Optimal Plant Design Research Articles

Optimization of CO2 capture plants with surrogate model uncertainties

CO2 capture plants can help reduce the cost of deploying capture systems across the globe. However, the CO2 variability and model uncertainty represent operational challenges to capture CO2 from different sources. This work proposes a framework for analyzing the optimal plant design considering different flue gas sources. We show a methodology to generate large data sets from optimization runs using rigorous models in Aspen Plus®. The efficiency of the approach allows its application to large-scale optimization problems, with an average CPU time per run of 176 s.We additionally build surrogate models (SMs) for the capital and operating costs of the capture plants, employing an iterative procedure to generate SMs using ALAMO. We systematically reject SMs with high uncertainty in the estimated parameters. This approach results in SMs with favorable bias-variance tradeoffs, enabling their effective application to optimization problems under uncertainty, as demonstrated with a pooling problem of CO2 streams.

Artificial neural Network-Based LCOH estimation for concentrated solar power plants for industrial process heating applications

A study was condacted to assess the feasibility of a 5MWt solar-only parabolic trough concentrated solar power (CSP) for industrial process heat (IPH) in various locations across Morocco. Furthermore, the initial plant design has been optimized by considering the number of full-load storage hours and the size of the solar field. Additionally, this study employs a Multi-Layer Perceptron (MLP) model, featuring five input parameters(storage hours, solar multiples,Direct Normal Irradiation, latitude, and longitude), two hidden layers, and one output parameter(levelized cost of heat (LCOH)). The model has been trained using Levenberg-Marquardt (LM) training algorithms. According to the study, the Errachidia site was found to be the most efficient location, generating 21983.56 MWhth at a LCOH of 3.31 ¢/kWth for the initial design. The optimal plant design was determined to have a solar multiple of 1.5 across all investigated sites. The statistical analysis revealed that the MLP model proposed in the study, which had 43 neurons in the hidden layer, demonstrates the lowest errors, including the root mean square error (0.0071) and the coefficient of variance (0.1852), and achieved the highest coefficient of determination value (0.99999) during the training process. Furthermore, the disparities in LCOH values between simulated data and the proposed MLP model across six different Moroccan cities, with errors ranging from 0.01% to −4.07%, confirm the reliability of the MLP model for predicting LCOH.

Profitability of hybrid power plants in European markets

As markets continue to evolve, hybrid power plants (HPPs) are attracting the interest of plant operators and industry players alike. With their ability to provide flexible dispatchability, HPPs are poised to become a competitive solution in future energy markets, particularly in grid ancillary markets (such as frequency control, reactive power control and black-start). Moreover, resource complementarities present in some sites have the potential to significantly improve production predictability of collocated HPPs. To assess the profitability of HPPs in European markets, a hybrid sizing algorithm is applied to three locations across Europe with distinct resource distribution characteristics. The sizing tool implemented includes several novel features that are not typically incorporated into existing sizing software currently available on the market. This includes turbine selection, a simplified physical design of the wind power plant, a surrogate wake model, and a simplified solar panel degradation model. NPV-over-CAPEX optimal plant designs, tend to favor hybridization, by either collocating wind and solar power with storage, or only solar power with storage. On the hand, minimizing the levelized cost of energy generally results in single-technology power plants. The results show significant potential for hybrid power plants in European markets. However, to truly demonstrate their economic competitiveness, the focus should shift from LCOE to other optimization metrics that consider revenue from different electricity markets.

Decision support tool for the construction and seasonal operation of farm-scale anaerobic digestion plants

Optimal plant design and management are critical components for the successful operation of farm-scale anaerobic digestion (AD) plants. However, this often proves challenging due to difficulties in designing and sizing the plant based on specific site conditions. The current investigation aims to address these difficulties by developing a universal decision support tool to assist in the optimal design and management of agriculture-based AD plants, accounting for site-specific practicalities and implications. The tool consists of various mathematical functions, which enable numerous simulations to be created and run. The developed tool was applied to a case study, located in Ireland, to test its usefulness, where the analysis showed the optimal, site-specific, plant design with key assessment indicators. For this case study, the feedstock availability assessment determined that the lignocellulose and non-lignocellulose biomass within a 10 km distance of the site. Based on the local energy demand of the area, the tool modelled an optimal AD plant design, including feedstock storage, digester volume, engine capacity, and digestate storage. The tool applied various technical, economic, and ecological assessment indicators to the plant to gauge its viability. Therefore, demonstrating the tool's usefulness in assisting stakeholders to make informed decisions and reducing costs by optimising plant design and performance.

Archimedes screw generator powerplant assessment and field measurement campaign

The Archimedes screw generator (ASG) is a cost-effective, eco-friendly micro-hydropower technology that can operate effectively in low head and moderate flow rate environments. The technology is well suited for distributed micro-hydropower and off-grid energy schemes, and its simple, robust design allows for local manufacturing and installation in developing regions. There are many successful ASG powerplants installed on- and off-grid, particularly in Europe. Archimedes screw generators are largely designed by experience. As such, there is a need for an accurate performance model in the literature to optimize plant design. There is a lack of experimental data in the literature, particularly for large-scale, real-world screw plants. This study sought to gather both qualitative and quantitative data and introduce it to the literature. The authors surveyed and measured various Archimedes screw powerplants in the United Kingdom. The measurements were then added to a larger dataset to validate a CFD simulation across a range of screw sizes: from a small 15 cm diameter laboratory screw to a large 3.5 m diameter powerplant. The authors plan to use the CFD simulation's accurate approximations to further the literature's ASG performance data for a wide range of screw sizes, shapes, and operational parameter.

Computational optimization of solar thermal generation with energy storage

Integrating renewable energy resources into power systems is essential for achieving sustainability targets. Concentrated solar power can incorporate thermal energy storage, which can provide larger storage capacities than other technologies. In this study, a comprehensive computational framework is developed for the modeling and optimization of a parabolic trough plant with storage. A Particle Swarm Optimization – Mesh Adaptive Direct Search hybrid scheme is applied to optimize hourly plant operations. A 16.5 MWe parabolic trough plant with thermal energy storage, simulated to be located in Daggett, California, is modeled and optimized. The optimal storage duration is four hours with a net present value of $37.40 million. The results show that scenarios including higher peaks in electricity prices yield an optimal storage duration of six hours and net present value of $63.70 million. Importantly, a plant operating under a Purchase Power Agreement without storage results in a 25% decrease in net present value, indicating that it is beneficial to utilize storage and sell at market prices rather than enter into a Power Purchase Agreement at fixed prices. Computational capabilities of the type developed in this study are essential for assessing the impact of future price scenarios on optimal plant design and operations.

Efficient prediction of concentrating solar power plant productivity using data clustering

Concentrating solar power (CSP) plants convert solar energy to electricity and can be deployed with a thermal storage capability to shift electricity generation from time periods with available solar resource to those with high electricity demand or electricity price. Rigorous optimization of plant design and operational strategies can improve the market-competitiveness and commercial viability; however, such optimization may require hundreds of annual performance simulations, each of which can be computationally expensive when including considerations such as optimization of dispatch scheduling, sub-hourly time resolution, and stochastic effects due to uncertain weather or electricity price forecasts. This paper proposes a methodology to reduce the computational burden associated with simulation of electricity yield and revenue for CSP plants over a single- or multi-year period. Data-clustering techniques are employed to select a small number of limited-duration time blocks for simulation that, when appropriately weighted, can reproduce generation and revenue over a single year or within each year of a multi-year period. After selection of appropriate data features and weighting factors defining similarity between time-series profiles, the methodology captured annual revenue within 2.3%, 1.7%, or 1.2% using simulation of 10, 30, or 50 three-day exemplar time blocks, respectively, for each of three single-year location/weather/market scenarios and five plant configurations ranging from low to high solar multiple and storage capacity. When applied to multi-year datasets, the proposed methodology can capture inter-year variability that is unavailable from typical meteorological year (TMY) datasets while simultaneously requiring simulation of less than a single year of data.

Determination of design parameters to minimize LCOE, for a 1 MWe CSP plant in different sites

The levelized cost of energy (LCOE) from a parabolic trough collector power plant is influenced directly by its design parameters. In the present paper, the impact of five design parameters (solar multiple, direct normal irradiation, cycle efficiency, mirrors efficiency and absorber efficiency) is investigated in order to minimize the LCOE for a 1 MWe power plant. The artificial neural networks (ANN) and the response surface methodology (RSM) are used to model the LCOE as function of these five parameters. The methodology is applied to six cities of Morocco, and the results show that the ANN is the best model to predict the LCOE using the best topology of 5-7-1 neurons. This structure allows satisfactory prediction of LCOE obtained with Mean Absolute Error (MAE) of about 2% and Maximum Error (ME) of 9%. A minimal cost of 8.74 cents is obtained for Tata city with a direct normal irradiation of 7.09 Wh/m2/day, by coupling ANN model to desirability function. The optimal plant design point is 1.94 for the solar multiple, 20.88% for the cycle efficiency, 0.8673 for the mirror efficiency and 0.8537 for the absorber efficiency.

Optimal Selection of Shale Gas Processing and NGL Recovery Plant from Multiperiod Simulation

The advent of unconventional oil and gas has resulted in a plethora of shale gas production, and it has been predicted to increase from 1 in 2000 to 40% in 2035 of the total US domestic gas produced. Since shale gas is both industrially economical and environmentally clean compared to oil or coal as a resource, many studies are focused on developing technologies to monetize shale gas. However, one of the key challenges in utilizing shale gas is its varied flow rate and uncertain compositional behavior. The flow rate of a shale gas well dwindles over a period of time, and the composition differs from well to well in the same shale play. This provides a challenge in designing a plant of optimum capacity for shale gas processing and natural gas liquids (NGL) recovery. In this study, the periodic variation in feed flow rate is addressed while designing shale gas processing and NGL recovery plant. Different shale gas flow rates as obtained from average shale gas rate declination curve of a shale play over the period of well life is used for the design of five different plants with varied capacities. The corresponding economics of different plants are compared to find optimal plant capacity. The base case scenario is further improved through incorporation of heat integration and optimization of individual units of the process using pinch and sensitivity analyses, respectively. Economic analysis of the optimum plant size is also performed with a constant feed flow rate over its entire life. From the economic analysis of various period studies, it is observed that the optimum plant design with highest ROI percentage has more than minimum ROI requirement both with declining and constant flow rate case, implying that the proposed processing plant is economically viable to implement. Finally, environmental impact of the base case and improved design is also discussed.

AATTENUATION—The Atmospheric Attenuation Model for CSP Tower Plants: A Look-Up Table for Operational Implementation

Attenuation of solar radiation between the receiver and the heliostat field in concentrated solar power (CSP) tower plants can reduce the overall system performance significantly. The attenuation varies strongly with time and the average attenuation at different sites might also vary strongly from each other. If no site specific attenuation data is available, the optimal plant design cannot be determined and rough estimations of the attenuation effect are required leading to high uncertainties of yield analysis calculations. The attenuation is caused mainly by water vapor content and aerosol particles in the lower atmospheric layer above ground. Although several on-site measurement systems have been developed during recent years, attenuation data sets are usually not available to be included during the plant project development. An Atmospheric Attenuation (AATTENUATION) model to derive the atmospheric transmittance between a heliostat and receiver on the basis of common direct normal irradiance (DNI), temperature, relative humidity, and barometric pressure measurements was developed and validated by the authors earlier. The model allows the accurate estimation of attenuation for sites with low attenuation and gives an estimation of the attenuation for less clear sites. However, the site-dependent coefficients of the AATTENUATION model had to be developed individually for each site of interest, which required time-consuming radiative transfer simulations, considering the exact location and altitude, as well as the pre-dominant aerosol type at the location. This strongly limited the application of the model despite its typically available input data. In this manuscript, a look-up table (LUT) is presented which enables the application of the AATTENUATION model at the site of interest without the necessity to perform the according complex radiative transfer calculations for each site individually. This enables the application of the AATTENUATION model for virtually all resource assessments for tower plants and in an operational mode in real time within plant monitoring systems around the world. The LUT also facilitates the generation of solar attenuation maps on the basis of long-term meteorological data sets which can be considered during resource assessment for CSP tower plant projects. The LUTs are provided together with this manuscript as supplementary files. The LUT for the AATTENUATION model was developed for a solar zenith angle (SZA) grid of 1°, an altitude grid of 100 m, 7 different standard aerosol types and the standard AFGL atmospheres for mid-latitudes and the tropics. The LUT was tested against the original version of the AATTENUATION model at 4 sites in Morocco and Spain, and it was found that the additional uncertainty introduced by the application of the LUT is negligible. With the information of latitude, longitude, altitude above mean sea level, DNI, relative humidity (RH), ambient temperature (Tair), and barometric pressure (bp), the attenuation can be now derived easily for each site of interest.

Triple-mode grid-balancing plants via biomass gasification and reversible solid-oxide cell stack: Concept and thermodynamic performance

Biomass-to-electricity or -chemical via power-to-x can be potential flexibility means for future electrical grid with high penetration of variable renewable power. However, biomass-to-electricity will not be dispatched frequently and becomes less economically-beneficial due to low annual operating hours. This issue can be addressed by integrating biomass-to-electricity and -chemical via “reversible” solid-oxide cell stacks to form a triple-mode grid-balancing plant, which could flexibly switch among power generation, power storage and power neutral (with chemical production) modes. This paper investigates the optimal designs of such a plant concept with a multi-time heat and mass integration platform considering different technology combinations and multiple objective functions to obtain a variety of design alternatives. The results show that increasing plant efficiencies will increase the total cell area needed for a given biomass feed. The efficiency difference among different technology combinations with the same gasifier type is less than 5% points. The efficiency reaches up to 50%–60% for power generation mode, 72%–76% for power storage mode and 47%–55% for power neutral mode. When penalizing the syngas not converted in the stacks, the optimal plant designs interact with the electrical and gas grids in a limited range. Steam turbine network can recover 0.21–0.24 kW electricity per kW dry biomass energy (lower heating value), corresponding to an efficiency enhancement of up to 20% points. The difference in the amounts of heat transferred in different modes challenges the design of a common heat exchange network.

Large Synchronous Motors as Drivers for Centrifugal Compressors in LNG Liquefaction Plants

Centrifugal compressors are major components in LNG liquefaction plants. Design considerations for optimal selection and proper sizing of compressor drivers include large starting torque requirements to enable compressor start from settle-out conditions and to prevent flaring the gas. Start-up capabilities of the motor driver can significantly impact the reliability and operability of the compressor train. The article reviews the requirements of API 617 and API 546, and addresses a number of important design and sizing criteria used to properly match the large centrifugal compressor and its synchronous motor driver. In the effort to optimize plant design and maintain the performance requirements, the article discusses optimization options, such as direct on line starting method to explore the motor rating limits, and the use of synchronous motors for power factor correction to eliminate or reduce the need for reactive power compensation by capacitor banks. This article presents a novel approach to show constant reactive power lines on traditional V curves. It also complements capability curves of synchronous motors with lines of constant efficiency.

Process simulation of high pH reverse osmosis systems to facilitate reuse of coal seam gas associated water

• First simulation of HERO process for CSG associated water treatment. • AqMB software comprehensively evaluated water treatment process options. • Identified major costs as acid consumption and power for air stripper. • Substantially improved economics by softening with sodium exchanged resin. This study has revealed the potential for a high pH reverse osmosis system to desalinate coal seam gas (CSG) (or coal bed methane (CBM)) associated water to facilitate beneficial reuse. To promote technology uptake in the CSG industry greater insight was required to understand the optimal water treatment plant design, the influence of solution composition, and process economics. Therefore, AqMB software was used to simulate the treatment of three CSG water compositions. Unit operations included: Filtration; Softening; Air Stripping; pH raising; and Reverse Osmosis. In all cases, the product water met reuse guidelines. However, process economics were less favourable as solution concentration increased. Hydrochloric acid regenerant for the cation resin was prohibitively expensive (up to A$17,554,687 per annum) and its use required air stripping to remove carbon dioxide formed in acidic conditions (power consumption 70 % of total process demand). Acid solutions also required large amounts of sodium hydroxide (up to A$4,546,682 per annum) to reach pH 10 wherein silica species were soluble. Use of a sodium chloride regenerated resin plus removal of the air stripping stage substantially improved economics to the point that the operational expenditure was predicted to be A$417 per ML of water treated; a value comparable to conventional reverse osmosis desalination units. AqMB software was advantageous in that it; simulated the entire water treatment process, not just reverse osmosis; generated robust predictions; and estimated operational costs. As an outcome, rapid interrogation of water treatment process designs was now possible.

Multi-scale membrane process optimization with high-fidelity ion transport models through machine learning

Innovative membrane technologies optimally integrated into large separation process plants are essential for economical water treatment and disposal. However, the mass transport through membranes is commonly described by nonlinear differential-algebraic mechanistic models at the nano-scale, while the process and its economics range up to large-scale. Thus, the optimal design of membranes in process plants requires decision making across multiple scales, which is not possible using standard tools. In this work, we embed artificial neural networks (ANNs) as surrogate models in the deterministic global optimization to bridge the gap of scales. This methodology allows for deterministic global optimization of membrane processes with accurate transport models – avoiding the utilization of inaccurate approximations through heuristics or short-cut models. The ANNs are trained based on data generated by a one-dimensional extended Nernst-Planck ion transport model and extended to a more accurate two-dimensional distribution of the membrane module, that captures the filtration-related decreasing retention of salt. We simultaneously design the membrane and plant layout yielding optimal membrane module synthesis properties along with the optimal plant design for multiple objectives, feed concentrations, filtration stages, and salt mixtures. The developed process models and the optimization solver are available open-source, enabling computational resource-efficient multi-scale optimization in membrane science.

Optimal design of geothermal power plants: A comparison of single-pressure and dual-pressure organic Rankine cycles

Four variants of the Organic Rankine Cycle (ORC) applied to geothermal power plants are optimized by means of numerical tools. These variants are: (i) the subcritical ORC with single-pressure heater (ORC/S/SC), (ii) the transcritical ORC with single-pressure heater (ORC/S/TC), (iii) the subcritical ORC with dual-pressure heater (ORC/D/SC), and (iv) the transcritical ORC with dual-pressure heater (ORC/D/TC). All the systems are recuperative and include a wet cooling tower. The objective function is the specific work output and design variables include operating pressures, mass flow ratios between the brine and the working fluid, superheaters effectiveness and cooling tower range. The systems are optimized for 20 different potential working fluids. The optimization is performed for inlet brine temperatures from 80 to 180 °C, and for ambient air wet bulb temperatures from 10 to 32 °C. The results show: (i) the superiority of ORC/D/TC for most of the cases, (ii) the relevance of using a dual-pressure heater at high sink temperature and low brine temperature, and (iii) the importance of choosing the right cooling tower range for an optimal power plant design.

Heat Integration and Batch Scheduling of Optimal Bioethanol Production

This paper examines bioethanol production from oil palm empty fruit bunches, 47,208 kgs of which could potentially produce 13,950 liters of ethanol per day, using the Aspen Plus Suite commercial software. Continuous ethanol production with efficiency, optimum plant design, process dynamics, and control were studied in a continuous process to minimize the loss of plant utility and prevent possible ethanol production failure, which can sometimes occur in real life. A Gantt chart was used for continuous production to obtain ethanol every 24.5 hours. The Gantt chart also increases equipment efficiency by up to 73.90%. Pinch analysis was conducted for minimal energy consumption at the plant. The proposed production without/with the heat exchanger network shows that the total costs per year are 1,343,861$/year and 886,253$/year, respectively. The implementation of a heat exchanger network can reduce costs by around 34.05% per year. The purification process is guaranteed by the controller of the column system due to its dynamic failure sensitivity. Details of the controller and the results in the dynamic mode are presented.

Comparison of Rotor Wake Identification and Characterization Methods for the Analysis of Wake Dynamics and Evolution

Optimal wind power plant design requires understanding of wind turbine wake physics and validation of engineering wake models under wake-controlled operating conditions. In this work, we have developed and investigated several different wake identification and characterization methods for analyzing wake evolution and dynamics. The accuracy and robustness of these methods, based on Gaussian function fitting and adaptive contour identification, have been assessed by application to a large-eddy simulation data set. A new contour-based method based on downstream momentum deficit has been considered. Uncertainties arising from wake-identification errors result in characterizations of the wake expansion, recovery, and meandering motion that differ by 19% of the rotor area, 4% of the freestream, and 15% rotor diameter, respectively.

Optimizing the SART process: A critical assessment of its design criteria

The SART (Sulfidization, Acidification, Recycling and Thickening) process has been successfully installed to recover cyanide and copper in seven worldwide gold cyanidation plants in the last 15 years. This technology has allowed the extraction of gold ores containing high grades of cyanide-soluble copper. However, this process has high capital costs and requires rigorous operational control, which have limited the installation of this process in the last 5 years. These drawbacks have not been overcome despite the successful operation of the SART process worldwide, limiting the interest of gold mining in performing new studies to support new developments in order to optimize plant design. In this framework, this study attempts to give a critical assessment of the design criteria of the SART process, based on experimental work focused on optimizing the SART process, challenging some conventional criteria typically considered in the design of SART plants. Thus, this study involves (i) a kinetic characterization, which allows determining the optimal residence time in the sulfidization reactor to be determined, (ii) oxidation/reduction potential (ORP) measurements with respect to the conversion, which suggest an adequate control logic that maximizes metal recovery in the sulfidization reactor, and finally (iii) settling tests to propose a correct criterion for recycling the settled slurry into the sulfidization reactor. These results show that it is possible to reduce the residence time in the sulfidization reactor by up to 1 min. Furthermore, it was established that the precipitation performance depends on the ORP value. Therefore, a control logic as a function of this parameter could be implemented. Finally, the settling results pave the way for establishing a new criterion for recycling the underflow slurry from the thickener to the sulfidization reactor, based on the copper concentration in the feed solution. Hence, this study proposes new design criteria for future SART plants, saving capital costs and optimizing the metal recovery control.

Optimisation in operating strategies for concentrating solar power plants

This research paper emanates from a directive of Eskom, South Africa’s power utility, to acquire a more full understanding of the technical, financial, and operational aspects of running a plant that is built on concentrating solar central receiver technology. Eskom owns the supply chain of electricity, this includes generation, transmission, distribution and management of the electric grid. The power utility identified the need to understand the operating capabilities of such plants within the South African electric grid. Therefore, to demonstrate the role and value Concentrated Solar Power (CSP) with storage offers to the electric grid, an array of operating strategies was identified during the design phase of Eskom’s Solar-1 project. These operating strategies explored in this paper are; Maximising Power Generation, Maximise Plant Revenue, Minimise Energy Dumping and Optimise Electric Grid (base load, peaking and auxiliary services). A Typical Meteorological Year (TMY) data set for the case study plant situated in Upington, South Africa is used to obtain the annual performance of the plant. In the ‘Maximise Power Generation’ operating strategy the developed model increases the power block efficiency by 1.8% utilising part-load operational optimisation instead of running the turbine at full load until depletion of the hot storage tank. This is achieved by reducing the total annual turbine stops by 50% to only 176 stops. The thermal energy saved from start-up of the power block is utilised for power generation instead. Results from the ‘Base Load’ operating strategy reduces the turbine stops by 85% to only 52 stops annually. Therefore, increasing the turbine online hours to 94.5%. The results from the ‘Minimise Energy Dumping’ operating strategy shows that it is possible to reduce energy dumping from 4.0% to 2.7% by effectively optimising the dispatch from the hot storage tank at selected times and increasing the power block output to 110%. As a result, an additional 7.4% of electric energy is generated from this saving. Under the ‘Optimise Electric Grid’ operating strategies, the complementary services of CSP to other non-dispatchable power generation technologies such as wind and photovoltaics are investigated. CSP with storage has the potential to increase the utilisation factor of distribution substations in grid constrained areas within the grid. This opportunity greatly increases the market of CSP and emphasises the value of thermal storage. However, plant design optimisation is required for this operating strategy as it greatly affects the plant performance of the case study plant. Although CSP with storage demonstrates continuous reliable power delivery to the electric grid it also provides complementary services to non-dispatchable generating technologies.

Diseño de una propuesta de una planta industrial para el aprovechamiento de residuos sólidos del proceso de curtido (Artículo Profesional)

En este artículo se presentará el diseño e implementación de una planta en una empresa industrial que permita el reaprovechamiento de los residuos sólidos (viruta de cuero) del proceso de curtido durante el proceso de elaboración de cueros para su utilización en un proceso complementario. Los problemas que se presentan al no contar con este tipo de procesos es incurrir en costos extras para tercerizar servicios para la disposición de estos desechos contribuyendo al aumento de la contaminación en la ciudad. Para implantar esta mejora será necesario realizar un análisis preliminar para poder elaborar la propuesta de diseño de planta más óptima que se adecúe a este proceso. La línea productiva está constituida por cuatro Unidades, la primera Unidad de Acondicionamiento de materia prima; la segunda Unidad de Mezclado y Moldeado; la tercera Unidad de Prensado y Secado, y finalmente la Unidad de Acabado.