| With the explosive growth of energy demand and the constant increase of oil price, the profitability and financial benefit of oil enterprises has reached the highest level in history in the recent years. However, the sustainable development of oil production has been neglected. On one hand, the phenomenon of "undermining" exploitation appears because some domestic oil production enterprises (oilfield companies) only pursue the output of crude oil. It is not only a waste of oil resources, but also brings negative effects to the sustainable production of oil, even threatens the survival and development of oil production enterprises. On the other hand, some oil production enterprises can't deeply realize the characteristics of oil production such as big initial investment, long period of effective time, high investment risks and so on;a great number of oil-production projects have been blindly proposed at the time that the influence of high oil price which has rocketed to the top in the last two years. There is a serious problem what difficulties the oil companies and their oil-production projects will be faced with if the oil price falls to the lever of two years before in the future 5-7 years.Though the reconstruction of it has been completed and it is basically in harmony with foreign companies, China's oil enterprises, as large state-owned enterprises, can't consider all the influencing factors of oil price in an all round way for the long-term impact of monopolistic policies of China's petroleum industry when they carry out the development strategies or plans of oilfield. The main purpose of this paper is to built and perfect the theory of oilfield development programming by researching the optimal decision of china's oil production enterprises to develop the"Decision Support System"of oilfield development programming, to provide the optimal decision of china's oil production enterprises with necessary theory and technology support, to study the simulation and forecasting of oil price from the perspectives of system theory, information theory and control theory, to investigate the optimal decision of oil enterprises with the variable oil price by the realization of the functional simulation and stochastic simulation of oil price system by computer, to establish various kinds of optimization models of oil production enterprises with deterministic oil price and stochastic oil price, and to obtain the optimal development decision of oil production enterprises with variable oil price through the solution of these models by means of computer.The Oilfield development programming can embody the strategies of oil enterprises (oilfield companies), while to build and to solve various optimization models of it is an indispensable mean to obtain the optimal oilfield development strategies. The oilfield development programming includes dynamic analysis, recovery calibration, production allocation, measures, cost, workload input and so on. The relational structure chart of the research contents of this paper: At present, the leading oilfield (oilfield company) considers much in dynamic analysis, `a little in the optimization models, but little in oil price when it establishes the annual or five-year programming of oilfield development. This paper mainly focuses on the research of various optimization models of Oilfield Company such as production allocation, measures, cost, workload input considering the deterministic or stochastic oil price. Obtain the optimal development programming or strategy of oilfield by establishing and solving various optimization models including a year or many years, single target or multi-target, single level or multi-level optimization models.According to the existing researches, oilfield optimal development programming or strategy is closely linked with oil price. Therefore, in order to get the optimal programming or strategy, at first it is necessary to investigate the dynamic simulation and forecasting of oil price, then to establish various optimization models of oilfield development according to the forecasting results, at last to solve these optimization models.This paper lets oil price be the system output and all influencing factors of oil price be the system input in the input-output perspective. It achieves the isomorphic function based on historical input- output information, and establishes an input-output associated relationship of the system by the methods of differential simulation and neural network. This input-output associated relationship can not only be used for the oil price forecasting, but also can provide a necessary theoretical foundation for the studying of controlling and driving oil price. In addition, this paper designs a stochastic simulator by Monte Carlo method which can stochastic forecast the oil price, finally obtains the risk decision, oilfield development programming with stochastic oil price and the optimal strategy of oil enterprises.There are six chapters in this paper.In Chapter 1, there is a general introduction of this paper. At first the importance, necessity, purpose and meaning of this paper is set forth. Then the processes and current situation of this research are reviewed, and the methods, principles and techno-way used in this paper are drawn according to the current research situation. Finally, the innovation points of this paper are enumerated.In Chapter 2, the oil price and its influence factors are analyzed. The influence factors of oil price are divided into there kinds, i.e., supply, demand and non-supply-demand factors based on the thorough analysis of the formation and development of oil price at home and abroad. Supply-demand factors mainly refers to the international politics, military, economic integrated situation, oil futures market and so on. There, supply and demand will be affected by many factors such as proven oil reserves, technological progress factor, oil reserve(stocks), GDP, the price of alternative energy sources and so on, while oil price will affect its influencing factors too, thus the relationship between oil price and its influencing factors is intricate and complex. Basing on the complex relationship between oil price and its influencing factors that using traditional analysis methods to study oil price is limited in many aspects, and it is difficult to accurately forecast the change of oil price in the future. That is just the reason why the researches in Chapters 3 are developed. For the preparations of oil price prediction, the chapter also studies the quantitative methods of oil price influencing factors;In Chapter 3, the functional simulation and forecasting of oil price system is investigated. Based on the input- output theory ,it lets all influencing factors (including all supply, demand and oil price influencing factors by supply and demand ) of oil price be the inputs (i.e., multiple-input) of oil price system and oil price be the output (i.e., single output) of oil price system. This chapter studies on oil price and its influencing factors from the perspectives of system theory, information theory and control theory. At first, a differential simulation model and a neural network model of oil price system are established. Based on the combination and improvement of the two methods it designs functional simulation system of the oil price system with time-varying systems, establishes an input-output associated relationship of the system, obtains oil price forecasting depending on simulation model of the system. Because of many influencing factors of oil price with great stochastic characteristics, the paper makes these as the stochastic processes. They are random variables at any point of time while the time unit is a year in this paper, and oil price (i.e., the function of random variables) is a random variable too. This chapter researches oil price from perspective of the function of random variables, obtains the probability distribution of oil price at any point of time in future and its software development by the stochastic simulation with Monte Carlo method according to the probability distribution characteristics of the influencing factors of oil price, and gains the forecasting oil price at that point of time by forecasting distributional characteristics(such as typical value, variance)of oil price in future. Based on deterministic or stochastic simulation, a detail design of software for oil price simulation system is given, and the simulation and forecasting of oil price system is realized by the VB.NET development tool. Finally, some operation examples of software system are given which indicate the forecasting oil price obtained by our method is much more accurate than that obtained by other methods on precision of history matching.In Chapter 4, various optimization decision models of oil enterprises are established. The ultimate purpose of this paper is not to simulate the oil price system or to forecast oil price, but to obtain the optimal programming (strategy) especially the optimal production of oil enterprises with forecasting oil price got by simulation. Therefore, this chapter investigates the production and operating activities of oil enterprises at first. Then it selective analyzes the output and its influencing factors, benefit output and non-benefit output, component of cost and so on. Finally, according to these analyses above, the important distinction between oil and other commonly productions in output, cost, benefit is shown. For commonly productions, the benefit increases with the increment of production as long as there has adequate market demand where the unit will not increase for the batch production. However, for oil enterprises, the unit cost will increase several times with the increment of production when the production of oilfield reaches a certain limit level even if there has adequate market demand, and this increased production will turn into non-benefit production. Therefore, various optimal decision models of oil enterprises under the deterministic oil price are established.Considering structure, they can be divided into single-layer optimization models and multiplayer optimization models from structure. Considering objective, they can be divided into"single-objective optimization models and multi-objective optimization models"from objective. Considering time, they can be divided into"yearly optimization models and multi-year optimization models". Considering decision rule, they can be divided into"maximal-output optimization models","lowest-cost optimization models"and"best-benefit optimization models".In Chapter 5, it studies algorithms of various optimization decision models of oil enterprises .At first; some algorithms and their principles for the optimization models of oil enterprises which are proposed in Chapter 4 are set forth, including SUMT method and the evaluation function method of multi-objective optimization. Then several kinds of genetic algorithms modified genetic algorithms and POWLL genetic annealing algorithms which can be applied to these especial optimization models are designed, and their software developments are realized by the VB.NET development tool. Finally, based on the actual data of domestic H-B oil field, the corresponding optimal decisions of several primary optimization models of oil enterprises with the deterministic oil price (the oil price can be altered as parameter) are obtained by our software, including the optimal construction of subentry production , the optimal construction of the corresponding cost of subentry production and the optimal construction of corresponding workload of subentry production, (which are also the most important part of the medium and long periods of development of oil field).In Chapter 6, the conclusion and further research: At first, the main conclusion is set forth and some suggestion of further direction of this research is given.The innovation of this paper is as follows: 1. Investigate the optimal development decision of china's oil production enterprises considering sufficiently the influencing factors of oil price, built and perfect the theory of oilfield development programming, develop the"Decision Support System"of oilfield development programming, and provide the optimal decision of china's oil production enterprises with necessary theory and technology support.2. The functional simulation and forecasting (with control) of oil price system. Based on the input-output thought of system and the quantitative description of the influencing factors of oil prices and variable step trained of BP network be introduced to differential simulation of parameter identification, it designs prices functional simulation system based on time-varying system. The simulation system not only has retained the characteristics of neural network's high precision fitting to the history, but also has the idea of differential simulation in the process of forecasting, fully thinking trending characteristics of output variables. This input-output associated relationship can not only realize the forecasting of oil price, but also can lay foundation for the control of oil price (or the driving of oil price). Even more, it opened up a new way for the research of the similar problems in ecnomics. Software called"oil price forecasting software system"is designed and developed in the VB.NET environment based on modularization design idea and embedded principles. A stochastic simulator of oil price is designed to simulate and forecast oil price randomly based on the further analysis of the stochastic characteristics of the influencing factors of oil price and the simulation of distribution and characteristics of oil price. 3. The establishment of various optimization models of oil production enterprises with deterministic and stochastic oil prices. Taking account into the unique upper and lower structural feature of China's oil production enterprises, the author firstly puts forward some new concepts of oil development programming such as"output constitution","output allocation"and"the structure of measure output"and so on, firstly applies bi-level programming for oil development programming, and builds s various optimization models of oil development programming including yearly and multi-year, single target and multi-target, single level and bi-level optimization models, which will successfully describe the problem of the planning and decision-making of China's oil production enterprises, especially the problem of the long-term strategic planning.4. A study of algorithms with special structure for the optimization models of oil development programming. At first, Powell Operator is defined in Genetic Algorithm and a hybrid genetic algorithm for solving the global optimal solutions of unconstrained optimization problems is obtained. Then the constrained optimization problems are processed by adaptive annealing penalty factors and penalty function, and it is proved the present algorithm is a promising approach for solving global optimization problems. The Powell algorithm can only get a local optimal solution, and it has low success probability in obtaining optimal solution by the repeated calculations of many given initial points. However, the new Powell genetic algorithm based on adaptive annealing penalty function can not only overcome above defects of Powell algorithm, but also significantly improve the probability of the genetic algorithm convergence to optimal solution. It will throw some light on the global optimization of nonlinear optimization theory by the application of the thought proposed above in the algorithms design of nonlinear optimization models which are multi-objective, multi-level and non-differentiable constraint optimizations. Further, it will have profound impact on the theoretical analysis and application of evolutionary algorithms, based on the improvement of genetic algorithm and annealing algorithm for bi-level models and the computer realization of those algorithms.
|
PDF Full Text Request