Managing Peak Electric
Demand Provides Considerable
Economy at Spring Manufacturer

By  JAMES WATTERS, President, Delaware Valley Utility Advisors, Lansdale, PA
ROBERT DELP, President, Springtec Corp., Kulpsville, PA
JAMES FRANCOEUR, Technical Director, Francoeur Systems, Lansdale, PA

Energy management systems do not have to be large, highly sophisticated, capital intensive projects. With modest equipment, an understanding of how peak demands impact on costs, and some operator training and involvement, significant savings can be achieved. This has been the experience at Springtec Corp.-a manufacturer of industrial springs in Kulpsville, PA. (suburban Philadelphia). Their products range from heavy duty wire springs used for military vehicles, to constant force stainless sheet metal springs used for cosmetic displays. Wire sizes range from .006 in. springs used for refrigerator controls, up to 5/8 inch. In 1992 Springtec installed equipment for monitoring peak demand (see Fig. 1). Demand readings and projections were displayed in the heat treating furnace area of the plant. The operator of the furnaces was given complete responsibility for scheduling his workload in a way that reduced the plant's
total power demand.

     Spring manufacturing area involving controlled power demand.

     Most plant engineers know the old adage, "don't start up all the big motors at one time." Peak demand plays a major role in most utility rate schedules. Some utility rates are structured so that there us a charge for each kilowatt of peak demand. With others there is no actual demand charge per se, but different blocks of energy are charged at different rates. The amount of peak demand each month determines the number of kilowatt hours charged at higher or lower rates. The impact of setting high demand peaks will vary with the power company, the rate schedule, the time of day that peaks are set, and the type of load factor that the plant experiences. As a result, the economic benefits from lowering demand will be very site-specific.
     .

Fig. 1 Equipment for monitoring power demand: (1) master controller with downloading port (Francoeur Systems); (2) electric meter (PECO Energy); (3) junction box (PECO Energy); (4) remote display (Francoeur Systems).
Spring manufacturing area involving controlled power demand.

     With Philadelphia Electric's (now PECO Energy) high tension rate schedules, an extra kilowatt at peak demand can cost as much as $25. Demand in kW is measured in half hour intervals, and the highest half hour of energy output for the month is used as the demand value. Prior to 1993, Springtec was experiencing demand peaks of about 225 kilowatts. There was some seasonality  to  peak  demands. Springtec's production levels were lower in the summer. A three-year history of peak demands is shown in Fig. 2
 

Fig. 2 Three year power demand history at Springtec Corp.
 

Fig. 3 Annealing ovens with baskets of springs for stress relieving.

     The largest loads in Springtec's shop are required for three annealing ovens used for low temperature stress relief. They are rated at 29 kW each. These ovens, Fig. 3, are operated by a single furnace operator who was also responsible for a vapor degreaser. The next largest loads are for surface grinders with 20 Hp motors. The plant operates with two shifts.
In reviewing electric power costs and usage, Delaware Valley Utility Advisors (DVUA) recognized that peak power demands were generally being set when more than one of the large ovens were being started up at the same time that both of the big grinders were operating. Fig. 4 shows the instantaneous demand for the shop on a typical day. Like many production facilities, it is common to set a demand peak an hour or two into the morning shift. Equipment is generally not started up immediately in the morning. A few machines are started up while others are being set up; that is, furnaces being loaded, etc. Springtec's normal demand peak occurs somewhere around 9:00 or 10:00 in the morning. There is a demand dip during the half hour that includes break time, followed by a surge as everyone goes back to work and production runs at maximum.

Fig. 4 Instantaneous power demand on a typical day at the spring manufacturer.
 

Instrumentation
     Demand monitoring equipment (from Francoeur Systems, Inc. of Lansdale, PA) was installed.  Philadelphia Electric (PECO Energy) installed a new power meter with a junction box for Springtec's use. The junction box provided a series of pulses and these signals were fed into a master controller. It calculated the projected demand for the current demand interval based on the information received up to that point in the half hour period.
     The monitoring equipment included a digital display, with numbers approximately two inches high located near the furnace operator. The display indicated the peak demand, which had already been set for the current month, and the projected demand for the current half-hour period.
     Red and yellow warning lights, included on the display panel, indicated when preset levels that had been programmed into the system were reached. This allowed the operator to know, five or ten minutes into the period, whether a new demand peak was going to be set. In addition to warning that energy management goals were going to be exceeded, the monitoring equipment also included a terminal port for downloading the information into a computer. Software (Francouer Systems) provided for retrieval of the demand data and downloading it into a PC (IBM) in a form that could be analyzed and manipulated using standard business software packages. Thus, actual profiles of demand data could be stored, recorded, and further analyzed.

Practical Use of Operating Data
     Responsibilities of the furnace operator, Gene Barndt, were to load and unload the furnaces with baskets of springs, program the furnaces, and record pertinent information for quality control purposes.
     Within a few weeks it became apparent to Mr. Barndt that he should try to run large loads early in the morning. Normally the big grinders did not start up until a couple of hours into the shift. The heaviest power draw on the heat treating furnaces was, of course, when they were coming up to temperature. The springs with the largest diameter wire can represent a furnace load several times that of a smaller diameter load. Load size can range as little as 2 lb. to as much as 1000 lb. Although it may seem unusual to run a cycle with 2 lb. of material, this load of .006 in. wire can represent as many as a quarter-of-a-million pieces.
     Like most manufacturing facilities, production tends to dictate what and when equipment and process runs. However, there was enough slack in furnace time to allow some flexibility in production scheduling. Again, with experience, it was possible to determine if the day was going to be one where high demands were being set, or a day where it was likely that a peak demand for the month would be registered. Sometimes based on judgment of electrical demands and production requirements, it was decided to delay a furnace load by half an hour or an hour, thereby achieving an economic advantage.

Fig. 5 Reduction in power demand peak (red) after initiating the demand monitoring program.

Results
     Equipment was installed to allow monitoring, displaying and analyzing demand patterns. As shown in Fig. 5, the results from this program were dramatic. After the few months in operation, peak demands dropped by 21%. Power level continued to remain at 170 to 185 kW-an average drop of 47 kW resulting in savings of $670 per month for this spring manufacturer.
     These savings were achieved with no disruption in production and very modest investment requirements, upon working with two consultants, an equipment vendor and a utility rate consultant. The key to the program was to provide operating personnel with the information necessary to manage power demands, and then simply allow them to do their job.
 

This article was published in the February, 1995 issue of Industrial Heating: The Journal of Thermal Technology.