Multi-stage production process

This chapter focuses on a new approach that allows for the comprehensive planning and optimization of multi-stage production processes - the quant-based combinatorial optimization. First, a distinction is drawn between classical approaches such as Linear Programming (LP) and the quant-based combinatorial approach. Before going into the special characteristics and requirements of the process industry the one model approach with quant-based combinatorial optimization is introduced. Then we will give two examples of how this new approach is applied to real life problems. 

The strategic and tactical model, for optimization of manufacturing and distribution networks, incorporates more than 100 customer zones or markets and dozens of product categories, manufactured in more than 250 production lines. Because of the diverse international nature of the problem, many international factors (such as domestic and international freights, transfer prices, taxes and duties) are considered. Some product categories have multi-stage production process and are processed in multiple echelons (Figure 8.3). Multiple productivity rates are considered for multi-product machines and products. The costs (facility overheads, production line fixed and variable costs and raw material consumption costs) are appropriately included in the optimization model. 

The possible mechanism of ionization, fragmentation of studied compound as well as their desoi ption by laser radiation is discussed. It is shown that the formation of analyte ions is a result of a multi stage complex process included surface activation by laser irradiation, the adsoi ption of neutral analyte and proton donor molecules, the chemical reaction on the surface with proton or electron transfer, production of charged complexes bonded with the surface and finally laser desoi ption of such preformed molecules. 

Tekmaheat Oy manufacture three kiln systems a compact chamber of 15-25m per batch, having an annual production capability of 1600-3000 m a drive-through chamber having a volume of 40-60 m, with an annual capacity of 5000-10 000 m and a multi-stage production line, in which the timber is pre-heated, dried, heat-treated and conditioned in a continuous throughput process, which has a capacity of 30 000-50 000 m per annum. 

To show how this complexity arises, let us consider the manufacture of a fictional, fine chemical product used in the pharmaceutical industry, called FCP1, by a multi-stage, batch process at a rate of 100 tonnes per year. 

New spatial forms of carbon - fullerenes, nanotubes, nanowires and nanofibers attract significant interest since the time of their discovery due to their unique physicochemical and mechanical properties [1-3]. There are three basic methods of manufacturing of the carbon nanomaterials (CNM) - laser evaporation, electric arc process, and catalytic pyrolysis of hydrocarbons. However, the multi-stage manufacturing process is a serious disadvantage for all of them. For example, the use of organic solvents (benzol, toluene, etc.) for separation of fullerenes from graphite soot results in delay of the synthesis process and decrease in the final product quantity. Moreover, some environmental problems can arise at this. 

LIMB [Lime/limestone injection into a multi-stage burner] A flue-gas desulfurization process used in Germany and Finland. Dry, ground limestone is injected directly into the combustion chamber. This reacts with the sulfur dioxide, and the dry particulate product is collected downstream together with the ash. The process is suitable only for those systems which limit the maximum combustion temperature by staging, in order to minimize the production of oxides of nitrogen. 

Figure 2.3 illustrates a process in which the outer layer of a product is quickly frozen as a congealed crust. These CRUSToFREEZE plants have a capacity between 1500 and 5000 kg/h and require 0.5 to 0.8 kg LN2 pro kg of product, which has to be frozen totally on a conveyor belt. Figure 2.4 shows the product exit of the plant in Fig. 2.3. The freeze drying of coffee and tea extracts, fruit pulps or small pieces of meat require a multi- stage pretreatment. The granulated end product from coffee and tea extracts should have a defined grain size, a desired color, and a predetermined density. Fruit pulps should become granulated, with the appearance of fruit pieces, while meat pieces should not stick together like a small meat ball, but be recognized as single pieces when presented in a meal.

Finally, the value chain network can have a single value-creation step or multiple value-creation steps in production and distribution where raw materials are processed through several production and distribution steps before being sold to the market. In the work, a multi-stage value chain network is considered. 

In order to reflect these lead times, the concept of a timestamp is introduced. Timestamp is used in computer science documenting the system time when a certain event or transaction occurs e.g. for logging events (N.N. 2007). In the context of future inventory value planning, the time-stamp marks the period, when the first raw material has reached a certain stage in the value chain network included into a specific product. In the example illustrated in fig. 57, the raw material is processed in the same period to be converted into product 1. Therefore, all four value chain steps indexed from one to four occur in the same period and have the same time-stamp one. Conversion into product 2, however, requires additional time caused by production lead times, safety inventory and/or transportation time, that the steps indexed with five and six have a time stamp of two. The timestamp reflects that the inventory value of product 2 is not based on the raw material costs from the same period but based on the raw material costs from the previous period in order to reflect the lead time. Consequently, value chain indices and timestamps are defined for all steps and can cover multiple periods reflecting that raw materials in a global complex multi-stage value chain network can take several months, until they are sold as part of a finished product to the market. 

---