CN-115879195-B - Efficient parallel optimization method and device considering practical engineering constraint of horizontal well

CN115879195BCN 115879195 BCN115879195 BCN 115879195BCN-115879195-B

Abstract

The invention relates to a high-efficiency parallel optimization method and device considering horizontal well practical engineering constraint, wherein the method comprises the following steps of S1 constructing a water displacement reservoir production optimization mathematical control model, S2 constructing a horizontal well constraint judgment method based on horizontal well engineering application conditions and an oil reservoir numerical model, S3 constructing a database based on a population in a Latin hypercube sampling initialization algorithm by utilizing the constraint judgment method of the horizontal well, S4 constructing a temporary population by utilizing excellent individuals in the database and constructing a Gaussian model based on the temporary population, S5 obtaining three sub-populations by utilizing different strategies based on a differential evolution algorithm and carrying out constraint treatment by utilizing the constraint judgment method of the horizontal well, S6 obtaining potential individuals of the sub-populations based on a pseudo-update strategy of the temporary population, S7 selecting an individual nearest to the current optimal individual based on Euclidean distance and carrying out parallel numerical simulation, and S8 updating the database.

Inventors

LIU FAN
PAN YUE
PAN CAIXIA
WANG KAI
TANG CHENYANG
SHEN ZIYANG
CAO CHENMING
Dai Qinyang
ZHOU WENSHENG
ZHENG WEI
YANG RENFENG
LI KE
SUN YIDAN
ZHANG KAI
LI JING
LIU CHEN

Assignees

中海石油(中国)有限公司
中海石油（中国）有限公司北京研究中心

Dates

Publication Date: 20260512
Application Date: 20221124

Claims (8)

1. The efficient parallel optimization method considering the actual engineering constraint of the horizontal well is characterized by comprising the following steps of: s1, constructing a water-flooding reservoir production optimization mathematical control model; s2, constructing a constraint judgment method of the horizontal well based on the horizontal well engineering application conditions and the oil reservoir numerical model; s3, carrying out constraint processing by using a constraint judgment method of a horizontal well based on the population in the Latin hypercube sampling initialization algorithm, calculating NPV in a water-drive reservoir production optimization mathematical model, and constructing a database; s4, forming a temporary population by using excellent individuals in the database, and constructing a Gaussian model based on the temporary population; S5, three sub-populations are obtained by using different strategies based on a differential evolution algorithm, and constraint processing is carried out by using a constraint judgment method of a horizontal well; s6, acquiring potential individuals of the sub-population based on a pseudo-updating strategy of the temporary population; S7, based on Euclidean distance, selecting an individual nearest to the current optimal individual, and performing parallel numerical simulation; s8, updating a database, outputting an optimal individual when the iteration stop condition is met, otherwise, returning to the step S4; the step S1 specifically comprises the following steps: According to the oil reservoir development requirement, with the aim of maximizing the economic net present value NPV, a water-drive reservoir production optimization mathematical control model is constructed, as follows: In the formula, Control variables for production optimization problems; Representing the objective function value, namely economic benefit NPV; A total time step for numerical simulation; Simulating the span size of the step length for the nth time value, wherein d is the number of control variables; is the annual attenuation rate; , And Representing the oil production speed and water production speed of the ith production well in the nth time step and the water injection speed of the ith water injection well in the nth time step respectively; , , The economic benefit of crude oil per unit volume, the cost required by wastewater treatment per unit volume and the injection cost of water per unit volume are respectively shown, and NP and NI are respectively the total number of production wells and water injection wells of the oil reservoir; the step S2 specifically comprises the following steps: S21, extracting the horizontal segment track characteristics of each horizontal well in the oil reservoir numerical model, approximating the horizontal segment track of the horizontal well to be a line segment from the plane, and aiming at any horizontal well Extracting grid coordinates corresponding to two end points of a horizontal segment in an oil reservoir numerical model And (3) with As its horizontal segment track mark ; S22 constructs a horizontal segment track identification set, , wherein, To determine the number of newly added horizontal wells that are to optimize the well location, The number of horizontal wells in the reservoir model is given; s23 identifying set for horizontal segment track Any two newly added horizontal wells or any one newly added horizontal well and the existing horizontal well And Using its corresponding horizontal segment track identification And (3) with Pretreatment before restraint, for horizontal well And With the result of pretreatment, exactly two diagonal lines are respectively horizontal wells And A rectangle connecting the two end points of the horizontal section; s24 pairs of two horizontal wells And Judging whether the formed rectangles overlap or not to determine whether constraint conditions are met or not; The step S3 specifically comprises the following steps: s31, generating by using Latin hypercube sampling method Well position scheme of horizontal well ; S32, carrying out well position constraint judgment of the horizontal well, wherein the judgment method is the same as that of the step S2; S33, if the individuals which do not meet the constraint exist, latin hypercube sampling is carried out on the individuals again until all the individuals meet the constraint; s34, carrying out parallel calculation on all individuals by using a numerical simulator to obtain NPV response values in the water-drive reservoir production optimization mathematical control model The initial database DB is constructed with S and y.
2. The efficient parallel optimization method according to claim 1, wherein step S4 specifically includes: s41 selecting the highest NPV response value in the database DB Construction of temporary populations of individuals ; S42 utilization Construction of Gaussian proxy model 。
3. The efficient parallel optimization method according to claim 2, wherein step S5 specifically includes: S51, generating three sub-populations based on three different strategies of a differential evolution algorithm ; S52 for three sub-populations Constraint processing is performed in the same manner as in steps S31 to S33.
4. The efficient parallel optimization method according to claim 3, wherein step S6 specifically includes: s61 Using Gaussian model Calculating target response values in each sub-population, and calculating fitness values based on the confidence lower bound LCB criterion; s62, for each sub-population, selecting an individual with the optimal response value Deposit in And update it to But not update ; S63 as The individuals stored in the medium reach If yes, proceeding to the next step, otherwise using the updated data Returning to step S5.
5. The efficient parallel optimization method according to claim 4, wherein step S7 specifically includes: S71 calculation Individuals of the Chinese to the present optimal individuals Is selected to be the m individuals with the smallest distance Parallel numerical simulation calculation is carried out to obtain corresponding NPV response values ; The number of iterations S72 increases by m.
6. An efficient parallel optimization device considering actual engineering constraints of a horizontal well, comprising: The first processing unit is used for constructing a water-drive reservoir production optimization mathematical control model and specifically comprises the following components: According to the oil reservoir development requirement, with the aim of maximizing the economic net present value NPV, a water-drive reservoir production optimization mathematical control model is constructed, as follows: In the formula, Control variables for production optimization problems; Representing the objective function value, namely economic benefit NPV; A total time step for numerical simulation; Simulating the span size of the step length for the nth time value, wherein d is the number of control variables; is the annual attenuation rate; , And Representing the oil production speed and water production speed of the ith production well in the nth time step and the water injection speed of the ith water injection well in the nth time step respectively; , , The economic benefit of crude oil per unit volume, the cost required by wastewater treatment per unit volume and the injection cost of water per unit volume are respectively shown, and NP and NI are respectively the total number of production wells and water injection wells of the oil reservoir; The second processing unit is used for constructing a constraint judgment method of the horizontal well based on the horizontal well engineering application conditions and the oil reservoir numerical model, and specifically comprises the following steps: S21, extracting the horizontal segment track characteristics of each horizontal well in the oil reservoir numerical model, approximating the horizontal segment track of the horizontal well to be a line segment from the plane, and aiming at any horizontal well Extracting grid coordinates corresponding to two end points of a horizontal segment in an oil reservoir numerical model And (3) with As its horizontal segment track mark ; S22 constructs a horizontal segment track identification set, , wherein, To determine the number of newly added horizontal wells that are to optimize the well location, The number of horizontal wells in the reservoir model is given; s23 identifying set for horizontal segment track Any two newly added horizontal wells or any one newly added horizontal well and the existing horizontal well And Using its corresponding horizontal segment track identification And (3) with Pretreatment before restraint, for horizontal well And With the result of pretreatment, exactly two diagonal lines are respectively horizontal wells And A rectangle connecting the two end points of the horizontal section; s24 pairs of two horizontal wells And Judging whether the formed rectangles overlap or not to determine whether constraint conditions are met or not; the third processing unit is used for carrying out constraint processing by utilizing a constraint judgment method of a horizontal well based on the population in the Latin hypercube sampling initialization algorithm and calculating NPV in a water reservoir production optimization mathematical model to construct a database, and specifically comprises the following steps: s31, generating by using Latin hypercube sampling method Well position scheme of horizontal well ; S32, carrying out well position constraint judgment of the horizontal well, wherein the judgment method is the same as that of the step S2; S33, if the individuals which do not meet the constraint exist, latin hypercube sampling is carried out on the individuals again until all the individuals meet the constraint; s34, carrying out parallel calculation on all individuals by using a numerical simulator to obtain NPV response values in the water-drive reservoir production optimization mathematical control model Constructing an initial database DB by S and y; the fourth processing unit is used for forming a temporary population by using excellent individuals in the database and constructing a Gaussian model based on the temporary population; The fifth processing unit is used for obtaining three sub-populations by using different strategies based on a differential evolution algorithm and performing constraint processing by using a constraint judgment method of the horizontal well; the sixth processing unit is used for acquiring potential individuals of the sub-population based on the pseudo-updating strategy of the temporary population; The seventh processing unit is used for selecting an individual closest to the current optimal individual based on the Euclidean distance and performing parallel numerical simulation; and the eighth processing unit is used for updating the database, outputting the optimal individual when the iteration stop condition is reached, and returning to the step S4 otherwise.
7. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the efficient parallel optimization method taking into account the actual engineering constraints of a horizontal well according to any one of claims 1-5.
8. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the efficient parallel optimization method according to any one of claims 1-5 taking into account the actual engineering constraints of a horizontal well when executing the computer program.

Description

Efficient parallel optimization method and device considering practical engineering constraint of horizontal well Technical Field The invention relates to a high-efficiency parallel optimization method and device considering the practical engineering constraint of a horizontal well, and belongs to the technical field of oil and gas field development. Background The horizontal well is an important means for developing complex oil reservoirs, compared with a vertical well, the horizontal well has a horizontal section extending in the reservoir and has a larger contact area with oil storage rocks, so that the petroleum yield and efficiency of the oil field are effectively improved, and the economic benefit is greatly improved. But the development cost of the horizontal well is very expensive and the risk is high. If the horizontal well is not properly deployed, a significant economic penalty may result. Thus, the deployment of horizontal wells is one of the important decisions in oilfield development. First, the well spacing must be close enough to help create as much of the stimulated reservoir volume as possible, however, the well spacing must also be wide enough to reduce fracture interference (inter-well interference and "frac impact") and excessive investment in oilfield development. In addition, the horizontal well horizontal section should extend along a better and thicker reservoir, perpendicular to the direction of maximum principal stress, so as to traverse more vertical fracture zones of fracturing, and expand the communication range of the reservoir, so as to obtain higher productivity. Since the 90 s of the 20 th century, sequential scholars have proposed a horizontal well position optimization method, and along with the development of numerical calculation technology and modern optimization algorithm, the optimization effect and efficiency are improved to a certain extent, and the horizontal well position optimization method is further developed. The differential evolution algorithm is proposed by Rainer Storn and KENNETH PRICE in 1997 on the basis of evolutionary ideas such as genetic algorithm, and is essentially a multi-objective (continuous variable) optimization algorithm (MOEAs) for solving the overall optimal solution in a multidimensional space. Compared with a genetic algorithm, the differential evolution algorithm has the advantages that an initial population is randomly generated at the same point, the fitness value of each individual in the population is used as a selection standard, and the main process also comprises three steps of mutation, crossover and selection. The difference is that the genetic algorithm controls the crossover of the father and the probability value of the filial generation selected after mutation according to the fitness value, and the probability of the individual with large fitness value being selected in the maximization problem is correspondingly larger. The variation vector of the differential evolution algorithm is generated by a parent differential vector, and is intersected with a parent individual vector to generate a new individual vector, and the new individual vector is directly selected with the parent individual. It is apparent that the differential evolution algorithm has a more remarkable approximation effect than the genetic algorithm. The well position optimization of the horizontal wells is required to be constrained because the horizontal wells cannot be intersected and the well spacing cannot be too close, so that the well position optimization meets the practical application of an oil field, namely the well position optimization of the horizontal wells is a constraint optimization problem. The differential evolution algorithm is an intelligent global optimization method, has very low requirements on the properties of the function, often only requires the objective function value to be calculated, does not require continuity, microminiaturization and other analytic properties, and is an algorithm based on group evolution, so that the constraint optimization problem can be solved by adopting the differential evolution algorithm. The key to solving the constraint optimization problem with the differential evolution algorithm is how to perform effective constraint processing, namely how to effectively balance the search in the feasible region and the infeasible region. In addition, the calculation resources required by numerical simulation of the actual model of the oil field are very expensive, a great amount of time is consumed in optimizing, and the calculation burden caused by numerical simulation can be effectively reduced by parallel calculation. Therefore, there is a need to provide a horizontal well constraint expensive optimization method for parallel algorithms. Disclosure of Invention Aiming at the technical problems, the invention provides the efficient parallel optimization method and the device considering the actual engineerin