* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Evaluation of a computer simulation of the radiant heat curing
Survey
Document related concepts
Copper in heat exchangers wikipedia , lookup
Heat equation wikipedia , lookup
Radiator (engine cooling) wikipedia , lookup
Cogeneration wikipedia , lookup
Intercooler wikipedia , lookup
Solar air conditioning wikipedia , lookup
Thermoregulation wikipedia , lookup
Thermal conduction wikipedia , lookup
Underfloor heating wikipedia , lookup
R-value (insulation) wikipedia , lookup
Dynamic insulation wikipedia , lookup
Transcript
University of Richmond UR Scholarship Repository Master's Theses Student Research 1988 Evaluation of a computer simulation of the radiant heat curing process for primary URD electrical cable Grover L. Stell Follow this and additional works at: http://scholarship.richmond.edu/masters-theses Part of the Business Administration, Management, and Operations Commons Recommended Citation Stell, Grover L., "Evaluation of a computer simulation of the radiant heat curing process for primary URD electrical cable" (1988). Master's Theses. Paper 1047. This Thesis is brought to you for free and open access by the Student Research at UR Scholarship Repository. It has been accepted for inclusion in Master's Theses by an authorized administrator of UR Scholarship Repository. For more information, please contact [email protected]. ;J UNIVERSITY OF RICHMOND LIBRARIESt.: 1111111111111111111111111111111111111111111111111111111111111111 3 3082 01027 9056 EVALUATION OF A COMPUTER SIMULATION OF THE RADIANT HEAT CURING PROCESS FOR PRIMARY ORD ELECTRICAL CABLE An independent research project submitted in partial fulfillment of the requirements for the MBA degree by Grover L. Stell, Jr. Executive MBA, Class Of 1988 University of Richmond March 11, 1988 TABLE OF CONTENTS .................................... 1 PRIMARY URD CABLE ..................... .......... 2 MODELLING THE PRODUCTION PROCESS ...... ......... 15 RESULTS ........ .. . ............................. 27 CONCLUSIONS .................................... 36 REFERENCES ... ... . .... ............. ............. 39 INTRODUCTION Stell, 1988 INTRODUCTION It is the intent of this paper to compare predictions of an in-house developed computer model, radiant heat curing process for plastic insulated cable, with facility data and gathered from actual manufacturing plant. product trial~ both from production First, in temperatures during a pilot full scale background explanations of the and chemical basis of the computer discussed in some detail. Then the model are The model predictions of processing are then product compared to results obtained during trials at the pilot facility. degree of compared cure data from actual plant to Finally, accuracy for the values generated by conclusions are drawn production the concerning the that should be predictions is desired. -1- examined if the Next, runs . is computer of the model predictions and suggestions areas the electrical and the manufacturing process are given. thermodynamic of in a the model. overall are improvement made in Stell, 1988 PRIMARY URD CABLE To get their product to the customer most companies -use a delivery channel consisting of manufacturers, wholesalers, various retailers and combinations the like. of The delivery system for electricity utilizes electrical cables as the connection customer. between the manufacturing Just distribution as there channels for are the many more utility types of standard and marketing goods and services depending on which system best fits the need, are many different types of electrical cables, such things customers as how much power must be are, the there depending delivered, who where they live, and how reliable the on the power must be. In this paper we are concerned primarily with a particular class of plastic insulated cables known as primary Underground Residential Distribution (URD) cable [2]. cables are typically buried in the ground to These provide intermediate link in the electricity distribution chain. general, cable the is amount of power that can be proportional current ratings. cross-sectional insulation, the to the product of conveyed its an In by the voltage and The larger the electrical conductivity and area of the conductor and the more power can be delivered by -2- thicker the the cable. Stell, 1988 Of course, there is the usual trade off that more capability means more cost. system can be For illustration purposes, compared to a water supply generator is analogous to a pumping station. like pipes, more flow higher pressures. can electrical system • . The The cables are the voltage like water pressure, and the current like water flow. carry an Pipes with larger cross sectional areas can and pipes with thicker Similarly, wall·s can handle cables with larger conductors carry more current and cables with thicker insulations can handle higher voltages. The components of a typical primary URD cable are in Figure 1. Working from the inside out, shown The first object NEUTRAL WIRES CONDUCTOR SHIELD INSULATION SHIELD INSULATION CONDUCTOR FIGURE 1: TYPICAL PRIMARY URD CABLE -3- Stell, 1988 is the metallic conductor. This is the supply pipe which the current going to the end user flows. aluminum cost, It is usually for this application due to the combination of light conductivity, size. through weight, and moderately high electrical approximately 62% of that of copper of cross-sectional area unit of measure is a circular abbreviated "cmil". is A cylindrical conductor whose one mil (0.001 inches) has an area of one thousand circular mils (26 kcmil) to conductor sizes less than 250 kcmil, (AWG) up. mil, diameter circular Typical sizes for primary URD cable range from 26 equal Conductors are designated by cross-sectional area and number of individual wires twisted together to ·make them The low mil. approximately 1000 kcmil. For the American Wire Gauge is usually used to describe the conductor size [19]. Table 1 below shows the common AWG sizes used for primary URD cable and the corresponding kcmil sizes (19]: . Table 1: Common primary URD conductor sizes AWG kcmil #2 #1 #1/0 #2/0 #3/0 #4/0 41.74 66.36 105.6 133.1 167.8 211.6 Conductors may be made up of from one to 61 wires. larger the number of wires, the more flexible -4- the The finished Stell, 1988 product used will be. for The more flexible strandings are the larger size conductors in order to usually make them called the easier to handle during installation. The layer conductor shield. conducting smooth uniformly the conductor is an extruded called of of semi-conducting over which the distributed is layer It serves two purposes [5]. surf ace performance to It plastic polyethylene. a closest [5]. partially cross linked First it provides electrical stresses It is vital to the the cable that this layer be long very electrical - service. close stress and thus failure bonding to the insulating plastic layer to avoid internal than in Second, the conductor shield provides a surface for between the two plastic layers [5]. the premature term smooth. Discontinuities on the surface can be sources of higher normal are These gaps can lead to electrical discharge which eventually integrity of the insulation layer gaps can and destroy create an electrical short circuit [5]. Just over insulation very the layer. conductor shield essential out" that homogeneous. before this Even the extruded This allows the conductor to be held high electrical potentials. "leaking is it gets layer very be It keeps the to the customer. microscopically small imperfections -5- power from It clean can at lead is and to Stell, 1988 distorted electrical stresses and subsequent early failure of a cable. Just as pipe walls must be thicker to higher water pressures, higher voltages. for accommodate insulation must be thicker to allow Typical thicknesses range from 0.175 inches 15000 volt applications to 0.345 inches for 35000 [4]. able Another must for the insulation material is that it be to withstand the heat generated by the conductor as carries current. generates The volts insulation proportional to the square of of primary which is up to 90 temperatures catalyst The electrical resistance of the conductor heat majority it URD capable c. cables of in the use operating current. today at This is achieved by have conductor blending with the thermoplastic polyethylene such that a when the material is subjected to heat during cable manufacturing, the insulation crosslinking, material imparts is cured. improved properties to the insulation [5]. This physical curing, and called electrical This curing process is the subject of the model which is under investigation herein. The last plastic layer is the insulation shield. the conductor shield, it it also has two purposes [SJ. Like First, provides a smooth surface to which the insulation can mated to avoid gaps and subsequent discharges [5]. neutral Secondly, detrimental be electrical in conjunction with the metallic wires it confines the high energy electrical -6- fields Stell, 1988 within the insulation and provides safety from shock hazard [5]. The concentric neutral wires shown in Figure 1 thought of as the return pipe in the water system They are typically made of copper due to its high conductivity corrosion and the fact that it can generally analogy. electrical has better resistance than aluminum in the presence of and minerals in the ground. The neutral wires be water both help the insulation shield provide containment for the electric fields and provide the return path for the electrical current to get back to the generating source. Figure 2 shows the typical stages in the processing of a primary URD crosslinked manufacturers material, dies cable with an aluminum conductor polyethylene insulation. Generally produce smaller size wires by drawing in extruded cable would start with the electrical rod as with successively smaller holes, together and the stranding operation to and then form it a raw through twist the them finished metallic conductor. The continuous extrusion/curing/cooling operation the plastic materials is diagrammed in more detail in 3. The conductor shield, insulation and insulation are each applied to the conductor in an extrusion Figure shield, process. Figure 4 shows a cutaway side view of a typical extruder -7- for for Stell, 1988 the application polyethylene, of a single in pellet form, layer is fed of plastic. The onto a rotating screw where it is pushed through the heated barrel and melted. The pumping and mixing action of the screw acts to homogenize the molten plastic and move it to the front of the extruder where it enters the crosshead, so called because in it the plastic angle turn makes BAUXITE DRE - - - - ' ,NA ALUMINUM a right -1 I I I I I PLASTIC + ALLfING CASTING ' RDl.t'NG COILING I -----· I I I I I I I I I the EXTRUSION -----, '+ . cur' I I PLASTIC LAYERS I I I -----· + NEUTRAL VIRES - - - - -1 FINISHED CABLE I TESTING -----· FIGURE 2. to STRANDING COOLING B..ECTRICAL ROD parallel -----.. ALUMINUM . CONDUCTOR -----· CRAVING METALLIC ALUMINUM ----- from melted ' I -----· MANUFACTURING STEPS FOR PRIMARY URD CABLE -8- Stell, 1988 extrusion screw to parallel to the incoming conductor. the melted plastic wraps around the conductor hollow tube Typically 3. over it as they exit crosshead forms· a together. the layout of the extruders is as shown in The conductor shield followed generally by insulation and SEMI-CONDUCTING PLASTIC PELLETS <INPUT> a is applied few feet of first, air by Figure itself, cooling. insulation shields are then applied by The two INSULATING PLASTIC PELLETS <INPUT> CONDUCTDR SHIELD EXTRUDER HlSULA TIOff EXTRUDER HEATING/CURING TUBE ALUMINUM CONDUCTOR <INPUT> the and Here CONDUCTOR PLUS SHIELDS ANO INSULATION <OUTPUT> CONDUCTOR WITH JNSULATJON SHIELD SHIELD COOLING TUSE EXTRUDER SEHI-CONOUCTJNG PLASTJC PELLETS <JNPUT> FIGURE 3. EXTRUSION, RADIANT HEAT CURING, COOLING PROCESS FOR URD CABLE -9- Stell, 1988 separate extruders feeding them through one common crosshead. Immediately as the metallic conductor and plastic layers emerge from the insulation-insulation shield crosshead enter the curing tube. As they travel through it, heated in a pressurized nitrogen atmosphere to a sufficient primarily to initiate the curing serves process. two purposes. First it they are temperature The nitrogen prevents gaseous ,.--- CONDUCTOR PLASTIC PELLETS CROSSHEAD HEATED BARREL SCREW MOTOR/ GEAR BOX MOLTEN PLASTIC FIGURE 4. PLASTIC EXTRUDER -10- they Stell, 1988 by-products of the curing process from creating voids within the insulation. inert bubbling up and Second, it provides·an atmosphere so unwanted chemical reactions do not place during the curing. In the radiant heat/cure the heat is provided by passing electrical the walls nitrogen. of the stainless steel process, current pipes take through enclosing the A typical system will have one to ten individually controllable heating zones to accommodate the varying heating requirements for different conductor size and plastic layer thickness combinations. As the product leaves the heated pipe, it passes directly into a water filled cooling pipe where it is kept at the same pressure as in the heating pipe. Before the can plastic exit the sufficiently pressurized cooled system, the must be due to Once cooling is to prevent voids from forming gaseous by products of the curing process. cable complete, the product exits the curing/cooling system through a water seal and is wound up onto reels for further we see that next the neutral processing. Returning wires are to Figure 2, twisted around the partially completed product. Finally, electrical and mechanical testing takes place before shipment to the customer to requirements are being met. -11- assure that specification Stell, 1988 There use the are currently two manufacturing facilities computer model discussed herein to radiant heat curing process lines. which simulate Each plant their has one production line similar in layout to that shown in Figure The lines primary There represent approximately two thirds of URD plant. are ten heated zones on each line with a total curing of cooling manufacturing capacity approximately portion at total each length cable the 3. two hundred feet of the individual lines per is line. The roughly three hundred feet long. In addition to the two production facilities making of the model, there is one research and development location with a scaled down radiant cure line for testing of and processes. totalling use products This pilot line has three heated curing zones around twenty feet and a cooling section about fifty feet long. Two dictate industry the properties cable. their requirements for publish mechanical standards and which electrical of crosslinked polyethylene used in The first of these, Association, publish associations primary the Insulated Cable URD Engineer's consists of a group of cable manufacturers that standards that may be referenced own purchase specifications [12]. by utilities The second the Association of Edison Illuminating Companies, -12- in group, is a group Stell, 1988 of cable users, which In primarily utilities, that publish standards any utility may reference or adopt as their addition especially to these industry own standards, the larger ones, write their own [4). utilities, specifications for physical and electrical properties and cable performance. One minimum requirement degree of of many of these specifications cure as measured by either a is solvent extraction test or a hot creep test [12). In the extraction test polyethylene weighed, a sample of the cured solvent boiled away, the less cured the material is. The more plastic ICEA limits the maximum plastic boiled away to thirty percent of the [12). plastic is boiled in a solvent, and weighed again to determine the amount of plastic dissolved away [12). weight a In the hot creep test a sample of initial the is heated in an oven while being held under cured tension [12). The amount of elongation and permanent stretch cannot exceed 175 degree of cure is a function of the time and temperature percent and 10 percent, respectively. As the to which the cable was subjected during curing these tests limit the maximum speed at which the cable can be processed [11). Another limit on the maximum processing speed length of time required to cool the cable. to prevent the formation of voids, cable the As stated before, the hottest part of insulation must be cooled sufficiently before -13- is the exiting Stell, 1988 the pressurized system. for this A generally accepted industry value maximum temperature at exit from the pressurized cooling system is approximately 200 degrees Fahrenheit. The hottest area normally occurs in the innermost layer since the cooling water cools the cable from the outside to the inside. Indirectly, two limitation on specification requirements maximum exit temperature. call for this First AEIC has requirements for maximum size and number of voids allowed the insulation limitations cable [4]. both AEIC and ICEA have on the amount of partial discharge allowed in [4,12]. generated Second, in Partial discharge is electrical within voids in the cable insulation when is applied to it. a noise voltage It is used as an indicator of the presence of voids. In the case of the radiant heat curing process, there is also a limitation on the minimum speed at which a primary URD cable can be processed. This is due to the high cable surface temperatures which can be encountered as a result the 750 to 850 degree Fahrenheit curing The polyethylenes show If temperatures. used for the insulation shield deterioration at approximately 575 degrees the long, pipe cable is allowed to remain in the of curing begin to Fahrenheit. pipe too the surface can heat beyond this temperature and cause damage to the insulation shield material. -14- Stell, 1988 Outside of the curing/cooling process limitations, there exist others that control how a primary URD cable line can maximum linear be operated. output production Among these are the minimum volumes of the extruders, and the speeds at which the other machines in the and maximum production line can be operated. MODELLING THE PRODUCTION PROCESS The algorithm used to model the radiant curing and water cooling portions of the primary URD cable manufacturing process is based on work reported by Boysen in 1970 [6]. describes a method curing/cooling process, In Boysen's model, of sections thickness for computer a similar only using steam as the heat source. the curing tube is divided into a along its length, is modelling He and the plastic divided into a number of annular number extrusion rings. For instance, if the pipe were one hundred feet long, it might be divided The into one hundred one foot increments for the length model. of the increment selected is influenced opposing factors. First, by it must be small enough to large temperature changes in any of the plastic rings as cable moves from one section of the heating tube to Too large a temperature change -15- would two avoid the another. invalidate the Stell, 1988 assumption Boysen makes that the temperature throughout incremental element of plastic is constant [6]. This in turn would severely affect the accuracy of the model. The second factor to be considered in choosing an appropriate length is calculation time. more of them there increment The smaller the increment, must be, and calculations that must be made. limits the subsequently the At some point the on the time that can be spent doing the the more practical calculations will force a lower bound on the increment size. The temperatures of each of the annular rings inside of the cable, in Boysen•s algorithm, at every section of the heating pipe. on the are recalculated The new temperature is based on the temperature of the ring as it exits the previous heating section, the amount of heat being conducted into and out of it by the ring inside and outside of it, own internal section. function energy over the time The time spent in each section, of the speed at which through the process. steam, change the and its in the spent of course, cable is it is a travelling The outer ring, the one exposed to the is assumed to always have a surface temperature equal to that of the steam, due to the condensation of the steam on the surface of the cable. The remainder of the heat transfer occurs similarly to the other rings. -16- The metallic conductor Stell, 1988 receives its heat from the inner plastic ring. The equations for the heat flow from Boysen are summarized below [6]: Equation 1) Heat Conduction from ring to ring Q= KA h.T/ ~ L Equation 2) Change in internal energy of material from heating section to heating ·section AQ/ 6.T Cp = (l/M) where: Q= heat flow from ring to ring aT= temperature difference A L= distance in direction of A T A= area normal to direction of plastic flow K= thermal conductivity of plastic M= mass of plastic ~ Q= heat flow difference cp = specific heat capacity of plastic As Boysen explains, if the temperatures of the cable components as they exit the extrusion operation and enter the curing along and cooling phases are known, the above with the concept of energy balance between and sections, equations, the rings can be used to calculate a new temperature for -17- Stell, 1988 each ring in each heating section [6]. The equation for the new temperature of a given ring is shown below [6]: Equation 3) Tn= To + (Q -Q )t/M/Cp 2 1 where: Tn= new temperature of ring T0 = old temperature of ring Q2= heat flow into/out of outer edge of ring o1= heat flow into/out of inner edge of ring t= length of time for ring to travel through heating increment M= mass of plastic flowing through heating increment Cp= specific heat capacity of plastic Once the calculated that has figured. cable temperature of all for a given heat section, taken the rings the degree place as a result of the heating has of been curing is The crosslinkable polyethylenes typically used for applications cure by means of the decomposition peroxide catalyst in a first order rate reaction [11]. first then order temperature, constant [9]. rate reaction the time required, at a to reduce the amount of catalyst by half Figure 5 shows a typical "half-life" time -18- of a In a given is a Stell, 1988 Half-life Time, min 1 0.5 0.3 0.1 0.05 0.03 0.01 0.005 0.003 0.001 .___ __,___._..___._-.L._._~_.___.___....__~_.___._---1"--'--.......__.___.---J J50 JOO 400 500 450 Temperature, Deg F FIGURE 5. versus With temperature curve for a the temperature calculated, 5. HALF LIFE TIME VS TEMPERATURE FOR TYPICAL CROSSUNKING PROCESS of crosslinkable each ring having already been the half-life time can be determined from Figure The number of half-lives is then readily dividing polyethylene. the time the plastic ring has calculated been at once the number of half-lives temperature of interest. known, the amount of peroxide decomposed period of interest can then be determined by relationship: -19- during the the by the is time following Stell, 1988 Eqµation 4: where: P= fraction of peroxide decomposed in N half lives N= number of half-lives Figure 6 is a graphical representation of this relationship. Once the amount of peroxide decomposed is known, of crosslinking is also known since they are the degree proportional [11]. Peroxide Decomposed, % of Remaining 100 90 80 70 60 50 40 30 20 10 2 4 J 5 6 7 Number of Half Lives FIGURE 6: PEROXIDE DECOMPOSED VS NUMBER HALF LIVES FOR FIRST ORDER RATE REACTION -20- Stell, 1988 An additional calculation related to degree of cure performed by the model. As was mentioned in section on industry and customer specification the is above requirements, the solvent extraction test requires a thirty percent maximum level for dissolved plastic. extractibles" estimate The model includes a calculation which is an empirical the amount of plastic that can be "percent attempt dissolved to after all processing has taken place. In Boysen•s model the temperature and peroxide calculations for the cooling zone are essentially the same as those for the heating zone, except the heat flows are reversed since the outer medium is now cooler than the cable. The model under study here generally utilizes the same assumptions and thermodynamic equations for heat transfer as explained by Boysen, once the heat has reached the surface of the cable. refinements. There are a few The minor differences, thermal conductivities for primarily the materials are allowed to vary with temperature. In addition, in an attempt to empirically account for differences actual an various between and predicted values of heat transfer and cure rate, artificial thermal resistance between the inner layer plastic and the metallic conductor was added to during it initial development. -21- the of model Stell, 1988 The primary difference in calculations occurs in the Recall that the method used to get the heat to the cable. Boysen model assumes that since the steam condenses and temperature steam (6]. is on directly the surface of the cable, rapidly that of the surface of the cable is the same as the the In the case of the radiant cure process, the heat transferred from the heated pipe to primarily by radiation. the cable surface Itaka et al have described typical radiant heat transfer equations as follows (13]: Equation 5) Wr= 4 4 S (T 2 -T1 ) A1 ----------------------1/E1 + A /A (l/E - 1) 1 2 2 where: Wr= heat transferred by radiation S= Stefan Boltzman constant Tl= absolute temperature of cable surface T2= absolute temperature of pipe surface area per unit length of cable A= 1 area per unit length of pipe A= 2 surface El= emissivity of cable E2= emissivity of pipe An additional mode of heat transfer to the cable is convection. Again, by Itaki et al have described typical heat -22- Stell, 1988 transfer equations for this situation [13]. The form is as follows: Equation 6) W = c where: W0 = heat transferred by convection K0 = constant, 14.9 for nitrogen gas P= gas pressure d 1= cable diameter d 2= pipe diameter T1= absolute cable surface temperature T2= absolute pipe temperature a= constant, 2.2 Another the Boysen model is the use of from one to ten insulated heating pipes, each of settings, between model variation which can that our model has from have its own and the use of a short, the allows length temperature variable non-heated last heating zone and the the and lengths to be input cooling zone. The by user. The the customarily used lengths are based on actual physical of the pipes less some allowance for non-insulated of the pipe where internal temperatures may be lower than in the insulated sections. -23- zone length sections considerably Stell, 1988 Why is it important to know how well a computer models the radiant electrical cable? applications Boysen model for The the identifies of curing process for a answer model and lies their three categories of curing process [6]. the in program production the potential possible benefits. applications One of of these for is a the prediction of optimum operating conditions for the production line [6]. Processing problems, product quality problems and productivity case. considerations are the primary issues Processing problems would include limitations imposed loss heating zone or the reduction in of a in this temporary due to equipment failures such output as the of an extruder to which the curing/cooling process must be matched. The model could be used to determine the cooling and temporary line speed conditions necessary to curing, match the limits. Product quality problems for which the model could be useful investigation tool would be where the cable failed meet specification requirements and would require or reprocessing. overheated cables surfaces and cables with internal voids as evidenced by partial measurements at the final electrical production. The model could be used to examine -24- testing to scrapping Examples of scrap generating problems with scorched a are undercooled discharge stage the of actual Stell, 1988 processing conditions used such as cure cooling water temperature, tube temperature, etc., as determined by production records, to determine if any variation in standard procedures that might problem. model be present were sufficient to caused the For product quality problems like undercuring, the could not only be used to have determine the possible reasons for the problem but also the conditions· necessary for potential scrap reducing remedies like recuring by passing the cable through the curing/cooling process again. As Bartnikas points out, due to the amount of capital typically required to build this type of production line, is economically essential that a company productivity by maximizing production speeds [5]. can maximize The model be used to assist in developing target production used to establish subsequent industrial engineering standard costs of production, it rates standards even for and products which have never actually been produced. A second application of cure calculation models cited by Boysen is the prediction of cure performance of new materials without the need for expensive plant trials [6]. This is particularly valuable when production capacity is limited and profit making production experiments on new products. with must be forgone to accommodate In the case of an organization pilot facilities where material characteristics can -25- be Stell, 1988 determined on small scale prototype equipment, the model can then be used to predict full scale production performance ·of the candidate materials. The final Boysen is the Here we are with the design of new production facilities or the prediction dealing of application optimum referred to by process design (6]. By being able to upgrade of current facilities. reasonably accurately predict production speeds for various combinations of curing and cooling lengths and temperatures, of these parts of the production process matched to the extruder outputs. the outputs be closely can This helps to minimize capital investment necessary to achieve rates. It entire cable also may help to maximize desired production productivity production facility since the the for an curing/cooling process is typically the bottleneck operation. The ability to predict production speeds through the use of the model has other potential applications in addition to those mentioned by Boysen. be made of production costs. For instance, Once the costs estimates can are predicted, business problems like how many of which products are best to make, what is the potential return on the capital investment, and what are the best ways in which to schedule production of orders, may be examined. -26- Stell, 1988 RESULTS Since is to the first step in calculating the degree of calculate logical that temperatures if the appropriate there were a temperatures, way to it actually cure seems measure inside the cable as it is being produced, this would be the place to start examining the capabilities of the model. Mitchell [14-16] and Robbins [17] report methods for measuring the temperature of the cable surface and of interface between the conductor shield and insulation the during experimental runs on a pilot radiant heat, dry cure extrusion line. The surface temperature is measured with an infrared pyrometer through quartz glass inspection ports at the end of each of the three heating zones on the line. This measurement method is subject to considerable variability and error since it depends on calibration of the system quality of the optics being used and the to the emissivity and surface condition of the material being measured [7]. The temperature is accomplished by placing a thermocouple on the surface of the conductor where the applied. measurement of the internal shield just before the cable enters the insulation and insulation shield A thermocouple measures temperature by crosshead layers are generating an electrical signal proportional to the temperature at which -27- Stell, 1988 it is being held (8]. display The signal generated must be sent to a device which converts it to a readable temperature. In trials described here, the thermocouple is attached to the readout device through a sufficient length of lead wire that the wire can be fed into the crosshead travels through the pilot production line. as the such cable In this way the temperature at the interface between the conductor shield and insulation can Measurement be read at any errors point for the type of along the thermocouples process. used in these studies are generally believed to be plus or minus four degrees Fahrenheit [3]. All of the trials reported by Mitchell [14-16] Robbins [17] were done on #1/0 AWG solid aluminum and conductors with 0.260 inches of crosslinked insulation. These trials are listed in represent Tables 2 through 9 as Tl AWG insulation and Trial T9 was on a #2 AWG compare the temperature data conductor collected 7 predictions of the computer model. the such inches wire as of aluminum Tables 2 through shield/insulation during They Trial TS was on a 19 strand aluminum conductor with .175 conductor with 0.175 inches of insulation. 6 T7. minor variations in processing variables line speed and cooling water temperature. #1/0 through trials interface with the Tables 2 through 4, show the errors in the calculated values through the three heating -28- Stell, 1988 zones are relatively small, less than ten percent. Tables 5 and 6 show the errors start increasing in the neutral zone and then thirty seven become quite significant, up to percent, after the cooling zone. TABLE 2: TABLE 3: Conductor Shield / Insulation Interface Temperature Data from Thermocouple Trials - End of Heat Zone 1 Trial Meas. Cale. Error Tl T2 T3 T4 T5 T6 T7 TS T9 264 256 263 251 249 251 250 270 258 270 265 266 270 270 269 269 274 267 2.3% 3.5% 1.1% 7.6% 8.4% 7.2% 7.6% 1.5% 3.5% Conductor Shield / Insulation Interface Temperature Data from Thermocouple Trials - End of Heat Zone 2 Trial Meas. Cale. Tl T2 T3 T4 TS T6 T7 TS T9 314 306 312 291 290 304 303 334 313 317 314 317 304 304 313 313 329 313 -29- Error 1.0% 2.6% 1.6% 4.5% 4.8% 3.0% 3.3% -1.5% 0.0% Stell, 1988 TABLE 4: TABLE 5: Conductor Shield / Insulation Interface Temperature Data from Thermocouple Trials - End of Heat Zone 3 Trial Meas. Cale. Tl T2 T3 T4 TS T6 T7 TS T9 349 340 343 336 33S 360 359 386 363 357 355 358 346 346 36S 368 379 360 Error 2.3% 4.4% 4.4% 3.0% 3.3% 2.2% 2.5% -l.S% -0.8% Conductor Shield / Insulation Interface Temperature Data from Thermocouple Trials - End of Neutral Zone Trial Meas. Cale. Error Tl T2 T3 T4 TS T6 T7 TS T9 34S 340 340 351 351 384 38S 386 378 363 361 364 384 3S4 417 417 409 395 4.3% 6.2% 7.1% 9.4% 9.4% 8.6% 7.S% 6.0% 4.5% -30- Stell, 1988 TABLE 6: Conductor Shield / Insulation Interface Temperature Data from Thermocouple Trials - End of Cooling Zone Trial Meas. Cale. Error Tl T2 T3 T4 TS T6 T7 TS T9 133 140 142 158 158 205 210 192 146 147 144 145 167 141 172 181 121 124 10.5% 2.9% 2.1% 5.7% -10.8% -16.1% -13.8% -37.0% -15.1% Tables 7 through 9 compare the surface temperature collected the during the trials with the calculated values from There is considerable variation in errors with model. values data ranging up to approximately twenty-two must be remembered, percent. though, that this particular measurement itself is subject to substantial errors. TABLE 7: It Cable Surface Temperature Data from Thermocouple Trials - End of Heat Zone 1 Trial Meas. Cale. Tl T2 T6 T7 425 435 510 510 400 397 427 428 -31- Error -5.9% -8.7% -16.3% -16.1% Stell, 1988 TABLE 8: Cable Surface Temperature Data from Thermocouple Trials - End of Heat Zone 2 Trial Meas. Cale. Tl T2 T6 T7 435 440 548 463 461 502 502 TABLE 9: 548 Error 6.4% 4.8% -8.4% -8.4% Cable Surface Temperature Data from Thermocouple Trials - End of Heat Zone 3 Trial Meas. Cale. Tl T2 T4 TS T6 T7 320 330 473 473 540 540 390 389 489 489 556 556 Error 21.9% 17.9% 3.4% 3.4% 3.0% 3.0% To test the ability of the model to predict cure results on products made on full scale factory production the products shown in Table 10 were selected for equipment, testing. They represent as wide a range of products as could be chosen given the production schedules in effect frame of this research. -32- during the time Stell, 1988 TABLE 10: Sample 1 2 3 4 5 6 7 8 9 10 Two samples. Product Descriptions for Cure Test Samples Description #4/0 AWG 19 wire aluminum, .175" insulation 500 kcmil 37 wire aluminum, .175" insulation 750 kcmil 61 wire aluminum, .175 11 insulation #4/0 AWG 19 wire aluminum, .220 11 insulation #1/0 AWG 19 wire aluminum, .260 11 insulation #1/0 AWG solid aluminum, .260 11 insulation #4/0 AWG 19 wire aluminum, .260 11 · insulation 500 kcmil 37 wire aluminum, .260 11 insulation #1/0 AWG solid aluminum, .345 11 insulation 750 kcmil 61 wire aluminum, .345 11 insulation sets of cure related tests were performed on The first was the normal solvent extraction the test, on a specimen taken from the inner twenty-five percent of the insulation thickness, specifications tests the required by industry as described previously. customer The results of and the model predictions are shown in Table table shows, there is considerable predicted and actual. in the As between In most cases the model predicts more #3, the model predicts substantially less cure. shown 11. variation cure than actually exists, but in some cases, has and a series of round-robin such as Sample As Adams [l] solvent extraction tests, the variation in actual test results is typically four percent. This accounts for only a small part of discrepancies seen here. -33- the large Stell, 1988 TABLE 11: The samples % Extractables, Inner 25% of Insulation Plant samples, Plant Measurements second were Meas. Cale. #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 17.4 20.2 21.1 16.7 16.6 18.0 18.6 16.8 17.6 18.3 13.5 11.3 36.0 16.6 19.2 18.6 12.5 11.3 11.3 12.4 set of cure tests performed done Chromatography) these, Sample using HPLC -22.4% -44.1% 70.6% -0.6% 15.7% 3.3% -32.8% -32.7% -35.8% -32.2% on the Performance samples of the uncured plastic are collected uncured cable Liquid analysis as described by Hercules (10]. cables are being manufactured. the (High Error the HPLC analysis is performed on samples to determine the concentration Then HPLC analysis is peroxide catalyst present. as In of the performed on the· cured samples, also to determine the level of peroxide present. The percent of original peroxide remaining in the cured samples is then calculated. The model results of the HPLC measurements and predictions Insulation insulation ring are one thickness. shown is the in Tables innermost The -34- measured 12 the through one-eighth values computer of shown 14. the are Stell, 1988 questionable since they don't show the expected decrease cure as the sample position gets nearer to the inside. very small sample quantities available for each likely contributed to the experimental error. which that may ring the time the ring samples were cut from them. The quite Another factor have affected the results is the length elapsed between the time the cables were in of produced · Robbins time and (18] suggests that the concentrations of the peroxide catalyst may tend His to equalize across the insulation wall as time best samples estimate of a time frame is a few passes. weeks. in this study typically were not cut up for four The to eight weeks after production. TABLE 12: Percent Remaining Peroxide, Insulation Ring 1 Plant Samples, Lab HPLC Measurements Sample Meas. 1 7 1 3 #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 2 1 4 5 0 0 -35- Cale. 0 0 77 19 28 24 5 0 0 0 Error -100% -100% 7600% 533% 1300% 2300% 25% -100% 0% 0% Stell, 1988 TABLE 13: Percent Remaining Peroxide, Insulation Ring 2 Plant Samples, Lab HPLC Measurements Sample Meas. #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 TABLE 14: Cale. 1 7 4 0 57 10 16 15 2 0 0 0 1 4 2 2 2 2 0 0 Error 471% -100% 5082% 186% 662% 838% 25% -100% 0% 0% Percent Remaining Peroxide, Insulation Ring 3 Plant Samples, Lab HPLC Measurements sample Meas. #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 Cale. l l Error 67% l 0 28 4 3 3900% -14% 2 2 2 0 0 5 7 0 0 0 0 -100% -100% 0% 0% 338% CONCLUSIONS Comparisons to reasonably accurate temperature measurements performed during experimental trials on a line show the model to have -36- conductor pilot shield/insulation Stell, 1988 interface temperature calculation errors ranging from percent to +10.5 percent. temperatures -16.9 to however, + 21.9 percent. The measurements to The solvent extraction data from production records show the model to be · substantially more deficient in percent extractibles, with probable errors ranging from -44.1 percent to +15.7 percent. HPLC from may contain considerable error due their inherent inaccuracy. plant surface during the trials indicate errors ranging percent themselves, Measurement of the cable -37.0 analysis were inconclusive due probability of high experimental error. therefore, to draw conclusions directly, presented herein, The results of to a substantial It is not possible, based on the about the ability of the model to percent remaining peroxide. It is probable, data predict however, that since the percent remaining peroxide calculation is dependent on the temperature calculations, the errors will be about the same at best. As a consequence of the results reported above, substantial caution should be exercised in the application of the model. comparisons It is probably most useful for making relative of the results of small changes in operating pa- rameters. There are numerous potential sources of the large errors exhibited by the model. Basically it is a -37- combination of Stell, 1988 theoretical thermodynamic equations and empirical attempting to provide a "best fit" model. parameters Either of these two areas contain possible problem sources. A possible theoretical problem might be for heat exchange surface. The between the curing pipe equations used radiation and natural convection. mode of include the equations and heat the exchange Natural convection is heat transfer when there is little or no the cable such as in the gas spacer cable Itaki, et al [13]. the surface described In the case of a radiant by relative movement between the gas pressurizing medium and the of cable heat by curing production line, however, the cable may be moving through the line at speeds in excess of one hundred feet per minute. It may be necessary to include a heat transfer component for the forced convection mode of heat transfer such as is done for bare electrical conductors installed typically outdoors and exposed to the wind [2]. Another transfer possible point. problem is the function used between the conductor shield and aluminum conductor. creates theoretical an heat the The model has built in an equation which artificial resistance to heat transfer at this This was originally done in an attempt to account for differences between measured and calculated temperatures -38- in Stell, 1988 early development work on the model. Perhaps another method of accounting for the differences might yield better results. For instance, increase in the model contains no provision strand shield material volume that for exists the on stranded conductors versus solid conductors due to the strand interstices. in Accounting for this material volume difference conjunction with other potential sources of improve the accuracy of the model, error might particularly for stranded conductors. Empirical parameters which might be sources of error are the variables and constants in the equations which either cannot be directly measured or derived from accepted physical constants. Such error sources would include items like pipe and cable emissivities, and the effective heated length the heating pipes which are not insulated over their length. -39- of entire Stell, 1988 REFERENCES [l] [2] Adams, James. "Round Robin Decalin Extraction Tests." Richmond, VA: Reynolds Metals Company, 1986 The Aluminum Association. Aluminum Electrical Conductor Handbook. Washington, D.C.: 1982. [3] American Society for Testing and Materials. Manual on the Use of Thermocouples in Temperature Measurement. Philadelphia, PA: 1981 [4] Association of Edison Illuminating Companies (AEIC). Specifications for Thermoplastic and Crosslinked Polyethylene Insulated Shielded Power Cables Rated 5 through 46 kV. New York: AEIC CSS-82, 1982. [5] Bartnikas, R., and K. D. Srivastava, eds. Elements of Cable Engineering. Waterloo, Ontario: Sandford Educational Press, 1980 [6] Boysen, R. L. "An Analysis of the Continuous Vulcanizing Process for Polyethylenes." Union Carbide Corporation Kabelitems Number 141. Bound Brook, N.J.: 1970. [7] Eshbach, Ovid w., and Mott Souders, eds. Engineering Fundamentals. Handbook of New York: John Wiley & Sons, 1975. [8] Faires, Virgil Moring. Thermodynamics. London: The MacMillan Company.: 1970. -40- Stell, 1988 [9] Gray, Harry B., and Gilbert P. Haight, Jr., Basic Principles of Chemistry. New York: w. A. Benjamin, Inc.: 1967. [10] Hercules Incorporated. "High Performance Liquid Chromatographic Assay of Di-cup(R) Dicumylperoxide. 11 Technical Data Bulletin ORC-207. Wilmington, Delaware: 1977. [11] Hercules Incorporated. "Di-Cup(R) and vul-CUp(R) Peroxides." Technical Data Bulletin ORC-lOlC. Wilmington, Delaware: 1970. [12) Insulated Cable Engineers Association (ICEA). Cross-linked Thermosetting Polyethylene Insulated Wire and Cable for the Transmission and Distribution of Electrical Energy. Washington, DC: ICEA S66-524, December 1984. [13) Itaka, Koshi, Tomoo Araki, and Takushi Hara. "Heat Transfer Characteristics of Gas Spacer Cables." IEEE Transactions on Power Apparatus and Systems (Sept/Oct 1978): 1579-1584. [14) Mitchell, J. L. "Dry Cure Extrusion Study: Hot Water In-line Anneal of VEPCO 1/0 35kV Cable." Reynolds Metals Company, 1981. -41- Richmond, VA: Stell, 1988 [15] Mitchell, J. L. "Extrusion Study: 1/0 VEPCO Cable-Pressure Cool at Different Quench Temperatures and stress Relieve." Richmond, VA: Reynolds Metals Company, 1982. [16] Mitchell, J. L. "Extrusion Study: 1/0 VEPCO Cable, Pressure Cool at Different Quench Temperatures and Stress Relieve, Trial 2. 11 Richmond, VA: Reynolds Metals Company, 1982. [17] Robbins, Bruce. "Drycure Extrusion Study In-line Anneal of 1/0 Sol. Aluminum VEPCO Construction using HFDA-0581 Conductor Shield." Richmond, VA: Reynolds Metals Company, 1981. [18] Robbins, Bruce. Manager, Materials Development, Reynolds Metals Company Electrical Division. Interview. Richmond, VA. Personal 12 Feb. 1988. [19] U.S. Dept. of Commerce. "Aluminum Wire Tables." National Bureau of Standards Handbook 109. Washington, D.C.: Government Printing Office, February 1972 -42-