Controlling the humidity of cupola blast air.

With the installation of a dehumidifier, U.S. Foundry lowered coke consumption requirements for a more cost-effective cupola operation.

The operation of a foundry cupola can be influenced by the weather; specifically, the ambient relative humidity can have a substantial effect on several performance measures. If the moisture content of the air entering the cupola is high, the operation is adversely affected, whereas on a dry day, when the air entering the cupola is low in moisture, beneficial effects can be noted. In many regions, operation the cupola is much more uniform during the winter, when humidity is generally low, compared to the summer season, when the moisture content of the air varies widely, often with sudden humidity fluctuations, within the course of a single day.

This article reviews the effect of blast air moisture content on cupola operations and discusses the remedy implemented by U.S. Foundry and Manufacturing Corp., Miami, and the cost impact experienced.

Reaction Results

From a chemical standpoint, water vapor in the blast air reacts with incandescent coke in the cupola according to the following water-gas reaction:

[H.sub.][O.sub.(gas)] + [C.sub.(coke)] + Heat = CO + [H.sub.2]

This ever-present reaction can have several negative effects on cupola operation and metal quality if left unrestricted. The reaction is endothermic; hence, it absorbs heat that would otherwise be available for melting, thereby decreasing the melt rate. Furthermore, the reaction consumes coke to form carbon monoxide (CO), wasting useful raw materials and forcing the need for additional combustion downstream while potentially depriving the iron of carbon (C).

The contact between water vapor and molten iron at high temperatures can produce objectionable results in the properties of iron. Both C and silicon content diminish in the presence of elevated moisture content, rendering lower tensile strength and increased chill. Since heat is necessary for the reaction, both spout and pouring temperatures are reduced at elevated water vapor levels. Industry handbooks have published detailed charts to illustrate the relationship between moisture and several properties [ILLUSTRATION FOR FIGURE 1 OMITTED]. While the presence of some water vapor is favorable, excess levels detract from many key measures, producing changes in the chemical composition of iron and related changes to the mechanical properties.

Moisture Decomposition

The C from coke decomposes water vapor in the blast air to form CO and hydrogen. This reaction can be shown to be endothermic by calculating the heat of reaction and referencing the heats of formation (Btu/lb mole) for each element.

[H.sub.2][O.sub.(gas)] [C.sub.(coke)] CO [H.sub.2] / 104,036-4280 49,549 0

Thus, the heat of reaction is calculated as:

49,549 Btu + 4280 Btu - 104,036 Btu = -52,210 Btu or -2898 Btu/lb [H.sub.2]O

indicating that heat must be supplied for the reaction. In the cupola, this is done through the combustion of coke. To make up for this loss in heat for melting and to replace the C consumed in the reaction, additional coke must be added to each charge, balanced with the water vapor present in the blast stream [ILLUSTRATION FOR FIGURE 2 OMITTED]. In the water-gas reaction, 1 mole of C (12 lb) is required to decompose 1 mole of water (18 lb). Thus, the decomposition of 1 lb of water vapor requires the burning of approximately 0.72 lb of coke (0.67 lb of c) to react with the water vapor and an additional 0.43 lb of coke (0.40 lb of C) to produce the heat required for the reaction. Consequently, every 1 lb of water vapor that enters the cupola consumes about 1.19 lb of coke.

Additionally, each grain of moisture per cubic foot of blast air replaces 0.28% of the blast volume. Naturally, supplementary blast air also must be supplied in order to burn the additional coke required, demanding an overall blast increase of approximately 1% for each grain per unit volume. A quick rule of thumb is that an increase of 1 grain of water requires 1 lb of coke/hr/100 cfm of blast and 1% increase in blast air volume. Hence, an increase in blast moisture of 3 grains at a blast rate of 9000 cfm would require the addition of 300 lb of coke/hr and an increase in blast air volume of 300 cfm. [This is not an exceptional amount of moisture; 3 grains of water/cu ft of air represents 45% relative humidity at 70F (21C).]

Exact computation of the offsetting coke addition under various climatic conditions is fairly simple using the parameters discussed above. Knowing the ambient (dry bulb) temperature and relative humidity, the total water vapor content or humidity ratio can be interpolated using a standard psychrometric chart. Considering that 1 lb of blast air will melt 1 lb of iron, the air required for a given melt output easily can be calculated, making sure to correct the air density for temperature and pressure. Multiplying the humidity ratio, expressed in pounds of water vapor per pound dry air, by the air required, in pounds, yields the water content of the blast air. This is then multiplied by the coke loss factor of 1.19, determining the total additional coke required in compensating for the moisture. Conversely, this is the potential coke saving if dehumidification is used.

Blast Dehumidification

Blast dehumidification involves the removal of water vapor (humidity) from the blast air stream in order to minimize the addition of offsetting coke and improve casting characteristics. This can be accomplished in a number of ways, the three most common being refrigeration. absorption and adsorption. The refrigeration method involves lowering the temperature of the blast air to condense excess moisture. The air is passed through a chilled water spray or across a cooled, finned coil, both requiring a refrigerating medium. Drawbacks to this include a large initial capital investment as well as high operating cost. considering the tremendous capacity required for cooling the large volume of air to a low enough temperature. Further. the need then arises to charge additional coke in order to reheat the air to temperature prior to dehumidification.

The absorption method involves the removal of water by chemical means, with an accompanying change in the chemical composition of the absorbing agent. This is usually accomplished by passing the air through a spray containing a hygroscopic salt in a water solution. The density and temperature of this solution determine the degree of moisture removal, and controlling these allows the moisture content of the treated air to be held constant. The solution is cycled through a regenerator, removing the moisture and stabilizing density and temperature. This method, though complex, allows for both the addition and extraction of moisture in order to maintain a constant humidity ratio in the blast stream.

The adsorption method also involves the removal of water by chemical means; however, there is no accompanying change in the chemical composition of the absorbing agent. The process air is passed through a bed of adsorbing material, usually an inert desiccant such as silica gel, and moisture is transferred from the air to the media. The adsorbent is then reconditioned with heat to drive off the moisture so it may be reused. The system contains a rotating wheel or alternating beds, allowing for the simultaneous treatment of both the saturated media and the blast air stream. This method, though quite practical for many installations, presents a fluctuating moisture loading of the outlet air, varying as a function of the inlet content, and does not allow for the hydration of blast air that is too dry.

Case Study

As a municipal casting manufacturer, U.S. Foundry and Manufacturing faces fierce competition with import castings and therefore sees cost reduction as a major consideration in capital improvements. Using this philosophy, the foundry replaced its existing cupola in 1990. After comparing the benefits to that of an electric furnace, a new cupola was chosen as the means to produce molten metal for its foundry operations, employing many modern and cost-efficient innovations. In the realm of blast conditioning, various enhancement methods were employed, including the use of a recuperative hot blast, an oxygen enrichment system and a blast air dehumidification system. Operating a foundry in South Florida provides many unique challenges in obtaining raw goods, such as coke, which must be brought in from as far as 800 miles away, adding significant shipping costs. As such, methods of coke reduction were considered imperative. Furthermore, with the notorious local rainy season and year-round high humidity, the foundry recognized the advantage of addressing the high moisture content of the blast air.

Daily minimum, maximum and average temperatures, as well as average daily relative humidity readings were collected for 1997 [ILLUSTRATION FOR FIGURE 3 OMITTED]. The lowest daily relative humidity reading of the year was 46% on January 18, with the highest reading of 96% recorded on December 4. For the purposes of analysis, monthly averages were taken in lieu of using the daily data, given the moderately minor fluctuations within the month. Additionally, no attention was paid to the fact that temperature and humidity both vary throughout the day, being that the two-shift operation spans this range adequately. Monthly average temperatures ranged from 68-83F (20-28C) with the overall average being 77F (25C), while monthly average relative humidity ranged from 71-81% with an average of 75% for the year.

While relative humidity remained fairly constant, absolute humidity (the actual measure of existent water grains per pound of air) varied widely, from a low of 74 grains in January to a high of 132 grains in July [ILLUSTRATION FOR FIGURE 4 OMITTED], during which the moisture in the blast air exceeds 1200 gal/day. The optimal moisture level in the cupola is 3 grains/cu ft of blast air, or 40 grains/lb of air at 70F (21C). At an average melt [TABULAR DATA FOR TABLE 1 OMITTED] rate of 14 tons/hr and a climate of 77F (25C) and 75% relative humidity, untreated blast air would introduce 417 lb of water/hr requiring the addition of approximately 500 lb of coke/hr.

The foundry elected to install a desiccant/adsorption-type dehumidifier for use in the new cupola project in 1990. This dehumidifier was sized for the maximum blast air demand of 9000 scfm, and was chosen for its efficient water removal, low operating and maintenance costs, and moderate initial investment. As the relative humidity level is nearly always high, the unit is not required to add moisture, making adsorption a good choice for the climate.

The model uses a 60-in. ceramic honeycomb rotor where synthesized silica gel is bonded to a ceramic substrate forming sealed air passages. A natural gas burner is used to preheat the air used to reactivate the silica gel for its next pass, driving the moisture off the rotor with 3000 scfm at a temperature of 280F (138C). The rotor uses a synthetic, non-corrosive, chemically inert silica gel with a high capacity for water removal without the need to drain liquid condensate. Technical performance characteristics were established assuming an ambient inlet temperature of 90F (32C) and a relative humidity of 50%, bearing moisture content of 106 grains/lb of air. The system then was rated for a moisture removal rate of 266.1 lb of water/hr when operating at a blast rate of 9000 cfm, resulting in an outlet temperature of 139F (60C) with moisture content of 60 grains/lb and reflecting a grain depression rating of 46. The physical size of the unit is 7 x 15 x 8 ft.

The cost of this dehumidification unit was $37,700 plus nominal installation expenses. Operating costs included gas consumption of 840 cfh facilitating a temperature rise of 190F (88C) at 3000 cfm and modest electrical costs associated with the rotor drive and reactivation blower totaling about $6/operating hr. Additionally, the desiccant wheel needed replacing after approximately 10,000 hr at a cost of about $8500 for a total operating cost of roughly $7/hr, with all other maintenance costs considered incidental. Thus, the annual operating cost of this unit for a 2-shift operation is $28,000.

Using the performance curves supplied by the vendor, the amount of water vapor still present in the cupola blast airstream after dehumidification can be extrapolated. These values can then be used to calculate coke savings, as shown in Table 1. As indicated, in the absence of a dehumidification unit, an average of 35 lb of coke would have to be added to each 1-ten charge to account for the moisture in the blast stream. At a melt rate of 14 tons/hr, this amounts to more than 980 additional tons of coke over the course of the year. For 1997, over 112,000 gal of water were eliminated, saving more than 550 tons of coke through the removal of moisture. This represents an annual benefit of 3-5 times the operating costs, depending on the prevailing cost of coke.

As anticipated, the coke usage in the cupola, represented by the iron-to-coke ratio, varied closely with the melt rate, while employing dehumidification [ILLUSTRATION FOR FIGURE 5 OMITTED]. There was little correlation between the iron-to-coke ratio and the moisture content of the treated blast air, except in the most humid summer months. This indicated that the unit met its objective of lowering the water content to the point where it had minimal effects on cupola operations. Given that and the cost savings, the unit has proven to be a valuable tool in helping the foundry control its melt process while simultaneously lowering costs. Even in less extreme climates, the advantages of cupola blast air dehumidification can be beneficial, if the blast air volume is high or if variations in iron properties must be finely controlled.

This article was adapted from a presentation at the 1998 AFS 2nd International Cupola Conference. Conference proceedings are available from AFS Publications at 800/537-4237.