ECO-optimization of pre-treatment processes in metal finishing

Abstract

Saving of bath chemicals and freshwater consumed within metal finishing systems and the reduction of wastewater produced can be important steps towards more sustainable processes. Furthermore, the application of regenerators is usually energy intensive. Accordingly, in an “eco–eco” (ECO) trade-off, importance should be given to both economic benefits as well as to several ecological measures. The approach presented here is the synthesis of an ECO-optimal reuse and recovery network (RRN) based on a simultaneous multi-objective system optimization model with dynamically chosen environmental objective. It includes a conventional cost objective and the potential environmental impact of the system. This second objective is represented by the maximal relative increase of indicator results. This, in turn, derives from a sensitivity analysis of life cycle inventory analysis (LCIA) integrated into mixed integer nonlinear programming (MINLP). This means, the LCA-rule, which claims that environmental indicators must not be aggregated, can be adhered strictly. In the case study, the synthesis procedure for eco-optimal design is applied to a large scale phosphating. Depending on the weighting ratio for the representative impact factor and the cost in the objective function, Pareto-optimal solution sets were calculated allowing to assess different eco-optimal RRN alternatives and to find a way out of the main dilemma between the contradicting behavior of energy demand and wastewater production. Up to 74$\%$ improvement in terms of environmental indicators could be realized at low cost increase.

Introduction

Minimizing environmental impacts is supposed to be achieved by reusing or recovery of the resources as much as possible. When the high quotes of recycling are aimed, this task requires hidden wastes such as energy consumption by the water, bath chemical recovery system and high costs (Cohen Hubal & Overcash, 1995). In the course of high energy consumption, there are both ecological burdens and economical charges. This dilemma indicates the need for multi-objective optimization debate allowing an ”eco–eco” (ecological and economic) trade-off.

This debate is investigated by means of two basic approaches; the first one, namely the performance of an impact assessment, such as like the standard life cycle assessment procedure followed by the evaluation of the most environmental benign system between the alternatives (Shonnard & Hiew, 2000) and the second one, in which more process integration methodology in the form of final comparative assessment is applied Bagajewicz, 2000, Alva-Arga’ez et al., 1998, Dantus and High, 1999.

In another multi-objective optimization methodology series of single objective optimization are performed on condition that all objectives except one are converted into constraints (Azapagic & Clift, 1999). A hybrid methodology application for minimization of cost and emissions, where a minimization of single objective optimization problems algorithm is developed, has been introduced through the work of Diwekar and Fu (2004). Further development, in which a combination of single objective and multi-objective optimization debated within a two layer algorithm for performing a hybrid method, is discussed in Kheawhom and Hirao (2004). It should be noted that, this combination of single and multi-objective optimization is supported by a computer-simulation model that handles the uncertainty using multi-period and stochastic optimization formulations. Additionally, hybrid form of both methodologies that integrates impact assessments, such as environmental considerations and socioeconomic factors with process synthesis by means of multi-objective optimization, was introduced by Erol and Thöming (2005).

The inherent disadvantage of hybrid methodologies is the requirement of quantitative weighting factors that are prone to individual interpretations and the subjectivity by ranking that is involved. However, in most instances both economic and environmental objectives are aggregated into a single objective function using the analytical hierarchy process (AHP) (Chen, Wen, Waters, & Shonnard, 2002). In hybrid approach, this consists of both quantitative and qualitative weighting.

All these methodologies provide relationships between mathematical programming and decision support systems leading to an expert system. An example of such an environmental decision support system is developed by Rizzoli and Young (1997), which integrates both the identification of the general attributes of the environment and the system sand simulation models. Similarly, to aid the decision making in the area of facility planning management, Han, Kim, and Adigüzel (1991) developed a combination of mathematical optimization model with a database management system as an expert system. In this work, an ECO-optimization approach is introduced that can lead to a decision support tool in modelling category of the expert systems, as it is categorized in Liao (2005).

Metal surfaces in process plants are usually coated with oxides, grease or dirt arising from prior operations such as working, storage and transport. This causes a need to have a pre-treatment in order to achieve a sufficiently cleaned surface which suites consequent process requirements.

Pre-treatment stages in metal finishing have significant energy, chemical and water consumption as mentioned in Weng, Wang, and Zhang (1998), especially for different phosphating processes. These are among the most widely used pre-treatment processes of metal finishing in industrial practice (cf. Fig. 1) for the purpose of improving the adhesion and service life of surface coatings under corrosive conditions on metal bodies prior to painting. For a uniform pickling effect, a prior degreasing and cleaning in aqueous alkaline or organic solutions is a necessity.

The principle of phosphating depends on a treatment with an aqueous solution of an inorganic acid, the so called pickling. This method converts the oxides tightly adherent to the metal surface to a soluble form that can be removed by a rinsing process. Usually in practice, pickling with phosphoric acid works in two stages. At first, the work piece is deeped into a 15–20 wt.% concentration, in which the formation of soluble iron–phosphate and removal of the impurities from the surface take place. This is then followed by passivation that occurs in a 1– 2 wt.% solution and forms a protective secondary and tertiary phosphate film on the metal surface.

The initial basic chemical reaction involved in phosphating process describes acid attack on the metal surface (cf. Eq. (1.1)). In this reaction, acid neutralization takes place (pH rises) with the increase in concentration of metal ions (${\text{M}}^{2+}$) (Cape, 1987). During the subsequent reaction, the production of metal ions (${\text{M}}^{2+}$) and the resulting consumption of acid cause a precipitation of divalent metal salts (cf. Eq. (1.2)).

$$2{\text{H}}^{+}+\text{M}\to {\text{H}}_2+{\text{M}}^{2+}$$

$$2{{\text{H}}_2{\text{PO}}_4}^{-}+3{\text{M}}^{2+}\to {\text{M}}_3{{\text{PO}}_4}_2+4{\text{H}}^{+}$$

These reactions can be accelerated and improved by introducing a mechanical and electrolytic action, which can unfortunately concurrently result in a very active surface, on which fresh rust very readily forms. To limit this effect a combination of pickling and passivating action is desired. On account of its film-forming properties, phosphoric acid is often a preferred pickling agent. The passivation film formed during phosphating provides at least a temporary protection against corrosion and a surface suited to an organic coating.

To achieve better phosphating coating and corrosion resistance, certain accelerators and additives (e.g. refiner agents, surfactants) are incorporated into the bath solution resulting in varying bath compositions. The type of bath accelerator used differs according to the applied phosphating technique, such as immersion and spray processes. Among the commonly used accelerators are nitrite, nitrate, chlorate and hydrogen peroxide. However, the most widely used accelerators are nitrite/nitrate and chlorate accelerators Rausch, 1990, Freeman, 1990. The relative importance of accelerators to the phosphating quality was pointed out by Sankara Narayanan (1996a), in which the detrimental effect of concentration fluctuations was discussed. As a consequence of the intensive implementation of phosphating as a pre-treatment process, the potential environmental burden of process occurs by the chemicals and large amounts of consumed water and energy; thus there is a need for a systematic assessment method for these potential environmental impacts (Weng et al., 1998).

In recent developments in metal finishing techniques, the phosphating formulations are modified by incorporation of additives like nickel and/or manganese ions to suit the needs of the electrophoretic paint finishing (Sankara Narayanan, 1996b). It has been observed that the inclusion of manganese and nickel ions in zinc phosphating bath causes the refinement of the crystal size and improvement of the corrosion resistance of the resultant phosphating coating. Other additives such as calcium modified zinc phosphating lead to the reduction in grain size and the improvement in compactness of the coating and corrosion resistance. Moreover, such modification results in the development of trication phosphating bath as an alternative to the recent conventional phosphating processes in metal finishing (Sankara Narayanan, 1996a).

Such a “trication phosphating”, low zinc phosphating with additional manganese and nickel ions, with immersion processing is selected as a case study. In this phosphating solution, products such as Zn(II), Ni(II), Mn(II), phosphoric acid, oxidation agents like chlorate (${{\text{ClO}}_3}^{-}$) and chloride (${\text{Cl}}^{-}$) ions are to be found (Brouwer, 1999).

In its current form, the considered process consists of an immersion zinc phosphating bath followed by three rinsing stages and a precipitation unit for metal ions (Zn, Ni etc.). Rinsing stages generate wastewater containing additives and their degradation products at relevant concentrations [mg/l] like ${\text{Zn}}^{2+}$: 115; ${\text{Ni}}^{2+}$: 57; ${\text{Mn}}^{2+}$: 59; ${\text{Na}}^{+}$: 430; ${{\text{H}}_2{\text{PO}}_4}^{-}$: 1500; ${{\text{ClO}}_3}^{-}$: 200; ${\text{Cl}}^{-}$: 300. Also for the precipitation stage there is great amounts of chemical consumption.

For a more sustainable phosphating process, the common rinsing system could be substituted with zero-water discharge reuse and recovery network (RRN) introduced by Thöming (2002). Since the zero-water discharge RRN are energy intensive, the aim of this work is to find a method to identify optimal process design with respect to wastewater production, energy demand and cost which leads to ECO-reuse and recovery networks (ECO-RRN).