forecasted fall makes it possible to integrate a plurality of PN parts, depending on the number of PN parts in the batch.

[0136] In this case, an additional selection criterion is the availability of sufficient space among the forecasted scraps CHk to produce a subset of parts PN to be, for example, produced on the same day, in a given order, in the same schedule, or in the same milestone. For this purpose, the set of forecasted scraps CHk of a milestone, a schedule or a set of schedules is considered for an optimization of a nesting of a subset of parts Pn-

[0137] According to one embodiment, the method of the invention makes it possible to take into account planning data associated with the PN part. This planning data allows to automatically check if the selected PN part from the first LISTi list is included in another PL planning; than the first PLi planning.

[0138] [Fig.2] represents this verification step in which the PN part extracted from the database BDi is included in another PL planning;. This other PL planning; may correspond to a planning of another final product or a manufacturing of a variant of the final product. When this planning data is associated with a PN part in the database BDi, it can be used in such a way as to predict a reintegration of at least one occurrence of this PN part in an available location of a forecast fall CHk.

[0139] One interest is to select each PN part according to a given criterion Ci in order to make the best use of each forecast fall CHk. This selection step is noted SELp

[0140] According to an exemplary embodiment, the correspondence indicator INDi makes it possible to associate a part PN with a forecasted fall CHk. In this example, the method of the invention makes it possible to search for which parts PN can be selected to be re-registered or re-nested in a given fall CHk. In this embodiment, the method of the invention makes it possible to determine which are the set of parts that best fill the forecasted fall CHk. An algorithm for optimizing the space defined by the forecasted fall is then implemented.

[0141] According to a first embodiment, each part PN of the list LISTi is associated with one or a plurality of forecast falls CHk when a correspondence indicator INDi has been generated.

[0142] We can note LISTik the list of compatible PN parts of a given forecast fall CHk identified in a milestone or a planning.

[0143] According to this embodiment, each forecast fall CHk identified in a milestone of a schedule or in a schedule is associated with a plurality of parts PN of the first list LISTi when a correspondence indicator has been generated.

[0144] It is understood that the invention also makes it possible to draw up a list LIST2 N of forecasted waste CHk for each part PN to be produced identified according to the correspondence indicator INDi

[0145] We can note LIST2 the list comprising all the LIST2N lists of forecast CHk scraps for each PN part to be produced identified according to the correspondence indicator INDk

[0146] The reintegration of a PN part into a given POA optimization plan can be carried out either in a first case when a forecast fall CHk corresponds to all the criteria, or in a second case following an algorithm optimization after processing all the forecast CHk drops of a milestone or a schedule to process cases where several solutions for reassigning a PN part are possible.

[0147] In the first case, the method of the invention comprises the selection SELi of a part PN to be produced associated with an exploitable forecast fall CHk to be reassigned within a milestone of a planning PLi, PL; in order to complete or regenerate an optimization plan POA.

[0148] According to the second case, the correspondence indicator INDi makes it possible to associate at least one forecasted fall CHk with a part PN. In this example, we seek which forecasted falls CHk correspond best for a given part PN.

[0149] We can consider a list of forecasted falls LIST2N or LIST2 N which is identified within at least one milestone of at least one planning or within a planning PLb For a given PN part, a list LIST2 N can then be filtered according to criteria making it possible to select the forecasted fall CHk to be associated with a given PN part.

[0150] Thus, this indicator makes it possible to establish reciprocal correspondences between PN parts from the list LISTi and the forecasted CHk scraps from the list LIST2. These correspondences make it possible, in a second step, to optimize a reuse of the PN parts within a PLi schedule or to optimize the forecasted CHk scraps identified for their reuse.

[0151] An advantage is to draw up a set of possibilities for reintegrating PN parts within a plurality of forecasted CHk scraps before implementing an algorithm aimed at optimizing this reintegration according to optimization criteria. Indeed, when several solutions exist for associating several forecasted CHk scraps with a PN part to be produced or for associating several PN parts to be produced, it is necessary to choose which forecasted CHk scrap is associated with a given PN part to be produced or conversely to choose which PN part to be produced is associated with a forecasted CHk scrap. Consideration of additional criteria

[0152] The optimization criteria may include the occurrence of a PN part, the dimensions of a PN part, etc. Indeed, the method of the invention makes it possible, for example, to prioritize the reintegration of large-sized PN parts to be produced. According to another example, a predicted drop location CHk making it possible to integrate a batch of identical PN parts may be prioritized with respect to another part to be integrated or a plurality of different parts.

[0153] The selection step SELi also makes it possible to take into account an additional criterion Ci associated with the part PN.

[0154] The criterion Ci may comprise different types of characterization of the part PN from the list LISTi or LISTiN or of a characterization of the product of which the part PN is part or of a planning PLi, PL;, in which the part PN is planned. For example, a criterion Ci comprises taking into account a first occurrence of a part extracted from the article database BDi within the first planning PLi or another planning PL;. In this case it is a characterization of the product or of the planning which is taken into account.

[0155] According to another example, the characterization may be linked to an intrinsic characterization of the part, such as its surface condition or an orientation or thickness constraint.

[0156] According to another example which can be combined with the previous one, a criterion Ci comprises taking into account a batching of identical PN parts from the article database BDi of parts to be produced.

[0157] According to another example which can be combined with the previous one, a criterion Ci comprises taking into account at least a first time interval between two milestones of the same planning PLi in which the part PN extracted from the article database BDi is associated. One advantage is to take into account a notion of temporality. Indeed, the method of the invention makes it possible to take into account a time horizon criterion beyond which one no longer seeks to reintegrate a part PN. Another advantage is to prioritize the reintegration according to the date on which the manufacturing need is expected.

[0158] According to another example which can be combined with the previous ones, a criterion Ci comprises taking into account at least one second time interval between two schedules PLi, PL; in which the part PN extracted from the article database BDi is associated with at least one milestone of each of the schedules.

[0159] According to another example which can be combined with the previous ones, a criterion of heterogeneity of the scraps produced is a criterion to be optimized. The objective of this optimization is to identify the redundant nature of the scraps produced. According to one embodiment, it is sought to minimize the heterogeneity of the scraps produced in such a way as to favor identical scraps produced. The scraps produced can come from forecast scraps which are not associated with PN parts to be re-nested or from new scraps resulting from a nesting of a PN part producing a new scrap.

[0160] According to another example which can be combined with the previous ones, the selection criterion favors the optimization of the dimensions of a set of forecast drops of a set of optimization plans compared to the dimensions of a particular forecast drop or of forecast drops of an optimization plan. In this, an entire planning comprising several milestones is rescheduled with the generation of new POA optimization plans. Classification

[0161] At the end of this selection step SELi, a classification of the forecast fall CHK is carried out. This step is noted CLASSi in [Fig.l]. The classification makes it possible to label the forecast falls according to different labels.

[0162] The classification can be performed by a machine learning algorithm such as a convolutional neural network. Any other learning function that can calculate a class membership prediction score from values forming attributes of the predicted fall CHk can be used. A cost function can then be implemented to train the learning function, in other words the machine learning model. Such a learning function can be learned with a data set whose class is known.

[0163] Other algorithms can be used such as any expert system based on a knowledge base, rules and an inference engine for example.

[0164] The score can also be calculated from rules making it possible to compare values of a set of measured, evaluated, calculated or predicted criteria with predefined threshold values or values from characteristics extracted from PN parts in the BDi article database.

[0165] According to an example, if a material criterion of the material profile within which a predicted fall CHk is identified is fulfilled, for example, the material is aluminum, a first value Vi = 1. If at least one part PN of an article base has dimensions compatible with a dimensional fall then a second value V2 = 1. Then if Vj = V2 = 1, the predicted fall CHk can have a first label LBi.

[0166] According to the same example, if the material is titanium, the first value Vi = 2. If at least one PN part of an article base has dimensions compatible with the dimensional drop CHk then the second value is always V2 = 1. Then, the predicted drop CHk can have a first label LB3.

[0167] Other rule possibilities can allow criteria values to be combined in order to associate labels with the predicted CHk falls.

[0168] A first label LBi corresponds to the forecasted falls CHk which can be reused so as to nest a part PN of the database BDi in the expected location of the forecasted fall CHk. Different re-nesting possibilities can be implemented in the method of the invention.

[0169] When a PN part is selected to be integrated into an optimization plan, different scenarios are possible. Figures 5 and 6 represent two scenarios different in which identical or different PN parts are integrated into a CHk forecast drop. [Fig.5] represents a scenario in which the larger PN parts have been reintegrated into the CHk forecast drops. The CHk' forecast drop in [Fig.4] has been divided into two portions defining two CHk drops that can be used by reintegrating two identical PN parts.

[0170] [Fig.6] represents an embodiment in which the predicted fall CHk' of [Fig.4] is used to reintegrate parts of smaller dimensions because they are included in a lot of parts.

[0171] According to one embodiment, the invention makes it possible to treat the forecast fall CH k' of [Fig.4] as a single forecast fall to which one or more PN parts are assigned. According to another embodiment, the invention makes it possible to treat the area occupied by the forecast fall CHk' of [Fig.4] as several forecast falls which are associated with several PN parts.

[0172] In both cases, the label of the forecast fall or subdivided forecast falls is noted LBi. The LBi label corresponds to a reuse of the forecast fall(s) in a current plan or a future plan.

[0173] This first label LBi can be addressed when at least one occurrence of a part PN is present in the first list LISTi, LISTik. This condition can be combined with a duration criterion. The forecast fall CHK is for example labeled with the label LBi if the selected part PN is planned in manufacturing in a limited time horizon, that is to say within a planning PLi or PL;. According to an example, a set of plannings PL; over a given duration.

[0174] A second label LB2 corresponds to forecasted scraps CHk which are not used to reintegrate a PN part to be produced during the production of a next PL planning; or a next milestone of a PL planning,. PL;. This second label LB2 corresponds to a label for preserving the forecasted scrap CHk. This case can occur when no PN part is selected which is compatible with the forecasted scrap CHk or when at least a first criterion Ci does not meet certain conditions even though other criteria of the PN part correspond to those of the forecasted scrap CHk, such as the dimension criteria of the location of the forecasted scrap CHk. This can be the case if no occurrence of the part is present in a milestone of a future PL planning.This can also occur when a part is only produced within a lot and there is not enough projected waste area to reintegrate all the parts in the lot.

[0175] A third label LB3 corresponds to forecast CHk falls that we wish to keep for their structural or dimensional property(ies) regardless of the availability of compatible length, dimension or surface of a PN part to be produced. In this case, a PN part from the first list LISTi LISTik> and whose criterion Ci is compatible with a forecast waste CHk is not sufficient to label the forecast waste as a source of material to be reused within the material profile PROk. In this case, the method of the invention makes it possible to value the forecast waste CHk according to a criterion linked to the quality of the material, its surface condition, the material used for example.

[0176] In the case of labeling by a third label LB3, of an identified forecasted fall CHk, the conservation decision can be made according to a criterion independent of the step of querying the first article database BDi making it possible to extract a geometric descriptor of a part PN. Indeed, according to an exemplary embodiment, only the dimensions of the forecasted fall CHk can be considered to label said fall according to the third label LB3. The invention therefore relates to a method of the invention in which the step of querying the article database BDi is an optional step.

[0177] A fourth label LB4 corresponds to forecasted scrap CHk for their scrapping. This label LB4 is for example assigned when the first criterion Ci includes an occurrence of selected parts less than a predefined threshold. In a borderline case, the threshold value is zero. It is understood that if there is no PN part to be produced that can be produced within a forecasted scrap CHk, there is no point in using this forecasted scrap CHk in an optimization plan. Finally, if this forecasted scrap CHk does not present any conservation interest for another use, its scrapping may be the most appropriate solution. A scrap with a length less than a predefined threshold can also be labeled with the fourth label LB4. The forecasted scrap CHk can therefore be scrapped.

[0178] According to one embodiment, other labels can be used to classify the forecasted waste CHk. According to one embodiment, the first label LBi comprises sub-labels making it possible to prioritize the use of the forecasted waste CHk in a future PL schedule. According to one example, a first sub-label LBn is assigned when the forecasted waste CHk must be reused for a PN part to be produced quickly, i.e. in a short time horizon, and therefore within a time limit below a given threshold. In the latter case, the association of a PN part with a schedule can be analyzed when such an association is created.

[0179] According to another example, a second sub-label LB^ is used to annotate a forecasted waste CHk according to a preference criterion for cutting one or more pieces to be reintegrated. This may correspond to the case of [Fig.5] and [Fig.6]. In both cases, the forecasted waste CHk is labeled with a label LBi. However, in [Fig.5], the reassignment of the forecast waste CHk is governed by the reintegration of large PN parts, which constitutes a priority criterion here. In this case, the forecast waste CHk can have a first label LBi2. In the case of [Fig.6], the subdivision criterion is more important, which constitutes another priority criterion here, and the forecast waste CHk can have another sub-label LBn. However, this last sub-labeling is not critical since the forecast waste CHk is reused in both cases.

[0180] According to one embodiment, the classification step comprises taking into account a storage indicator of the scraps already produced, for example having a label LB2, LB3. To this end, querying a scrap database BD2 makes it possible to obtain this indicator. One advantage is to assign a label to a scrap taking into account a quantity of scraps already present.

[0181] According to one embodiment, the assignment of a conservation label LB2, LB3 comprises the association of a conservation duration with a forecast or actual drop, the data characterizing said drop and duration being recorded in the database BD2. An advantage is to trigger an automatic action at the end of the collapse of the conservation duration. The action can be the assignment of a new label or the search for a schedule comprising a part to be produced corresponding to the data characterizing the recorded drop, or even the association with a new conservation duration to keep or not keep the drop.

[0182] According to one embodiment, a fifth label LB5 is assigned to a forecast fall CHk. This fifth label LB5 makes it possible to annotate a forecast fall CHk to be re-cut for later use. In the latter case, the label LB5 is a temporary label assigned to a forecast fall CHk which will have a new label LB2 once cut.

[0183] One advantage is to keep the waste actually produced close to a cutting machine so that the waste produced is re-cut initially and then archived in a dedicated room for its conservation later.

[0184] According to one embodiment, a sixth label LB6 is associated with a forecasted fall CHk restoring a surface condition of the latter. One advantage is to allow a surface treatment to be carried out on the fall produced for its reuse according to one of the labels LBi, LB2, LB3. The label LB6 can therefore be assigned cumulatively with another label assigned to a forecasted fall CHk. The label of a forecasted fall LBb LB2, LB3 can be modified after the surface treatment has been carried out. New planning

[0185] When a part PN is selected from the list LISTi, LISTik to be assigned to a location of a forecasted waste CHk of an optimization plan POA of a milestone of a first planning PLi, the first planning PLi is modified. This modification leads to generating a new planning PLi'. The new planning PLi' may include a modification of a single milestone or of a plurality of milestones depending on the number of optimization plans PO A which have been impacted by the reassignment of a part PN to be produced. The modification of a milestone may include a new quantity of parts PN to be produced within a material profile PROk from a new optimization plan generated using the spaces defined by the forecasted waste CHk 2.

[0186] If an assignment of a PN part within an optimization plan of a given milestone impacts a subsequent milestone comprising another POA optimization plan, then the new PLi' planning includes the update of all the milestones. Intermediate optimization plan

[0187] According to one embodiment, an intermediate optimization plan PO; is generated whose length corresponds to the length of the forecast fall CHk. This intermediate optimization plan PO; comprises a plurality of parts PN whose profiles are integrated so as not to overlap and so as to be included in the available length of the forecast fall CHk.

[0188] In a second step, a new optimization plan POB is generated comprising the superposition of the first optimization plan POA and the optimization plan PC). An advantage of this solution is that it does not modify the optimization plan(s) already optimized according to certain constraints. An advantage is that it allows the implementation of the method of the invention without modifying the optimization software of the existing optimization plan.

[0189] The new POB optimization plan replaces the old optimization plan of a PLi or PL planning milestone.

[0190] According to one embodiment, a new optimization plan POB is generated. The new optimization plan POB comprises the arrangement of all the parts Pi initially nested in the optimization plan POA and the new identified parts that can be arranged in the corresponding forecasted scraps CHk. In order to carry out this nesting of at least one new part PN, a nesting or optimization software component can be implemented to regenerate an optimization plan from all the identified parts PN of the database BDi. Parts database - BD1

[0191] Advantageously, the method of the invention comprises a step aimed at updating MAJ2 the BDi database. This update comprises the removal of the occurrences of PN parts that have been selected to be integrated into a new POB optimization plan of a milestone of a given planning PLI, PL;. Waste base - BD2

[0192] Furthermore, the method of the invention comprises an update MAJi of a fall database BD2 in which at least one labeled forecast fall CHk is recorded.

[0193] The BD2 falls database includes the forecast falls CHk identified by their identifier during their CARi characterization. One advantage is to store forecast falls CHk that can change status from “forecast fall” to “actual fall” because they have been produced, for example when they have a LB2 label or an LB3 label. According to one embodiment, the actual falls of a LB2 or LB3 label are made available via a third-party service accessible to at least one user.

[0194] A user interface can then allow a user to access a plurality of actual falls. The actual falls are potentially stored or reused in a PLi or PL schedule. To this end, the user interface makes it possible to sort, choose and select an actual fall that interests a user and that possibly meets certain constraints. The user interface therefore makes it possible to access sorting functions or to display selection criteria. When the data relating to the forecast or actual falls are recorded on a server, the service can be accessible via a communication interface accessible from a data network such as the Internet or an intranet.

[0195] The method of the invention makes it possible to record data specific to forecasted or actual falls even when they have been used by a rearrangement of the optimization plan through the integration of new PN parts. One advantage of recording data relating to an actual fall that no longer exists is to account for the savings made using the method of the invention. The savings can be financial, but also environmental by measuring the reduction in the carbon footprint or the reduction in energy consumption.

[0196] Indeed, a predicted CHk fall which is avoided makes it possible to reduce the subsequent management of residual material. Fall rate

[0197] According to one embodiment, the method of the invention comprises a calculation of a waste rate Te for a given material profile category PROk. The waste rate Te is for example calculated on a set of optimization plans POA of one or more plannings PLb PL;.

[0198] The drop rate corresponds, in one example, to the percentage of unretained material of a given PROk material profile. In another example, the drop rate Te is expressed as a percentage of residual material length.

[0199] According to one embodiment, the drop rate is used to calculate a desired length of a material profile category PROk.

[0200] An advantage of this embodiment is to optimize on the one hand the reuse of the CHk scraps and on the other hand the optimal length of the PROk material profiles to be considered at the input of a PLi, PL; planning.

[0201] According to one embodiment, the method of the invention implements for a given schedule, an optimization algorithm making it possible to minimize the CHk drop rate of the fourth label LB4 for a set of material profiles having different lengths. According to one embodiment, the algorithm comprises an inference or regression operation in order to calculate a minimum CHk drop length having the fourth label LB4.

[0202] This embodiment can also be applied to other labels or sub-labels of forecasted offcuts or to a group of labels. According to one embodiment, the algorithm is applied to a given number of predefined lengths of material profile. This case is interesting when it is known that a reduced number of sizes of material profiles PROk are available, for example, from a catalog. According to another embodiment, the algorithm is carried out on a continuous range of values of sizes of material profiles. This case is interesting when it is possible to define the input length of a material profile PROk as soon as it is machined.

[0203] According to one embodiment, a yield rate is calculated instead of a drop rate. In this case, the aim is to optimize the value of the yield rate. It is possible to apply an algorithm aimed at optimizing all of the labels LBi, LB2, LB3 or to apply an algorithm favoring the optimization of a single label such as the label LBi.

[0204] According to one embodiment, the optimization of the waste rate Te or the yield rate can take into account an optimization criterion which is for example the surface condition or the material of the profile. That is to say that one seeks to optimize in a certain way a profile of such type or such type. One will seek for example to optimize the waste rate Te according to the label LB3 for a material profile having a surface treatment of a first type and one will seek to optimize the waste rate Te according to the label LBi for a material profile having a surface treatment of a second type.

[0205] When a drop rate Te is calculated for a given profile category, this drop rate Te can be optimized for a profile family. In this case, we call, a category or family of profile, material profiles having for example at least one physical property in common such as a surface state, a density or an identical material.

[0206] A surface condition may include a characterization of a coating, paint, or machine surface treatment such as a direction or orientation of rolling or other tool surface treatment.

[0207] By "category of profiles", we can also understand alternatively or in combination, a profile having differences in length less than a given threshold, such as a threshold of 5%, with respect to an average length of this set.

[0208] Another alternative or cumulative condition for defining profiles of the same category may be the taking into account of at least one characteristic dimension of the section of each profile of the set lower or higher than a given threshold with respect to a reference characteristic dimension.

[0209] The reference value can be obtained by calculating an average of the profiles of this set.

[0210] According to one embodiment, the method of the invention comprises a step of simulating at least one planning according to a set of profile categories PROk. In this case, the simulation comprises an optimization of the determination of a material profile category PROk so as to minimize the waste rate Te of said simulated planning.

[0211] According to one embodiment, the simulation makes it possible to carry out different optimizations by taking into account different criteria to infer and determine an optimized solution. For example, the optimization can take into account a given number of parts of a simulated schedule, a number of identical parts of a simulated schedule, a heterogeneity or homogeneity factor of a set of parts arranged within a POA optimization plan, or even a statistical distribution of parts according to a given characteristic such as their length.

[0212] According to one embodiment, the method of the invention comprises the calculation of a statistical indicator corresponding to a quantification of a type of forecast or actual waste produced on a set of schedules. According to one example, this statistical indicator makes it possible to generate a criterion taken into account when labeling a waste to be produced. One advantage is to take into account the actual waste already produced and to redirect the production of other waste that can be kept. System

[0213] Finally, the invention relates to an STi system represented in [Fig.7] comprising at least one computer for executing the steps of the method. According to one embodiment, the computer is a server accessible from a NETp data network. According to another embodiment, the SERVi server is a local server accessible from the industrial site from which the user interface is located.

[0214] According to one embodiment, the different databases, including BDi and BD2, are hosted within a remote data server. According to another case, these databases are not hosted on the same server. Other remote servers may be used within the framework of the invention, in particular servers allowing access to the service offered by the implementation of the method of the invention. By way of example, a user authentication server may be implemented within the framework of the invention.

[0215] The computer or the plurality of calculation units executing the method of the invention comprises at least one calculator Ki making it possible in particular to carry out the operations of comparison and labeling of the forecasted falls CHk. The computer comprises at least one communication interface INT i making it possible to receive data such as the nesting plans, the schedules, and the part identifiers. According to one example, different input interfaces INTi make it possible to receive the different data processed in the method of the invention.

[0216] The system of the invention also comprises a computer having an INT2 output making it possible to transmit the data produced by the calculator such as the data characterizing the predicted falls CHk and the labels.

[0217] According to one example, the calculator Ki also performs the operations aimed at producing a new optimization plan. However, according to another example, this operation is performed by another remote server (not shown) capable of producing a new optimization plan integrating the new designated and characterized PN parts which have been selected from the list LISTi.

[0218] The computer advantageously comprises at least one memory Mi for storing data received or produced by the computer such as the list LISTi or labeling or characterization data for forecast falls CHk or parts PN. List of advantages

[0219] An advantage is to take into account the thickness of the material profile when applying filters to a criterion for selecting the parts to be re-nested.

[0220] One advantage is to optimize the potential of an area to be reused within a material profile.

[0221] An advantage is to qualify the fall as much as possible for better exploitation of the latter.

[0222] One advantage is that it allows available space to be reallocated to optimize material production.

[0223] One advantage is to keep a database up to date with changes to the schedule. Advantageously, a part to be produced selected to be re-nested in a milestone is preferentially removed from the list of parts to be produced.

[0224] According to one embodiment, the first article database comprises a set of parts pre-sectioned to be produced in at least one schedule.

[0225] One advantage is to define a subset of parts to be produced that can be extracted according to a product membership criterion or a production time window criterion.

Claims

1. Claims Computer-implemented method for classifying predicted material waste in an industrial manufacturing process comprising: • Acquisition (ACQi) of a first planning (PLi) of a set of parts (P;) to be cut, said parts to be cut (P;) being intended to be cut within a set of material profiles, each material profile each defining a material element (PROk) extending along a main dimension; • Acquisition (ACQ2) of at least one first optimization plan (POA) of parts to be cut (P;) for each profile (PROk), each part to be cut (P;) within a profile (PROk) comprising geometric descriptors defining at least one given section (Si), a given length (Li) said given length being less than the length of the profile (PROk); • Characterization (CARi) of at least one forecast fall (CHk) defining at least one zone of the optimization plan (OPA) not occupied by the parts to be cut (Pi), said characterization (CARi) comprising at least: • an identifier (IDC) of each forecast fall (CHk); • dimensions (Kc) of each forecast fall (CHk) including at least one fall length (Lck), at least one data characterizing the section of the fall (Sck); • Querying (REQi) a first article database (BDi) comprising a set of parts (Pn) to extract the geometric descriptors of said parts (Pn); • First comparison (COMPi) between the dimensions (Kc) of the at least one forecast fall (CHk) and the dimensions of at least one part (Pn) extracted from the first article database (BDi), said first comparison (COMPi) producing a correspondence indicator (INDi); • Determination of a first list (LISTi) of compatible parts (Pn) from the value of the correspondence indicator (INDi); • Selection (SELi) of at least one part (Pn) from the first list (LISTi) according to at least one given criterion (Ci); • Classification (CLASSi) of each forecasted waste (CHk) according to a plurality of labels (LBi, LB2, LB3, LB4) from a calculation of a characteristic score of said forecasted waste (CHk), said score being calculated in particular from at least the first criterion (Ci); • Generation (GENi) of a new first planning (PLi') comprising the generation of at least one new optimization plan (POB) for at least one material profile (PROk) comprising at least one new part (PN) selected from the first list (LISTi) within zones corresponding to a forecasted waste (CHk) having a reuse label (LBi);• Update (MAJi) of a falls database (BD 2) in which at least one labeled predicted fall (CHk) is recorded.;

2. Method according to claim 1 characterized in that it comprises the calculation of a waste rate (Te) for a given material profile category ({PROk}), the waste rate (Te) being calculated on a set of optimization plans (POA) of one or more plannings (PLi), said waste rate (Te) being expressed as a percentage of residual material length, said waste rate being used to calculate a desired length of a material profile category ({PROk}).

3. Method according to any one of claims 1 to 2 characterized in that a given profile category comprises a set of material profiles (PROk) having • length differences of less than 5% with respect to an average length of this set and; • at least one characteristic dimension of the section of each profile of the set is less than a characteristic dimension of the average of a dimension corresponding characteristic of the section of the profiles of the set, the difference between the chosen characteristic dimension of the section of a profile of the set with the average characteristic dimension of the section being less than 10%.

4. Method according to any one of claims 1 to 3 characterized in that it comprises a simulation of at least one planning according to a set of profile categories ({PROk}), said simulation comprising an optimization of the determination of a material profile category (PROk) so as to minimize the waste rate of said planning.

5. Method according to claim 4 characterized in that the simulation optimization comprises taking into account: the number of parts in a schedule, the number of identical parts in a schedule, a factor of heterogeneity / homogeneity / variability from one part to another, statistical distribution of parts (length, see peak / Gaussian).

6. Method according to any one of claims 1 to 5 characterized in that the step of: • Generation (GENi) of a new first planning (PLi') comprises the generation of at least one new optimization plan (POB) associated with a new material profile (PROk) comprising a dimension different from the first material profile (PROk), said new optimization plan (POB) comprising a forecast drop (CHk) whose surface is smaller than the forecast drop (CHk) of the first optimization plan (POA).

7. Method according to any one of claims 1 to 6 characterized in that the step of: • Generation (GENi) of a new first planning (PLi') comprises the generation of a plurality of new optimization plans (POB) associated with a new material profile (PROk) comprising a dimension different from the first material profile (PROk), said new optimization plans (POB) together comprising a surface of forecast falls (CHk) whose surface area is smaller than the entire surface area of the forecast falls (CHk) of the first optimization plans (POA).

8. Method according to any one of claims 1 to 7, characterized in that it comprises the calculation of a statistical indicator corresponding to a quantification of a type of forecast fall produced on a set of schedules (PLi, PL;), said statistical indicator making it possible to generate a criterion taken into account when labeling a fall to be produced.

9. Method according to any one of claims 1 to 8 characterized in that the classification (CLASSi) comprises: • an assignment of a first label (LBi) to the forecasted scraps (CHk) for their reuse when the first criterion (Ci) comprises: • an occurrence or a lotting of parts selected in the first list (LISTi) in the first planning plan (PLi) or another planning plan (PL;) greater than a predefined threshold and; • a first or a second time interval less than a predefined duration; • an assignment of a second label (LB2) to the forecasted scraps (CHk) for their conservation when the first criterion (Ci) comprises: • an occurrence or a lotting of parts selected in the first list (LISTi) in the first planning plan (PLi) or another planning plan (PL;) greater than a predefined threshold and • a first or second time interval greater than a predefined duration and / or; • an assignment of a third label (LB3) to the forecast offcuts (CHk) for their conservation when the first criterion (Ci) corresponds to structural, dimensional or quality properties of the forecast offcut corresponding to a predefined criterion and / or; • an assignment of a fourth label (LB4) to the forecast offcuts (CHk) for their disposal when the first criterion (Ci) includes:; • an occurrence of selected parts less than a predefined threshold and; • a first or second time interval greater than a predefined duration.

10. Method according to any one of claims 1 to 9, characterized in that the first planning (PLi) comprises at least one milestone corresponding to a cutting forecast of at least one material profile (PROk) forming a step of an industrial production process for a plurality of parts to be produced.

11. Method according to any one of claims 1 to 10 characterized in that the optimization plan (POA) comprises a set of identifiers of parts to be produced, and for each identifier geometric descriptors of a part to be manufactured which is arranged within a characteristic length representing the length of a material profile (PROk) represented within the optimization plan (POA).

12. Method according to any one of claims 1 to 11, characterized in that the dimensions (Kc) of each forecast fall (CHk) include in particular a given length of the forecast fall (CHk) and characteristic dimensions (Kcc) of a section of the forecast fall (CHk).

13. Method according to any one of claims 1 to 12 characterized in that a set of offcuts (CHk) is generated from a processing of all the zones of the optimization plan not occupied by the parts to be manufactured (P;), said processing comprising a segmentation step (SEGi) of the lengths of each forecast offcut from the cutting lines of the parts to be manufactured in the optimization plan (POA) and the edges of the optimization plan.

14. Method according to claim 13 characterized in that the step of segmenting (SEGi) the lengths of each predicted fall (CHk) comprises a first characterization (CAR10) of the fall taking into account the dimensions of the predicted fall (CHk), the position of the predicted fall (CHk), a surface state of the predicted fall (CHk) and / or the material of the predicted fall (CHk).

15. Method according to any one of claims 1 to 14 characterized in that the first comparison (COMPi) comprises the comparison of the length of a predicted drop (CHk) with the length of a part identified in the article database (BD^.

16. Method according to any one of the preceding claims, characterized in that the comparison step (COMPi) comprises comparing the dimensions of a section of a predicted fall (CHk) with the dimensions of a section of a part selected in the article database (BDi).

17. Method according to any one of claims 1 to 16 characterized in that the first comparison (COMPi) comprises the selection of a set of articles from the article database (BDi) having at least one characteristic dimension of the section of said part less than a characteristic dimension of the section of the fall, the difference between the characteristic dimension of the section of each selected part and the characteristic dimension being less than 10% of the value of the characteristic dimension.

18. Method according to any one of the preceding claims, characterized in that the first criterion (Ci) comprises: • taking into account a first occurrence of a part extracted from the article database (BDi) and associated with a first schedule (PLi) or with another schedule (R) and / or; • taking into account a batching of identical parts extracted from the article database (BDi) and associated with at least one material profile (PROk) of the first schedule (PLi) and / or; • at least a first time interval between two milestones of the same schedule (PLi) in which the part extracted from the article database (BDi) is associated with each of the two associated milestones, each milestone being associated with a different material profile (PROk) and / or; • at least a second time interval between two schedules (PLi, PL;) in which the part extracted from the article database (BDi) is associated with at least one milestone of each of the schedules.;

19. Method according to claim 9 taken in combination with any one of claims 1 to 18 characterized in that the assignment of the second label (LB2) or the third label (LB3) includes the association of a retention period with a forecast drop having the second or third label (LB2,LB3).

20. Method according to any one of the preceding claims, characterized in that the generation of the new optimization plan (POB) comprises the superposition of the first optimization plan (POA) and an intermediate optimization plan (POi), said intermediate optimization plan (PO;) comprising at least one new part to be produced selected from the first list (LISTi) and positioned and oriented within zones corresponding to a forecast fall having a reuse label (LBi).

21. Method according to any one of claims 1 to 20 characterized in that the database of falls (BD2') comprises the falls having the second label (LB2) and / or the third label (LB3) so as to provide access to said falls thus labeled to a user via a communication interface accessible from a data network.

22. A computer program product comprising instructions which, when the program is executed by a computer, cause the latter to implement the steps of the method according to any one of claims 1 to 21.

23. System (STi) for generating a labeling of a forecasted fall characterized in that it comprises: • at least one communication interface (INTO) for acquiring the planning data and the optimization plans and the parts from the first database, • a memory (MEMO) for recording said data and • a calculator (KO) for generating a labeling of said forecasted falls according to the method of any one of claims 1 to 21, and in that it further comprises an output communication interface (INT2) for transmitting the updated data of the labeled falls to computer equipment.

Images