Process units built prior to 1973, the year of the drastic rise in energy costs, were generally designed on a low capital cost investment basis for maximum rates of return. Energy saving equipment was included in the investment if it obviously improved the return on investment. No extensive engineering was directed at energy in the design phase.
In the current period of high energy costs, economics still dictates how much energy a new plant design can conserve. But the incentive to expend more engineer-ing time in the design phase to optimize the process with maximum energy conserva-tion has increased. Likewise, there is the economic incentive to return to older operating plants and retrofit them with additional energy saving equipment.
Similarly, years ago, plant operators had been instructed to minimize off specification production. They achieved this and reduced the amount of scrutiny and effort needed to operate the unit by producing a purer product than necessary. This results in an increase in energy usage. This section of the manual will cover changes in plant operation with minimal capital investments to reduce the energy required to produce one pound of product.
Your operating procedures were probably written before the large increase in energy cost drew attention to energy conservation as one primary objective. In addition, the operators are probably using the procedures only as a guide and have developed their own procedures based upon ease of operation.
3-A-1. Reducing the Reflux Ratio of Columns
The optimization of the reflux ratio of the distillation column can produce significant energy savings. The investigation can start by checking the operating manual and column performance specifications for the design conditions, including the reflux ratio. If the design conditions are no longer valid due to changes in feed composition or product requirements, it is recommended that a vigorous distillation calculation be made. If the calculations are very difficult, you can make use of commercial computer programs made available through various computing service bureaus (see section 4-B). The design reflux should be compared with the actual ratios controlled by each shift operator. The daily laboratory analyses of the column products are compiled and compared with the design specifications. If the column is operated at a reduced production rate, the design reflux rate should be calculated for this reduced rate.
It is extremely difficult to change people, even more difficult when it requires more work effort without visually seeing the results. If one operator was found who operated the column at a lower reflux ratio than the others, you might get the confidence of the operators by getting all the operators to maintain this ratio. If you merely write a note in the unit's operating log leaving instructions, you will probably not be successful in lowering the reflux ratio. You must work closely with the superintendent, foremen, and operators instilling confidence as you show the energy savings resulting from their efforts. If the operating depart-ment has monthly meetings for the supervisory people, you can use it as a forum to present your objectives, how you plan to approach them, and request their support and assistance. Later you can report progress and discuss problem areas.
Steam or fuel usage per pound of product can be tabulated daily along with reflux ratio, product purity, etc. and compared with column performance before the change. The savings in energy can be converted to a monetary value and reported to the operating people. As an alternate you might represent the energy savings as barrels of imported oil per year.
As the reflux ratio is reduced, a point will be reached at which the operators are overworked and having difficulty in maintaining product pur-ity. This is the opportunity to show your concern to the operators by backing off on the ratio.
3-A-2. Lowering Product Specifications
Sometimes, product specifications can be lowered. Who decided on the present product specifications? Are they justifiable? For example, the sales group may have had the product purity increased to justify selling more product and beating the competitors. The buyer may require a purity in excess of his real needs. Higher purity product requires more energy to be consumed per pound of product. Since the sales depart-ment has probably expressed an optimistic opinion as to the value of higher product specifications in the market place, an economic analysis based upon their opinions would most likely say to make no specification changes. A better approach may be to analyze the specification requirements for each type of user of the chemicals and determine if the higher specification is required. A different selling technique may retain the customer even if product specifications are lowered to save energy.
If the product from the column is feed to another unit in the plant, then the effect of lowering the purity on the other unit must be determined. Thus, the energy conservation project requires the additional collection and tabulation of operating data. A statistical approach may be required to fully interpret the results of changes due to the variability of the processes by changes in other parameters.
3-A-3. Lowering Pumping Costs.
When making an inspection of the unit for an energy audit, you should note any operation of two centrifugal pumps in parallel. Within the distill-ation unit, you can have reflux pumps, product pumps, feed pumps, pumpa-rounds, etc. with spares. Other examples are cooling water pumps in the water cooling tower and cooling pond systems.
If the pumping system was designed for one pump and the operator places the spare pump in service, too, he has not doubled the flow rate. Instead, each pump provides one half of the developed system flow rate and each operates at the identical head. To understand this, let us assume a centrifugal pump characteristic curve as shown in Figure 3-1. At 100 gpm of flow, one pump produces 130 ft of head. If identical pumps are on stream, the flow is 100 + 100 or 200 gpm at 130 ft of head. The characteristic curve for two pumps was developed this way and is also shown in Figure 3-1. The actual flow rate through the piping system is set by the intersection of the pump curve with the system head curve. Referring to Figure 3-1, the flow rate is 160 gpm with one pump operating and 172 gpm with two pumps on stream. In the latter case, each pump is handling one half the flow or 86 gpm.
The efficiency of centrifugal pumps varies with flow rate. Thus, pumps are selected in the design phase to operate at or near their highest efficiency. As seen in Figure 3-1, the pumping efficiency decreased from 46.5% at 160 gpm to 34% at 86 gpm. Assuming an electric motor efficiency of 95%, the energy used in both cases is determined as follows:
For one pump operating
For two pumps in parallel
By increasing the flow 7.5%, the energy requirements increased 60%. As an alternate to two pumps, the size of the impellers could be increased to handle the 172 gpm of flow with one pump. Assuming an efficiency of 47%, the energy required is:
Thus, 17.4 - 12.6 or 4.8 Hp was conserved. In section 5 of this report, the concept of investment equivalent for energy savings is developed. This is the amount of capital that can be invested to save a unit of energy. If new impellers were placed in the two pumps (one pump is the spare), the impellers would likely be expensed (if the motors were changed, the new motors would probably be capitalized). How long would it take to recover the expense of purchase and installation of the two impellers if the pump operated at 172 gpm with 0.95 on stream time? Assuming the cost of electricity at 3.0 cents per KWHr and the replacement expense of $800, the payout is:
(X) (.95)(4.8)(.746)(.030)= $800
where X = hrs
X = 7839 hrs or 0.9 years
Management should be receptive to this expenditure.
As chemical plants expand by adding more process units, additional cooling water is probably required. Usually, the existing cooling water lines are not replaced with larger lines, but additional pumps are added to handle the increased flow requirements. Suppose a new pump was purchased with an impeller that gives a higher head to compensate for the higher system pressure drop. The impellers of the existing pumps are replaced with larger diameter impellers. This is a minimal capital cost pumping install-ation, but what about energy usage?
As an example, Figure 3-2 shows the pump characteristics and system curves for a cooling water pumping system before and after expansion. Flow was increased from 1500 gpm to 2250 gpm. At the original flowrate, pumps A and B operated at 750 gpm each at 70 ft of head and probably at the best efficiency for these pumps. With the expansion, flowrate is at 2250 gpm at a head of 108 ft. At 108 ft of head, pumps A and B handle 1150 gpm or 575 gpm each. The efficiency of the two pumps probably dropped. Frictional energy increased 38 ft. The following calculations assume a $0.03 per KWHr of electricity:
Operating cost before change
Pumping cost = (55.8)(.746)(0.03)(24) = $29.98/day
Pumping cost per day per gpm = $.020
Operating cost after change
Pumping cost = (136.5)(.746)(.03)(24) = $73.33
Pumping cost per day per gpm = $.033
The pumping cost per gpm has increased 65% in addition to the capital costs, not a very efficient modification. Before making the pumping change, it may be possible to reduce frictional energy losses. The existing distribution system should be traced and pressure drop calculations made for sections of the system that appear to have high pressure drops. Maybe a short section of pipe could be replaced with a larger size. Maybe the proposed tie-in point for the cooling water to the new process could be moved closer with a small increase in piping costs, but a significant lowering of frictional energy losses.
Another possible way to cut energy usage is to limit cooling water flow through the exchangers. It is doubtful that the operating procedure covered this aspect. If flow is not throttled, the flow through an exchanger is determined by the DP available from the pumping system and the frictional energy losses in the exchanger and piping. For example, an unthrottled flow showed 8 psi across the exchanger. Design flow was for 800 gpm with a 5 psi
drop across the exchanger. Since flow is approximately proportional to the square root of the pressure drops, the flow rate isor 1000 gpm. An inexpensive type butterfly valve with a manual lock positioner could be in-stalled to throttle the flow to 800 gpm, saving 200 gpm of cooling water.
If a cooling water system operated at 6000 gpm and 50 psig before the exchanger flows were throttled and 5000 gpm at 55 psig after the throttling, how much energy was saved? Let us assume there is an improvement in efficiency from 0.50 to 0.52.
Horespower before change
Horespower after change
Electrical savings
Savings = (43) (.746) (24) (365) (.95) (.03)
= $8000 per year
Even better savings may be gained by changing impellers, etc. to give 5000 gpm at 50 psig or less. If a process fluid is being cooled by cooling water to l00ºF, but a fluid temperature of 120ºF is acceptable, it may be possible to use less cooling water or cooling water preheated by another source, thereby reducing cooling water flow.
Flow of liquids through piping transfer lines is generally controlled by the use of throttling valves. Past design practice has been to design the control valve to take from 25% to 50% of the system pressure drop. This gives the control valve a rangeability of approximately 50 to 1. The valve has converted work energy derived from electricity into frictional heat. Most processes don't require this much rangeability so a larger control valve with less pressure drop could replace the original valve, the rangeability being reduced say to ll to 1. Of course, energy savings can only occur if the pressure in the line is reduced, possibly by reducing the diameter of the pump impeller. The electric motor should also be replaced with one of lower horsepower that meets power requirements. Just installing a new control valve will be useless as the valve will throttle down until flow is controlled to the original point. Shinskey, in the "Control Systems Can Save Energy" article graphically discusses this energy saving idea.
3-A-4. Lowering Steam Usage
One of the most talked about energy wasters is steam leakage from "bad" steam traps and leaking fittings. Steam traps are blamed for being inefficient or worn out and causing as much as 10% of the generated heat from steam to be lost. Is this true or just a sales method to sell more traps? It turns out that steam leaks cause a significant energy loss.
Mr. Goyette, in his article "Estimating the Costs of Steam Leaks", (see appendix 7-C) shows the cost effect of steam leakage from various size holes (1/8", 1/16", and 1/32") in a 150 psig steam system. The cost was based upon incremented steam costs. An example showed that a 1-inch union was found leaking at a loss of $3000 per year. The repair cost was $50 or a six day payout. Of all the energy savings steps that the Tenneco plant did, Mr Goyette said the single largest contributor was steam-leak repairs. Steam traps will wear out. Armstrong Machine Works claim that the inexpensive disk type steam trap wears out in 6 months and should be replaced that frequently. If condensate is recovered, leaking traps can cause an excessive return temperature and cause failure of the condensate return pumps. Severe water hammer can occur as hot steam contacts conden-sate that has cooled below the temperature of the steam.
The following steps are recommended for saving energy in your steam condensate distribution system and starting an effective steam energy management program:
- Develop an estimate of the cost of steam leaks based upon your plant costs similar to the Goyette article described above. A method for demonstrating visually to plant people what these losses are can be made.
- Run a survey, recording all leaks, size, cost, and location.
- Check the operation of all installed disc traps used for drips and steam tracing. If found leaking, consider replacing with a more efficient type trap. Before replacing, check installation design and confirm trap size (not over or undersized).
- Check installation and operation of steam traps used on equipment using the sound detection method, the pyrometer method, or the glove method. The installation should be checked for proper trapping. Items checked include strainer, check valve, back pressure, orifice, and inert gas venting. Improper venting can cause a severe reduction in heat transfer rate.
- Check vent valves on steam jacketed equipment and kettles for proper operation (removal of inerts without steam loss).
- Start a preventative maintenance program to maintain the steam distribution system in excellent condition. If manpower is not available in maintenance, you can have the operating people maintain a simple log for their area of responsibility.
- Steam trap manufacturers will be happy to furnish information to assist in your energy saving program to reduce steam losses, but use your own economic costs to decide whether to replace, repair, or redesign the system.
There is insufficient published information to say that 10% of the steam is wasted by steam traps, but some major chemical companies have invested large amounts of manpower and money to replace or revise steam trapping systems in their plants.
3-A-5. Process Heaters
The Texas Industrial Commission has developed a manual specific to boiler and process heater efficiency. Consequently, our discussion of process heaters will be very limited, briefly covering the reduction of excess combustion air and reduction of stack temperature with small capital investment.
Control of Excess Air
According to Mr. A. M. Woodard, (see article, "Reduce Process Heater Fuel", in appendix 7-C), over half the total fuel consumption for refineries is for process heaters, the remaining for steam generation. These fired heaters can be improved from an energy efficiency viewpoint by reducing the amount of heat in the stack discharge. With the advent of the more accurate and simpler oxygen analyzers, the control of excess air in a fired heater can be automatically or manually controlled by the operator. Mr. Woodard's article details a method of sampling the flue gas, monitoring and controlling the system. Four systems are described, but system 3 is recommended. This consists of locating the draft and oxygen analyzer readouts in the control room, too. The operator can then monitor and control the operation of the heater or heaters with ease and comfort. Two safeguards are built into the system. stop installed to prevent full closure. failure, the positioner opens the damper. The damper has a mechanical If there is an instrument air failure, the positioner opens the damper. A simple stepwise procedure for heater adjustment is given on the last page of the article.
A target excess oxygen for the oxygen recorder with remote manual damper control was given in the article as 4.0% for gas and 4.5% for oil firing. More recently, manufacturers are indicating the oxygen can be controlled at 2%. The decision to go this low must be based upon the risk of temporarily going below stoichiometric conditions with possible explosion when the heater returns to excess oxygen conditions. Based upon figure 1 of Mr. Woodard's article, substantial reductions in heat input are accomplished by this approach. This modification will probably cost less than $5000, yet show considerable savings.
Recovery of Heat from Stack Gases
The amount of heat extracted from burning a fuel can be related to the flue gas or stack temperature. The extracted heat is defined as the heat absorbed by the process stream being heated and the losses from the furnace casing (generally around 2%). Thus the percent heat extracted is:
Heat available in Btu/lb of fuel at the Flue Gas temperature (FGI), divided by the Heat Content of the Combustion Fuel in Btu/lb times 100.
The lower heating value (LHV) of the fuel is used for efficiency calculations. The flue gas temperature depends upon the design condition of the convection section of the heater and the physical condition of the convective tubes. A reasonable FGT is the inlet process fluid temperature plus approximately 150º F. If your inlet fluid is at 300º F, the FGT is approximately 450º. A check of the FGT for your heater may show 500º F. Thus, your convective tube section may have lost some of its heat transfer ability by loss of fins on the tubes. This becomes a replacement expense.
If an exchanger (or reboiler or condenser) used to recover heat from a hot stream is slowly losing the amount of heat recovered because of fouling, when do you shutdown? This decision can be based upon maximizing heat recovered or minimizing the loss in profits. Three cases are described below:
Case 1---Decision based upon energy conservation
Given: An exchanger used to recover waste heat is rated at 11,000,000 Btu/Hr when clean before fouling. This exchanger slowly loses its heat transfer capability and the loss is estimated to be 10,000 Btu/Hr per day. A 12 hour shutdown is required to replace the tube bundle.
Find: Frequency of shutdown to maximize the energy recovery. Assume a 3500 day period of time.
A) At start of day l, heat transfer rate =
At the end of day l, heat transfer rate is
or 10,990,000 Btu/Hr
B) Let C = number of repairs during the 3500 day period. The heat recovered for any given day, X of the cycle is
The heat recovered for an entire cycle is
For the 3500 day period, total heat recovered is
Case 2---Decision based upon maximum profit, production rate not affected.
Given: Same conditions as Case 1
Each
Btu is worth $2
Each shutdown costs $10,000 in maintenance and $20,000 in profits.
Find: Frequency of shutdowns to maximize dollar savings
A) Savings =
B)
C)
Case 3---Decision based upon maximum profit, production rate affected by loss of heat transfer.
Given: Same conditions as Case 1 and 2, but production capacity is reduced by .05% per day. Each .05% loss in rate is $20 per day (20,000 x x .0005) in profits.
Find: Frequency of shutdowns to maximize dollar savings.
A) Savings =
B)
or
C)
Summary
Case 1 Max Energy Case 2 Max Profit Case 3 Max Profit Btu's total No. of cycles 105.5 9.9 24.4 Days per cycle 33 355 143
It is doubtful that management would agree to shutdowns every 33 days to maximize energy savings when the maximum profit occurs at 355 days (Case 2), or 143 days (Case 3). However, as the energy cost increases, the frequency of exchanger cleaning will increase for Cases 2 and 3. In a real plant, the assumptions of linear losses of heat transfer and production may not be true, but the principles of handling the decision making are still valid.
