Performance Forecasting Research Articles

Unified relative permeability model and waterflooding type curves under different levels of water cut

Understanding the relative permeability of multiphase flow in porous media is crucial for the efficient development of waterflooding reservoirs. The semi-log linear relationship between oil-water relative permeability ratio (kro/krw) and water saturation (Sw) has been commonly used to derive waterflooding type curve—an empirical and practical tool for the forecast of production performance. However, recent experiments suggest that owing to the highly dispersed oil in water phase, the kro/krw vs Sw correlation deviates from a straight line in a semi-log plot at high water-cut stage (>90%). Considering the linear model of ln(kro/krw) vs Sw fails to describe the multiphase flow behavior for average water saturation greater than ~65%, we present a unified mathematical model to account for different water saturations and develop two type curves to predict the waterflooding performance. Data from previous literatures, numerical simulations, and actual oilfields are employed to examine the capability of the proposed method. In comparison with traditional Type-A and -B curves, in which the correlation of ln(kro/krw) vs Sw is linear, the proposed model provides remarkably accurate results. Although the nonlinear relationship occurs only at high water saturations, neglecting this feature of the relative permeability curve can cause a significant error in the prediction. This work provides a comprehensive model of ln(kro/krw) vs Sw and new Type-A and -B curves for reservoirs at high water-cut stage, which are beneficial for accurate performance forecasts for the entire duration of waterfloods.

Modeling of expanded granular sludge bed reactor using artificial neural network

In present study, a three-layer backpropagation neural network (BPNN) model was developed to predict the performance of an expanded granular sludge bed (EGSB) reactor. Six related variables such as influent chemical oxygen demand (COD) concentration, hydraulic retention time (HRT), alkalinity (ALK) concentration, pH, volatile fatty acid (VFA) concentration and oxidation reduction potential (ORP), were selected as inputs of the model. All input values were converted to the range (−1, 1) before passing them into the network. Activation function of hidden layer and output layer were “tansig” and “purelin” individually. Several comparisons were conducted to obtain an optimal network structure. Dividerand function was chosen to divide the operating data into training group, testing group and validation group. The Levenberg Marquardt algorithm (trainlm) was found as the best of the ten training algorithms. Other model parameters such as number of neurons in the hidden layer (X1), initial adaptive value (X2) and initial value of weights and biases (X3) were optimized using response surface methodology (RSM). The optimum conditions for minimum mean squared error (MSE) were as follows: X1 (12), X2 (6.0) and X3 (1.0). The precision of optimum ANN model was assessed by means of various statistics such as MSE, determination coefficient (R2), coefficient of variation (CV) and MSE. The result indicated that the proposed ANN model exhibited superior predictive accuracy for the forecast of COD removal performance by EGSB system. Finally, the results of connection weights method demonstrated that VFA concentration (50.37%) had a remarkable impact on reactor performance.

Uncertainty Quantification for History-Matching Problems

This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 181611, “Uncertainty Quantification for History-Matching Problems With Multiple Best Matches Using a Distributed Gauss-Newton Method,” by Guohua Gao, SPE, Jeroen C. Vink, SPE, Chaohui Chen, SPE, Mohammadali Tarrahi, SPE, and Yaakoub El Khamra, Shell, prepared for the 2016 SPE Annual Technical Conference and Exhibition, Dubai, 26–28 September. The paper has not been peer reviewed. It is critically important for decision making and an extremely challenging task to quantify the uncertainty of model parameters and production forecasts properly after conditioning to production data. This paper presents a novel approach to generate approximate conditional realizations using the distributed Gauss-Newton (DGN) method together with a multiple local Gaussian approximation technique. Results are compared with those obtained from other approaches, such as randomized maximum likelihood (RML), ensemble Kalman filter (EnKF), and Markov-chain Monte Carlo (MCMC). Introduction It is well-known that oil- and gasfield development is a high-risk venture. Uncertainties originating from geological models (e.g., structure, stratigraphy, channels, and geobodies) are coupled with uncertainties of reservoir models (e.g., distribution of permeability and porosity in the reservoir) and uncertainties of economic parameters (e.g., oil and gas prices and costs associated with drilling and other operations). It is critically important to properly quantify the uncertainty of such parameters and their effect on production forecasts and economic evaluations. Recently, multiobjective-optimization techniques have been developed to maximize expectations of some economic indicators (e.g., net present value) and, at the same time, to minimize associated uncertainty or risk. Because of limited access to the subsurface reservoir (e.g., it is impossible to measure the permeability and porosity at the location of each gridblock of a simulation model), reservoir properties have quite large uncertainties. In a greenfield, before starting production, available data are limited to static data or hard data such as core data, well-log data, well-testing data, and sometimes seismic data. Hence, it is necessary to generate multiple realizations of structural and petrophysical models with limited information, on the basis of prior estimates of the relative probability of each realization. Then, uncertainty in reservoir performance or production forecasting can be quantified (e.g., with design of experiment or with other techniques). Generally, production profiles generated by running reservoir-simulation models with these prior realizations or unconditional realizations do not honor production data. Uncertain model parameters, such as permeability and porosity, potentially in each gridblock, have to be tuned so that the modified reservoir models are able to generate production profiles that are compatible with production data (e.g., by use of assisted history matching).

Fuzzy modeling of wave-shielding under consideration of cost-effectiveness for an efficient reduction of uncertainty

The shielding of waves is an appropriate method to protect human beings, livestock, as well as structures and commercial goods from the harmful effects of dynamic loads. Areas of application of the concept of shielding range from ocean and coastal engineering, acoustics, to civil as well as mechanical engineering. Materials that can be effective in constructing shielding structural components and facilities are primarily cementitious composites as well as various categories of non-cementitious composites. However, the planning and production processes behind these low-cost materials is often characterized by uncertain information regarding their mechanical properties and even their dimensions. This leads to uncertainty in the forecasting of their overall performance, as well as their effectiveness in shielding incoming elastic, acoustic and water waves. In this paper, a new approach is presented which allows for improving these performance forecasts. In the proposed interdisciplinary procedure used here, the cost-effectiveness fuzzy analysis is connected to an appropriate mechanical model describing the shielding of elastic waves. This analysis procedure leads to a reduction of epistemic uncertainty. This can be achieved provided additional monetary investment is made, as proposed for the first time here, in order to improve the production quality of these shielding materials.

Development of a data-driven forecasting tool for hydraulically fractured, horizontal wells in tight-gas sands

Tight-gas sand reservoirs are considered to be one of the major unconventional resources. Due to the strong heterogeneity and very low permeability of the formation, and the complexity of well trajectories with multiple hydraulic fractures; there are challenges associated with performance forecasting and optimum exploitation of these resources using conventional modeling approaches. In this study, it is aimed to develop a data-driven forecasting tool for tight-gas sands, which are based on artificial neural networks that can complement the physics-driven modeling approach, namely numerical-simulation models. The tool is designed to predict the horizontal-well performance as a proxy to the numerical model, once the initial conditions, operational parameters, reservoir/hydraulic-fracture characteristics are provided. The data-driven model, that the forecasting tool is based on, is validated with blind cases by estimating the cumulative gas production after 10 years with an average error of 3.2%. A graphical-user-interface application is developed that allows the practicing engineer to use the developed tool in a practical manner by visualizing estimated performance for a given reservoir within a fraction of a second. Practicality of the tool is demonstrated with a case study for the Williams Fork Formation by assessing the performance of various well designs and by incorporating known uncertainties through Monte Carlo simulation. P10, P50 and P90 estimates of the horizontal-well performance are quickly obtained within acceptable accuracy levels.

Volatility Forecast of Financial Performance Based on the Logit and LS-SVM Mixed Model

In order to predict the volatility of the financial performance of the listed companies more effectively, constructed the Logit and LS-SVM mixed model. In the empirical study, to financial performance volatility of Chinese listed companies as the object of study, with 5-fold cross-validation accuracy as the evaluation criteria, it is shown that the prediction accuracy using combined system constructed by the mixed model is significantly higher than that of single model, confirmed the validity of the model.

Improving stock index forecasts by using a new weighted fuzzy-trend time series method

We propose using new weighted operators in fuzzy time series to forecast the future performance of stock market indices. Based on the chronological sequence of weights associated with the original fuzzy logical relationships, we define both chronological-order and trend-order weights, and incorporate our proposals for the ex-post forecast into the classical modeling approach of fuzzy time series. These modifications for the assignation of weights affect the forecasting process, because we use jumps as technical indicators to predict stock trends, and additionally, they provide a trapezoidal fuzzy number as a forecast of the future performance of the stock index value. Working with trapezoidal fuzzy numbers allows us to analyze both the expected value and the ambiguity of the future behavior of the stock index, using a possibilistic interval-valued mean approach. Therefore, using fuzzy logic more useful information is provided to the decision analyst, which should be appropriate in a financial context. We analyze the effectiveness of our approach with respect to other weighted fuzzy time series methods using trading data sets from the Taiwan Stock Index (TAIEX), the Japanese NIKKEI Index, the German Stock Index (DAX) and the Spanish Stock Index (IBEX35). The comparative results indicate the better accuracy of our procedure for point-wise one-step ahead forecasts.

Direct forecasting of reservoir performance using production data without history matching

The conventional paradigm for predicting future reservoir performance from existing production data involves the construction of reservoir models that match the historical data through iterative history matching. This is generally an expensive and difficult task and often results in models that do not accurately assess the uncertainty of the forecast. We propose an alternative re-formulation of the problem, in which the role of the reservoir model is reconsidered. Instead of using the model to match the historical production, and then forecasting, the model is used in combination with Monte Carlo sampling to establish a statistical relationship between the historical and forecast variables. The estimated relationship is then used in conjunction with the actual production data to produce a statistical forecast. This allows quantifying posterior uncertainty on the forecast variable without explicit inversion or history matching. The main rationale behind this is that the reservoir model is highly complex and even so, still remains a simplified representation of the actual subsurface. As statistical relationships can generally only be constructed in low dimensions, compression and dimension reduction of the reservoir models themselves would result in further oversimplification. Conversely, production data and forecast variables are time series data, which are simpler and much more applicable for dimension reduction techniques. We present a dimension reduction approach based on functional data analysis (FDA), and mixed principal component analysis (mixed PCA), followed by canonical correlation analysis (CCA) to maximize the linear correlation between the forecast and production variables. Using these transformed variables, it is then possible to apply linear Gaussian regression and estimate the statistical relationship between the forecast and historical variables. This relationship is used in combination with the actual observed historical data to estimate the posterior distribution of the forecast variable. Sampling from this posterior and reconstructing the corresponding forecast time series, allows assessing uncertainty on the forecast. This workflow will be demonstrated on a case based on a Libyan reservoir and compared with traditional history matching.

Prospects of application of artificial neural networks for forecasting of cargo transportation volume in transport systems

The purpose of research – to identify the prospects for the use of neural network approach in relation to the tasks of economic forecasting of logistics performance, in particular of volume freight traffic in the transport system promiscuous regional freight traffic, as well as to substantiate the effectiveness of the use of artificial neural networks (ANN), as compared with the efficiency of traditional extrapolative methods of forecasting. The authors consider the possibility of forecasting to use ANN for these economic indicators not as an alternative to the traditional methods of statistical forecasting, but as one of the available simple means for solving complex problems. Materials and methods . When predicting the ANN, three methods of learning were used: 1) the Levenberg-Marquardt algorithm-network training stops when the generalization ceases to improve, which is shown by the increase in the mean square error of the output value; 2) Bayes regularization method - network training is stopped in accordance with the minimization of adaptive weights; 3) the method of scaled conjugate gradients, which is used to find the local extremum of a function on the basis of information about its values and gradient. The Neural Network Toolbox package is used for forecasting. The neural network model consists of a hidden layer of neurons with a sigmoidal activation function and an output neuron with a linear activation function, the input values of the dynamic time series, and the predicted value is removed from the output. For a more objective assessment of the prospects of the ANN application, the results of the forecast are presented in comparison with the results obtained in predicting the method of exponential smoothing. Results . When predicting the volumes of freight transportation by rail, satisfactory indicators of the verification of forecasting by both the method of exponential smoothing and ANN had been obtained, although the neural network showed the best result (the average relative forecast error was 8.97% for ANN and 11.21% for the method of exponential smoothing, respectively). This can explained by the fact that the temporal dynamic range of the values of the volumes of cargo transportation by this type of transport, for the period under review, has a nonlinear but uniformly changing character. In the case of forecasting the volumes of cargo transportation by road, the time series of initial values for the reporting period is simultaneously non-linear and unevenly changing. This explains the large values of forecasting errors by the method of exponential smoothing (the average relative forecast error of 47.47% for methods of exponential smoothing ); the forecast error with ANN was 13.97%, therefore the results of the prediction obtained by the method of exponential smoothing are considered unsatisfactory, and for ANN – satisfactory. The conclusion . The results of the study confirm the feasibility of using trained artificial neural networks in forecasting the volumes of freight traffic with different cargo flows that have the initial statistical data of which have an uneven nonlinearly changing character in the time dynamic series. A sufficiently high verification in the application of ANN for difficult-to-forecast indicators of the transport process confirms the practical significance of the application of this method in the modeling of the logistics network.

The Impact of the Deviation in Sustainability Reporting from Implementation on Future Performance and Analysts' Forecasts

The Impact of the Deviation in Sustainability Reporting from Implementation on Future Performance and Analysts' Forecasts

Cognitive Data-Driven Proxy Modeling for Performance Forecasting of Waterflooding Process

Assessment of diverse operational constraints and risk appraisal associated with reservoir heterogeneities are essential foundation of production optimization and oil field development scenarios. Water-flooding performance evaluation that comprises comprehensive numerical simulations is typically cumbersome in terms of time and money, which is not reasonably appropriate for practical decision making and future performance forecasting. Cognitive data-driven proxy modeling practices, which incorporate data-mining techniques and machine learning concepts, offer a fascinating substitute for explicit models of the underlying process that can be instantaneously reassessed, especially for extremely nonlinear system forecasts. In this paper, an exploratory data analysis is applied to create a comprehensive data set from Water-flooding actual field data, which entails different characteristics labeling reservoir heterogeneities and other pertinent operational constraints. Artificial neural network (ANN) is applied as a cognitive data-driven proxy modeling effort to predict Water-flooding production in heterogeneous reservoirs. This study presents the great potential of cognitive data-driven proxy modeling techniques for practical applications and as a feasible add on for investigating a huge quantity of real field data efficiently. In addition, the suggested methodology can be incorporated directly into most present reservoir development decision making routines.

The Health Haven (a)

The financial manager at the medical devices start-up Health Haven had to develop a revenue forecast that was aggressive enough to meet required performance targets but realistic enough to be honest and achievable given a reduced budget. He wanted to lead from the middle of his organization by providing an acceptable performance forecast on which Haven could deliver. This case is supported by a teaching note for instructors which contains links to supplemental video for students. Excerpt UVA-OB-1062 Rev. Jan. 26, 2016 The Health Haven (A) Eric Hunt had a great deal to accomplish in very little time. He had only recently transitioned to his new role as senior finance manager of the Health Haven (Haven) from a more junior finance- and business-operations support position. Although he had been with the company for nearly four years, the level of responsibility he now faced was new to him. Not only was the experience of leadership new, but the challenge of preparing this year's budget was far more complex than he had ever thought possible. His predecessor, Steve Knight, had acted in ways that had eroded the trust his associates had in his company's leadership—including Hunt. The circumstances were dicey, there were multiple parties to please, and much was at stake. Company History Knight founded Haven in Durham, North Carolina. He was an emergency department physician who had discovered several useful, low-tech ideas for improving patient care in the course of his work. Because his improvement ideas were not owned outright as the intellectual property of his employer, Knight was able to explore the commercial viability of nine of his ideas using $ 100,000 in start-up capital from family and friends. Eventually, one of his ideas seemed destined to be a hit. His most promising product, the ShoulderBand, was an innovative harness system for supporting a shoulder joint, so it could heal without limiting mobility. It was especially useful to those with mild-to-moderate shoulder pain who also had hypermobile joints, which tended to hyperextend. The hyperextension could delay the healing process. Because the ShoulderBand was classified as a device and not a drug, it did not have to go through a lengthy FDA approval process. Instead, it required laboratory testing and a clearance process that was less expensive and cumbersome to obtain than FDA drug approval. Knight employed two part-time researchers to explore the commercial viability of this and his eight other medical-device ideas. Knight was close, but he had not yet been able to raise the money necessary to scale operations and secure accounts with key customers such as hospital systems and other providers. Although a true philanthropist and decorated physician in his own right, Knight was also modest. He did not have the personality for the external advocacy required to grow the company and had distaste for what he thought was incessant self-promotion. Two years in, he crossed paths with Ray McDonald, a physician from his days at medical school. McDonald's career had paradoxically wound its way into the world of medical-related corporate management, and Knight thought he had the connections, the financial acumen, and the salesman-like personality to take ShoulderBand to the next level as Haven's CEO. . . .

Solid Modeling and Dynamic Analysis of Mechanisms of Press-forging Machines

Based on the example of AV1818 machine for cold die forging the paper considers the matters of complex use of solid modeling systems (Autodesk Inventor Professional) and the dynamic analysis of mechanisms (DAM) to solve problems associated with the performance forecasting of forge-and-pressing equipment at the stage of their design. The dynamic processes in the mechanisms, primarily, the dynamic forces and positioning accuracy of units are defined based on the models representing systems of point masses connected with elastic ties. Basic parameter computation for the objects of models (units of mechanisms or their fragments): inertia (mass or inertia moment) and elasticity (rigidity at strain-compression, bend or torsion) is a rather complicated task solved with the methods of theoretical mechanics and resistance of materials. The procedure suggested in the work allows increasing the accuracy of calculations and decreasing their complexity at the expense of the final element method (FEM) to solid models of these objects.

Comparison of Simulated and Monitored Energy Performance Indicators on NZEB Case Study Eco Silver House

Building energy performance forecasting is a well-known challenge. Several studies show that, when in operation, heating, ventilation and air conditioning (HVAC) systems in buildings do not perform exactly as predicted because of climate differences, user and operational profiles deviations, and furthermore due to technical changes and imperfections that occur in construction process. In engineering practice it is common to use monthly method (EN ISO 13790) for calculation of residential building energy performance. However, if this method is used in a commercial tool or in a national methodology for energy performance certification (Directive 2010/31/EU), it is usually customized with a number of default values, simplifications, and predefined weather data to facilitate the work or architects and engineers. Mismatching of calculated energy performance (either at design stage or for a building as built) and metered energy use of a building is becoming a problem in the real-estate market. More accurate building energy modelling based on hourly dynamic simulation method is a promising solution to diminish the discrepancies of calculated and metered energy use. The paper investigates the calculated and metered energy use of monitored apartments in a highly energy efficient apartment building Eco Silver House committed to meet the national nearly zero energy buildings (nZEB) requirements. The monitoring of 20 apartments (some fully in use, some not occupied but moderately heated) was done in the first heating season after the completion of the building. For two selected flats the detailed comparison of calculated energy use with two monthly method based tools (PHPP and national) and with an hourly simulation by IDA ICE software is presented. Shortcomings and advantages of the applied calculation methods are discussed, and the importance of the dynamic simulation in nZEB design is pointed out. Inter alia, by dynamic simulation the energy performance of two simulated flats in transient boundary conditions the compliance of the apartment energy indicators in design and in operation stage was proven. The case study derives from FP7 EE-Highrise demonstration project – Eco Silver House (ESH) - a multi-residential high-rise energy efficient building of 128 flats, located in the city center of Ljubljana, Slovenia (www.ee-highrise.eu).

Learning, Visualizing, and Assessing a Model for the Intrinsic Value of a Batted Ball

We present a multidisciplinary approach for learning, visualizing, and assessing a model for the intrinsic value of a batted ball in baseball. The new methodology addresses one of the most fundamental problems in baseball analytics. Traditional outcome-based statistics for representing player skill on batted balls have been shown to have a low degree of repeatability due to the effects of multiple confounding variables, such as the defense, weather, and ballpark. New sensors have created the opportunity to define batted-ball descriptors that are invariant to these variables. We exploit this opportunity by using a Bayesian model to construct a continuous mapping from a vector of batted-ball parameters to an intrinsic value defined using a linear weights representation for run value. A kernel method is used to learn nonparametric estimates for the component probability density functions in Bayes theorem using a set of over 100 000 batted-ball measurements, while cross validation enables the model to adapt to the size and structure of the data. Properties of the mapping are visualized by considering reduced-dimension subsets of the batted-ball parameter space. The approach separates the intrinsic value of a batted ball at contact from its outcome and, as a result, allows the definition of batted-ball statistics for batters and pitchers that are less subject to systematic bias and random variation than traditional statistics. We use Cronbach’s alpha to show that statistics derived from batted-ball intrinsic values have a higher reliability than the traditional outcome-based statistics and that this leads to more accurate estimates of player talent level that can be used for performance forecasting.

Study of Factors of Low-carbon Operation (LCO) of Manufacturing Enterprises and Their Implementation Paths

Low-carbon psychological consciousness, enterprise’s image expectation, cost and performance forecast, the context combining institution and technology and public guidance are five factors that have enormous influence on LCO pattern of enterprises. Among these factors, the first one belongs to the category of consciousness, the second and the third ones fall into the category of action while the last two deal with the environment. According to the correlations among awareness-circumstance—action, an integration model is set up and the mechanism between each factor to the low-carbon operation model is explored. The research also provides targeted policy thinking and implementation path for the enterprise low-carbon operation.

Analysis and Forecast of Tracking Performance of Hong Kong Exchange-Traded Funds: Evidence from Tracker Fund and X iShares A50

This study examines the tracking performance of two Hong Kong exchange-traded funds (ETFs): Tracker Fund and X iShares A50. The turnover of these two ETFs was more than half the total turnover of the 141 ETFs in the Hong Kong market during 2005–2013. Tracking performance is assessed using pricing deviation, which is found to be nonzero and predictable. This indicates that the premium paid by investors is of considerable economic interest. The significant differences in the tracking performance of physical ETFs and synthetic ETFs highlight the relative inability of synthetic ETF to track the market. Additionally, we document the existence of co-integration between the ETF prices and stock market prices. An econometric model is estimated to forecast the pricing deviation, which shows different price dynamics between the two ETFs, but an absence of arbitrage opportunities. The time series regression model of pricing deviation is significantly influenced by market value, dividend yield, trading volume, bid-ask spread, and market risk. The size of the regression coefficients indicates that synthetic ETFs have relatively poor ability to track the market during market fluctuations.

Performance Forecasting of Share Market using Machine Learning Techniques: A Review

<p>Forecasting share performance becomes more challenging issue due to the enormous amount of valuable trading data stored in the stock database. Currently, existing forecasting methods are insufficient to analyze the share performance accurately. There are two main reasons for that: First, the study of existing forecasting methods is still insufficient to identify the most suitable methods for share price prediction. Second, the lack of investigations made on the factors affecting the share performance. In this regard, this study presents a systematic review of the last fifteen years on various machine learning techniques in order to analyze share performance accurately. The only objective of this study is to provide an overview of the machine learning techniques that have been used to forecast share performance. This paper also highlights a how the prediction algorithms can be used to identify the most important variables in a share market dataset. Finally, we could have succeeded to analyze share performance effectively. It could bring benefits and impacts to researchers, society, brokers and financial analysts.</p>

Wear Performance Forecasting of Chopped Fiber–Reinforced Polymer Composites: A New Approach Using Dimensional Analysis

ABSTRACTSolid particle erosion of polymer matrix composites is a complex process in which wear occurs from the target surface by impingement of rigid sand particles in an air medium. The rate of material removal (RMR), also referred to as the erosion rate, mainly depends on target material parameters and the erosion conditions such as impact angle, impact velocity, and erodent size. A new semi-empirical model for prediction of the erosion rate of polymer matrix composites has been developed using a dimensional analysis technique based on Buckingham's π theorem. The predictive model analytically rests upon parameters related to chopped glass fiber composites, erodent (target material properties), and operating variables that mainly affect the erosion process of chopped glass fiber–vinyl ester resin composites. The forecasting ability of the predictive model has been assessed and verified by experimental investigations for chopped glass fiber–reinforced vinyl ester resin (VGF) composites. Validation of the theoretical erosion rates obtained from the predictive model showed that they were in good agreement with the experimentally determined erosion rates, where the average error range was estimated to be ∼10 to ∼20%.

Influence of spatial distribution of sensors and observation accuracy on the assimilation of distributed streamflow data in hydrological modelling

ABSTRACTThe aim of this study is to assess the influence of sensor locations and varying observation accuracy on the assimilation of distributed streamflow observations, also taking into account different structures of semi-distributed hydrological models. An ensemble Kalman filter is used to update a semi-distributed hydrological model as a response to measured streamflow. Various scenarios of sensor locations and observation accuracy are introduced. The methodology is tested on the Brue basin during five flood events. The results of this work demonstrate that the assimilation of streamflow observations at interior points of the basin can improve the hydrological models according to the particular location of the sensors and hydrological model structure. It is also found that appropriate definition of the observation accuracy can affect model performance and consequent flood forecasting. These findings can be used as criteria to develop methods for streamflow monitoring network design.