Method to design piping of an industrial chemical plant

Abstract

It is disclosed a method to design piping an industrial plant; position information of items and ends of the pipelines at the items are read; -a grid (1) of discrete coordinates (s, 2-20, 22-27, d) is processed, each coordinate having adjacent discrete coordinates (13, 15, 11, 17, 5, 23) along X, Y, Z directions, representing connection positions (2-20, 22-27) of connectable sections of pipelines, wherein distances (3) between adjacent discrete coordinates correspond to lengths of straight sections of the pipelines. An oriented graph (100) is processed from the grid (1), wherein n logical nodes (g1-g11, gs, gd) representing spatial orientations in the X, Y, Z directions of a connectable section of the pipeline at the discrete coordinate (14) are generated; and for each logical node (g1) -i) at least one oriented arc (a2g1) with a logical node (a2) of an adjacent discrete coordinate (13) representing the same spatial orientation is drawn; -ii) a direction is given to the oriented arc (a2g1), representing a flow direction and a weight is associated to the distance between the adjacent discrete coordinates (13, 14); and -iii) internal oriented arcs (g1gd) are drawn between the logical node (g1) and logical nodes of the discrete coordinate (14), where respecting spatial and process constraints of the plant; iv) -weights are also associated to the internal oriented arcs, having a value, if the arc is between logical nodes with different (XY, XZ, YZ) spatial orientations, or a lower value, if the arc is between logical nodes with same (XX, YY, ZZ) spatial orientations; and the optimal piping is searched in the oriented graph minimizing weights.

Description

[0001] Title: Method to design piping of an industrial chemical plant

[0002] Field of application

[0003] The present invention relates to the field of methods for designing piping of an industrial chemical plant of the type including a plurality of equipment and items, wherein each pipeline of the plant includes connectable sections, elbows and valves. Each pipeline extends from one end to an opposite end. In particular, the invention relates to a method of the type cited above wherein a processor is used to implement the design of the piping in accordance with established rules of art.

[0004] Prior art

[0005] The methods to design piping of an industrial plant include manually drawing the piping by means of 3D tools, following the process schemes represented in the Piping and Instrumentation Diagrams (P&ID).

[0006] The P&IDs are typically drawn manually using specific electronic tools to prepare schematic diagrams of the plant’s piping, equipment, instrumentation, and control systems. Some of these tools are called “intelligent” because the process information is attributed to each item and can be called up later to develop other linked documents.

[0007] The P&ID tools are used to prepare schematic diagrams of the piping, equipment, instrumentation, and control systems of the plant. The diagram includes the interconnection of equipment, the instrumentation, the piping sizes and specifications, and the flow direction. Equipment includes, for instance, vessels, tanks, pumps, compressors and heat exchangers. Instrumentation includes sensors, flow meters, pressure gauges and level indicators. Control systems include valves and other components for regulating and controlling the processes in the plant.

[0008] P&IDs tools are crucial for understanding the operation and maintenance of the processes and they serve as a communication tool between engineers and operators involved in the design, construction, and operation. However, the diagrams are schematic drawings not in scale, not representative of the dimensions of pipelines, instrumentation, and equipment and therefore not sufficient for estimating the impact of pipelines on the costs of the plant.

[0009] On the other hand, 3D CAD (Computer-Aided Design) tools support the manual design of three-dimensional models and drawings of the plant, starting from the P&ID diagrams. Software packages, such as AutoCAD Plant 3D, Aveva E3D, and Smart Plant 3D are tailored to meet the needs of plant engineers, but also require hundreds of thousands of man-hours of engineers from multiple disciplines for designing complex piping, for instance, for a new fertilizer plant. It is estimated that approximately 20 percent of the total engineering manhours budget is required for designing the 3D model for piping.

[0010] For reducing this impact, attempts have been made to simplify design by means of automatic processing: positions of equipment, instrumentation and control systems of the plant, as available from P&ID diagrams and 3D layouts, are reported into a 3D model, which is then searched by an algorithm to find the best path in terms of length (shortest length, ultimately lowest weight of the piping) and number (the lowest) of elbows. The scope of these algorithms is to find the best piping for reducing the cost of material and accelerating estimation time.

[0011] However, due to the size and complexity of the model, the algorithmic approach is computation wise unacceptable: time increases exponentially with the number of parameters to process, these last depending on the size of the model, on the degrees of freedom for directions of pipelines in the model and, last but not least, on a plethora of constraints, to cite some the direction of flows, the minimal height from the ground, the orientation at exits from or entry into the equipment, making the processing impossible, at least in absence of a disproportionately expensive computing power.

[0012] The problem at the base of the present invention is provide an automated method for designing piping that adheres to all engineering rules, suitable for reducing design and estimation time while assessing the impact of piping on the plant's cost and identifying or approaching the optimal connection.

[0013] Summary of the invention

[0014] The idea of solution at the present invention is to associate a tridimensional discrete model to a space available for the arrangement of piping of a plant, and processing an oriented graph from the tridimensional model, wherein a variable number of logical nodes of the oriented graph are configured for each discrete coordinate of the tridimensional model, the logical nodes implementing predetermined degrees of freedom of possible directions of the piping but also other process and spatial constraints at the discrete coordinate in the graph.

[0015] Different discrete coordinates with different process and / or spatial constraints may be associated with a different number of logical nodes.

[0016] Position of the items (equipment, instrumentation, control system) support structures and racks are predefined in the tridimensional discrete model, as taken from already available P&ID diagrams and 3D CAD drawings, and piping is designed as an optimal connection among the items by means of an algorithm, searching into the oriented graph after its processing.

[0017] The applicant found that this solution drastically reduces computation time with respect to prior art methods, from exponential to polynomial, due to the decrease in resources spent for processing.

[0018] The technical problem above identified is solved by a computer implemented method to design piping of an industrial chemical plant according to claim 1.

[0019] Preferred embodiments of the computer-implemented method are given in dependent claims from 2 to 15.

[0020] The steps of the method as claimed are executed by technical means. Technical means include computerized means, such as one or more processor(s) and one or more memories. Steps of reading information and constraints on the plant are therefore implemented as inputs to the above-mentioned technical means. Steps of processing a grid, processing an oriented graph and searching for an optimal connection are also implemented by one or more algorithms executed by the processor(s). The technical effect of identifying the optimal connection of the industrial plant is achieved by executing the claimed method steps by means of said technical means.

[0021] Brief description of the drawings

[0022] Fig. 1 is a grid representing a 3D model in which a connection between a discrete coordinate s and a discrete coordinate d at items to be connected in an industrial plant is indicated by arcs representing a pipeline.

[0023] Fig. la schematically represents connectable sections of a pipeline.

[0024] Fig. 2 is another view of the 3D model of Fig. 1, where some labels have been used to identify some discrete coordinates.

[0025] Fig. 2a is a 2D view representing only part of the discrete coordinates of fig. 2 identified by labels.

[0026] Fig. 2b schematically represents a step of processing a portion of an oriented graph from the discrete coordinates of fig. 2a.

[0027] Fig. 3a and 3b schematically represent another step of processing the portion of the oriented graph of fig. 2b.

[0028] Fig. 4a and 4b schematically represent a step of processing the portion of the oriented graph of fig. 2b, for endpoints of the pipeline.

[0029] Fig. 5a and 5d are examples of pipelines not respecting constraints (fig. 5a-5c) and respecting constraints (fig. 5d).

[0030] Fig. 6 schematically represents seven discrete coordinates A-G.

[0031] Fig. 7 schematically represents a model of logical nodes according to the present invention, applied to one (G) of the seven discrete coordinates A-G of fig. 6, i.e. the discrete coordinate 1.

[0032] Fig. 7a is another representation of the model of Fig. 7. Fig. 8 schematically represents the model according to the present invention, applied to another one (B) of the seven discrete coordinates A-G of Fig. 6, i.e. the discrete coordinate B.

[0033] Fig. 8a schematically represents the model according to the present invention, applied other (F, D, E) of the discrete coordinates of Fig. 2.

[0034] Fig. 8b schematically represents the model according to the present invention, applied to other (s, 4, 2, 10) of the discrete coordinates of Fig. 2.

[0035] Fig. 9 schematically represents a step of deleting logical nodes from the models applied to the discrete logical nodes G and B, based on process and spatial constraints.

[0036] Fig. 10 schematically represents logical nodes of discrete coordinates D, F after applying the model and after discarding another logical node (DG) based on process and spatial constraints, the figure also represents oriented arcs between remaining logical nodes of adjacent discrete coordinates.

[0037] Fig. 10a is another schematic representation of Fig. 10.

[0038] Fig. 11 schematically represents internal oriented arc of logical nodes of discrete coordinate G.

[0039] Fig. 12 schematically represents a step of assigning a weight to the oriented arc of fig. 10.

[0040] Fig. 13 schematically represents a step of assigning a weight to the internal oriented arc of fig. 12.

[0041] Detailed description of the invention

[0042] A method to design piping of a chemical industrial plant according to the present invention is disclosed with reference to Figures from 1 to 4b.

[0043] With the expression “chemical industrial plant” it is meant, for instance, a plant where raw materials are processed into chemical products through various physical and chemical processes. These plants are designed for large-scale production and typically include equipment such as reactors, distillation columns, heat exchangers, and storage tanks. Examples of industrial chemical plants are Petrochemical Plants, for processing crude oil and natural gas into fuels, plastics, and chemicals; Pharmaceutical Plants, for manufacturing medicines and active pharmaceutical ingredients; Fertilizer Plants, for producing ammonia, urea, and other fertilizers for agriculture; Polymer & Plastic Plants to make synthetic materials like polyethylene, polypropylene, and PVC, Specialty Chemical Plants for producing adhesives, coatings, dyes, and other specialized products; Food & Beverage Processing Plants, using chemical processes for food additives, preservatives, and flavouring agents.

[0044] Key operations in the chemical industrial plant are: Reactions, i.e. chemical transformations occurring in reactors; Separation and Purification, such as distillation, filtration, or other methods to isolate the desired product; Mixing and Blending, for combining ingredients to achieve specific formulations, but also Packaging and Storage, to prepare products for distribution while ensuring safety.

[0045] According to an embodiment, the method as disclosed in the following description is to design the piping for the chemical industrial plant, in particular for producing ammonia or urea.

[0046] The industrial chemical plant is of the type including a plurality of items, comprising equipment, such as vessels, tanks, pumps, compressors, heat exchangers, instrumentation sensors, flow meters, pressure gauges, level indicators and control systems, such as valves. The term “items” is used in the following description to address any of the above-mentioned components of the industrial chemical plant, and in particular to indicate any component that is destined to be connected to another component of the industrial plant by means of a pipeline.

[0047] The industrial chemical plant further includes structures and buildings, and one or more pipe racks, in general a plurality of main racks and a plurality of sub-racks, which are structures that serve to distribute large quantities of pipes within the plant. A pipeline must be designed to connect equipment according to the process requirements outlined in the P&IDs. It must do so with the shortest path to reduce costs, while respecting the safety and maintainability of the plant. It must include all necessary devices, such as valves or instruments, that are perfectly accessible and removable. It must be integrated with the same criteria that other pipelines have. Finally, it must have a route that is easily supportable, allowing for the attachment of supports to the existing structures.

[0048] A pipeline cannot extend through a solid structure and cannot extend through an item, whereas it preferably extends through racks rather than outside them.

[0049] The items, structures and equipment arrangement are already indicated in a P&ID diagram and on a 3D CAD model. Detailed information on the items is already available, including, for each item, a shape (at least in abstract term, such as “cylinder” or “parallelepiped”), a volume, an orientation, a height in the 3D CAD model, and one of more connection regions where the items have to be connected to one or more pipelines. Orientation is with respect to a reference system (for instance a cardinal system North, South, East, West, Up, Down), height is for example the height of a base of the item with respect to the ground or reference level and the connections region(s) is a region(s) on the outer surface of the shape where one end of the pipeline has to be connected.

[0050] The information on the items, racks and structures, as given from the 3D CAD tool, includes position information in a 3D space of the items, racks and structures and position information of the ends of the pipelines at the items. In an embodiment, the method as disclosed receives in input from the 3D CAD tool the structures, racks and items as objects having said detailed information.

[0051] Piping has to be drawn according to the method of the present disclosure automatically, among other things, to estimate the cost of the piping. Indeed, due to the number of items and complexity of the plant, manual design of piping is very time consuming. The best piping between all the items to be connected has to be found to reduce the cost. A pipeline between two items may potentially assume many different shapes, depending on how straight sections and elbows of the pipelines are connected. In this respect, the pipeline includes a plurality of connectable sections. A connectable section may be a straight section or an elbow.

[0052] To reduce the cost, the overall length of the straight sections forming the pipeline has to be reduced: this reduction substantially corresponds to a reduction in the weight of (the material necessary to manufacture the) piping. Moreover, the overall number of elbows has to be reduced, since the cost of an elbow is higher than the cost of a straight line.

[0053] More precisely, each pipeline has two opposite ends: it extends from one end, where it is connected to one item (at the predetermined connection region thereof), to the opposite end, where it is connected to another item (at the predetermined connection region of said another item) .

[0054] Between the opposite ends of the pipeline, a plurality of connectable sections of the pipeline are arranged, also said intermediate connectable sections, straight sections and / or elbows. Of course, nothing prevents two items from being connected by means of a single connectable section with no intermediate connectable sections.

[0055] To design the pipelines, not only items, structures and racks arrangement have to be considered, but also spatial (or design) constraints and process constraints. Constraints may include (just to cite some) a minimum height of the piping, orientation of a connectable section at the exit from or entry into the item, flow direction into the sections, size of the pipeline, etc.

[0056] Setting constraints is part of an engineering activity, resulting from past experiences of engineers on other plants. Based on experiences gained, the applicant established a best practice for pipeline arrangement; the constraints are stored in a configuration file and are applicable to design piping of new plants to comply with the best practice.

[0057] Here below some examples of constraints are given in natural language whereas a possible encoding thereof in the configuration file is given as an example further below. Example of constraints (natural language) :

[0058] - Minimum height of the ground floor set at value, for instance 100'000 mm.

[0059] - Minimum overhead height set at a value, for instance 102'100 mm.

[0060] - Floors in the main rack set at a value, for instance 104'000 mm for utility pipes and 106'000 mm for process pipes.

[0061] - Planes in sub racks set at a value, for instance 103'000 mm for utility pipes and 105'000 mm for process pipes.

[0062] - Release height inside racks and sub racks equal at a value, for instance 1,000 mm.

[0063] - North orientation set as a y-axis, East orientation set as a x-axis and Up orientation set as a z-axis.

[0064] - Pipe-equipment distance set at a value, for instance at least 300 mm. If the pipeline passes over the equipment, the distance is greater than 1,000 mm.

[0065] - Entering the rack as soon as possible (shortest length) is preferred.

[0066] - Overhead height is the cheapest outside of racks and sub racks; the cost then increases as the distance from the height of the overhead increases.

[0067] - Before changing direction on the x-y plane, change height on the z axis.

[0068] - If the item (equipment) is vertical, the pipeline connection towards the rack is at overhead height; if the item (equipment) is horizontal, the pipeline connection towards the rack is at the most convenient height.

[0069] - Exit from horizontal item (equipment) is from the top towards the short side to avoid passing over equipment.

[0070] - Avoid passing a hose over the appliances.

[0071] - Within the same rack level, the larger tubes go outwards, while the smaller ones go inwards.

[0072] - The critical pipelines and / or pipelines that are more expensive are laid first. The priorities are i) expensive material, (ii) rating, (Hi) diameter, (iv) elevation from the ground. Pipelines for utilities (e.g., cooling circuit; type of fluids: water, steam, air) have a low priority and are therefore the last to be laid.

[0073] - Joint pipelines with different sizes are split in half.

[0074] - If the nozzle position is not specified in the input, the center of the equipment is taken.

[0075] - The headers are located on the main rack.

[0076] - Battery limits must be positioned at the edge of the main piperack, according to the general layout.

[0077] - Control valve group placed as close as possible to the related equipment. Descending to a height of 500mm from the floor to insert the valves, so that they are accessible to humans. Subsequently, get up again on the working surface of the structure.

[0078] - For horizontal equipment, the control valve group can be placed adjacent to the equipment. For vertical equipment, the control valve group can be placed in front of the structural column of the closest structure, in the direction of the rack.

[0079] - All valves must be placed on the same side of the equipment, leaving the other side free for instruments.

[0080] - Distance between the outside of the equipment and the outside of the valves equal to at least 300mm.

[0081] - Pipeline exiting a structure below raised floor running North-South at 1.2m below the raised floor.

[0082] - Pipeline exiting a structure below raised floor running East-West at 1.2 +1.2 m below the raised floor

[0083] - A pipeline exiting a structure remains at its height as it heads towards the rack, without descending to overhead height.

[0084] - Pump template: if the number of pumps is even, delivery and suction pipelines are placed symmetrically.

[0085] - Template ammonia reactor placed inside a structure defined below.

[0086] - The pipelines coming out of the nozzles in the upper part of the reactor are placed at 90° to each other and have a fixed pattern: first they extend away from the reactor, then they extend down and return towards the reactor at an angle of 45°; then they extend toward the bottom along the side of the reactor until the first structure floor encountered. Here, they extend along the side of the structure towards the side of the rack, where the appropriate valves are inserted at a height of 500 mm above the structure floor. Finally, the lines extend towards the rack as one line.

[0087] - On pipe racks, supports for pipes without loops: every 6 m "rest" and "guide" type supports are alternated. The first and last supports in the rack must be of the "rest" type. Where necessary, position two consecutive "guides".

[0088] - Supports for pipes with loops: once the "rest" type supports have been placed laterally, alternate first the "guide" type support and then the "stop" type support symmetrically with respect to the position of the loop.

[0089] - If “Ext Fin Type” (External Finished Type) in the line list shows that it is externally insulated, i.e. it is equal to IH, IC, AF, TE or IP, the rule for positioning the tubes in the pipe rack and structures is: half steel shape + 100mm shoe + half pipe. When “Ext Fin Type” is N or P, the bottom of the tube rests directly on the iron so the rule becomes: half steel shape + half tube.

[0090] The constraints as mentioned above are those considered best practice by the applicant. However, nothing prevents the possibility that more or less constraints or different constraints than those disclosed above are pre-defined, encoded in the configuration file, and applied to find the piping of a new plant.

[0091] The configuration file is given as input to the method of the present invention, along with the above-mentioned information on the plant (the position of items, racks, and structures).

[0092] According to the method as disclosed, the 3D space is modeled into a discrete coordinate system. Modeling the 3D space into a discrete coordinate system includes processing a grid 1 of a plurality of discrete coordinates.

[0093] Fig. 1 represents schematically an example of the grid 1.

[0094] The discrete coordinates s, 2-20, 22-27, d include a discrete coordinate 14 having adjacent discrete coordinates 13, 15, 11, 17, 5, 23 along X, Y, Z directions of the 3D space.

[0095] The discrete coordinates s, d represent positions of the ends s and d of the pipeline to be identified for connecting two items. In other words, the connection regions of two items to be connected in the plant are at s and d. s and d are known from the P&ID and 3D tool.

[0096] Discrete coordinates 2-20-22-27 between s and d represent possible connection positions 2-20, 22-27 of the connectable sections of the pipeline.

[0097] Each connectable section of the pipeline has opposite connection portions for attachment to other connectable sections and a middle portion, straight or curved, extended between the connection portions. Fig. la summarizes this terminology.

[0098] The discrete coordinates 2-20, 22-27 in the grid 1, therefore, represent possible positions for the end portions of the connectable sections of the pipeline. Each connectable section has a central axis and the discrete coordinates 2-20, 22-27 represent the possible positions of the axis of the connectable section at the connection portion thereof.

[0099] Distances 3 between adjacent discrete coordinates in grid 1 correspond to the lengths of the straight sections.

[0100] In processing the grid, a distance between adjacent discrete coordinates in the 3D space is set at a predetermined value in a range. The range is for instance from 0.5 meters to 10 meters. If a fixed value of 1 metre is set, adjacent discrete coordinates are all spaced by 1 meter in the grid 1.

[0101] A discrete coordinate adjacent to another discrete coordinate 14 is defined as the closest discrete coordinate along an axis X, Y, Z with respect to the discrete coordinate 14.

[0102] In this respect, however, although the disclosure is given with reference to adjacent discrete coordinates having a fixed distance, the grid 1 processing may be adaptive, meaning that distance(s) between discrete coordinates may vary depending on volumetric regions of the 3D space. At a first volumetric region, for instance, intended to include several pipelines of small diameters and / or including many items and / or at positions very close to or at the items, distance(s) between adjacent discrete coordinates may be smaller than the distance(s) at a second volumetric region intended for arrangement of pipelines with greater size and / or not including items.

[0103] Preferably, the adjacent discrete coordinates close to or onto the items have lower distances than adjacent discrete coordinates more distanced therefrom. This is due to the fact that the availability of several discrete coordinates, one close to the other, in the 3D space at the (outer surface of the shape representing the) items, support drawing connections of the item to a plurality of pipelines in case a plurality of pipelines have to be arranged one close to the other and attached to the item for multiple connections.

[0104] As said, in order to reduce the cost of the piping, an optimal connection of the plurality of items has to be identified by reducing the weight of the straight sections and the number of elbows while satisfying all the constraints of the plant.

[0105] According to the present disclosure, the optimal connection is searched on an oriented graph 100 which is processed starting from the grid 1. The oriented graph allows to minimize the cost of the pipeline and, at the same time, to satisfy process and spatial constraints, with fewer computational resources than prior art methods.

[0106] Hereafter, steps for processing the oriented graph 100 are disclosed. Some discrete coordinates of the grid 1 are indicated in Fig. 2 with labels (label G, A, B ...in fig. 2 correspond respectively to discrete coordinates 14, 13 and 15 ...in fig. 1). Fig. 2 is the same grid of Fig. 1 but with discrete coordinates marked by letters instead of numbers.

[0107] The discrete coordinates G, A, B, C, D, E, F are also represented, for clarity, in Fig. 6, separately from other discrete coordinates. This Figure is referred by to explain in detail the method steps for processing the oriented graph 100, in particular to explain how it is processed a part of the oriented graph 100 originating from discrete coordinates A-G. Although these method steps are explained taking discrete coordinate G and its adjacent coordinates only, the skilled person may appreciate that same method steps may be processed for all the discrete coordinates of the grid 1.

[0108] Discrete coordinate G has adjacent discrete coordinates A, B, C, D, E and F along X, Y, Z directions of the 3D space. The discrete coordinates A, B, C, D, E and F represent potential positions of connectable sections of the pipeline. Discrete coordinate G has six adjacent discrete coordinates A, B, C, D, E and F along the three axis X, Y, Z of the 3D space.

[0109] However, according to the method disclosed, other discrete coordinates may have fewer adjacent discrete coordinates; this may be the case of discrete coordinates at a border of the 3D space or close to a structural element, such as a wall, through which the pipeline cannot run; for instance, discrete coordinate 2 of Fig. 2 has four adjacent discrete coordinates (s, E, C, 3) and discrete coordinate 3 of Fig. 2 has three adjacent discrete coordinates (2, 12, 6).

[0110] According to the method disclosed, a model is associated to the discrete coordinate. Still considering discrete coordinate G and with reference to Fig. 7, the model defines

[0111] -a logical node representing a flow-to each of the adjacent discrete coordinates to the discrete coordinate G and

[0112] -a logical node representing a flow-from each of the adjacent discrete coordinates to the discrete coordinate G. More in particular: logical nodes GA, GB, GC, GD, GE, GF represent a flow from discrete coordinate G to, respectively, discrete coordinate A, B, C, D, E, F; logical nodes AG, BG, CG, DG, EG, GF represent a flow to discrete coordinate G from, respectively, discrete coordinate A, B, C, D, E, F.

[0113] Twelve logical nodes GA, GB, GC, GD, GE, GF, AG, BG, CG, DG, EG, GF are generated for discrete coordinate G since it has six adjacent discrete coordinates, as mentioned above.

[0114] In order to simplify the meaning and function of the logical nodes of discrete coordinate, each logical node of discrete coordinate G in Fig. 7a is also identified by direction, orientation and flow, as follows. The discrete coordinate G may be considered as the center of a Cartesian system with X, Y, Z axis. “Direction” may be the direction of one of the axis, X, Y, Z. Different logical nodes may be on the same direction X but with different orientation (+X, -X). “Orientation” means the positive or negative side of a certain direction X; a logical node may be associated to a direction X having positive orientation +X and a logical node may be associated to the same direction X but with a negative orientation -X; moreover, a logical node at a certain direction and orientation (+X, -X) may represent an incoming flow (in) and another node at the same direction and orientation (+X, -X) may represent an outgoing flow (out) .

[0115] With reference to Fig. 7a (discrete coordinate G), logical node FG has direction / orientation +x and is for incoming flow, logical node GF has direction / orientation +x and is for outgoing flow, GD has direction / orientation +z for outgoing flow, DG has direction / orientation +z for incoming flow, BG has direction / orientation +y for incoming flow, GB has direction / orientation +y for outgoing flow, EG has direction / orientation -x for incoming flow, GE has direction / orientation -x for outgoing flow, CG has direction / orientation -z for incoming flow, CG has direction / orientation -z for incoming flow, GA has direction / orientation -y for outgoing flow, AG has direction / orientation -y for incoming flow. For a discrete coordinate having fewer than six adjacent discrete coordinates, fewer than twelve logical nodes are created.

[0116] Given a number Nadjc of adjacent discrete coordinates to a certain discrete coordinate, the number of logical nodes for said certain discrete coordinate is

[0117] U = 2 * Nadjc + Kend where Nadjc is the number of adjacent discrete coordinates and

[0118] Kend is equal to two, if the discrete coordinate is associated to one end or the opposite end of the pipeline or

[0119] Kend is equal to 0, otherwise.

[0120] The end and opposite end of the pipeline of Fig. 2 are associated, respectively, to discrete coordinate s and to discrete coordinate d. Each of discrete coordinate s and d have only three (Nadjc = 3) adjacent discrete coordinates, respectively, coordinates 4, 10, 2 (adjacent to s) and 24, B, 12 (adjacent to d). Accordingly, (ns= na = 2 * 3 + 2 = 8) eight logical nodes are generated for each one of discrete coordinate s and d. The logical nodes for discrete coordinate s are represented in Fig. 8b, where Ss and Sd are the logical nodes generated because discrete coordinate is the end of the pipeline.

[0121] Accordingly, if the discrete coordinate is associated to the end of the pipeline and has six (Nadjc = 6) adjacent discrete coordinates, fourteen (n = 2 * 6 +2 = 14) logical nodes are generated for the discrete coordinate. Again, as another example, discrete coordinate B has 5 adjacent coordinates, 24, 6, 12, 18, G (see Fig. 2). Therefore, this step of the method creates ten logical nodes for discrete coordinate B.

[0122] In Fig. 8 are represented only four logical nodes BG, GB, 24B, B24 for discrete coordinate B, for clarity of explanation and since the focus is mainly given in the part of the graph involving discrete coordinates A-G.

[0123] The same apply to discrete coordinates F and D: they have 5 adjacent coordinates (Fig. 2), respectively, 16, 18, 8, 26, G (for F) and 22, 24, 26, 20, G (for D); therefore, this step of the method creates ten logical nodes for each of discrete coordinates F and D.

[0124] In Fig. 8a are represented only two logical nodes, FG, GF, for discrete coordinate F and two logical nodes GD, DG for discrete coordinate D. For same reasons (easy of explanation, focus on part of the graph), only one logical node is represented in Fig. 8a for discrete coordinate E.

[0125] Relevant is however to note that, at the end of this step, a different number of logical nodes (also fewer than ten) may be created for certain discrete coordinates, for instance based on proximity of the discrete coordinates to walls of structures where it is impossible arranging the pipeline and therefore useful to provide connectability (logical nodes). In particular, the discrete coordinates at the border of the 3D space or proximate to walls, structures or other bodies or volumes where the connectable sections cannot run, are associated to fewer logical nodes than those (twelve logical nodes) associated to the discrete coordinate G.

[0126] The above-mentioned step of generating the logical nodes is repeated for all the discrete coordinates.

[0127] Subsequently, another step of the method provides, for some discrete coordinates, removal of some or all of the logical nodes. This removal implements spatial and process constraints.

[0128] For instance, among the logical nodes GA, GB, GC, GD, GE, GF, AG, BG, CG, DG, EG, GF generated for discrete coordinate G, those logical nodes (GA, AG, CG, GC, EG, GB, BG in Fig. 9) not complying with spatial and process constraints at the discrete coordinate (G) are discarded. After this step are therefore kept only the remaining logical nodes GE, GF, FG, GD, DG. These are the logical nodes that satisfies the process and spatial constraints at the discrete coordinate G.

[0129] An example of process and spatial constraints, and therefore possible reasons why logical nodes GA, AG, CG, GC, EG, GB, BG of discrete coordinate G are removed, is given below, without a limiting purpose.

[0130] Process constraints: gas GS must be led above height HGS = 4 meters.

[0131] Spatial constraints: enter a rack as soon as possible. If discrete coordinate G is at a height of 5 meters but adjacent discrete coordinate E is not, i.e. it is for instance at 3 meters, logical node GE of discrete coordinate G, the one potentially connecting in the model the adjacent discrete coordinate E for an outgoing flow from discrete coordinate G to E, is discarded; logical node EG of discrete coordinate G, instead, which is the one potentially connecting the adjacent discrete coordinate E for an incoming flow from discrete coordinate E to G, is kept. This logical node indeed satisfies said process constraint at G since it may be used for carrying gas GS from below 4 meters to above 4 meters.

[0132] Moreover, if a rack at the height of discrete coordinate G (5 meters) is closer to discrete coordinate D than to discrete coordinates A, B, C, logical nodes AG, GA, BG, GB, CG, GC of discrete coordinate G are discarded whereas logical nodes GD, DA of discrete coordinate G are kept, since said spatial constraints provides that the pipeline enters the rack as soon as possible.

[0133] At the end of this step, the discrete coordinate G is associated only to the nodes GF, EG, GD, DG, EG as represented in Fig. 9.

[0134] This step is repeated for all the discrete coordinates. In Fig. 9 is for instance represented discrete coordinate B which, after removal of logical nodes GB, BG does no longer offer possibility of connection with discrete coordinate G. Logical node DG of discrete coordinate D is also discarded (Fig. 9).

[0135] At the end of this step of configuration, for each discrete coordinate, only logical nodes satisfying process and spatial constraints are available. The logical nodes are not still connected and the following steps of the method are provided to this purpose.

[0136] An oriented arc

[0137] -TF (Fig. 10) is drawn from the logical node GF of the discrete coordinate G to the logical node GF of the adjacent discrete coordinate F, if the logical node GF represents the flow-to the adjacent discrete coordinate F and / or an oriented arc FF from the logical node FG of the adjacent discrete coordinate F to the logical node FG of the discrete coordinate G, if the logical node FG represents a flow-from the adjacent discrete coordinate F.

[0138] As shown in Fig. 10, at the end of this step, a graph is drawn having eight logical nodes, i.e. the logical nodes GF, FG of discrete coordinate F; the logical nodes GF, FG, GD, DG, EG of discrete coordinate G; the logical node GD of discrete coordinate D and the logical node EG of discrete coordinate E; and four oriented arcs TF, FF, TD, FE.

[0139] This graph is reported also in Fig. 10a with another nomenclature where logical nodes GF, FG of discrete coordinate F are GFF, FGF, logical nodes GF, FG, GD, DG, EG of discrete coordinate G are GFG, FGG, GDG, DGG, EGG, logical node GD of discrete coordinate D is GDD and logical node EG of discrete coordinate E is EGE.

[0140] Internal oriented arc (il; i2) are also drawn. These arcs connect at least one couple of logical nodes of the discrete coordinate that are associated to two different adjacent coordinates, one logical node of the couple representing a flow-from one of the two different adjacent coordinates and the other logical node a flow-to another one of the two different adjacent coordinates.

[0141] For instance, still taking in consideration discrete coordinate G of Fig. 11 , internal arc i 1 connects logical node EG of discrete coordinate G with logical node GF of discrete coordinate G and internal arc i2 connects logical node FG of discrete coordinate G with logical node GD of discrete coordinate G.

[0142] A weight is set for the oriented arc. The weight is representative of a distance between the discrete coordinate and the adjacent discrete coordinate. This step is represented in Fig. 12. A weight is also set for the internal oriented arc. This step is represented in Fig. 13. This weight is a first value (0) if the two different adjacent coordinates are on a same direction or a second value greater than the first value if the two different adjacent coordinates are on a different direction. The weight of internal oriented arc takes into account the disadvantage of using elbows.

[0143] The piping from a first discrete coordinate at one item to a second discrete coordinate at another item is determined by searching the lowest connection between a logical node of the first discrete coordinate and a logical node of the second discrete coordinate in the graph 100 formed by all the logical nodes, oriented arcs and internal oriented arcs.

[0144] The portion of the oriented graph of Fig. 6 is represented in a further nomenclature in Fig. 3b where -as apparent, for instance- gn corresponds to FGG, gio corresponds to GFG , fi2 corresponds to FGF and fg corresponds to GFF, etc... .

[0145] The portions of the graph of Fig. 3b does not however represent same process and spatial constraints of Fig. 10: no logical nodes is for instance discarded in association with the discrete coordinate 14. The following disclosure is indeed focused on the step of generating the logical nodes and the step of drawing arc, for instance considering the no logical nodes is discarded at discrete coordinate 14 (since all the process and spatial constraints are satisfied without discarding) .

[0146] According to the method disclosed, for each discrete coordinate 14:

[0147] - 14 logical nodes gi-gi2, gs, gd (fig. 3a, 3b) are generated, representing spatial orientations in the X, Y, Z directions of a connectable section of the pipeline at the discrete coordinate 14.

[0148] The spatial orientation is the orientation of the connectable section with respect to a reference system, for instance, the cardinal system (or the Cartesian system). For instance, logical node g7 (corresponding to DGG), which is associated with discrete coordinate 14, means that a connection section at discrete coordinate 14 may be oriented upwards, in the Z axis (with the opening of the connection section upwards); logical node gio, always associated to discrete coordinate 14, means that the connection section at discrete coordinate 14 may be oriented towards East, in the X axis (i.e. with the opening of the connection towards East); logical node gi, means that the connection section at discrete coordinate 14 is oriented to North, in the Y axis (with the opening towards North).

[0149] For each logical node of the n logical nodes gi-gi2, gs, gd, the following steps are processed.

[0150] One oriented arc a^gi is drawn (processed) between the logical node gi and one logical node a2 of an adjacent discrete coordinate 13 (here it is considered that case that also such logical node a2 of the adjacent discrete coordinate 13 is available, i.e. not discarded).

[0151] As shown in 3D representation of figure 2, discrete coordinate 13 is adjacent to discrete coordinate 14. Also, adjacent discrete coordinate 13 includes logical nodes ai.aia, only in part represented in fig. 3a and 3b. The logical node a2 of the discrete coordinate 13 represents the same spatial orientation of the logical node gi, i.e. orientation on Y axis or North. Due to the same orientation, oriented arc a2^gi represents a possible connection between the adjacent discrete coordinates 13, 14, meaning that a connection section may be considered for arrangement between discrete coordinates 13, 14.

[0152] Same procedure is applied for other logical nodes, for instance gs, c? to process other arcs gs^c? representing a possible connection between the adjacent discrete coordinates 14, 5 (see also fig. 1, where in may be appreciated that discrete coordinates 14, 5 correspond to G and C of Fig. 2).

[0153] A direction and a weight are associated with all the oriented arcs, including arc a2^gi. The direction represents a flow direction. For instance, arc is from a2 to gi since these logical nodes represent a flow that goes from a2 to gi

[0154] In the same way, a direction is associated to the oriented arc between the logical nodes of discrete coordinate 14 and the logical nodes of the other adjacent discrete coordinates. The weight is set to the distance between the adjacent discrete coordinates. For instance, if discrete coordinate 14 is distanced one meter from discrete coordinate 13, weight of arc a^gi is also set to one meter.

[0155] The oriented arcs represent connections between logical nodes of discrete coordinate 14 and logical nodes of adjacent discrete coordinates 13, 15 which are discrete coordinates associated to physical points in the 3D space distanced from discrete coordinate 14. Logical node a2 is an external logical node with respect to logical node g2, since the first logical node a2 belongs to a discrete coordinate 13 different from the discrete coordinate 14 to which the latter logical node g2 belongs.

[0156] On the other hand, logical nodes gi-gi2 are all associated with discrete coordinate 14, and therefore, they are internal logical nodes (internal to the same discrete coordinate). Connections between internal logical nodes gi-gi2 are used to represent the shape (elbow or straight section) of the connection sections. For instance, a connection between logical nodes g3 and g4 represents a straight section since g3 and g4 are on (represent a) same direction Y (fig. 3b); instead, a connection between logical nodes g3 and gio or g6 represents an elbow, since g3 is on a direction Y different from the direction represented by gio (direction X) or go (direction Z) .

[0157] Based on the above, at least one internal oriented arc gi^g2 (Fig. 3a), gii->gs (Fig. 3b) is drawn between internal logical nodes of the discrete coordinate 14.

[0158] Each internal oriented arc gi^g2 is also associated to a weight. The weight of an internal oriented arc has a value if the internal oriented arc is between logical nodes representing spatial orientations in different directions XY, XZ, YZ or a lower value if the internal oriented arc is between logical nodes with spatial orientations in the same direction XX, YY, ZZ. This different weight represents a higher cost of elbows respect to a straight section; therefore, elbows have to be avoided in a subsequent step of searching the optimal path in the oriented graph, if preferred choices (going straight) are available. The value may be 1 and the lower value may be 0. However, other values may be adopted.

[0159] At the end of the steps mentioned above, the oriented graph 100 including a plurality of logical nodes (internal and external to each discrete coordinate) and a plurality of arcs (internal or not) is complete.

[0160] The step of searching the optimal connection is made subsequently, on the oriented graph 100, by:

[0161] - identifying, for each end and opposite end of a pipeline at items to be connected, the oriented arcs and internal oriented arcs connecting the logical node corresponding to the discrete coordinate s at one end of the pipeline to the logical node corresponding to the discrete coordinate d at the opposite end of the pipeline, that are suitable to

[0162] - minimize the sum of weights of the oriented arcs and internal oriented arcs.

[0163] The expression “searching the optimal connection” means “searching the plurality of pipelines necessary to connect all the items to be connected with the lowest price (weight)”.

[0164] The spatial and process constraints are taken into account due do the fact that the logical nodes of the oriented graph already intrinsically implement (satisfy) the constraints.

[0165] For instance, if a constraint is in the form (natural language)

[0166] “an item at the end of the pipeline may be entered only from the above”, the configuration file may encode the constraints as

[0167] “if (discrete coordinate == d) than select only logical nodes from Z direction”, wherein d is the end of the pipeline.

[0168] An algorithm used to determine the optimal connection is faster than those adopted in the prior art since logical nodes not satisfying spatial and process constraints are discarded. The configuration of the logical nodes at the discrete coordinates is adaptive. Which logical nodes to discard and keep, and therefore the number of logical nodes for each discrete coordinate, is determined locally, at the discrete coordinate.

[0169] This approach allows to significantly reduce computation time in the search. In particular, the applicant noted that, by paying a little more in terms of memory consumption to store the oriented graph (it is more memory consuming than an un-oriented arc and due to the number of logical nodes potentially generated), a much greater benefit is reached in terms of time for processing the optimal connection with respect to the prior art methods. According to the method as claimed, indeed, the time for processing increases in polynomial order with the increase of the number of logical nodes whereas in the prior art method the time for processing increases in exponential order with the increase of the number of nodes.

[0170] Time may be further reduced by increasing distances among the adjacent discrete coordinates, especially in the second volumetric regions, i.e. at a certain distance from the items, and therefore reducing the number of logical nodes created.

[0171] For instance, at a second region farther than 3 meters from an item, the distance between adjacent discrete coordinates may be set at a value between 5 and 15 meters. In an embodiment, a distance between the discrete coordinates in the 3D space is set at a value in a predefined range, the range being from 0.5 meters to 10 meters. In other embodiments, more than two volumetric regions are defined, and different distances are set for any volumetric regions.

[0172] Still with reference to Fig. 3a, 3b, it may be appreciated that the step of generating n logical nodes includes generating two logical nodes gn, gn for each direction X, Y, Z:

[0173] - one logical node gn of the two gn, gn logical nodes is associated to a first direction with a positive orientation +X, representing connection from East (or top in fig. 3b) of the discrete coordinate 14, and the other one gn of the two logical nodes gn, gn is associated to the first direction with a negative orientation -X, representing connection from West (or bottom in fig. 3b, i.e. differently from the example given above where X was the direction associated to the height), - one logical node g4 of the two g4, g3 logical nodes associated to a second direction with positive orientation +Y, representing connection from a side (North) and the other one g3 of the two g4, g3 logical nodes associated to the second direction with negative orientation -Y, representing connection from an opposite side (South in fig. 3b),

[0174] - one logical node gs of the two gs, g7 logical nodes associated to a third direction with positive orientation +Z, representing connection from upward (up) and the other one g7 of the two gs, g7 logical nodes associated to the third direction with negative orientation -Z, representing connection from the downward (fig. 3b)

[0175] In an embodiment, generating two logical nodes for each direction may be superfluous; for instance, when the discrete coordinate is at ground level, the logical node in the Z direction upwards is necessary, but not the logical node in the Z direction downwards, since the pipelines cannot extend underground level. In this case the logical node in the Z direction downwards is never created. However, in an embodiment, both the logical nodes in the Z direction are processed and then the logical node in the Z direction downwards is discarded, in such a case, the configuration file also includes one constraint of the type:

[0176] “If discrete coordinate is at ground level, discard the logical nodes representing Z direction downward”.

[0177] The step of generating n logical nodes includes further generating two logical nodes for each of the connections from the bottom, top, side, other side, front and back, wherein

[0178] - a first one gn, g7, g3, g9, gs, gi (fig. 3a, 3b) of the logical nodes is for connection representing process flow incoming, respectively, to the bottom, top, side, other side, front and back and

[0179] - another one gio, g6, g2, gi2, gs, g of the logical nodes represents process flow for connection outgoing, respectively, from the bottom, top, side, other side, front and back.

[0180] Among the information taken in input from the P&ID and 3D tool, the direction of a process flow from and to an item is available. This means that one of the two logical nodes representing the incoming or outgoing flow is disregarded based on the information on the item as read from the P&ID and 3D tool.

[0181] In a preferred embodiment, the step of generating n logical nodes generates fourteen logical nodes gi-gi2 if the discrete coordinate 14 is at a position of one of said ends s, d, i.e. at the connection region of an item. Among the fourteen logical nodes gi-gi2, gs, gd , nodes gsand gd are specific for the end and opposite end of the pipeline. Fig. 4a and 4b represent node G in case it was, respectively, an end s of the pipeline where a flow originates or an opposite end of the pipeline where the flow terminates (i.e. in this case G no longer relates to fig. 2). In such case, the internal arc exiting from logical node gsmay go in any direction, i.e. towards any one of logical nodes g2, g6, gio, gs, g4, gi2 representing outgoing flows and the internal arc entering into logical node gd may enter from any one of logical nodes gi, gs, go, g7, g3, gn representing ingoing flows. The cost of both said internal arcs is null.

[0182] For intermediate discrete coordinates, the step of generating n logical nodes may generate up to twelve logical nodes gi-gi2 if the discrete coordinate is at a position of a connectable section between the end and the opposite end (i.e. in this case, G is as drawn in fig. 2, since intermediated between for instance A and B) .

[0183] In a preferred embodiment, fourteen logical nodes are generated for any discrete coordinates, independently from the fact that the coordinate is intermediate or at the end of the pipelines, but logical nodes gsand gd are not connected for intermediate discrete coordinates. This is indeed the case represented in fig. 3a, 3b, where logical nodes gsand gd are processed for discrete coordinate 14, although the discrete coordinate G (in fig. 2) is intermediate.

[0184] In this embodiment, and in the worst scenario where process and spatial constraints do not reduce the number of logical nodes, for each non oriented arc of the grid 1, two oriented arcs are generated in the oriented graph 100, and for each node in the grid 1 fourteen nodes are generated in the oriented graph 100 and 34 internal oriented arcs. Also, with this choice, the Applicant appreciated a significant reduction in time for processing the optimal connection.

[0185] Here below are given some examples for implementing constraints through the logical nodes and arcs, giving the contains in natural language (a to c) and in a pseudo language (a’ to c’) for possible encoding in the configuration file. Of course, these are only examples, not exhaustive or limitative.

[0186] Natural language: a) Entering the item at destination of a pipe always from the top; b) Horizontal pipelines have to be higher than two meters; c) Before changing direction on the x-y plane, change height on the z axis.

[0187] Pseudocode in the configuration file: a’) if (discrete coordinate == d) than discard the logical node entering the discrete coordinate d on Z direction from the bottom (-Z in) and use internal oriented arcs on Z direction from the top (+Z in); b’) if (z component of discrete coordinate is lower than 2 meters) than discard the logical nodes on X and Y directions if the discrete coordinate is lower than 2 meters and use internal oriented arcs on Z direction; c’) if (logical nodes is go or gio or gn or gi2 does not use internal arc from gi to g4).

[0188] As appreciated for this example, also the way in which arc are drawn may implement constraints. In example c’) no one of logical nodes g9, gio, gn, gi2, gi to g4 has been discarded but internal arc from gi to g4 are not drawn to avoid changing direction on the x-y plane, before changing height on the z axis. Since this internal arc is not drawn, the algorithm to search may be faster. In an embodiment, the process of drawing internal arcs and external arcs is subject to verification of constraints in the configuration file, meaning that an arc that does not satisfy the constraints is never drawn even if logical nodes of adjacent discrete coordinates and logical nodes of the discrete coordinate are available (i.e. they have not been discarded). In another embodiment, the process of drawing an internal arc and the external arc is carried out, meaning that the arc is drawn between logical nodes of adjacent discrete coordinates (nodes that have not been discarded) or between logical nodes of the discrete coordinate, but subsequently the arc is discarded subject to verification of constraints in the configuration file. Arcs discarded before searching the graph.

[0189] Fig. 5a to 5d represent a pipeline between s and d.

[0190] The pipeline in fig. 5d is longer than the pipelines of Fig. 5a to 5c but differently from these last satisfies physical and process constraints. Indeed:

[0191] Pipeline {s, 2, 3, 12, d} of Fig. 5a is shorter than pipeline {s, 10, 11, 12, d} in same fig. 5a but no one of them satisfy the following constraints (given in natural language for easy of explanation) :

[0192] Cl) before changing direction from Y direction to X direction (or vice versa), turn to Z direction.

[0193] This constraint is implemented by not drawing the internal arc.

[0194] Pipeline in Fig. 5b satisfies constraint Cl) but fails to satisfy following constraint:

[0195] C2) equipment at the end of the pipeline has to be entered from the top (from discrete coordinate 24).

[0196] This constraint is implemented by discarding a logical node (to avoid entering from the bottom). Pipeline in Fig. 5c satisfies constraints Cl) and C2) but fails to satisfy following constraint:

[0197] C3) horizontal section of the pipeline (on axis X, Y) has to be higher than a minimum height.

[0198] This constraint is implemented by discarding some logical nodes (those moving horizontally at a height lower than the minimum height. All requirements Cl)-C3) are satisfied by the pipeline of Fig. 5d.

[0199] All these requirements are already satisfied when the searching in the oriented graph.

[0200] According to the method as disclosed, once the optimal connections have been automatically found, a further step is provided to draw the piping in the 3D model. This is a great advantage for engineers who may appreciate compliance of the piping to the constraints and, only if desired, change some path. In any case, for the purpose of estimating the cost of the piping, further optimization (if possible) at this stage may be skipped.

[0201] Advantageously, according to the present disclosure, each node of the grid representing 3D spatial component of piping (the discrete coordinates) is replaced with a new node with n connections, preferably fourteen connections. This replacement allows to treat the problem of searching the optimal piping at polynomial complexity instead of exponential complexity. For instance, a bidirectional Dijkstra’s algorithm may be adopted.

Claims

CLAIMS1. Computer implemented method to design piping of a chemical industrial plant including a plurality of items, each pipeline including connectable sections, comprising at least one straight section and / or at least one elbow, and extending from one end at one item to an opposite end at another item of said plurality of items, the method including the following steps processed by a processor:- reading information on the plant, including position information in a 3D space of the items and position information of the ends of the pipelines at the items, and reading constraints of the plant, including spatial and process constraints of the pipelines;- processing a grid including a plurality of discrete coordinates (A-G, ) of the 3D space, each discrete coordinate (G) having one or more adjacent discrete coordinates (A-F) along X, Y, Z directions of the 3D space, and said discrete coordinates (A-G) representing positions of said ends and connection positions of said connectable sections, where the method for each discrete coordinate (G): i) associate a model to the discrete coordinate (G), the model defining a logical node (GAG, GBG, GCG, GDG, GEG, GFG) representing a flow-to and a logical node (AGG, BGG, CGG, DGG, EGG, FGG) representing a flow-from each of said one or more adjacent discrete coordinate (A-F); ii) among the logical nodes of the model, discarding those logical nodes (GAG, AGG, GBG, BGG, GEG, GCG, CGG) not complying with spatial and process constraints at the discrete coordinate (G), and configuring only the remaining logical nodes (GDG, GFG, FGG, DGG, EGG) , where the step of configuring includes for each (GFG , FGG) of the remaining nodes (GDG, GFG, FGG, DGG, EGG) : a) drawing an oriented arc (TF) from the logical node (GFG) of the discrete coordinate (G) to a logical node (GFF) of the adjacent discrete coordinate (F), if the logical node (GFG) represents the flow-to the adjacent discrete coordinate (F) or drawing an oriented arc (FF) from the logical node (FGF) of the adjacent discrete coordinate (F) to thelogical node (FGG) of the discrete coordinate (G), if the logical node (FGG) represents a flow-from the adjacent discrete coordinate (F); b) drawing an internal oriented arc (il; i2) connecting at least one couple (GFG-EGG) of logical nodes of the discrete coordinate (G) that are associated to two different adjacent coordinates (F, E), one logical node (EGG) of the couple (GFG-EGG) representing a flow-from one (E) of the two different adjacent coordinates (F, E) and the other logical node (GFG) a flow-to another one (F) of the two different adjacent coordinates (F, E); c) setting a weight (10; 20) to the oriented arc (TF) representative of a distance between the discrete coordinate (G) and the adjacent discrete coordinate (F), and a weight (0; 2) to the internal oriented arc at a first value (0) if the two different adjacent coordinates (F, E) are on a same direction (Z) or a second value (2) greater than the first value if the two different adjacent coordinates (F, D) are on a different direction (Z, X), wherein the piping from a first discrete coordinate at one item to a second discrete coordinate at another item is determined by searching the lowest connection between a logical node of the first discrete coordinate and a logical node of the second discrete coordinate in the graph formed by all the logical nodes, oriented arcs and internal oriented arcs.

2. Method according to claim 1 wherein a number of logical nodes configured at one discrete coordinate (G) after said step i) of associating the model at said one discrete coordinate (G) and before said step ii) of discarding logical nodes at said one discrete coordinate (G) is different from a number of logical nodes configured at another discrete coordinate (B) after said step i) of associating the model at said another discrete coordinate (B) and before said step ii) of discarding is logical nodes at said one discrete coordinate (B).

3. Method according to claim 1 wherein a number of logical nodes configured at one discrete coordinate (G) after said step i) of associating the model at said one discrete coordinate (G) and before said step ii) ofdiscarding logical nodes at said one discrete coordinate (G) is equal to a number of logical nodes configured at another discrete coordinate (B) after said step i) of associating the model at said another discrete coordinate (B) and before said step ii) of discarding is logical nodes at said one discrete coordinate (B).

4. Method according to claim 3 wherein a number of logical nodes configured at one discrete coordinate (G) after said step ii) of discarding logical nodes at said one discrete coordinate (G) is different to a number of logical nodes configured at another discrete coordinate (B) after said step ii) of discarding is logical nodes at said one discrete coordinate (B) .

5. Method according to claim 1 wherein said step a) of drawing an oriented arc (TF) from the logical node (GFG) of the discrete coordinate (G) to the logical node (GFF) of the adjacent discrete coordinate (F) or an oriented arc (FF) from the logical node (FGF) of the adjacent discrete coordinate (F) to the logical node (FGG) of the discrete coordinate (G) is stopped and the internal oriented arc is not drawn if said oriented arc (TF, FF) does not satisfy said spatial and process constraints.6 Method according to claim 1 wherein said step b) of drawing an internal oriented arc (il; i2) connecting at least one couple (GFG-EGG) of logical nodes of the discrete coordinate (G) that are associated to two different adjacent coordinates (F, E) is stopped and the internal oriented arc is not drawn if said oriented arc (TF, FF) does not satisfy said spatial and process constraints.

7. Method according to claim 1 wherein a number of oriented arc exiting from logical nodes associated to said discrete coordinate (G) and / or entering to the logical nodes associated to discrete coordinate (G) is different from a number of oriented arc exiting from logical nodes associated to another discrete coordinate (F) and / or entering to logical nodes of said another discrete coordinate (F).

8. Method according to claim 1 wherein a number of internal oriented arcs between logical nodes associated with said discrete coordinate (G) is different from a number of internal oriented arc between logical nodes associated with another discrete coordinate (G) .

9. Method according to claim 1, wherein said step i) of generating n logical nodes generates at least twelve logical nodes if said discrete coordinates has six adjacent discrete coordinates.

10. Method according to claim 9, wherein said step i) of generating n logical nodes generates fourteen logical nodes if said discrete coordinate has six adjacent discrete coordinates and if said discrete coordinate is associated to said one end or to said opposite end at another item.

11. Method according to claim 1 , wherein said step i) of generating n logical nodes generates n = 2 * Nadjc + Kend logical nodes, wherein Nadjc is the number of adjacent discrete coordinates and Kend is equal to two, if said discrete coordinate is associated to said one end or to said opposite end or Kend is equal to 0, otherwise.

12. Method according to claim 1, wherein said step i) of generating n logical nodes generates less than twelve logical nodes if said discrete coordinates has less than six adjacent discrete coordinates.

13. Method according to claim 1, further including the step of processing a configuration file encoding the spatial constraints and the process constraints.

14. Method according to claim 1, wherein said step of searching is carried out by a bidirectional Dijkstra’s algorithm.

15. Method according to claim 1 wherein the industrial chemical plant is a plant for producing ammonia or urea.

Images