In the past few years, the numerical solution of non-linear equations for natural gas network optimization has proven to be useful. This paper presents:a sensibility analysis for the most important flow equations defining the key parameters in the optimization process, anda software "GasNet" based in the Stoner method, improved with several techniques to solve the equations system, allowing until an error of 50% in the initial values which avoids the divergence and accelerates the convergence. GasNet is a powerful tool for design and optimization because it has advantages such as many routines for handling the specific gravity of the natural gas mixture composition; in addition it includes correlations used to obtain other properties. It also has graphic utilities that make easier the detection and solution of problems. With information from the network as: flow rate, input and/or output pressures, features of the pipelines, compressors, valves and regulators, etc., we perform a sensibility analysis in order to get which flow equation: Weymouth, Panhandle, AGA, etc., is best at predicting the natural gas flow behavior in the pipeline (also this adjustment is applied to the PVT correlations and the equations of state by determining the Z-factor). The variation between the predicted and measured values with two parameters was estimated using the coefficient of multiple correlation and the X2-distribution. The optimization process is performed with two criteria: the maximum operational permissible pressure and the gas flow velocity. These criteria allow for security in the facilities and suppress the harmful bottle-neck. Two field examples are considered to illustrate the applicability of the proposed algorithms: distribution of a high pressure natural gas network in the Mexico Valley, Mexico and a network in Michigan State, USA. Introduction In the design or modification of any gas network the main goal is to get capacity to satisfy the demand in the present and future in the best way. It is a fact that the components of a gas network such as compressors, valves, pressure regulators and pipes are very expensive in acquisition and operation; therefore it is important to obtain a good design of the system to maximize the profits. The gas transport systems frequently are connected conforming a network where the flow regime is mostly transient, however in practice, the design, control and optimization problems are solved on the basis of steady state flow with reasonable results. In this paper, we will use equations that assume steady - state flow. The basic mathematical model assumes a gas network with two elements: nodes and nodes connectors; the nodes are defined as the points where pipes start or ends, locations where pipe material or diameter changes, and entrance or exits of gas to the system. The nodes are usually represented as circles in a network diagram and the connectors as lines representing the pipes and regulators. Flow equations are commonly used for the analysis of fluid behavior in pipes. Practical flow equations for steady state gas network analysis are those of Weymouth, Panhandle, American Gas Association AGA, etc.. The most popular solution method was proposed by Stoner1. The application of Stoner technique results in a non-linear equation system, whose common solution methods are the Newton Raphson (N-R), Hardy-Cross and Linearization. The N-R method is an iterative technique which starts with initial values for the unknowns. It is the most commonly used method. With guess values, N-R begins to estimate corrections to get the final solution after several iterations. The N-R method converts a non-linear problem in a succession of linear problems. For solving linear systems we use a technique named Conjugate Gradient Method developed for Hestenes and Stiefel2, modified by us to take some advantage in the solution process. This innovation was proved in several field cases with successful results.

This paper presents a new methodology for the automated parameter estimation from well test data, based on type curve matching using the Signal Theory. This procedure solves the inverse problem faster than the conventional techniques, with the additional advantage that the results are not affected by noisy data. The new procedure has been proved with hundred of synthetic and field cases, and it can be used for the interpretation of all the different tests currently of common use in the field. In the present paper this technique is applied to three tests, all already published in the literature. The methodology derived in the study surpasses the currently available matching theory based on non linear regression, since it requires a smaller computing time and its assured convergence (selection) of the correct (best) model that describes the physical conditions of the formation on the vicinity of the well. On the other hand, the conventional regression methods require large computing times for the solution of the system of non linear equations; in addition, they are affected by the presence of noise in the recorded pressure response and usually present convergence problems when the initial solution for the unknown parameters is not close enough to the searched solution. The conventional type curve matching procedure inherently introduces interpretation subjectivity and the possibility of errors, because of the close similarity of the different pressure responses, and to its visual solution approach, which currently is not well understood; thus, it is not possible to develop fully efficient codes that could emulate this human function. Based on a comprehensive set of type curves, the technique proposed in this paper allows an automated match of the pressure response, without the need of the visual effort of the analyst. The Signal Theory has been modified for the automated interpretation of well test data. The new matching correlation derived in this work, based on rules of shift, multiply and sum, solves the matching of pressure data in an improved way. Introduction Reservoir Management is usually defined as an integrated continuous process, requiring a full cooperation and Teamwork between various disciplines to use the resources available to optimize the exploitation of the reservoir. Reservoir management requires economic evaluation and analyses of the asset and associated projects throughout the life of the reservoir, always observing the safety and environmental regulations. The petroleum engineer has had for a few decades numerical reservoir simulators, which must be properly adjusted to match the behavior of the integrated system (reservoir, production string and surface facilities). Observation of model performance under different producing conditions, aids in selecting an optimum set of producing conditions for the reservoir. Today we are now in a situation where the capability of the simulators to use reservoir data frequenty exceeds the availability of data. With regard to the estimation of the properties of the formation, there are several methods available; among them well test analysis has demonstrated to be an excellent source of characterization dynamic data, representing average values in the drainage region of the well. Other sources for the estimation of properties, like well logging and petrophysical laboratory studies, consider far smaller rock volumes. It has been common in the past that advances in a particular discipline are used (transferred) to other areas resulting in a faster overall development. Petroleum engineering has successfully used knowledge from other fields of study, such as heat conduction and geohydrology, borrowing solutions for some problems and applying them for related conditions, especially for well test analysis.

This work presents a new simple method to detect a.linear impermeable barrier by analysis of transient pressure data. This technique is based on the desuperposition method (negative superposition) discussed by some authors and considers the calculation the pressure change caused by the presence of the barrier. The pressure change is analyzed to estimate the distance between the well and the barrier by using the type curve matching technique. Type curves are provided for both drawdown and build tests. The advantage of the technique presented in this work is that pressure data can be analyzed even if the second semilog straight line, whose slope is twice the slope of the first semilog straight line, is not reached. Examples of application are discussed for illustration.

No abstract

The objective of this paper is to present a process for improving the planning of gas field development. We discuss how static and dynamic characterization can be combined to help optimize gas field development. The main concepts, methodologies, and results are shown for an actual Mexican gas field. Static characterization centred on a series of seismic amplitude maps constructed from 3D seismic interpretation. Dynamic data included production data and initial pressure gradients which were useful in delineating individual reservoirs and establishing hydraulic communications between certain reservoirs. The seismic amplitude maps, modified by considering the dynamic data, improved the evaluation of reservoir quality, the estimation of drainage areas, original gas-in-place, and proved reserves. A strategy for the optimal field development was designed by using this combination of seismic amplitude maps modified with information from logs, cores, production, and pressure data. Introduction The subject gas field is located in the central area of the Veracruz basin southeast of Veracruz, Mexico. The field was discovered in 1921 with Well 1, which was drilled by a foreign company. The field is formed by many lenticular sandstones containing gas at abnormal pressures. The first producer well (Well 3) was completed in 1962 in Tertiary sandstones. The field has had a total of 24 wells drilled, in addition to Well 1. Fourteen wells are now gas producers (Wells 3, 4, 5, 6, 201, 402, 403, 404, 405, 406, 412, 415, 420, and 436), nine wells have watered out (Wells 10, 12, 13, 15, 101, 407, 414, 428, and Ma-1), and one well was lost because of mechanical failure (Well 102). Currently, the gas field is comprised of three main producing sandstones: the sandstones at the base of the Lower Pliocene (body "E" located at 1,600 - 1,680 m or 5,249–5,512 feet of depth) which began development in November 1969 with Well 5; the sandstones of the Upper Miocene (body "G" located at 2,050 - 2,250 m or 6,726 - 7,382 feet of depth) which began development in August 1966 with Wells 3, 4, and 6; and, the sandstones of the Late Medium Miocene (body "M" located at 2,500 - 2,700 m or 8,202 - 8,858 feet of depth) which began development in August 1988 with Well 201. Table 1 shows the well names, the reservoir, and fluid data for each producing sandstone. In 1999, a series of 3D seismic surveys were performed covering an area of 240 km2 (59,305 acres). The interpretation of the 3D seismic surveys allowed the construction of several seismic amplitude maps. These maps were used for detecting significant volumes of gas related to high seismic amplitude areas, while establishing geological models and delimiting stratigraphic features. The seismic amplitude maps were calibrated with reservoir and fluid properties as well as production data obtained through productive wells from different sandstones. Using these modified maps then led to an improved development plan for the field. The fundamental objective of this work is to present the methodology and results of the teamwork aspect of this integrated reservoir management study.