A method and device for planning an oiling path

Abstract

The application provides a kind of oiling path planning method and device, and a kind of computer readable storage medium.The planning method includes the following steps: according to the process flow diagram of crude oil tank farm, directed graph model is established, and the attribute data and state data of process equipment and / or pipeline in the crude oil tank farm are saved to corresponding node and / or corresponding link;Adjacency list is established to store the attribute data, state data, index information of the pointing node of each node and link information connected to the pointing node;Obtain at least one subtask starting node information, termination node information and ideal pump flow data;According to multiple optimization objectives, the directed graph model and the adjacency list, the objective function of the multi-objective optimization problem is established, and the feasibility constraint conditions of path search and path combination are set;And based on the feasibility constraint condition, multi-objective evolution operation is carried out to determine at least one non-dominated path planning scheme.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of pipeline transportation of oil products, and particularly relates to a method for planning an oil delivery path, a device for planning an oil delivery path, and a computer readable storage medium. BACKGROUND

[0002] Pipeline transportation is one of the main transportation methods of tank farm oil products, and its scheduling process mainly includes storage and transportation, pipeline transportation, oil product blending, etc. The oil delivery path planning is an important part of the oil scheduling task, which is related to the feasibility, convenience and economy of the oil scheduling task. However, the oil delivery path planning currently mainly relies on the personal rich professional knowledge and experience of the scheduling personnel to manually formulate, and the following problems exist:

[0003] 1) At present, most of the domestic crude oil tank farms have the phenomenon of multiple types of oil products, complex pipeline network structure, widely distributed storage tanks, and large number of process equipment, and the process flow diagram also has the phenomenon of complex connection relationship of process equipment such as storage tanks, oil pumps and valves and numerous pipeline intersections, which causes the problems of complex operation, long time consumption, difficulty in balancing efficiency and cost, etc. for the manual scheduling of the scheduling personnel.

[0004] 2) The current process flow diagram cannot display the real-time state of the equipment, the scheduling personnel are difficult to select appropriate process equipment and pipelines, and the actual state of all pipelines and equipment cannot be taken into account, which is easy to cause pump flow mismatch, equipment damage, pipeline conflict, etc.

[0005] 3) The oil scheduling task generally includes multiple sub-tasks, involves multiple storage tanks and blending heads, and the number of feasible path planning schemes is huge, so it is difficult for the scheduling personnel to select the best path planning scheme.

[0006] 4) At present, the domestic crude oil tank farm has the phenomenon of multiple types of oil products, complex pipeline network structure, widely distributed storage tanks, and large number of process equipment, which leads to high complexity of the oil delivery path planning task and great optimization difficulty.

[0007] In order to overcome the above-mentioned defects existing in the prior art, there is an urgent need in the field for a planning technology for an oil delivery path, which is used to overcome the dependence of the oil delivery path planning on personnel knowledge and experience, reduce the operation difficulty, shorten the planning time, and improve the optimization degree of the oil delivery path planning. SUMMARY

[0008] The following gives a brief summary of one or more aspects to provide a basic understanding of these aspects. This summary is not an exhaustive overview of all contemplated aspects, and neither is it intended to identify key or critical elements of all aspects nor to delineate the scope of any or all aspects. Its only purpose is to present some concepts of one or more aspects in a simplified form, as a prelude to the more detailed description given later.

[0009] In order to overcome the above-mentioned defects existing in the prior art, the present application provides an oil delivery path planning method, an oil delivery path planning device, and a computer readable storage medium, which can overcome the dependence of oil delivery path planning on personnel knowledge and experience, reduce the operation difficulty, shorten the planning time, and improve the optimization degree of oil delivery path planning.

[0010] Specifically, the above-mentioned oil delivery path planning method provided by the first aspect of the present application comprises the following steps: establishing a directed graph model according to a process flow diagram of a crude oil tank area, and saving attribute data and state data of at least one process equipment and / or at least one pipeline in the crude oil tank area to corresponding nodes and / or corresponding links of the directed graph model; establishing an adjacency list according to modeling data of the directed graph model to store attribute data, state data, index information of a pointing node, and link information connecting the pointing node of each node; obtaining starting node information, terminating node information of at least one subtask in an oil product scheduling task, and ideal pump flow data of oil product movement; establishing a target function of a multi-objective optimization problem according to a plurality of optimization objectives of the oil product scheduling task, the directed graph model and the adjacency list, and setting a first feasibility constraint condition of path search and a second feasibility constraint condition of path combination; and performing multi-objective evolution operation based on the first feasibility constraint condition and the second feasibility constraint condition to determine at least one non-dominated path planning scheme.

[0011] Further, in some embodiments of the present application, the plurality of optimization objectives comprises reducing total path length, improving matching degree of pump flow, and reducing total power of pumps. The step of establishing a target function of a multi-objective optimization problem according to a plurality of optimization objectives of the oil product scheduling task, the directed graph model and the adjacency list comprises: determining the target function of the multi-objective optimization problem as:

[0012]

[0013]

[0014]

[0015] wherein F1 represents total path length of a path planning scheme, F2 represents matching degree of pump flow in the path planning scheme, and F3 represents total power of pumps in the path planning scheme.

[0016] Further, in some embodiments of the present application, the process equipment comprises storage tanks and oil pumps. The nodes comprise storage tank nodes and oil pump nodes, and the first feasibility constraint of the path searching comprises that each node in the path of each subtask is passed only once, and that only one storage tank node and one oil pump node are included in the path of each subtask.

[0017] Further, in some embodiments of the present application, the second feasibility constraint of the path combination comprises that the paths of the subtasks do not pass through the same nodes and / or the same pipelines.

[0018] Further, in some embodiments of the present application, the step of performing multi-objective evolutionary operation based on the first feasibility constraint and the second feasibility constraint to determine at least one non-dominated path planning scheme comprises: constructing and initializing an initial population and an external population; constructing a plurality of weight vectors and weight nearest neighbor sets thereof according to the number of optimization objectives, and assigning each subproblem of the multi-objective optimization problem with a weight vector; randomly selecting two indexes from the weight nearest neighbor set corresponding to the subproblem, and performing crossover and mutation on the corresponding two parent path planning schemes to generate a child path planning scheme of the subproblem; judging whether the child path planning scheme is better than the current path planning scheme of the subproblem; and in response to the judgment result that the child path planning scheme is better than the current path planning scheme, replacing the current path planning scheme with the child path planning scheme in the initial population, and updating the external population with the non-dominated child path planning scheme.

[0019] Further, in some embodiments of the present application, the step of constructing and initializing an initial population and an external population comprises: from the starting nodes of each subtask, sequentially setting the search direction of the current node according to the first feasibility constraint to generate a plurality of random paths connecting the corresponding termination nodes; combining the random paths of each subtask according to the second feasibility constraint, and adding the path planning scheme obtained by the combination to the initial population as a random individual; and constructing an external population, and adding the non-dominated schemes in the initial population to the external population.

[0020] Further, in some embodiments of the present application, the step of constructing a plurality of weight vectors and weight nearest neighbor sets thereof according to the number of optimization objectives comprises: generating N weight vectors uniformly distributed on a hyperplane using a standard boundary crossover method, wherein, H represents the number of partitions of each dimension objective, and m is the number of objectives; and calculating at least two nearest neighbor weight vectors of each weight vector, and storing the indexes of the at least two nearest neighbor weight vectors in the corresponding weight nearest neighbor set B{i}.

[0021] Further, in some embodiments of the present application, the step of assigning each sub-problem of the multi-objective optimization problem a weight vector comprises: decomposing the multi-objective optimization problem into N sub-problems using the MOEAD algorithm, and assigning each sub-problem a weight vector.

[0022] Further, in some embodiments of the present application, the step of randomly selecting two indexes from the weight nearest neighbor set corresponding to the sub-problem, and performing crossover and mutation on the corresponding two parent path planning schemes to generate a child path planning scheme of the sub-problem comprises: randomly selecting index k and index l from the weight nearest neighbor set of the i-th sub-problem, to take out the corresponding parent path planning schemes S k and S l from the parent population; and performing crossover and mutation on the paths of each sub-task of the parent path planning schemes S k and S l respectively, and / or performing crossover on the path combination of the parent path planning schemes S k and S l to generate a child path planning scheme of the sub-problem.

[0023] Further, in some embodiments of the present application, the step of determining whether the child path planning scheme is better than the current path planning scheme comprises: setting an ideal point Z = (0, 0, 0); calculating the Chebyshev distance of the ideal point Z to the child path planning scheme and the current path planning scheme respectively; and in response to the determination result that the Chebyshev distance of the ideal point Z to the child path planning scheme is less than the Chebyshev distance of the ideal point Z to the current path planning scheme, determining that the child path planning scheme is better than the current path planning scheme.

[0024] Further, in some embodiments of the present application, the planning method further comprises the following steps: in response to the determination result that the child path planning scheme is not better than the current path planning scheme, the sub-problem retains the current path planning scheme.

[0025] Further, in some embodiments of the present application, the planning method further comprises the following steps: determining whether the current evolution generation number reaches a preset maximum generation number; in response to the determination result that the current generation number does not reach the maximum generation number, reselecting parent path planning schemes for each sub-problem of the multi-objective optimization problem to perform the next round of crossover and mutation iteration; and in response to the determination result that the current generation number reaches the maximum generation number, outputting the external population.

[0026] In addition, the oiling path planning device according to the second aspect of the present application includes a memory and a processor. The processor is connected to the memory and is configured to implement the oiling path planning method according to the first aspect of the present application.

[0027] In addition, the computer readable storage medium according to the third aspect of the present application has computer instructions stored thereon. When the computer instructions are executed by a processor, the oiling path planning method according to the first aspect of the present application is implemented. BRIEF DESCRIPTION OF DRAWINGS

[0028] The above features and advantages of the present application will be better understood by reading the detailed description of the embodiments of the present application in conjunction with the attached drawings, in which:

[0029] Figure 1 A flowchart of the oiling path planning method according to some embodiments of the present application is shown.

[0030] Figure 2 A process flow diagram of a crude oil tank farm according to some embodiments of the present application is shown. DETAILED DESCRIPTION

[0031] The present application is described in detail by specific embodiments, and other advantages and effects of the present application can be easily understood by those skilled in the art from the disclosure. Although the description of the present application will be introduced in conjunction with preferred embodiments, it does not mean that the features of the present application are limited to the embodiments. On the contrary, the purpose of introducing the present application in conjunction with the embodiments is to cover other options or modifications that can be extended based on the claims of the present application. In order to provide a better understanding of the present application, many specific details will be included in the following description. The present application can also be implemented without using these details. In addition, in order to avoid confusion or obscure the focus of the present application, some specific details will be omitted in the description.

[0032] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connecting" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be direct connection, or indirect connection through intermediate medium, or internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0033] It is to be understood that, although the terms "first", "second", "third", and the like can be used herein to describe various components, regions, layers and / or sections, these components, regions, layers and / or sections should not be limited by these terms. These terms are only used to distinguish one component, region, layer and / or section from another component, region, layer and / or section. Therefore, a first component, region, layer and / or section discussed below can be referred to as a second component, region, layer and / or section without departing from some embodiments of the present application.

[0034] As described above, pipeline transportation is one of the main transportation methods of tank farm oil products, and its scheduling process mainly includes storage transfer, pipeline transportation, oil product blending and other links. The oil delivery path planning is an important part of the oil scheduling task, which is related to the feasibility, convenience and economy of the oil scheduling task. However, the oil delivery path planning currently mainly relies on the personal rich professional knowledge and experience of the scheduling personnel to manually formulate, which has problems such as relying on personnel knowledge and experience, high operation difficulty, long planning period, and difficult to improve the optimization degree of oil delivery path planning.

[0035] In order to overcome the above-mentioned defects existing in the prior art, the present application provides an oil delivery path planning method, an oil delivery path planning device, and a computer readable storage medium, which can overcome the dependence of oil delivery path planning on personnel knowledge and experience, reduce the operation difficulty, shorten the planning time, and improve the optimization degree of oil delivery path planning.

[0036] In some non-limiting embodiments, the above-mentioned oil delivery path planning method provided by the first aspect of the present application can be implemented by the above-mentioned oil delivery path planning device provided by the second aspect of the present application. Specifically, the planning device is configured with a memory and a processor. The memory includes but is not limited to the above-mentioned computer readable storage medium provided by the third aspect of the present application, and the computer instructions are stored on the memory. The processor is connected to the memory and is configured to execute the computer instructions stored on the memory to implement the above-mentioned oil delivery path planning method provided by the first aspect of the present application.

[0037] Please refer to Figure 1 and Figure 2 . Figure 1 The flowchart of the oil delivery path planning method according to some embodiments of the present application is shown. Figure 2 The process flow diagram of the crude oil tank farm according to some embodiments of the present application is shown.

[0038] As Figure 1 shown, in some embodiments of the present application, the specific implementation steps of the oil scheduling-oriented oil delivery path planning method are as follows:

[0039] S1: refer to Figure 2The tank area process flowchart is shown. The intersection of the storage tank, oil pump, part of the valve, and different pipeline in the figure is taken as a node, and the pipeline between the nodes is taken as a link to establish a directed graph model. Here, most of the valves in the process flowchart do not affect the path search direction, and such valves can be regarded as the attribute of the pipeline to which they belong, which is saved in the pipeline parameter. However, for the valves 10165, 10174, 10175, and 10176 in Figure 2 The existence of the valves 10165, 10174, 10175, and 10176 will affect the path search direction, so the valves 10165, 10174, 10175, and 10176 can be modeled as nodes, and restrictions are imposed to prevent the path from changing between the two pump lines to form unnecessary detours during path search. In addition, the planning device can also save the entity equipment and pipelines corresponding to the nodes and links and their attribute and state data in the model, and save the model as a json file for the path search algorithm to call.

[0040] S2: Read the modeling data saved in the json file to build the adjacency table of the directed graph. Specifically, the adjacency table of each node in the above directed graph model needs to save the state of the node (for example: whether it is available), the attribute of the node (for example: the flow, head, and power of the pump), the node pointed to by the node, the link of the pointed node, and the attribute of the link (for example: valve, pipeline length).

[0041] S3: Read the oil scheduling task. Specifically, the scheduling task of the entire crude oil tank area can include several subtasks. Each subtask further includes task information such as the starting node and the terminating node of the oil delivery path, and the ideal flow requirement q of oil movement.

[0042] S4: Establish the objective function of the multi-objective problem according to the optimization goal of the path planning task, and set the first feasibility constraint of the subtask path search and the second feasibility constraint of the subtask path combination.

[0043] S5: Use the multi-objective evolutionary algorithm (MOEAD) based on decomposition to search for a path planning scheme, and output a non-dominated path planning scheme set after the algorithm completes the search. Here, each scheme in the output non-dominated path planning scheme set can meet the requirements and restrictions of oil movement, and the scheduling personnel can select the path planning scheme that best meets the oil movement from the non-dominated path planning scheme set according to experience, preference, and specific requirements of the oil scheduling task.

[0044] Further, in some embodiments, the objective function of the multi-objective problem and the constraint condition of the path planning scheme combination in the above step S4 are as follows:

[0045] S4.1: Objective function:

[0046]

[0047]

[0048]

[0049] wherein, F1 represents the total length of the path planning scheme, F2 represents the matching degree of the flow of the path planning scheme, and F3 represents the total power of the pump selected by the path planning scheme. N_Paths represents the number of subtasks in the path planning task, and the i-th subtask path r i comprises n i nodes, represents the distance between the j-th node v i and the j+1-th node v j . j+1 represents the length of the path r i . Q i represents the flow parameter of the oil pump in the i-th subtask path, q i represents the ideal flow of the i-th subtask, and Power i represents the power parameter of the oil pump in the i-th subtask path.

[0050] S4.2: Subtask path search constraint condition:

[0051] (1) Any node in the subtask path can only be passed once to prevent loops in the same path.

[0052] (2) The oil product cannot enter other storage tanks during movement from the starting storage tank, and the oil product can only be sent to the blending area by one oil pump, that is, one subtask path can only pass through one storage tank and one oil pump.

[0053] S4.3: Path planning scheme combination constraint condition:

[0054] As described above, the oil scheduling task includes several subtasks of oil movement. In order to prevent the mixing of oil products in the pipeline, the same nodes or pipelines cannot be passed between different subtask paths in the path planning scheme.

[0055] Further, in some embodiments, the process of searching for the path planning scheme in step S5 described above using the decomposition-based multi-objective evolutionary algorithm as the optimization algorithm is as follows:

[0056] S5.1: Initialize population:

[0057] Generate subtask paths: use depth-first search to search for subtask paths, and prune the paths by constraint conditions to ensure the feasibility of the oil movement path. ​

[0058] Construct initial population: search a path for each subtask by depth first search and save it in the path set R{i}. Then, select subtasks in turn and take a path R j from the path set R{i} of the selected subtask j The nodes passed by are {V1, V2, V3, …, V n}. In {V1, V2, V3, …, V n}, the adjacency list of nodes V a , V b , V c , … contains multiple nodes, and a node V x is randomly selected. The node pointed to by V j in the path R x is V y , and a node V x different from V y is randomly selected from the adjacency list of V z . A path from node V z to the end point is generated, replacing the path from node V j to the end point in the path R y , and a new path Here, if the path is not in the path set R{i}, then is added to the path set R{i}. Then, one path is randomly selected from each of the remaining path sets of subtasks and combined with the newly generated path If the combined path planning scheme satisfies the feasibility constraint, the path planning scheme is added to the initial population as an individual, and the construction of new path planning schemes is stopped when the number of individuals reaches the set population number.

[0059] Construct external population: construct an external population for saving non-dominated path planning schemes in each generation of population, calculate the objective function values of the path planning schemes in the initial population and normalize them. Then, sort the path planning schemes according to the function values, find the path planning schemes that are not dominated by other schemes in the population from the initial population, called non-dominated path planning schemes, and update the non-dominated path planning schemes in the external population.

[0060] S5.2: Construct uniform weight vector and weight nearest neighbor set: use the standard boundary crossing method to generate uniformly distributed weight vectors on the hyperplane, and the number of weight vectors N is:

[0061]

[0062] Where H represents the number of partitions for each dimension of the objective, controlling the number of weight vectors, and m is the number of objectives. Then, the T nearest neighbor weight vectors for each weight vector are calculated based on Euclidean distance, and the indices of these nearest neighbor weight vectors are stored in the nearest neighbor set B{i}.

[0063] S5.3: Weight Allocation: Based on the decomposition method of the MOEAD algorithm, the above multi-objective optimization problem is decomposed into N sub-problems, where each sub-problem corresponds to a path planning scheme, and each sub-problem is assigned a weight vector.

[0064] S5.4: Offspring Generation: Randomly select two indices k and l from the nearest neighbor set B{i} corresponding to the weight vector of the i-th subproblem, and extract the corresponding path planning scheme S from the parent population. k and S l Perform crossover and mutation operations. Since the path planning scheme of the present invention has many feasibility constraints and the generated offspring schemes have low feasibility probabilities, the crossover probability can be set to 1 and the mutation probability can be set to 0.5 to improve the feasibility probability of the generated offspring schemes.

[0065] In some embodiments, the crossover and mutation operations employed in this invention can be implemented simultaneously based on two different crossover operations and one mutation operation:

[0066] (1) Randomly select the j-th subtask and extract the path planning scheme S respectively. k and S l Path of the j-th subtask and Find the nodes that are common to both paths, and randomly select a node V from them. Swap the paths. and The path following node V. Place the two intersecting paths back into the path planning scheme S. k and S l At this point, if the new path planning scheme meets the constraints, it is saved as... and Otherwise, the original path planning scheme S will be changed. k and S l Save as and And return the path planning solution. and

[0067] (2) Randomly select the h-th subtask and calculate the path planning scheme. and The path swapping begins from the h-th subtask. If the generated path plan satisfies the constraints, it is saved as... and Otherwise, save the original path planning scheme and as and and return the path planning scheme and

[0068] After that, the planner randomly selects the nth subtask path from the path planning scheme or or The planner can update the path or in the manner of randomly modifying the path search direction in the population initialization process, and put the updated path back into the path planning scheme or At this time, if the generated path planning scheme satisfies the constraint condition, it is saved as and and the mutation operation is completed. Otherwise, the original path planning scheme and is saved as and and the path planning scheme and is returned to complete the mutation operation.

[0069] After that, the planner can randomly select one of the path planning schemes and as the child path planning scheme S child .

[0070] S5.5: Population update: Since the above multi-objective optimization problem is a minimization problem, the planner can set the ideal point as Z = (0, 0, 0), normalize the objective function values of the sub-problem path planning scheme S i and the generated child path planning scheme S child , and calculate the Chebyshev distance between the normalized objective function values and the ideal point. The formula for calculating the Chebyshev distance is as follows:

[0071]

[0072] where λ i is the weight vector of the ith sub-problem, and f j is the jth objective function value. Here, if the Chebyshev distance value of the path planning scheme S child and the ideal point Z is small, the planner can replace S child with S i ​As a sub-problem, the planning device can reserve the original path planning scheme S for the sub-problem i .

[0073] Further, for the update of the external population, the planning device can use the path planning scheme S child to perform a non-dominated comparison with the path planning schemes of the external population, and remove the path planning schemes in the external population dominated by the path planning scheme S child . At this time, if the path planning scheme S child is not dominated by any path planning scheme in the external population, the offspring path planning scheme is added to the external population.

[0074] S5.6: Determine whether the algorithm is ended: if the current generation number reaches the maximum generation number, stop the search, and the planning device outputs the external population. Otherwise, the planning device can return to step S5.4 to continue searching for path planning schemes.

[0075] Further, in some embodiments, the scheduling task of the crude oil tank area includes the following two sub-tasks:

[0076] Task 1: Tank T11 -> crude oil blending area-C, ideal flow rate: 300 m 3 / h;

[0077] Task 2: Tank T15 -> crude oil blending area-A, ideal flow rate: 500 m 3 / h.

[0078] At this time, in order to simplify the calculation of path length, the planning device can assume that the path length between adjacent nodes in the path is 1, express the square sum of the difference between the pump flow rate of the sub-task path and the ideal flow rate as the square sum of the flow rate difference, and express the total power as the sum of the pump power of the sub-task path. Here, the path planning scheme [T11, N12, N20, N23, N28, N29, N40, N46, P102, N43, N56, N62, N75, crude oil blending area-C] represents the sub-task path: T11 -> N12 -> N20 -> N23 -> N28 -> N29 -> N40 -> N46 -> P102 -> N43 -> N56 -> N62 -> N75 -> crude oil blending area-C.

[0079] Afterwards, the planning device can set the initial population number to 105, the maximum number of generations to 1000, and construct an external population to save the non-dominated path planning schemes of each generation. Further afterwards, the planning device can generate 105 weight vectors using the standard boundary crossover method, and correspondingly generate one path planning scheme for the initial population with each weight vector. Further afterwards, the planning device can calculate the 5 nearest neighbor weight vectors of each vector according to the Euclidean distance, and store the index of the nearest neighbor weight vector into the weight nearest neighbor set B{i} to construct the initial population as shown in Table 1:

[0080] Table 1 Path planning schemes of initial population

[0081]

[0082]

[0083]

[0084] Further afterwards, the planning device can calculate the function values of the path planning schemes in the initial population, compare the dominance relationship between the path planning schemes, and update the non-dominated path planning schemes to the external population to construct the external population as shown in Table 2:

[0085] Table 2 Path planning schemes of external population

[0086]

[0087]

[0088] Afterwards, the planning device can decompose the path planning problem into 105 sub-problems according to the decomposition strategy of the MOEAD algorithm to correspond to 105 weight vectors. Further, the planning device can randomly take out two indexes from the weight nearest neighbor set of each sub-problem in turn, take the path planning scheme corresponding to the index pair as the parent, generate two new path planning schemes through crossover and mutation with a crossover probability of 1 and a mutation probability of 0.5, and randomly select one path planning scheme as the child. Further, the planning device can update the external population using the child, compare the dominance relationship between the child and the path planning schemes in the external population, and eliminate the path planning schemes dominated by the child from the external population. Here, if the child is not dominated by any path planning scheme in the external population, the planning device can add the child to the external population. Further, the planning device can set the ideal point as Z = (0, 0, 0), update the population, calculate the function value of the path planning scheme corresponding to the sub-problem and the function value of the child path planning scheme, and calculate the Chebyshev distance between the function value of the two path planning schemes and the ideal point, respectively. Here, if the child path planning scheme is better than the path planning scheme corresponding to the sub-problem, the planning device can update the child to the path planning scheme corresponding to the sub-problem. Otherwise, the planning device can keep the original path planning scheme. The updated population and external population are shown in Table 3 and Table 4:

[0089] Table 3 Path planning schemes of the updated population

[0090]

[0091]

[0092] Table 4 Path planning schemes of the updated external population

[0093]

[0094]

[0095] Afterwards, the planning device can repeat the above population updating steps to evolve for 1000 generations, and output the path planning schemes in the external population as the result after the search is completed. The search result of the path planning is shown in Table 5:

[0096] Table 5 Path planning schemes of the output result

[0097]

[0098]

[0099] According to the above description, the path planning method, the path planning device, and the computer readable storage medium provided by the present application can achieve the following beneficial effects:

[0100] (1) By modeling the crude oil tank farm process flow diagram as a directed graph and saving the attributes and status of pipelines and process equipment in the model, this invention can avoid the situation where it is difficult to take into account the current status of pipelines and process equipment when manually planning the path.

[0101] (2) By using depth-first search to search for subtask paths and adding constraints to the subtask paths and path planning schemes, the present invention can ensure the feasibility of the path planning schemes.

[0102] (3) By constructing an objective function for a multi-objective optimization problem, this invention can simultaneously consider path length, pump flow matching, and oil product movement costs while planning paths, and search for a set of non-dominated path planning schemes for dispatchers to choose from. Furthermore, this invention can quickly plan feasible paths for dispatching tasks and can search for the path scheme closest to the Pareto front from a large number of feasible path schemes, thereby improving path planning efficiency while reducing the cost of dispatching schemes.

[0103] Although the methods described above are illustrated and depicted as a series of actions for the sake of simplicity, it should be understood and appreciated that these methods are not limited by the order of the actions, as some actions may occur in a different order and / or concurrently with other actions from the illustrations and descriptions herein or not illustrated and described herein but which may be understood by those skilled in the art, according to one or more embodiments.

[0104] The prior description of this disclosure is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to this disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not intended to be limited to the examples and designs described herein, but should be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An oil application path planning method characterized by comprising: The method comprises the following steps: According to the process flow diagram of the crude oil tank farm, a directed graph model is established, and attribute data and state data of at least one process equipment and / or at least one pipeline in the crude oil tank farm are saved to corresponding nodes and / or corresponding links of the directed graph model, wherein the attribute data comprises flow and power of a pump, the state data comprises available state of a pipeline, the nodes comprise a storage tank node, an oil pump node and a pipeline intersection, and the links comprise pipelines between the nodes; An adjacency list is established according to modeling data of the directed graph model, so as to store attribute data, state data, index information of a pointing node and link information connecting the pointing node of each node; Starting node information and ending node information of at least one subtask in an oil product scheduling task and ideal pump flow data of oil product movement are obtained; According to a plurality of optimization objectives of the oil product scheduling task, the directed graph model and the adjacency list, a target function of a multi-objective optimization problem is established, and a first feasibility constraint condition of path searching and a second feasibility constraint condition of path combination are set, wherein the plurality of optimization objectives comprise reducing total path length, improving matching degree of pump flow and reducing total power of a pump; and Multi-objective evolution operation is performed based on the first feasibility constraint condition and the second feasibility constraint condition, so as to determine at least one non-dominated path planning scheme.

2. The planning method of claim 1, wherein, The step of establishing the target function of the multi-objective optimization problem according to the plurality of optimization objectives of the oil product scheduling task, the directed graph model and the adjacency list comprises: The target function of the multi-objective optimization problem is determined as: Wherein, F1 represents total path length of a path planning scheme, F2 represents matching degree of pump flow in the path planning scheme, and F3 represents total power of a pump in the path planning scheme.

3. The planning method of claim 1, wherein, The process equipment comprises a storage tank and an oil pump, the first feasibility constraint condition of path searching comprises: Each node in a path of each subtask is passed only once; and Only one storage tank node and one oil pump node are included in the path of each subtask.

4. The planning method of claim 1, wherein, The second feasibility constraint condition of path combination comprises: Paths of each subtask do not pass through the same node and / or the same pipeline.

5. The planning method of claim 1, wherein, The step of performing multi-objective evolution operation based on the first feasibility constraint condition and the second feasibility constraint condition to determine at least one non-dominated path planning scheme comprises: An initial population and an external population are constructed and initialized; According to the number of optimization objectives, a plurality of weight vectors and a weight nearest neighbor set thereof are constructed, and each subproblem of the multi-objective optimization problem is allocated a weight vector; Two indexes are randomly selected from the weight nearest neighbor set corresponding to the subproblem, and two parent path planning schemes are crossed and mutated to generate a child path planning scheme of the subproblem; It is judged whether the child path planning scheme is better than a current path planning scheme of the subproblem; and If the child path planning scheme is better than the current path planning scheme of the subproblem, the child path planning scheme is added to the external population. In response to a result of judging that the offspring path planning scheme is better than the current path planning scheme, the current path planning scheme is replaced by the offspring path planning scheme in the initial population, and the non-dominated offspring path planning scheme is used to update the external population.

6. The planning method of claim 5, wherein, The step of constructing and initializing the initial population and the external population comprises: According to the first feasibility constraint condition, a search direction of a current node is randomly set from a starting node of each subtask to generate a plurality of random paths connecting corresponding termination nodes; According to the second feasibility constraint condition, the random paths of each subtask are randomly combined, and a path planning scheme obtained by the combination is added to the initial population as a random individual; and An external population is constructed, and non-dominated schemes in the initial population are added to the external population.

7. The planning method of claim 5, wherein, The step of constructing a plurality of weight vectors and weight nearest neighbor sets according to the number of optimization objectives comprises: N weight vectors are generated on the hyperplane using a standard boundary intersection method, wherein, H represents the number of partitions of each dimension of the target, and m is the number of targets. At least two nearest neighbor weight vectors of each weight vector are calculated, and indexes of the at least two nearest neighbor weight vectors are stored in a corresponding weight nearest neighbor set B{i}.

8. The planning method of claim 7, wherein, The step of assigning each subproblem of the multi-objective optimization problem with a weight vector comprises: The multi-objective optimization problem is decomposed into N subproblems by using the MOEAD algorithm, and each subproblem is assigned with a weight vector.

9. The planning method of claim 5, wherein, The step of randomly selecting two indexes from the weight nearest neighbor set corresponding to the subproblem, and performing crossover and mutation on the corresponding two parent path planning schemes to generate an offspring path planning scheme of the subproblem comprises: Randomly select index k and index I from the weight nearest neighbor set of the ith sub-problem to take out the corresponding parent path planning scheme S from the parent population k and the parent path planning scheme S l ; and to the parent path planning scheme S k and the path of each subtask of the parent path planning scheme S l is respectively crossed and mutated, and / or the paths of the parent path planning scheme S k and the parent path planning scheme S l are crossed to generate the offspring path planning scheme of the subproblem.

10. The planning method of claim 5, wherein, The step of judging whether the offspring path planning scheme is better than the current path planning scheme comprises: An ideal point Z=(0, 0, 0) is set; Chebyshev distances of the ideal point Z to the offspring path planning scheme and the current path planning scheme are respectively calculated; and In response to a result of judging that the Chebyshev distance of the ideal point Z to the offspring path planning scheme is less than the Chebyshev distance of the ideal point Z to the current path planning scheme, it is determined that the offspring path planning scheme is better than the current path planning scheme.

11. The planning method of claim 5, wherein, Further comprising the following steps: In response to a result of judging that the offspring path planning scheme is not better than the current path planning scheme, the current path planning scheme is retained by the subproblem.

12. The planning method of claim 5, wherein, Further comprising the following steps: It is judged whether a current evolution generation number reaches a preset maximum generation number; In response to a result of judging that the current generation number does not reach the maximum generation number, parent path planning schemes are selected again for each subproblem of the multi-objective optimization problem to perform an iteration of next round of crossover and mutation; And In response to a result of judging that the current generation number reaches the maximum generation number, the external population is output.

13. An oil application path planning device characterized by comprising: It comprises: a memory; And a processor connected to the memory and configured to implement the method for planning an oil delivery path according to any one of claims 1-12.

14. A computer readable storage medium having stored thereon computer instructions, wherein, The computer instructions are executed by the processor to implement the method for planning an oil delivery path according to any one of claims 1-12. The computer instructions are executed by the processor to implement the method for planning an oil delivery path according to any one of claims 1-12.

Images