Claims: 1. A processor implemented method (200) of monitoring an Acoustic Wave Separator (AWS), comprising: collecting (202) turbidity measurements from the AWS and laboratory data, as input data, via one or more hardware processors; converting (204) the turbidity measurements to cell concentration for each of a plurality of chambers in the AWS, using a soft-sensor data from a soft sensing model, via the one or more hardware processors; determining (206) Cell Separation Efficiency (CSE) of each of the plurality of chambers in the AWS, based on the cell concentration, using a prediction model, via the one or more hardware processors; and optimizing (208) a flow rate and an acoustic power of one or more of the plurality of chambers to improve the determined CSE, using an optimization model, via the one or more hardware processors. 2. The method as claimed in claim 1, wherein the prediction model is re-tuned when a measured deviation between an experimental value of the CSE and the CSE determined by the prediction model exceeds a threshold of deviation. 3. The method as claimed in claim 1, wherein the optimization model is updated when a plurality of trajectory profiles of a plurality of manipulated variables associated with the determined CSE, obtained after optimizing the flow rate and the acoustic power, deviate from a plurality of reference trajectory profiles beyond an accepted threshold. 4. The method as claimed in claim 3, wherein a root-cause analysis is performed to determine one or more root-causes for the determined deviation, the root-cause analysis comprising: monitoring one or more Key Performance Indicators (KPIs) associated with the plurality of manipulated variables for any variation from a pre-defined range of values; monitoring a plurality of process parameters associated with the one or more KPI for which the variation from a pre-defined range of values is identified; determining one or more corrective actions based on a historical data, wherein the historical data comprises information on a corrective action taken for specific deviations in past instances; and recommending the determined one or more corrective actions. 5. The method as claimed in claim 1, wherein the input data comprises real-time data and non-real-time data. 6. The method as claimed in claim 1, wherein the CSE of each of the plurality of chambers is determined based on a measured value of the flowrate, inline turbidity measurements of feed stream, and a measured value of the acoustic power, with the cell concentration. 7. The method as claimed in claim 1, wherein optimizing the flow rate and the acoustic power comprises: generating (302) an optimization problem dynamically as a function of the flow rate, and the acoustic power; and determining (304) one of a dynamic optimization and a static optimization, as an optimization technique to be used, wherein performing the dynamic optimization, comprises: identifying a plurality of constraints affecting optimization of the flow rate and the acoustic power, from the generated optimization problem; identifying a control and prediction horizon, for the optimization problem, using the plurality of constraints; obtaining a trajectory of optimal set points of the flow rate and the acoustic power, based on the identified control and prediction horizon; and generating recommendation (306) based on the obtained trajectory of optimal set points to improve the determined CSE. 8. A system for monitoring an Acoustic Wave Separator (AWS), comprising: one or more hardware processors (104); an I/O interface (106); and a memory (102) storing a plurality of instructions, wherein the plurality of instructions cause the one or more hardware processors to: collect turbidity measurements from the AWS and laboratory data, as input data; convert the turbidity measurements to cell concentration for each of a plurality of chambers in the AWS, using a soft-sensor data from a soft sensing model; determine Cell Separation Efficiency (CSE) of each of the plurality of chambers in the AWS, based on the cell concentration, using a prediction model; and optimize a flow rate and an acoustic power of one or more of the plurality of chambers to improve the determined CSE, using an optimization model. 9. The system as claimed in claim 8, wherein the system re-tunes the prediction model when a measured deviation between an experimental value of the CSE and the CSE determined by the prediction model exceeds a threshold of deviation. 10. The system as claimed in claim 8, wherein the system updates the optimization model when a plurality of trajectory profiles of a plurality of manipulated variables associated with the determined CSE, obtained after optimizing the flow rate and the acoustic power, deviate from a plurality of reference trajectory profiles beyond an accepted threshold. 11. The system as claimed in claim 10, wherein the system performs a root-cause analysis to determine one or more root-causes for the determined deviation, by: monitoring one or more Key Performance Indicators (KPIs) associated with the plurality of manipulated variables for any variation from a pre-defined range of values; monitoring a plurality of process parameters associated with the one or more KPI for which the variation from a pre-defined range of values is identified; determining one or more corrective action based on a historical data, wherein the historical data comprises information on a corrective action taken for specific deviations in past instances; and recommending the determined one or more corrective actions. 12. The system as claimed in claim 10, wherein the system collects real-time data and non-real-time data as the input data. 13. The system as claimed in claim 10, wherein the system determines the CSE of each of the plurality of chambers based on a measured value of the flowrate, inline turbidity measurements of feed stream, and a measured value of the acoustic power, with the cell concentration. 14. The system as claimed in claim 10, wherein the system optimizes the flow rate and the acoustic power by: generating an optimization problem dynamically as a function of the flow rate, and the acoustic power; and determining one of a dynamic optimization and a static optimization, as an optimization technique to be used, wherein performing the dynamic optimization, comprises: identifying a plurality of constraints affecting optimization of the flow rate and the acoustic power, from the generated optimization problem; identifying a control and prediction horizon, for the optimization problem, using the plurality of constraints; obtaining a trajectory of optimal set points of the flow rate and the acoustic power, based on the identified control and prediction horizon; and generating recommendation based on the obtained trajectory of optimal set points to improve the determined CSE. , Description:FORM 2 THE PATENTS ACT, 1970 (39 of 1970) & THE PATENT RULES, 2003 COMPLETE SPECIFICATION (See Section 10 and Rule 13) Title of invention: METHOD AND SYSTEM FOR OPTIMIZATION OF ACOUSTIC WAVE SEPARATION IN BIOMANUFACTURING SYSTEMS Applicant: Tata Consultancy Services Limited A company Incorporated in India under the Companies Act, 1956 Having address: Nirmal Building, 9th Floor, Nariman Point, Mumbai 400021, Maharashtra, India The following specification particularly describes the invention and the manner in which it is to be performed. TECHNICAL FIELD The disclosure herein generally relates to Acoustic Wave Separation, and, more particularly, to a method and system for improving Cell Separation Efficiency (CSE) in the Acoustic Wave Separation. BACKGROUND The process of biomanufacturing involves culturing of host cell population in a nutrient rich growth medium, facilitating cell replication and product formation. Upon achieving a desired amount of product quality and quantity, the host cells need to be removed from the culture medium to extract the product of interest. The process of isolating target protein from the culture medium is called cell harvesting (also referred to as cell clarification and used herein interchangeably). The common method in practice for the cell harvesting is centrifugation, where continuous spinning separates the protein from the host cells themselves. Choosing centrifugation as a choice for cell harvesting involves some risks. The high shear force due to centrifugation can rupture the cells and intracellular components exude into the medium. The additional debris in the medium will deteriorate the purity of the batch and affect downstream chromatography efficiency. All these downsides of centrifugation become severe at high cell density. Likewise, depth filtering is another harvesting technique, which is adversely impacted at high cell densities. Therefore, these conventional methods may not be suitable for continuous biomanufacturing process. Acoustic wave separation technology (AWS) removes host cells and debris from the cell culture medium continuously and it can handle high volumes of medium without considerable effort in scaling and cleaning. AWS eliminates the volume limitation with conventional methods, thus making it itself suitable for continuous biomanufacturing. The AWS technology uses non-contact mode of operation where the cell separation is achieved using acoustic waves. The cells suspended in the medium move towards the node or antinode of the standing acoustic wave and form agglomerates, and in the due course, settle down due to gravitational force. However, various parameters that affect efficiency of the AWS may have to be controlled to obtain a desired/target performance levels, which the state-of-the-art systems fail to do. SUMMARY Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems. For example, in one embodiment, a processor implemented method of monitoring an Acoustic Wave Separator (AWS) is provided. In this method, initially turbidity measurements from the AWS and laboratory data, are collected as input data, via one or more hardware processors. Further, the turbidity measurements are converted to cell concentration for each of a plurality of chambers in the AWS, using a soft-sensor data from a soft sensing model, via the one or more hardware processors. Further, a Cell Separation Efficiency (CSE) of each of the plurality of chambers in the AWS is determined, based on the cell concentration, using a prediction model, via the one or more hardware processors. Further, a flow rate and an acoustic power of one or more of the plurality of chambers are optimized to improve the determined CSE, using an optimization model, via the one or more hardware processors. In another aspect, a system for monitoring an Acoustic Wave Separator (AWS) is provided. The system includes one or more hardware processors, an I/O interface, and a memory storing a plurality of instructions. The plurality of instructions cause the one or more hardware processors to collect turbidity measurements from the AWS and laboratory data, as input data. The turbidity measurements are converted to cell concentration for each of a plurality of chambers in the AWS, using a soft-sensor data from a soft sensing model, via the one or more hardware processors. Further, a Cell Separation Efficiency (CSE) of each of the plurality of chambers in the AWS is determined, based on the cell concentration, using a prediction model, via the one or more hardware processors. Further, a flow rate and an acoustic power of one or more of the plurality of chambers are optimized to improve the determined CSE, using an optimization model, via the one or more hardware processors. In yet another aspect, a non-transitory computer readable medium for monitoring an Acoustic Wave Separator (AWS) is provided. The non-transitory computer readable medium includes a plurality of instructions, which when executed, cause the monitoring of the AWS via the following steps. Initially turbidity measurements from the AWS and laboratory data are collected as input data via one or more hardware processors. Further, the turbidity measurements are converted to cell concentration for each of a plurality of chambers in the AWS, using a soft-sensor data from a soft sensing model, via the one or more hardware processors. Further, a Cell Separation Efficiency (CSE) of each of the plurality of chambers in the AWS is determined, based on the cell concentration, using a prediction model, via the one or more hardware processors. Further, a flow rate and an acoustic power of one or more of the plurality of chambers are optimized to improve the determined CSE, using an optimization model, via the one or more hardware processors. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles: FIG. 1 illustrates an exemplary system for Acoustic Wave Separator (AWS) monitoring, according to some embodiments of the present disclosure. FIG. 2 is a flow diagram depicting steps involved in the process of AWS monitoring by the system of FIG. 1, according to some embodiments of the present disclosure. FIG. 3 is a flow diagram depicting steps involved in the process of performing dynamic optimization by the system of FIG. 1, according to some embodiments of the present disclosure. FIG. 4 depicts an example implementation of the system of FIG. 1, for the AWS monitoring, according to some embodiments of the present disclosure. FIG. 5 is an example diagram depicting architecture of the AWS being monitored by the system of FIG. 1, in accordance with some embodiments of the present disclosure. FIG. 6 is an example diagram depicting architecture of chambers in the AWS of FIG. 5, in accordance with some embodiments of the present disclosure. FIG. 7 is an example diagram depicting the step of determining CSE of the AWS, based on CSE of chambers in the AWS, in accordance with some embodiments of the present disclosure. FIGS. 8A and 8B are example graphs representing impact of the flow rate and acoustic power determined by the system of FIG. 1 on efficiency of the AWS, in accordance with some embodiments of the present disclosure. DETAILED DESCRIPTION OF EMBODIMENTS Exemplary embodiments are described with reference to the accompanying drawings. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. Wherever convenient, the same reference numbers are used throughout the drawings to refer to the same or like parts. While examples and features of disclosed principles are described herein, modifications, adaptations, and other implementations are possible without departing from the scope of the disclosed embodiments. It is intended that the following detailed description be considered as exemplary only, with the true scope being indicated by the following claims. The process of biomanufacturing involves culturing of host cell population in a nutrient rich growth medium, facilitating cell replication and product formation. Once a desired amount of product quality and quantity are achieved, the cells need to be removed from the culture medium. The process of isolating target cell from the culture medium is called cell harvesting (also referred to as cell clarification and used herein interchangeably). The most common method in practice for this operation is centrifugation, where continuous spinning separates the target cells themselves. While choosing centrifugation as a choice for cell harvesting, it involves some risks also. The high shear force due to centrifugation can rupture the cells and intracellular components exude into the medium. The additional debris in the medium will deteriorate the purity of the batch and downstream chromatography efficiency. All these downsides of centrifugation become severe at high cell density. Likewise, depth filtering is another harvesting technique, which again struggles at high cell densities. Therefore, these conventional methods may not be suitable with continuous biomanufacturing process. Acoustic wave separation technology (AWS) continuously removes cells and debris from the medium, it can handle high volume of medium without additional effort in scaling and cleaning. AWS eliminates the volume limitation with conventional method, thus makes itself suitable for continuous biomanufacturing. The AWS technology uses non-contact mode of operation where the cell separation is controlled using acoustic waves. The cells suspended in the medium move towards the node or antinode of the standing acoustic wave and forms agglomerates, and in the due course it settles down due to gravitational force. The disclosure herein provides a system and method for control and optimization of acoustic wave separator, where 3D standing waves are used for enhanced cell clarification. The key performance parameters (KPI) of the AWS equipment comprises cell separation efficiency (CSE), throughput, and the power consumed. The process should be optimized to achieve desired CSE for a given flow rate and acoustic power. The optimization can be performed by analyzing real time and non-real time variables and controlling the manipulated variables. The real-time data from AWS equipment comprises operations data such as inline turbidity measurements, flowrates, temperatures, power consumption, motor currents, motor Rotations Per Minute (RPM), and so on, measured for various pumps, flows, acoustic chambers, etc. The input data may also comprise environment data such as ambient temperature, atmospheric pressure, and the like. The non-real time data includes data from LIMS (laboratory information management system). Laboratory data consists of parameters such as cell size distribution, density, impurity levels, turbidity, and the like. A typical set-up of the AWS operated in fed-batch mode is given in FIG. 5. The equipment is divided into four units. Acoustic wave generator integrated Five peristaltic pumps (one feed and four concentrate pumps) Four disposable acoustic chambers with temperature and fluid level sensors. Five on-line turbidity meter units. The online-turbidity meters measure the real-time turbidity of the cell culture broth (CCB). The disposable acoustic chambers are connected in series along with the turbidity unit such that each acoustic chamber is preceded by a dedicated turbidity meter. The chamber is equipped with a piezo electric crystal, the acoustic waves produced by the piezoelectric transducer are reflected by a reflector. The generated wave and reflected wave together form an acoustic standing wave. The cells get agglomerated at the standing wave nodes. When the mass of the agglomerate crosses a critical value, it settles down due to gravity. Each acoustic chamber is equipped with peristaltic pump to retrieve the concentrate slurry. The temperature of CCB was maintained at the required temperature with a jacketed surge tank to avoid the temperature increase above a critical temperature during the overall acoustic clarification process, which may otherwise lead to protein aggregation. Before commencing the clarification process, the gamma-irradiated acoustic chambers are connected to a turbidity meter unit with tubing and the feed and concentrate tubing’s are connected to the respective tanks in a sterile environment. CCB is pumped from the feed surge tank to acoustic chambers using a feed pump. As the CCB passes through feed turbidity unit, the turbidity value is recorded. The material then enters the acoustic chamber from two middle ports and is allowed to fill up to minimum critical volume to ensure that the entire acoustic zone is immersed with the cell culture fluid, following which the acoustic power is turned on. Subsequently, the acoustic agglomeration starts to occur and cell agglomerates settle at the bottom of the chamber concentrate slurry collector. The cell slurry takes a while to settle down after initiating the acoustic process, and hence the concentrate pump is switched on after short time interval. Both acoustic power and concentrate pump flowrate values can be decided based on the turbidity readings. The harvested cell culture fluid (HCCF) then exits the chamber and is pumped to the next turbidity meter. The same process is repeated for the remaining three chambers until the final HCCF starts collecting after leaving the fourth chamber. Architecture of an acoustic wave chamber (alternately referred to as ‘acoustic chamber’) equipped with a level sensor, temperature sensor and a cooling jacket is depicted in FIG. 6. Considering the importance of controlling key process parameters and monitoring key performance indicators in real time, the system 100 elaborated in the embodiments of present disclosure act as a digital replica of the AWS with a dynamic optimization and control system which addresses controlling and dynamic optimization of the AWS process in real time. Referring now to the drawings, and more particularly to FIG. 1 through FIG. 8B, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments and these embodiments are described in the context of the following exemplary system and/or method. FIG. 1 illustrates an exemplary system for AWS monitoring, according to some embodiments of the present disclosure. In an embodiment, the system 100 includes a processor(s) 104, communication interface device(s), alternatively referred as input/output (I/O) interface(s) 106, and one or more data storage devices or a memory 102 operatively coupled to the processor(s) 104. The system 100 with one or more hardware processors is configured to execute functions of one or more functional blocks of the system 100. Referring to the components of system 100, in an embodiment, the processor(s) 104, can be one or more hardware processors 104. In an embodiment, the one or more hardware processors 104 can be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the one or more hardware processors 104 are configured to fetch and execute computer-readable instructions stored in the memory 102. In an embodiment, the system 100 can be implemented in a variety of computing systems including laptop computers, notebooks, hand-held devices such as mobile phones, workstations, mainframe computers, servers, and the like. The I/O interface(s) 106 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface to display the generated target images and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular and the like. In an embodiment, the I/O interface (s) 106 can include one or more ports for connecting to a number of external devices or to another server or devices. The memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes. Further, the memory 102 includes a database 108 that stores all data associated with the AWS monitoring being performed by the system 100. For example, the database 108 stores the configurable instructions that are executed to cause the one or more hardware processors 104 to perform various steps associated with the AWS monitoring. The database 108 may further store all data that is collected as input for determining the Cell separation Efficiency (CSE) of the AWS. The database 108 may further store information on the determined CSE and corresponding values of flow rate and acoustic power (also referred as manipulated variables and interchangeably used herein). Functions of the components of the system 100 are explained in conjunction with the flow diagrams in FIG. 2 and FIG. 3, the example implementation given in FIG. 4, and the AWS architecture given in FIG. 5. FIG. 2 is a flow diagram depicting steps involved in the process of AWS monitoring by the system of FIG. 1, according to some embodiments of the present disclosure. When the system 100 is deployed to monitor an AWS plant, the system 100 establishes connection with the AWS through one or more suitable interfaces provided by the I/O interface 106. In an embodiment, the data with respect to working/performance of the AWS, in terms of a plurality of parameters, is stored in a server connected to the AWS. In this scenario, the system 100 may query the server to get required input data. After establishing the connection, at step 202, the system 100 collects turbidity measurements from the AWS and experimental data as input data. In various embodiments, the input data may be real-time data as well as non-real-time data, wherein the real-time data may be from the AWS, and the non-real-time data generated using laboratory experiments (hence may be referred to as laboratory data). In an embodiment, the system 100 collects the input data at pre-defined intervals of time (for example, once in every 1 min, once in every 5 min, and so on). In another embodiment, the system 100 continuously collects the input data. Also, in addition to the turbidity measurements, the real-time input data from the AWS may also include information on parameters such as but not limited to flowrates, temperatures, power consumption, motor currents, and motor RPM. The input data may also comprise environmental variables such as ambient temperature, atmospheric pressure and disturbance variables such as feed concentration and feed temperature. The laboratory data may include information on parameters such as but not limited to cell size distribution, density, and contamination levels. The system 100 may perform pre-processing of the input data to format the data to suit data quality, and data type/format requirements, to make the data suitable for further processing. The preprocessing of the input data comprises identification and removal of outliers, imputation of missing data, and synchronization and integration of a plurality of variables from one or more data sources using the frequency of various units in the plant. A sampling frequency of the real-time and non-real-time data may be unified to, for example, once every 1 min, where the real-time data is averaged as necessary and the non-real-time data is interpolated or replicated as necessary. Pre-processed data generated by pre-processing the input data is then further processed by the system 100 as part of the AWS monitoring. Further, at step 204, the system 100 converts the turbidity measurements to cell concentration, for each of a plurality of chambers of the AWS. The system 100 uses a soft-sensor data from a soft sensing model to convert the turbidity measurements to cell concentration. Turbidity is a measure of relative cloudiness in a liquid. It is an optical characteristic and is a measure of the amount of light that is scattered by cells in the liquid. The higher the intensity of scattered light, the higher the turbidity. In-line turbidity sensors are used to measure the turbidity of feed and turbidity in each acoustic chamber. These sensors measure the incident intensity and transmitted intensity of the scattered light and provides the optical density measurements (OD) using the formula, OD= ?log?_10 I_0/I --- (1) where, ’ I_0’ is incident intensity and ‘I’ is transmitted intensity. However, these turbidity measurements need to be converted to concentration (c) to understand the operations better. One of the ways to convert OD to concentration is by using Beer-Lambert law. However, the parameters in the equation needs to be calibrated for solutions with different cell sizes and materials. Apart from Beer-Lambert law, various other empirical models are also used for converting OD measurements to concentration. For example, a second order polynomial equation is calibrated with experimental OD measurements to predict the concentration where assumptions of Beer-Lambert law are not valid. Therefore, the system 100 uses these models or available correlation to convert the turbidity measurements received in real-time to concentration in each chamber. The soft sensing model uses the pre-processed data, and the soft- sensor data to generate information on the cell concentration. The soft-sensor model includes physics-based, data-driven or rule-based soft sensors, which converts the inline turbidity measurements to the cell density or cell concentration in AWS chamber. The soft sensing model may further combine the experimental data and the pre-processed data to obtain an integrated data, which is further used to for performing the AWS monitoring and associated steps. Further, at step 206, the system 100 determines a Cell Separation Efficiency (CSE) of each chamber of the AWS, based on the cell concentration obtained at step 204. At this step, the system 100 processes the cell concentration using a prediction model, to determine the CSE. The prediction model predicts the CSE using feed flowrate, inline turbidity measurements of feed stream, and acoustic power, along with the cell concentration. The prediction model may be a data-based model or physics-based model or a combination of both, which in turn is used for determining the CSE. Data-based models such as machine learning and deep learning models in the model repository are used for the CSE prediction. The physics-based model includes models like population balance model to determine number density of cells. A comprehensive predictions model using the data based and/or the physics based model can determine the CSE with high accuracy. The prediction model used in the present disclosure makes use of population balance model to account for cell aggregation. The population balance models could be any one of two-dimensional PBM (accounting for surface area and volume), one dimensional PBM (accounting for volume or number density alone), QMOM (Quadrature method of moments) or a simple monodisperse model. PBMs can help in accurate prediction of CSE. The CSE can be determined for individual chamber or for all chambers as required. The CSE determined for individual chambers can together contribute to overall CSE of the AWS, as depicted in FIG. 7. At this step, the prediction model fetches a population balance model (PBM) from a model repository in the memory 102, and predicts number density of cells and thereby calculates the CSE. The physics based PBM captures phenomena of agglomeration of cells and breakage of cell cluster. The number conservation of cells in the chamber can be described as: (?n(v,t))/?t=Birth-Death+Input-Output where n is the number density of the cells and v is the volume of cells. The population balance model calculates the change in number density function. PBM is formulated based on the size range of cells. Any cell with size exceeding the critical size undergoes settling due to gravity, therefore the PBM model for cells above critical size does not require breakage phenomena to be accounted for. The PBM model for multiple size ranges are solved to obtain a transient number density of the cells. The steps involved in calculating the number density using PBM and in turn the CSE are: Formulation of PBM based on the size range of cells by fetching appropriate models from the model repository. Solving the PBM using numerical techniques and determining the transient distribution of cell number density. Converting number density of the cells to mass of the cells. Evaluate the CSE, where CSE is defined as the ratio of the mass of cells settling due to acoustic agglomeration to the mass of cells entering the feed. In an embodiment, the system 100 verifies accuracy with which the prediction model determined the CSE, by comparing the determined value of CSE with an experimental value of CSE. The experimental value of CSE is obtained via laboratory experiments. If difference/deviation between the determined value and experimental value of CSE is exceeding a threshold, the system 100 performs a self-learning of the prediction model. Self-learning of the prediction model includes performing one or more of a) modifying a plurality of physics-based or data-based models associated with the prediction model, based on the detected deviation, b) fine-tuning one or more model parameters associated with the prediction model, c) retraining the prediction model, and d) re-building the prediction model. Modifying the models may involve calibrating a plurality of soft-sensors used for converting turbidity measurements to cell concentration or mass of cells whenever there is a change in feed quality and/or cell size ranges. Similarly, the parameters in data based and physics-based models (such as PBM models) used to predict CSE, also need tuning when there is a change in feed material or AWS design itself. In case of data-based models, the model predictions are sensitive to changes in input conditions (such as feed material and cell size/type changes) and process regime (for which the data was collected and the model was trained). Whenever there is change in input conditions or regime of AWS operations, the real-time data from new operating conditions is stored in database and is used later for training these data-based models. Further, at step 208, the system 100 optimizes the flow rate and acoustic power to improve the CSE. In various embodiments, the system 100 uses one or a static optimization and a dynamic optimization to optimize the flow rate and acoustic power. Optimization aims at improving the CSE by suggesting corresponding optimal trajectories of feed flowrate and acoustic power such that by optimizing/controlling the flow rate and the acoustic power, the CSE determined at step 206 can be improved further. Various steps involved in the process of optimization are depicted in FIG. 3 (method 300). At step 302, the system 100 generates an optimization problem dynamically as a function of the flow rate, and the acoustic power measured from the AWS. Further at step 304, the system 100 determines one of a dynamic optimization and a static optimization, as an optimization technique to be used. In case of dynamic optimization, a trajectory of optimal set-points of flowrate and acoustic power is obtained dynamically when the AWS is running/operating, and the flowrate and acoustic power may be controlled/adjusted during the working. However, in case of static optimization, a set-point of the manipulated variables i.e., the flowrate and acoustic power are determined for a run, till next implementation is obtained. The dynamic optimization involves the following steps. Initially, a plurality of constraints affecting the optimization are identified from the generated optimization problem. A few examples of the constraints are feasible or critical volume of each chamber, critical flow rate, feasible ranges of acoustic power, and bounds for KPI’s and so on. The constraints may be defined by an authorized user, based on requirements. Further, a control and prediction horizon are identified for the optimization problem, using the plurality of constraints. In case of dynamic optimization, the control and prediction horizon values are obtained from the knowledge database. The prediction horizon is the duration of for which the optimization model predicts the future values of manipulated variables and control horizon is the duration for which the optimized manipulated variables are implemented for control. Further the system obtains a trajectory of optimal set points of the flow rate and the acoustic power, based on the identified control and prediction horizon. Further, at step 306, the system 100 may generate recommendations based on the obtained trajectory of optimal set points to improve the determined CSE. At this stage, an optimization model further optimizes various key performance indicators (KPIs) such as throughput, power consumption, and so on to determine the trajectory and in turn the optimum set points for the acoustic power and the flowrate to maximize or minimize the same. These cases of optimization are identified based on operational requirements. For example, for an operation of the AWS where there is a strict limit on the product quality, the optimization problem is solved where CSE is maximized by constraining the throughput above a threshold value. In case of maximizing the profit of operations, an optimization problem is solved in which throughput is maximized and cost is minimized by constraining the CSE above the threshold value. A few example use case scenarios are mentioned below. Case 1: Maximize the CSE by optimizing the values of feed rate and acoustic power in their practical limits such that throughput does not go below a certain threshold value. ¦(Maximize@(feedrate,acoustic power)) CSE subject to throughput>?throughput?_min ?feedrate?_min?CSE?_min ?feedrate?_min?throughput?_min CSE>?CSE?_min ?feedrate?_min