Simulation of Pb(II) and Ni(II) Adsorption in a Packed Column: Effects of Bed Height, Flow Rate, and Initial Concentration on Performance Metrics

Abstract

Numerous studies have been conducted employing various techniques to remove pollutants from water bodies. Among these techniques, adsorption a surface phenomenon that utilises adsorbents derived from agricultural residues has shown considerable potential for the removal of contaminants such as heavy metals. However, most of these investigations have been carried out at the laboratory scale, with limited efforts directed towards predicting the performance of these systems at an industrial level. Accordingly, the present study aims to model a packed bed column at industrial scale for the removal of Pb(II) and Ni(II) ions from aqueous solutions, employing biomass derived from oil palm residues as the adsorbent material. To achieve this, Aspen Adsorption was used as a modelling and simulation tool to evaluate the impact of bed height, inlet flow rate, and initial concentration through a parametric assessment. This evaluation incorporated the Freundlich, Langmuir, and Langmuir–Freundlich isotherm models in conjunction with the Linear Driving Force (LDF) kinetic model. The results indicated that the optimal operating parameters included a column height of 5 m, a flow rate of 250 m³/day, and an initial metal concentration of 5000 mg/L. Moreover, all models demonstrated removal efficiencies of up to 94.6% for both Pb(II) and Ni(II). An increase in bed height resulted in longer breakthrough and saturation times but led to a reduction in adsorption efficiency. Conversely, higher flow rates shortened these times yet enhanced efficiency. These findings underscore the potential of computational modelling tools as predictive instruments for evaluating the performance of adsorption systems at an industrial scale.

1. Introduction

Various anthropogenic activities have led to a decline in the quality of water bodies, resulting in significant challenges related to the supply of drinking water and the treatment of wastewater, due to the high volumes of effluent containing hazardous contaminants. Among these, heavy metals are considered some of the most dangerous pollutants owing to their low biodegradability, high toxicity at trace concentrations, and tendency to bioaccumulate in living organisms [1]. Lead (Pb), for instance, is associated with numerous adverse health effects and is extensively utilised in industries such as battery manufacturing, painting, and welding [2]. Similarly, Nickel (Ni) poses considerable health and environmental risks when exposure occurs at elevated concentrations over prolonged periods; it is commonly employed in the production of alloys, coins, jewellery, batteries, and household appliances [3]. Exposure to these heavy metals has been linked to a range of health conditions, including gastrointestinal disorders, respiratory obstruction, dermatitis, skeletal deformities, and hepatic dysfunction, among others [4]. Considering their hazardous nature, regulatory bodies have established stringent limits for their presence in water. For instance, the World Health Organization (WHO) recommends maximum allowable concentrations of 0.01 mg/L for Pb(II) and 0.07 mg/L for Ni(II) in drinking water [5].

Adsorption is a surface-based technique employed for the removal of contaminants from water. It is characterised by being cost-effective, environmentally sustainable, and relatively simple to design, making it one of the most promising technologies for the elimination of heavy metals [6]. This process is generally categorised into two types: physisorption (physical adsorption), in which the contaminant adheres to the surface of the adsorbent material through non-specific Van der Waals forces, and chemisorption (chemical adsorption), which involves the formation of stronger attractive forces such as ionic or covalent bonds as a result of chemical reactions [7]. Various adsorbents derived from agricultural residues have been developed and demonstrate high efficiencies in removing heavy metals. However, these investigations have predominantly been conducted at a laboratory scale. As a result, there is a growing need to develop predictive approaches capable of estimating the performance of adsorption processes at larger, industrial scales.

To this end, computational tools have been employed to simulate adsorption processes initially developed at the laboratory scale and to extrapolate their performance to industrial-scale applications. One such tool is Aspen Adsorption, a software package developed by AspenTech, designed for the modelling and simulation of adsorption systems. Its primary functions include simulating contaminant removal from solutions, substance separation, and purification processes, among others. Aspen Adsorption enables a wide range of operations, including the design of adsorption-based systems aimed at optimising and maximising process efficiency, as well as the modelling of mass transfer, heat transfer, and fluid dynamics phenomena. Accordingly, this study aims to model a packed bed column at industrial scale using computational tools and a parametric evaluation approach to remove Pb(II) and Ni(II) from aqueous solution, employing residues from Elaeis guineensis as the adsorbent material.

2. Materials and Methods

2.1. Physical Properties, Parameters, and Mathematical Models Required for Aspen Adsorption

The first step in simulating the packed column adsorption process involves selecting the list of components and defining the physical property package using the Aspen Properties^® V12 software (Aspen Technology, Inc., Bedford, MA, USA) database to facilitate the removal of the target contaminants. The Non-Random Two-Liquid Electrolyte (ELECNRTL) property method was selected, as it is suitable for aqueous solutions containing electrolytes at both low and high concentrations, provided that no vapour phase is present in the mixture [8]. Once the components and physical properties are defined, the packed column configuration must be established in order to simulate the adsorption process. This requires the specification of various parameters required by the software. For this purpose, data from previous studies involving industrial-scale adsorption in packed columns for heavy metal removal were used as a reference. The column configuration includes a column diameter of 1 m, a bed porosity of 0.67 m³ voids/m³ bed, a total void porosity of 0.4 [9,10], an adsorbent bulk density of 589 kg/m³ [11], and a mass transfer coefficient of 2.3 × 10⁻⁴ s⁻¹ [12]. In addition, appropriate mathematical models and operating conditions must be defined to simulate the adsorption process accurately. Aspen Adsorption provides various configuration tabs that enable users to set up the system accordingly. The configuration steps and associated considerations are summarised in

Table 1. Considerations for simulations.

Tab	Considerations
General	Upwind Differencing Scheme 1 (UDS1) discretisation method with a number of nodes of 10
Material/momentum balance	Convection is in the liquid phase Velocity is constant No pressure drop is present
Kinetic model	Linear Driving Force Kinetic Model (LDF)
Isothermal model	Langmuir isothermal model Freundlich isothermal model Langmuir-Freundlich isothermal model
Energy balance	Isothermal process

.

2.2. Mathematical Fundamentals

The UDS1 method, also known as the first-order upwind differencing scheme, is a numerical approach used to solve mass transport equations. It is based on a first-order Taylor series expansion [13] and is described by Equation (1)

$${\displaystyle \frac{\delta {\Gamma }_i}{\delta z}}={\displaystyle \frac{{\Gamma }_i-{\Gamma }_{i-1}}{\Delta z}}$$

(1)

where ${\Gamma }_i$ is the concentration of the contaminant in the fluid phase (mol/m³), i represents the position in the adsorption bed at which the equation is being solved, $\Gamma$_i–1 refers to the preceding position in the direction of flow, and Δz is the distance between two nodes in the bed discretisation.

The adsorption isotherms describe how the adsorption process occurs when an adsorbate meets an adsorbent. Three isotherm models were selected for this study. The Langmuir model is one of the fundamental models used to describe adsorption on solid surfaces. It is especially suitable for systems in which adsorption occurs as a monolayer and where the adsorbed molecules do not interact significantly with one another [14]. This model is expressed by Equation (2)

$$q_e={\displaystyle \frac{q_{max}b{C}_e}{1+b{C}_e}}$$

(2)

where q_e is the adsorption capacity of the contaminant (mg/g), $q_{max}$ is the maximum adsorption capacity (mg/g), b is the Langmuir constant (L/mg), and C_e is the equilibrium concentration of the contaminant in solution (mg/L).

The Freundlich model is widely applied to describe adsorption in systems where the process is not limited to a single molecular layer on the adsorbent surface. It is therefore suitable for systems in which adsorbate molecules can form multiple layers, and it also accounts for adsorption on heterogeneous surfaces [15]. This model is expressed by Equation (3)

$$q_e=k_f{C}_e^{1/n}$$

(3)

where $q_e$ is the adsorption capacity of contaminant (mg/g), $k_f$ is the Freundlich constant indicating adsorption capacity ((mg/g) (mL/g)ⁿ), $C_e$ is the concentration of contaminant in solution at equilibrium (mg/L), and 1/n is the effect of initial concentration on adsorption capacity. The Langmuir–Freundlich model is an extension of both the Langmuir and Freundlich models, and it is used to describe adsorption in systems where both monolayer adsorption (as assumed in the Langmuir model) and multilayer adsorption (as in the Freundlich model) coexist [16,17]. This model is expressed by Equation (4)

$$q_e={\displaystyle \frac{q_{max}{\left(K\ast C_e\right)}^n}{1+{\left(K\ast C_e\right)}^n}}$$

(4)

where $q_e$ is the adsorption capacity of contaminant (mg/g), $q_{max}$ is the maximum loading capacity (mg/g), $C_e$ is the concentration of contaminant in solution at equilibrium (mg/L), K is the affinity constant or adsorption energy [(L/mg)ⁿ], and n is the surface heterogeneity factor.

To describe the rate at which contaminant removal occurs, the software employs kinetic models to predict the removal rate, evaluate the efficiency of the selected adsorbent, and understand the underlying mass transfer processes. Among the models available in Aspen Adsorption, the LDF (Linear Driving Force) model was selected for this study. This model assumes that the driving force for mass transfer is represented by a linear function of the difference between the current and equilibrium concentrations of the component in either the solid or liquid phase [18]. It is described by Equation (5)

$${\displaystyle \frac{\partial w_k}{{\partial }_t}}=MT{C}_{sk}\left(w_k^{\ast }-w_k\right)$$

(5)

where $w_k$ is the concentration adsorbed on the solid for component k (mg/g), $w_k^{\ast }$ is the amount that should be adsorbed if the system were in instantaneous equilibrium with the fluid phase (mg/g), and $MT{C}_{sk}$ is the mass transfer coefficient (1/s).

2.3. Parametric Evaluation

To assess the effect of varying individual parameters on the adsorption process, a parametric evaluation was carried out by considering three scenarios: variation in column height, variation in inlet flow rate, and variation in initial contaminant concentration. Each parameter was altered independently, while the remaining two were kept constant. The adsorption simulations were conducted under dynamic conditions at 1 atm and a temperature of 30 °C. The values used for the constants of the isothermal models used for the development of the simulation are as follows:

• For Freundlich with Pb(II) we used a $k_f$ = 8.380 ((mg/g) (mL/g)ⁿ) and 1/n = 0.921, while with Ni(II) we used a $k_f$ = 1.252 ((mg/g) (mL/g)ⁿ) and 1/n = 0.975 [19,20]
• For Langmuir with Pb(II) we used a $q_{max}$ = 162.630 mg/g and b = 0.054 L/mg, while with Ni(II) we used $q_{max}$ = 47.930 mg/g and b = 0.039 L/mg [19,20]

On the other hand, for the Langmuir–Freundlich isotherm model, the values of the aforementioned model parameters were used. To determine the influence of bed height on the process performance and breakthrough time, the height of the column was varied across 3 m, 4 m, and 5 m [21], while the inlet flow rate and initial concentration remained fixed. Similarly, the effect of inlet flow rate was evaluated using flow rates of 100, 150, and 200 m³/day [22], maintaining constant bed height and initial concentration. Lastly, the influence of initial metal concentration was assessed using values of 2000, 3500, and 5000 mg/L for both contaminants [23,24], with the bed height and inlet flow rate held constant.

3. Results and Discussion

3.1. Data Obtained from Simulations of the Oil Palm-Packed Column for Removing Pb (II) and Ni (II)

Multiple simulations of the packed column were conducted using Aspen Adsorption as a computational tool, applying various parametric configurations of bed height, initial metal concentration, and inlet flow rate.

Table 2. Obtained data from Pb(II) adsorption simulation.

Model	Concentration (mg/L)	Bed Height (m)	Flow Rate (m3/day)	R.T (min)	S.T (min)
Freundlich–LDF	2000	3	250	161	1165
			200	218	1549
			150	275	1927
		4	250	204	1455
			200	275	1927
			150	347	2412
		5	250	275	1928
			200	370	2568
			150	467	3193
	3500	3	250	161	1168
			200	218	1545
			150	275	1930
		4	250	204	1451
			200	275	1930
			150	348	2400
		5	250	275	1930
			200	371	2556
			150	467	3195
	5000	3	250	161	1170
			200	218	1547
			150	275	1931
		4	250	204	1453
			200	275	1931
			150	347	2401
		5	250	275	1931
			200	371	2557
			150	468	3180
Langmuir–LDF	2000	3	250	161	1167
			200	218	1565
			150	275	1945
		4	250	204	1470
			200	275	1945
			150	347	2424
		5	250	275	1946
			200	371	2577
			150	467	3205
	3500	3	250	161	1172
			200	218	1549
			150	275	1945
		4	250	204	1456
			200	275	1945
			150	347	2423
		5	250	275	1945
			200	371	2577
			150	467	3210
	5000	3	250	161	1178
			200	218	1561
			150	275	1956
		4	250	203	1466
			200	275	1955
			150	347	2429
		5	250	275	1951
			200	371	2596
			150	468	3230
Langmuir-Freundlich–LDF	2000	3	250	161	1172
			200	218	1564
			150	275	1947
		4	250	204	1469
			200	275	1947
			150	347	2412
		5	250	275	1939
			200	371	2567
			150	467	3193
	3500	3	250	161	1176
			200	218	1561
			150	275	1950
		4	250	204	1461
			200	275	1950
			150	347	2429
		5	250	275	1947
			200	371	2567
			150	467	3194
	5000	3	250	161	1177
			200	218	1558
			150	275	1951
		4	250	204	1464
			200	275	1948
			150	347	2425
		5	250	275	1952
			200	371	2567
			150	468	3193

and

Table 3. Obtained data from Ni(II) adsorption simulation.

Model	Concentration (mg/L)	Bed Height (m)	Flow Rate (m3/día)	R.T (min)	S.T (min)
Freundlich–LDF	2000	3	250	161	1165
			200	218	1549
			150	275	1927
		4	250	204	1455
			200	275	1927
			150	347	2412
		5	250	275	1928
			200	370	2568
			150	467	3193
	3500	3	250	161	1168
			200	218	1545
			150	275	1930
		4	250	204	1451
			200	275	1930
			150	348	2400
		5	250	275	1930
			200	371	2556
			150	467	3195
	5000	3	250	161	1170
			200	218	1547
			150	275	1931
		4	250	204	1453
			200	275	1931
			150	347	2401
		5	250	275	1931
			200	371	2557
			150	468	3180
Langmuir–LDF	2000	3	250	161	1165
			200	218	1549
			150	275	1927
		4	250	204	1455
			200	275	1927
			150	347	2412
		5	250	275	1928
			200	370	2568
			150	467	3193
	3500	3	250	161	1168
			200	218	1545
			150	275	1930
		4	250	204	1451
			200	275	1930
			150	348	2400
		5	250	275	1930
			200	371	2556
			150	467	3195
	5000	3	250	161	1170
			200	218	1547
			150	275	1931
		4	250	204	1453
			200	275	1931
			150	347	2401
		5	250	275	1931
			200	371	2557
			150	468	3180
Langmuir-Freundlich–LDF	2000	3	250	161	1172
			200	218	1564
			150	275	1947
		4	250	204	1469
			200	275	1947
			150	347	2457
		5	250	275	1954
			200	371	2618
			150	467	3266
	3500	3	250	161	1176
			200	218	1561
			150	275	1950
		4	250	204	1461
			200	275	1950
			150	347	2429
		5	250	275	1947
			200	371	2604
			150	467	3267
	5000	3	250	161	1177
			200	218	1558
			150	275	1951
		4	250	204	1464
			200	275	1948
			150	347	2425
		5	250	275	1952
			200	371	2606
			150	468	3251

present the results obtained for the breakthrough time (R.T.) and saturation time (S.T.) in the competitive adsorption of Pb(II) and Ni(II). Breakthrough time (R.T.) is defined as the moment at which the adsorbent material begins to lose its retention capacity for the contaminants, while saturation time (S.T.) corresponds to the point at which the adsorbent reaches its maximum adsorption capacity. The simulations were performed using the Langmuir, Freundlich, and Langmuir–Freundlich isotherm models, each combined with the Linear Driving Force (LDF) kinetic model. Table 2 presents the results related to the removal of Pb(II).

Table 3 shows the results for the removal of Ni(II).

3.2. Effect of the Change in the Initial Concentration of Pb(II) and Ni(II)

The effect of varying the initial concentration of Pb(II) and Ni(II) on the adsorption process was assessed using concentration levels of 5000, 3500, and 2000 mg/L, while maintaining a constant inlet flow rate of 250 m³/day and a fixed adsorption column height of 3 m. Figure 1 illustrates the breakthrough curves for the Freundlich (Figure 1a,b), Langmuir (Figure 1c,d), and Langmuir–Freundlich (Figure 1e,f) isotherm models, all coupled with the Linear Driving Force (LDF) kinetic model. As the initial pollutant concentration increased, the breakthrough time (R.T.) remained constant, whereas the saturation time (S.T.) showed slight variations depending on the model used. This behaviour results from the higher pollutant load entering the column, which increases the availability of active adsorption sites, thereby accelerating the establishment of adsorption equilibrium [21]. The simulation data revealed that, Freundlich–LDF model, an initial concentration of 5000 mg/L yielded R.T. and S.T. values of 161 and 1170 min, respectively, for both Pb(II) and Ni(II). At 3500 mg/L, the values were 161 min (R.T.) and 1168 min (S.T.), while at 2000 mg/L, they were 161 min (R.T.) and 1165 min (S.T.). For the Langmuir–LDF model, the breakthrough time remained 161 min across all concentrations for both metals. However, the saturation times for Pb(II) were 1167 min (2000 mg/L), 1172 min (3500 mg/L), and 1178 min (5000 mg/L), while for Ni(II), the values were 1165 min, 1168 min, and 1170 min, respectively. In the case of the Langmuir–Freundlich–LDF model, the R.T. and S.T. values for both Pb (II) and Ni(II) were 161 and 1172 min at 2000 mg/L, 161 and 1176 min at 3500 mg/L, and 161 and 1177 min at 5000 mg/L, respectively. In general, variation in the initial concentration did not significantly influence the breakthrough time, and only minor differences in saturation times were observed among the models used and the concentration levels applied. This behaviour can be attributed to several factors, including the number of active sites available on the adsorbent, the strong adsorbent–adsorbate affinity—which may have facilitated the rapid attainment of adsorption equilibrium—and the operating conditions, among others. Furthermore, the adsorption efficiencies achieved for all models (Langmuir–LDF, Freundlich–LDF, and Langmuir–Freundlich–LDF) were consistently 94.6% for both Pb(II) and Ni(II), indicating that any of the tested models is suitable for describing the adsorption process presented in this study.

3.3. Effect of Changing the Inlet Flow Rate

The effect of varying the inlet flow rate to the adsorption column was assessed at values of 150, 200, and 250 m³/day, while maintaining a constant initial concentration of 5000 mg/L and a fixed column height of 3 m. The impact of this parameter on the adsorption process was evaluated. Figure 2 presents the breakthrough curves corresponding to the Freundlich (Figure 2a,b), Langmuir (Figure 2c,d), and Langmuir–Freundlich (Figure 2e,f) isotherm models, each coupled with the Linear Driving Force (LDF) kinetic model. An increase in adsorption efficiency was observed as the inlet flow rate increased; however, both the breakthrough time (R.T.) and saturation time (S.T.) decreased. This trend is attributed to the higher velocity of the inlet stream, which leads to a more rapid saturation of the adsorbent’s pores and thus a reduction in the overall contact time for adsorption [22]. The data obtained from the simulation showed that, for Pb(II), under the Freundlich–LDF model, an inlet flow rate of 150 m³/day resulted in R.T. and S.T. values of 275 and 1931 min, respectively. For 200 m³/day, R.T. was 218 min and S.T. were 1547 min, while at 250 m³/day, R.T. dropped to 161 min and S.T. to 1170 min. Similarly, under the Langmuir–LDF model, flow rates of 150, 200, and 250 m³/day yielded R.T. and S.T. values of 275 and 1956 min, 218 and 1561 min, and 161 and 1178 min, respectively. The Langmuir–Freundlich–LDF model showed comparable results, with R.T. and S.T. values of 275 and 1951 min (150 m³/day), 218 and 1558 min (200 m³/day), and 161 and 1177 min (250 m³/day). On the other hand, for Ni(II) the values obtained reveal the same trends observed—for the Freundlich–LDF model, R.T. and S.T. were 275 and 1930 min at 150 m³/day, 218 and 1545 min at 200 m³/day, and 161 and 1168 min at 250 m³/day. Under the Langmuir–LDF model, values of 275 and 1931 min, 218 and 1547 min, and 161 and 1170 min were obtained, respectively. Finally, for the Langmuir–Freundlich–LDF model, the values were 275 and 1951 min; 218 and 1558 min; and 161 and 1177 min for flow rates of 150, 200, and 250 m³/day, respectively. Moreover, the adsorption efficiencies obtained for Pb(II) and Ni(II) across all models were 91.19% at 150 m³/day, 93.3% at 200 m³/day, and 94.6% at 250 m³/day, suggesting that all three isotherm–kinetic model combinations are capable of adequately describing the adsorption process analysed in this study.

3.4. Effect of Column Height Change

The effect of varying the height of the adsorption column on the performance of the adsorption process was analysed using bed heights of 3, 4, and 5 m, while maintaining the initial contaminant concentration and inlet flow rate constant at 5000 mg/L and 250 m³/day, respectively. Increasing the bed height led to longer rupture times (R.T) and saturation times (S.T) but resulted in a slight decrease in overall process efficiency. Figure 3 presents the breakthrough curves for the Freundlich (Figure 3a,b), Langmuir (Figure 3c,d), and Langmuir–Freundlich (Figure 3e,f) isotherm models coupled with the LDF kinetic model. This behaviour can be attributed to the greater volume of adsorbent material at increased bed heights, which provides more active sites and slows the rate at which they are occupied, thereby prolonging the adsorption process [25]. This trend is supported by the results presented in Table 2 for Pb(II). For the Freundlich–LDF model, R.T and S.T values were 161 and 1170 min at 3 m, 204 and 1453 min at 4 m, and 275 and 1931 min at 5 m, respectively. The Langmuir–LDF model yielded R.T and S.T values of 161 and 1178 min at 3 m, 203 and 1466 min at 4 m, and 275 and 1931 min at 5 m. For the Langmuir–Freundlich–LDF model, the corresponding values were 161 and 1177 min at 3 m, 204 and 1464 min at 4 m, and 275 and 1952 min at 5 m. A similar pattern was observed for Ni(II), as shown in Table 3. For the Freundlich–LDF model, R.T and S.T times were 161 and 1170 min at 3 m, 204 and 1453 min at 4 m, and 275 and 1931 min at 5 m. In the case of the Langmuir–LDF model, values of 161 and 1170 min at 3 m, 204 and 1453 min at 4 m, and 275 and 1931 min at 5 m were recorded. For the Langmuir–Freundlich–LDF model, the results were 161 and 1177 min at 3 m, 204 and 1464 min at 4 m, and 275 and 1952 min at 5 m, respectively. Regarding process efficiency, adsorption efficiencies for both Pb(II) and Ni(II) were 94.6% at 3 m, 92.9% at 4 m, and 91.2% at 5 m across all isotherm models, indicating that any of the applied models is suitable for describing the adsorption process investigated in this study.

3.5. Comparison with Other Results Found in the Literature

The data obtained from the simulation of the packed bed using oil palm residues at industrial scale were compared with findings reported in the literature. It is important to note that this comparison is only of relative significance, as each study was conducted under different operating conditions (inlet flow rate, initial concentration, and bed height), and with adsorbent materials distinct from those used in the present work. The simulation results demonstrated that employing oil palm residues as packing material in an industrial-scale column yields reasonable performance in the removal of Pb(II) and Ni(II) from aqueous solutions. A comparison of the values reported in the literature with those obtained in this study is presented in

Table 4. Comparison of result with literature.

Parameter	Pb(II)	Cd(II)	Cu(II)	Pb(II)/Ni(II)
Adsorbent	Olive tree pruning	Dolochar	Olive stone	Oil palm
Initial concentration (mg/L)	100	55	583	5000
Inlet flow rate (m3/day)	128.04	75	43.2	250
Bed height (m)	2.26	0.65	1	3
Rupture time (min)	201.6	666	9828	161/161
Saturation time (min)	503	-		1178/1177
Source	[26]	[27]	[8]	This study

.

3.6. Model Limitations and Assumptions

3.6.1. Dependence of Constants of Isothermal Models

The values of the constants for the isotherm models used in this study were obtained from scientific literature. These constants depend on several factors that influence the adsorbent–adsorbate system, such as the following:

• The specific characteristics of the selected biomass, the pollutant being removed and its chemical properties, the experimental conditions established in each particular study, and the chemical interactions between the adsorbent material and the contaminant.
• This dependence on specific factors may limit the direct applicability of these values to the system studied in this work, potentially affecting the accuracy of the simulation results.

3.6.2. Restricted Operating Conditions

The study was conducted under isothermal conditions, without considering the effect of temperature on the adsorption process. Additionally, the parametric sensitivity analysis was limited to varying only a few key parameters within a predefined range, such as inlet flow rate, bed height, and initial concentration. Other important parameters—such as bulk density, mass transfer coefficient, porosity, and column diameter—were kept constant. This restriction in operating conditions and parameter variation limits the generalisability of the results obtained, thereby constraining a comprehensive evaluation of the dynamic behaviour of the process under different operating scenarios.

4. Conclusions

This research presents an innovative approach to the modelling and simulation of industrial-scale adsorption columns using agro-industrial waste—specifically oil palm residues—for the removal of contaminants such as Pb(II) and Ni(II) in competitive aqueous systems. The results obtained provide valuable quantitative data that contribute to the advancement of industrial adsorption processes.

The parametric evaluation conducted enabled the assessment of the impact of key variables, including bed height, inlet flow rate, and initial concentration, on process efficiency. The findings indicate that increasing bed height prolongs breakthrough and saturation times but reduces adsorption efficiency. Conversely, higher inlet flow rates enhance adsorption efficiency but shorten the saturation time of the biomaterial. It was also observed that initial concentration has no significant effect on adsorption efficiency.

These results provide a robust foundation for the design and optimisation of industrial effluent treatment systems, allowing for the prediction of system behaviour prior to implementation. An economic assessment and techno-economic resilience analysis are planned for future studies to evaluate the feasibility and sustainability of the optimal column configuration.