Study on Rapid Screening Method for Different Chemical Flooding Methods in Heavy-Oil Reservoirs

Abstract

Heavy-oil reservoirs exhibit a high water–oil mobility ratio. During cyclic steam stimulation or water flooding in the later stages, severe fingering occurs, making it difficult to produce the remaining oil. Chemical flooding methods such as polymer flooding, surfactant–polymer flooding, weak gel flooding, and gel flooding have achieved significant enhanced oil recovery (EOR) effects in the development of high-water-cut oilfields in China. However, the reservoir applicability conditions for each chemical flooding method differ. How to quickly select the appropriate chemical flooding method based on reservoir conditions remains a challenge. This paper uses the basic parameters of a heavy-oil reservoir in Shengli Oilfield as a reference and establishes numerical simulation models for different chemical flooding methods. Then, using the permeability variation coefficient as an indicator to evaluate reservoir heterogeneity, the suitable permeability variation coefficient ranges for different chemical flooding methods are obtained and used as the first-level decision method. Subsequently, based on the differences in temperature and salt tolerance of each chemical flooding method, the applicable ranges for different chemical flooding methods are determined and used as the second-level decision method. Through this two-level decision-making process, the suitable chemical flooding development method for a target reservoir can be rapidly identified, providing support for the efficient development of heavy-oil reservoirs using chemical flooding. The findings are based on a typical heavy-oil reservoir model from Shengli Oilfield; the specific thresholds presented should be calibrated accordingly when applied to reservoirs with different characteristics.

1. Introduction

According to the BP Statistical Review of World Energy, global heavy-oil reserves account for approximately 10–15% of proven petroleum reserves. In the current context of persistently tight energy supply, the efficient development of heavy-oil reservoirs is crucial. Due to the high viscosity and low mobility of crude oil, heavy-oil reservoirs present significant extraction challenges. Traditional development methods (such as conventional cyclic steam stimulation and water flooding) often face severe fingering phenomena in the middle and late stages of development in heavy-oil reservoirs, making it difficult to effectively mobilize the remaining oil [1,2,3,4,5].

Chemical flooding methods such as polymer flooding, surfactant–polymer (SP) flooding, weak gel flooding, and gel flooding have been field-tested and achieved great success in the development of numerous heavy-oil reservoirs in China, including Shengli Oilfield and Daqing Oilfield [6,7,8,9]. Among them, polymer flooding can effectively increase aqueous phase viscosity, reduce the water–oil mobility ratio, and expand the swept volume of water flooding [10]. SP flooding adds surfactants to polymer flooding, significantly reducing oil–water interfacial tension, thereby expanding the swept volume while improving displacement efficiency [11]. Weak gel flooding and gel flooding involve injecting chemicals that gel in situ underground. Since the injected chemicals first enter high-permeability thief zones, they effectively plug these channels after gelling, forcing subsequent injected water into low-permeability areas that are not swept during the water flooding stage, thereby significantly expanding the swept volume [12,13,14,15]. To intuitively illustrate the distinct mechanisms and application conditions of these EOR methods, their key features are summarized in

Table 1. Comparison of Different Chemical Flooding Methods.

Method	Primary Mechanism	Primary Application Condition	Typical Implementation Field
Polymer flooding	Improves mobility ratio by increasing aqueous phase viscosity.	Reservoirs with moderate heterogeneity, improving volumetric sweep efficiency.	Daqing Oilfield, China; partial blocks in Permian Basin, the United States; Pelican Lake Oilfield, Canada; Mangala Oilfield, India
Surfactant–polymer flooding	Reduces interfacial tension and improves mobility ratio simultaneously.	Reservoirs with high remaining oil saturation, targeting both sweep and displacement efficiency.	Shengli Oilfield, China; Daqing Oilfield, China; (Liaohe Oilfield, China
Weak gel/gel flooding	Selectively plugs high-permeability thief zones, diverting flow to unswept areas.	Heterogeneous reservoirs with severe channeling or fractures, achieving profile modification.	Shengli Oilfield, China; Liaohe Oilfield, China; Bohai Oilfield, China; Prudhoe Bay, the United States

.

With numerous chemical flooding methods available, selecting the appropriate one becomes critical. Many researchers have studied the reservoir applicability ranges for various chemical flooding methods. Jouenne et al. [16] tested the stability and viscosity-enhancing performance of polymers under high-temperature and high-salinity conditions, obtaining the applicable reservoir range for polymer flooding. Chen et al. [17] measured the properties of polymer and surfactant at 90 °C and 57.67 g/L salinity and used micromodel etching to study the EOR effect of SP flooding under high temperatures and high salinity. Wu et al. [18], Salunkhe et al. [19], and Zhang et al. [20] used core flooding experiments and micromodel etching to study the plugging effect and compatibility of weak gels on high-permeability streaks. Yin et al. [21] used physical experiments to study the influence of crosslinker pairs and temperature on gelation time and evaluated the EOR effect in heterogeneous reservoirs through core flooding experiments. Lei et al. discussed the performance of gels under harsh conditions such as high temperature and high salinity, providing strategies for gel type selection and optimization.

Although the aforementioned research has studied the applicability conditions of different chemical flooding methods, due to the variety of chemical flooding methods available, how to quickly select the appropriate method for a target reservoir to maximize oil recovery remains a problem for many reservoirs. Therefore, this paper takes a typical reservoir in Shengli Oilfield, China, as an example. First, numerical simulation models for different types of chemical flooding were established. Based on this, combined with economic considerations, the applicable reservoir ranges for each chemical flooding method were determined, and a two-level rapid decision-making method was established to achieve rapid optimization of chemical flooding methods. The innovation of this paper lies not merely in identifying these factors, but in (1) establishing quantitative, simulation-derived cutoff values for reservoir heterogeneity; (2) proposing a structured, two-level decision-making hierarchy; and (3) integrating a practical economic indicator to define the final application boundaries. This approach transforms common knowledge into an actionable engineering tool.

2. Chemical Flooding Mathematical Model

2.1. Chemical Flooding Types and Component Settings

The numerical simulations in this study were conducted using a mature, in-house chemical flooding reservoir simulator that has been extensively validated and applied in Shengli Oilfield. Regarding flow characterization, the model accounts for the non-Newtonian properties of polymer solutions, such as shear-thinning effects, through the specific viscosity models presented later in this section. This paper primarily considers four development methods: polymer flooding, SP flooding, weak gel flooding, and gel flooding. All numerical simulation models include oil and water phases and oil and water components. Additionally, the polymer flooding model includes a polymer component, monovalent ion component, and divalent ion component. The SP flooding model includes a polymer component, surfactant component, monovalent ion component, and divalent ion component. The weak gel flooding and gel flooding models primarily include a crosslinked polymer component and crosslinker [22].

2.2. Physicochemical Mechanism Characterization Models

(1) Gel Formation Model

Both weak gel flooding and gel flooding involve the crosslinking reaction between crosslinked polymer components and crosslinkers under reservoir conditions to form a weak gel or gel. The crosslinking reaction is represented by the following Arrhenius formula [23]:

$$r_k=r_{rk}{\displaystyle \prod _{i=1}^{n_c}\left(\varphi {\rho }_j{S}_j{x}_{ji}\right)j=w,o}$$

(1)

where $r_k$ is the reaction rate, kg/m³/s; $r_{rk}$ is the reaction rate constant, 1/s; $\varphi$ is porosity; ${\rho }_j$ is the density of phase j, kg/m³; $S_j$ is the saturation of phase j; and $x_{ji}$ is the mole fraction of component i in phase j.

(2) Viscosity Characterization Model

The viscosity of the chemical solution primarily considers the effects of concentration, salinity, and shear. The characterization model is as follows:

$${\mu }_p={\mu }_w+{\displaystyle \frac{{\mu }_p^0-{\mu }_w}{1+{\left(\raisebox{1ex}{$\gamma $}\!\left/ \!\raisebox{-1ex}{${\gamma }_{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}$}\right.\right)}^{n-1}}}$$

(2)

where ${\mu }_p$ is the chemical solution viscosity, Pa·s; ${\mu }_w$ is the aqueous phase viscosity, Pa·s; ${\mu }_p^0$ is the zero-shear-rate viscosity, Pa·s; $\gamma$ is the shear rate, 1/s; ${\gamma }_{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}$ is the shear rate at which the chemical solution viscosity reduces to half, 1/s; and n is an exponent.

Considering formation water salinity, the model for calculating ${\mu }_p^0$ is the following:

$${\mu }_p^0={\mu }_w\left[1+\left(A_1{c}_p+A_2{c}_p^2+A_3{c}_p^3\right){\displaystyle \frac{1}{1+{\left(\raisebox{1ex}{$c_s$}\!\left/ \!\raisebox{-1ex}{$c_{s,0.5}$}\right.\right)}^{n_s}}}\right]$$

(3)

where $A_1$, $A_2$ and $A_3$ are coefficients; $c_p$ is the chemical concentration, kg/m³; $c_s$ is the salt concentration, kg/m³; $c_{s,0.5}$ is the salt concentration at which the chemical solution viscosity reduces to half, kg/m³; and $n_s$ is an exponent.

(3) Relative Permeability Interpolation Model

Surfactant injections alter the capillary number, thereby changing endpoint values (residual oil saturation, irreducible water saturation) and the shape of relative permeability curves. The mechanism of surfactants can be characterized through relative permeability curve interpolation. The interpolation method can be expressed as follows:

$$k_{rl}=k_{rl}^0{\left({\displaystyle \frac{S_l-S_{rl}}{1-{\displaystyle \sum _{l=o,w}{S}_{rl}}}}\right)}^{nl}$$

(4)

$$k_{rl}^0=k_{ro}^{0l}+{\displaystyle \frac{S_{rl}^l-S_{rl}}{S_{rl}^l-S_{rl}^h}}\left(k_{rl}^{0h}-k_{rl}^{0l}\right)$$

(5)

$$n_l=n_l^l+{\displaystyle \frac{S_{rl}^l-S_{rl}}{S_{rl}^l-S_{rl}^h}}\left(n_l^h-n_l^l\right)$$

(6)

where $k_{rl}$ represents the relative permeability of phase l; S_rl^l and S_rl^h represent the residual saturation of phase l at low and high capillary numbers, respectively; k_rl^0l and k_rl^0h represent the endpoint relative permeability values of phase l at low and high capillary numbers, respectively; and n_l^l and n_l^h represent the relative permeability exponents of phase l at low and high capillary numbers, respectively.

(4) Component Adsorption/Retention Model

Adsorption satisfies the Langmuir isotherm adsorption equation [24]:

$${\widehat{c}}_i={\widehat{c}}_{i\max }{\displaystyle \frac{b{c}_i}{1+b{c}_i}}$$

(7)

where ${\widehat{c}}_i$ and ${\widehat{c}}_{i\max }$ are the adsorption concentration and saturated adsorption concentration of component i, respectively, kg/m³; and b is an experimentally determined coefficient.

(5) Aqueous Phase Permeability Reduction Factor Model

The aqueous phase permeability reduction factor characterizes the reduction in water phase permeability caused by the adsorption of polymer or gel onto reservoir rock. This coefficient can be characterized as follows:

$$R_k=1+\left(R_{k\max }-1\right){\displaystyle \frac{{\widehat{c}}_i}{{\widehat{c}}_{i\max }}}$$

(8)

where $R_k$ and $R_{k\max }$ are the residual resistance factor and the maximum residual resistance factor, respectively.

3. Adaptability Study of Different Chemical Flooding Methods

A numerical simulation model was established based on the basic parameters of a block in Shengli Oilfield. The average reservoir thickness of the simulated block is 24 m, with an area of approximately 6.02 × 10⁵ m². Based on the oil viscosity of 75 mPa·s at reservoir conditions and standard industry classifications, the crude oil in this study is classified as conventional heavy oil. The block employs a five-spot well pattern, with 4 injection wells and 10 production wells within the simulation area. The daily water injection rate per injection well is 168 m³/d, with a production-injection ratio of 1:1. Chemical flooding commences when the water cut reaches 90%, with a slug size of 0.3 PV. Key parameters for the block and chemicals are shown in

Table 2. Key parameters of the simulation model.

Parameter	Value	Parameter	Value
Reservoir thickness, m	24	Oil viscosity, mPa·s	75
Average porosity	0.31	Net gross ratio	0.7
Average permeability, mD	2000	Total injection rate, m3/d	672
Coefficient of Variation	1.0	Surfactant concentration, %	0.4
Polymer concentration, mg/L	1500	Strong gel, mg/L	1500
Weak gel, mg/L	1500	$c_{s,0.5}$, kg/m3	2.2
rrk, 1/s	0.15	ns	2.0
γ1/2, 1/s	6.0	A2	1.4
A1	1.4	A3	−0.001
Srwl	0.24	Srwh	0.0
Srol	0.25	Sroh	0.0
nwl	2.0	nwh	1.00
nol	2.0	nohhhh	1.0
b	4.0	Rkmax	2.0
Parameter	Value	Parameter	Value

. It is important to note that while the base model parameters are representative of a typical Shengli heavy-oil reservoir, the screening framework itself was developed through a parametric study. The first-level decision involved systematically varying the permeability variation coefficient, while the second-level decision was determined across a range of temperatures and salinities. The oil viscosity of 75 mPa·s was selected as it represents a median value for the target reservoirs of chemical EOR in Shengli Oilfield (typically 20–150 mPa·s), a choice made to enhance the relevance of the results for this common application scenario.

Gel-based chemicals can achieve strong plugging and are, therefore, suitable for reservoirs with stronger heterogeneity, while polymer and SP flooding are suitable for reservoirs with relatively weaker heterogeneity [25,26,27,28]. Considering the main differences among the chemical flooding methods, a two-level decision-making method [29] is proposed for screening chemical flooding methods. The first-level decision primarily considers the impact of reservoir heterogeneity on the development effectiveness of each chemical flooding method. The Sequential Gaussian Simulation method [30] was used to establish permeability-heterogeneous reservoirs, and the permeability variation coefficient of the reservoir was adjusted using proportional scaling. Simulation studies of different chemical flooding methods were conducted to determine the applicable permeability variation coefficient boundaries for each method. The second-level decision primarily considers the differences in temperature and salt tolerance among the chemical flooding methods. As reservoir temperature and salinity increase, the viscosity-enhancing or gel-forming performance of chemicals weakens, leading to a decline in their EOR effectiveness. Due to differences in chemical cost and operating expenses of different chemical flooding methods, the minimum amount of incremental oil that must be produced per ton of chemical injected (ton of incremental oil per ton of chemical, TIOTC) also varies. This paper uses the TIOTC values for different chemical flooding methods obtained through financial accounting at Shengli Oilfield as the evaluation index for the second-level decision. It is important to clarify that these TIOTC lower limits are the break-even thresholds derived from a comprehensive techno-economic analysis conducted by Shengli Oilfield. This internal evaluation incorporates the critical economic variables mentioned by the reviewer, including project Net Present Value (NPV) calculations, sensitivity analyses on chemical costs and oil price scenarios, operational risk factors, and a full accounting of capital and operational expenditures. While a full description of this complex economic model is beyond the scope of this technical paper, the fundamental methodology has been described in our previously published work [31]. The resulting lower limits of TIOTC for polymer flooding, SP flooding, weak gel flooding, and gel flooding are 14.8, 16.7, 17.1, and 18.2 tons of oil per ton of chemical, respectively.

Figure 1 shows the well locations of the established numerical simulation model and the permeability distribution with a permeability variation coefficient of 1.0. The PEBI (Perpendicular Bisection) unstructured grid was used to discretize the simulation domain. This grid allows adaptive refinement around wells to ensure high-precision simulation in near-well regions with high-pressure gradients, while using coarser grids in regions farther from wells to accelerate simulation speed.

The first-level decision method is illustrated using polymer flooding and weak gel flooding as examples. Figure 2 shows the change in water saturation during the water flooding stage. It can be observed that due to strong reservoir heterogeneity, water channeling occurs to some extent in all well groups during water flooding development. For example, in the W11 well group, at a water cut of 90%, the water saturation between W11 and wells W1/W3 is significantly higher than between W11 and wells W6/W4. Figure 3 shows the chemical concentration distribution after injecting 0.2 PV of the chemical agent. It can be seen that due to the relatively poor profile control effect of polymers, the injected polymer still primarily enters the dominant channels formed during water flooding, and its distribution near the wells remains uneven. SP flooding builds upon polymer flooding by adding a surfactant to improve displacement efficiency. Since the surfactant has a minimal impact on volumetric sweep, the polymer distribution for SP flooding (Figure 3b) is nearly identical to that of the standard polymer flood (Figure 3a). In contrast, weak gel has better profile control ability compared to polymer; after entering dominant channels, it forms strong resistance, forcing subsequently injected chemicals and water into relatively low-permeability reservoir layers. Therefore, as shown in Figure 3c, the weak gel concentration distribution around the wells is relatively uniform after injecting 0.2 PV of weak gel. The standard gel system possesses even stronger plugging capability than weak gel, resulting in the most uniform near-wellbore distribution, as depicted in Figure 3d.

Simulation results show that polymer flooding, SP flooding, weak gel flooding and gel flooding have different adaptability to reservoir heterogeneity. By adjusting the permeability variation coefficient and simulating, the relationship between incremental oil recovery and permeability variation coefficient for different chemical flooding methods was obtained, as shown in Figure 4. It can be seen that the incremental recovery factor for all chemical flooding methods decreases as the reservoir permeability variation coefficient increases; however, the rate of decline is not uniform. To identify the boundary of applicability, we have defined a quantitative threshold as the point beyond which the performance degrades rapidly. This threshold is identified as the last data point where the negative slope of the line segment to the subsequent point becomes more than double the average negative slope of all preceding segments. Taking polymer flooding as an example, this sharp acceleration in performance degradation occurs after a permeability variation coefficient of 0.96. Therefore, the applicable range for polymer flooding is defined as 0~0.96. The analysis shows that SP flooding and polymer flooding share the same applicability regarding reservoir heterogeneity, with a suitable permeability variation coefficient range of 0 to 0.96. Weak gel flooding and gel flooding are designed for reservoirs with stronger heterogeneity, with applicable permeability variation coefficient ranges of 0 to 1.04 and 0 to 1.4, respectively.

The second-level decision primarily considers the differences in temperature and salt tolerance among chemical flooding methods. As reservoir temperature and salinity increase, the viscosity-enhancing or gel-forming effects of chemicals weaken, leading to a decline in their enhanced oil recovery (EOR) effectiveness. Due to variations in chemical costs and operational expenses across different chemical flooding methods, the minimum incremental oil production required per ton of chemical injected—termed ton of incremental oil per ton of chemical (TIOTC)—also differs. This study employs the TIOTC values for various chemical flooding methods derived from financial accounting at Shengli Oilfield as the evaluation metric for the second-level decision. The lower TIOTC limits for polymer flooding, SP flooding, weak gel flooding, and gel flooding are 14.8, 16.7, 17.1, and 18.2 tons of oil per ton of chemical, respectively.

Figure 5 illustrates the variation in ton-incremental-oil-per-ton-chemical (TIOTC) with temperature and salinity for different chemical flooding methods. As shown in Figure 5a, the TIOTC values for different methods continuously decrease with increasing temperature. However, due to the inferior temperature tolerance of polymer compared to gel, the TIOTC for polymer flooding and SP flooding decline rapidly under high-temperature conditions, whereas that for weak gel and gel flooding exhibits an approximately linear decline. When salinity increases, the TIOTC values for all chemical flooding methods demonstrate a near-exponential decrease. Using the lower TIOTC limits of 14.8, 16.7, 17.1, and 18.2 tons of oil per ton of chemical as evaluation criteria, the maximum temperature tolerance limits for polymer flooding, SP flooding and weak gel flooding are determined to be 79 °C, 75 °C, 83 °C and 96 °C, respectively, while their maximum salinity tolerance limits are 18,000 mg/L, 13,000 mg/L, 25,500 mg/L and 42,000 mg/L, respectively. Ultimately, through the application of the two-level decision-making process, rapid screening of chemical flooding development methods for target reservoirs can be achieved, as shown in Figure 6.

4. Framework Validation Against Field Practices

To validate the practical relevance and logical soundness of the two-level decision framework proposed in this study, its screening principles were compared against several well-documented, successful chemical EOR field applications. It is important to emphasize that the objective of this analysis is not to directly validate the universal applicability of the specific numerical thresholds calibrated from the Shengli Oilfield model, but rather to confirm that the framework’s core decision-making logic aligns with the industry’s best practices.

(1) Polymer Flooding in Daqing Oilfield, China [32]: Validation of the First-Level Principle

The Daqing Oilfield is the world’s most successful polymer flooding application. Its target reservoirs are characterized by moderate heterogeneity (with a permeability variation coefficient typically ranging from 0.6 to 0.8) and benign conditions (approx. 45 °C and <7000 mg/L salinity). According to our framework’s first-level principle, a mobility control technology like polymer flooding is the correct choice for this type of reservoir with non-extreme heterogeneity. The immense success at Daqing strongly supports the validity of using heterogeneity as the primary criterion for decision-making.

(2) Surfactant–Polymer Flooding in Gudong Field, Shengli [33]: Validation of the Model’s Local Calibration

The Gudong Field, part of the Shengli complex, is geologically similar to the reservoir used in our numerical model. Its successful application of SP flooding under conditions of permeability variation coefficient ~0.75, temperature ~70 °C, and salinity ~12,000 mg/L aligns well with the screening windows calculated by our framework. This case serves as a “type-block” validation, demonstrating that for a Shengli-like reservoir, the specific numerical thresholds generated by our model are reasonable and provide effective guidance.

(3) Gel Treatment in Prudhoe Bay, USA [34]: A Contrasting Validation of the First-Level Principle

The gel treatment program at Prudhoe Bay is a benchmark for conformance control. The field is characterized by extreme heterogeneity with high-permeability thief zones resulting from long-term injection. Following our framework’s logic, such reservoirs are unsuitable for standard polymer flooding and require a profile modification technology like gel treatment to block these thief zones. The success of this project powerfully validates the first-level decision logic from the opposite perspective: the degree of heterogeneity is the fundamental basis for distinguishing between mobility control and profile modification technologies.

In summary, this comparative analysis with world-class projects demonstrates that the decision-making logic embedded in our two-level framework is consistent with successful industry strategies. While the specific numerical boundaries would require recalibration for different reservoir types and economic conditions, the framework itself stands as a robust and practical screening methodology.

5. Conclusions

(1) Due to the relatively poor profile control effect of polymers, the injected polymer still primarily enters the dominant channels formed during the water flooding stage, resulting in an uneven distribution of the polymer near the wells in the reservoir. Weak gel exhibits better profile control ability compared to the polymer; after entering dominant channels, it forms strong resistance, forcing subsequently injected chemicals and water into relatively low-permeability reservoir layers. The applicable permeability variation coefficient range for polymer flooding is 0~0.96. Weak gel flooding is suitable for reservoirs with stronger heterogeneity, with an applicable permeability variation coefficient range of 0~1.04.

(2) As temperature increases, the TIOTC value for both polymer flooding and weak gel flooding decreases continuously. However, the temperature tolerance of the polymer is poorer compared to that of a weak gel. Therefore, at high temperatures, the TIOTC value for polymer flooding decreases rapidly, while that for weak gel flooding decreases approximately linearly. When salinity increases, the TIOTC values for both chemical flooding methods decrease approximately exponentially. The maximum temperature tolerance for polymer flooding and weak gel flooding is 79 °C and 83 °C, respectively, while the maximum salinity tolerance is 18,000 mg/L and 25,500 mg/L, respectively.

(3) Using the permeability variation coefficient as an indicator to evaluate reservoir heterogeneity, the suitable permeability variation coefficient ranges for different chemical flooding methods were obtained and used as the first-level decision method. Then, based on the differences in temperature and salt tolerance among the chemical flooding methods, the applicable ranges for different chemical flooding methods were determined and used as the second-level decision method, establishing a rapid decision-making method for selecting chemical flooding methods for target reservoirs.

(4) The core contribution of this work is the establishment of a transferable two-level screening methodology. It is crucial to note that the quantitative thresholds presented are specifically calibrated for the clastic heavy-oil reservoir model representative of Shengli Oilfield and are based on operational parameters that reflect the region’s optimized best practices. Consequently, while the framework’s logic is broadly applicable, applying it to reservoirs with fundamentally different geological characteristics—such as carbonate reservoirs with their complex fracture and VUG systems—would require a dedicated recalibration of the decision boundaries. Therefore, future work should build upon this foundation by extending the framework to other geological settings and by conducting a more rigorous uncertainty quantification. This would involve a comprehensive sensitivity analysis of key operational and geological parameters, ideally using stochastic methods like the Monte Carlo simulation, to develop more robust and generalized screening charts.