Locating Defective Transmission-Line Insulators

Locating Defective Transmission-Line Insulators

Details of the "Buzz-Stick" Method That Can Be Used on Live High-Voltage Lines to Detect

Faulty Insulators of Suspension or Pin Type, Together with the

Construction and Use of Special Tools Required

BY T. F. JOHNSON, JR.

General Electrical Superintendent, Georgia Railway & Power Company

A SIMPLE, economical and accurate test, which is called the "buzz-stick" method, has been developed by the writer and extensively used under widely varying conditions on live transmission lines of all types for locating defective insulators of either the suspension or the pin type. For locating defective suspension insulators it can be used at any voltage provided that there are not fewer than two insulators per suspension or dead end. While the minimum number of insulators is limited to two, there is no limit to the maximum number per string, provided that the operating voltage of the line is not less than about 20 per cent of the operating voltage for which the line is insulated. For example, the "buzz-stick" method can be used on a line built and insulated for 110,000 volts but operated at 22,000 volts. However, if the same lines were operated at 13,000 volts or 11,000 volts, it would be difficult to obtain accurate results. The method applies equally well to all makes of suspension-type insulators, both single-part and multiple-part, and will indicate the particular type of defectiveness which is characteristic of each make. The condition of suspension-type insulators can be determined all the way from perfectly good, through the different degrees of defectiveness, to totally bad. In other words, it will locate all insulators which will puncture before they will flash over.

The accuracy of this method can be checked by the use of the three methods commonly in use-the megger, the oscillator and the 60-cycle test, Any insulator shown bad or defective will probably pass any one of the above tests, or perhaps any two, but it will probably not pass the three. However, it is entirely possible for an expert tester to locate defective insulators which will not show defective on any of the tests mentioned. These insulators, if left in the line, will go dead in a period of time which clearly indicates that they were defective in spite of the showing made on other tests.

The "buzz-stick" method is also an accurate test for locating defective multiple-part pin-type insulators or for locating any defective unit part of a multiple-part pin-type insulator. It will work on pin-type insulators of any design now made except single-part pin-type insulators and those designs of multiple-part insulators in which the cement holding the insulator together is entirely covered by the porcelain parts. There are, however, a large number of insulators which at first sight seem to belong to the latter class in which the cement can be easily reached by properly designed points on the testing tools. The "buzz-stick" method will locate, with the exceptions stated, any defective multiple-part pin-type insulator regardless of the degree, small or great, of the defectiveness. It will also locate any defective unit part or parts in any combinations and in any degree of defectiveness in any multiple-part pin insulator on any transmission line carrying any voltage.

To use this test method for locating faulty suspension or pin-type insulators the line must be alive. It is a part of the system developed and used by the writer for maintaining all classes of transmission lines without disturbing the operation of these lines. Naturally the use of this method requires study and practice, but the knack is soon acquired and the testing rapidly done by a good lineman.

Details of the ways the tests are made and of the construction of the special tools required are given in the paragraphs that follow for locating defective insulators of either suspension or pin type.

TESTING SUSPENSION INSULATORS WITH A

"BUZZ STICK"

The instrument or tool used when testing out suspension insulators is a treated stick about 8 ft. or 10 ft. (2.4 m. or 3 m.) long and 11 in. (3.1 cm.) in diameter. On the end of this stick is fitted an adjustable wire fork, as shown in Fig. 1, with both prongs or horns electrically connected together. This tool is called the "buzz stick," and the method of testing the insulators is called "buzzing" them. There are two steps in the "buzzing" of a string of insulators-first, what is called feeling them out; second, what is called shorting them out. The feeling-out operation always precedes the shorting-out operation and determines the general condition of the string. While it does not definitely determine the exact condition of the string, it gives a fair indication and shows whether or not the shorting-out operation can be applied without the danger of flashing or puncturing the string of insulators. The shorting-out operation always follows the feeling-out operation except in those cases where the feeling out of the string has indicated that it would be dangerous to apply it. The shorting-out operation accurately locates the defective insulator unit.

FEELING OUT A STRING OF INSULATORS

When feeling out a string of suspension insulators, the feeling point on the horn A of Fig. 1 of the "buzz stick" is touched to the line conductor and drawn slowly away. There is then produced a distinct buzzing sound which continues to be audible, in the case of 110,000 volts, until the feeling point is some 21 in. or 3 in. (6 cm. or 7.5 cm.) away. This operation is repeated on the insulator cap No. 1 (Fig. 2). The sound will be less distinct and will follow the feeling point at a shorter distance. When the operation is repeated for insulator caps Nos. 2 to 8 (Fig. 2) in succession, it will be found that the sound is less for each cap as the distance from the line increases, until on cap No. 6 or 7 there is very little or no sound and the sound, from cap No. 8 is louder than from No. 6 or No. 7. Such results indicate a perfectly good string of insulators.

Assume now that insulator No. 2 is totally bad-in other words, dead. In this case the sound from caps No. 1 and No. 2 will be the same and cap No. 6 or No. 7 will produce a sound. In case insulator No. 6 is bad, then caps No. 5 and No. 6 will produce the same sound and the sound from cap No. 7 will be very much increased.

Assume that insulators Nos. 2, 3 and 4 are dead. Then the sounds from caps Nos. 1, 2, 3 and the sounds from caps Nos. 4, 5, 6, 7 and 8 will be very much magnified. In other words, the sound from the cap of a dead insulator is the same as the sound from the cap of the next insulator to it nearest the line. In any of the above assumed cases it would be safe to apply the shorting-out process. If all insulators are dead except two, to apply the shorting-out process would knock the line out.

It is a common occurrence on lines that are three or four years old to find strings of insulators to which the feeling-out operation will indicate that it is unsafe to apply the shorting-out operation. This is essentially the case in dead ends. In these cases it is advisable to count the whole string as defective, remove it and separate the good insulator units from the defective ones by other tests. The feeling-out operation is for the purpose of determining, without the possibility of flashing or puncturing the string, if there is a sufficient number of good insulator units left in the string to stand the line potential when one of the good insulator units is shorted out during the shorting-out process. On 110,000-volt lines there must be at least three good units to stand the shorting-out process. On lower voltage there must be at least two good units, except on voltages such as 11,000 and 13,000, where there are only two insulator units per string. Then only one unit need be good.

SHORTING OUT A STRING OF INSULATORS

In applying this test touch either one of the horns of the "buzz stick" (Fig. 1) to the line conductor. Hold this horn in contact with the line and make and break contact with the other horn with cap No. 1 (Fig. 2). There will be produced a snappy spark on making and breaking contact with cap No. 1. Then hold either of the horns in contact with cap No. 1 and make and break contact with the other horn with cap No. 2. A snappy spark will be produced on making and breaking contact with cap No. 2, but the intensity of the spark will be less than when the horn was held in contact with the line conductor and the other horn touched to cap No. 1. Repeating this operation across each insulator unit in the string, it will be found that the intensity of the spark will diminish as the distance from the line increases until insulator No. 8 is shorted out, when the spark across this unit will be found to be more than across unit No. 7. These conditions indicate a perfectly good string.

Assume now that unit No. 2 is dead. When this unit is shorted out as described there will be absolutely no spark. If unit No. 6 is dead likewise, no spark will be produced. With dead insulators in any number and any combination the results will be the same; that is, absolutely no spark across the dead unit.

Assume the case that insulator unit No. 5 (Fig. 2) is slightly defective-that is, the value of the flashover voltage is only slightly greater than the voltage which would puncture it-and that all other units in the string are good. Under these conditions the spark obtained from cap No. 5 will be less than when insulator unit No. 5 is good and the spark from cap No. 6 will be greater. This is so for any insulator in the string, no matter how slightly defective. Bad, or dead, and very defective insulators will have to be removed before the slightly defective insulators in the same string can be detected. In the case of lines from which insulators are removed and replaced with the lines out of service, the worst insulators are located on the first test with the "buzz stick" and the slightly defective insulators are located on subsequent tests made after the worst insulators have been removed. When the lines are maintained while under operation, the tester precedes the maintenance crew and locates and marks with paint, by means of a brush on the end of a stick, the most defective insulators. The maintenance crew follows the tester, replacing the marked insulators while the line is alive. After they have done this they "buzz" the string out and at this time locate and replace the slightly defective units. A line which is maintained dead will require, as a general rule, more frequent testing than when maintained at operating voltage.

THE SILENT INSULATOR IN A STRING

In every string of insulators, regardless of the make, the number of insulators in the string or the voltage on the line, there is one insulator in the string the cap of which will give the minimum spark both in the feeling-out process and the shorting-out process. This insulator is generally the second one from the cross-arm. As soon as the tester has tested a sufficient number of strings on a given line to locate this silent cap and knows the character of the spark and sound given off by it when the string is entirely good, it is no longer necessary for him to test in detail each string. By touching the feeling point of the "buzz stick" to this cap he can easily determine whether the string is entirely good or not. If the spark from the cap is normal, he can pass on; if it is abnormal, he should "buzz" the entire string out and definitely locate and mark the defective insulators. A tester soon becomes expert and will be able by touching the cap of the silent insulators to tell the number of defective units in the string, although it is always necessary to "buzz" the entire string out to locate the units definitely.

It is obvious that in this case the line voltage must be low, such as 22,000 volts or 13,000 volts or 11,000 volts, and there will always be at least one good unit. Since there are only two units, the one which is good and the one which is bad must be determined accurately by the feeling-out process. A trial will show that this is easily done. When the bad one is determined by the feeling-out process that unit can be then checked by the shorting-out process. There is no need to test the second unit, nor is it possible. A tester can best acquire his skill on longer strings and consequently higher voltages before he tries to test the shorter strings.

The use of the "buzz stick" is accompanied by some dangers until the tester acquires skill. These dangers are due to the likelihood of an inexperienced tester flashing a string and knocking the line out. To avoid these dangers the tester should test the lowest string on the tower first and should always stand well below the string he is testing. When this is not possible, he should stand with the wind blowing from him to the string. He should wear colored glasses to protect his eyes. If these precautions are taken, the arc if one is started, will always travel away from the tester and there will be no danger of his standing under a string which is about to fail. However, when a man becomes expert he may dispense with these precautions, except that he should always test the lowest string first.

"BUZZ STICK" METHOD OF DETERMINING PROPER

NUMBER OF SUSPENSION INSULATORS

When a source of electrical energy is at hand of the same pressure as that of the line upon which it is proposed to use the insulators, the "buzz-stick" method can be employed to determine the proper number of suspension insulators to be used. To illustrate, assume that it is desired to determine how many suspension-type insulator units it is necessary to put in series on a transmission line operating at 110,000 volts.

In every string of insulators of the proper length and number for the voltage of the line upon which they are used there is one insulator the cap of which will give the minimum spark during the feeling-out process. This insulator is called the silent insulator. If the string is too long, there will be more than one silent insulator. The number of silent insulators there are depends on the amount of over-insulation of the line. The silent insulator is not always absolutely silent-whether it is or not depends on the make of the insulator-but it isn't all makes the insulator the cap of which gives the minimum spark of any cap in the string during the feeling-out process.

Now, to find the proper number of insulators for the above-assumed case apply 110,000 volts to a string of insulators and add or take units from the string until there is one silent insulator, which should be the second or third one from the ground. This string would then be sufficient for 110,000 volts. The same procedure would apply for any voltage, regardless of the make of insulator or whether the insulators were to be used on suspension or dead ends.

For dead ends the number of units which should be used in addition to give a greater factor of safety is simply a question of personal opinion and the frequency of testing the lines. It is undoubtedly better to have one too many insulators in a string than too few.

New insulators can be tested before using them by means of a rack built to hold any number of strings of insulators and by applying voltage to one end of the strings and grounding the other ends and then "buzzing" them.

The method of testing insulators of more than one part is exactly similar to the testing of single-part insulators as already described. The cement between the various parts acts at all times like so much metal as far as the "buzz stick" is concerned. Knowing this, the test can be made between the cement and the cap of the insulator, or between the cement and the pin of the insulator, or between cement and cement. Care should be taken in making this test that one horn of the "buzz stick" is actually in contact with the cement, otherwise the indications will be deceptive.

It is advisable never to send less than two men to test the insulators with the "buzz stick" on any line, one to do the testing and one to do the marking or painting of the defective insulator. The defective insulator should be marked while the tester has his "buzz stick" across it, and the men should check each other; otherwise it is surprising how often the wrong insulator will be marked.

It has been found that a brilliant red paint is the most desirable color to use for marking and that it should be very thick. One large mark on the head of an insulator is used to indicate a totally dead unit, two marks on the head indicate a defective insulator, one large mark on one of the porcelain parts indicates that particular part to be dead, and two small marks indicate that particular part to be defective. The markings of the testers should, whenever possible, be so placed and of such size as to be easily seen from the ground.

COST OF TESTING WITH "BUZZ STICK"

By averaging the cost of testing 100,000 units of various types on a variety of lines it was found that the cost per unit, including all expense except royalties, was 8.3 mills. Adding royalties the cost would be, at a maximum, 3 cents per unit, and at a minimum for a system having a large number of insulators about 1.5 cents per unit. The use of the "buzz stick" does not entail any loss due to the shutting down of the line, but even with the elimination of this point the method compares favorably with any other method of testing insulators. A zero record on line outages due to insulator failures can be obtained with very inferior insulators. The more inferior the insulators the more frequent should be the test and removals, the better the insulators the less frequent the tests and removals; but in both cases a zero record of interruptions due to insulator failures is possible. The writer would like to point out that absence of interruptions due to insulator failures does not mean a perfectly operating line. Transmission-line outages are sometimes the results of causes other than insulator failures, but any one will grant that if insulator failure outages are eliminated transmission lines will give more nearly continuous service.

REPLACING DEFECTIVE INSULATORS ON LIVE LINE

Although the "buzz stick" opens a way to locate defective insulators while lines are in regular operation, it does not provide a way to replace these defective insulators while the lines are in regular operation. However, the writer has methods for this purpose for use on any type of line. These methods have been in use about five years, and where they are in use both the testing and the insulator replacements, as well as all other transmission-line maintenance, are carried on without interfering in any way with the operation of the system, or the operation of the system delaying or causing inefficiency in the maintenance of the lines.

The following figures are the result of actual tests with the "buzz stick" and megger:

As an example of what the use of the "buzz-stick" method of locating defective suspension insulators will accomplish on a given line, reference can be made to the case of the Atlanta-Tallulah Falls line of the Georgia Railway & Power Company. During the year previous to the adoption of the "buzz-stick" method by that company the Atlanta-Tallulah Falls line was out twenty-nine times owing to insulator failures. For the year subsequent to the adoption of the "buzz stick" there was not a single interruption due to insulator failures. The company has not had a single insulator failure on lines which have been tested with the "buzz-stick" method. Insulators are tested and removed on these lines once a year; consequently the results obtained form a good basis for the opinion that interruptions due to insulator failures have been reduced to a negligible quantity.

The tools required in applying this method are of two types, as shown in Figs. 3 and 4. Fig. 3 is called a feeling stick. Two sticks of this kind are needed when the insulators under test have more than two parts, but only one stick of this kind is needed when the insulators under test have only two parts. It consists of a stick about 1 1/8 in. (3 cm.) in diameter and about 8 ft. (2.4 m.) long, composed of suitable insulating material, at one end of which is fixed a sharp metal point of suitable shape, the shape depending on the type of the insulators to be tested. This metal point is known as the feeling point. Fig. 4 is what is called a shorting stick. Not more than one shorting stick is ever needed. It also is a stick about 14 in. in diameter and about 8 ft. long, of suitable insulation material, with a continuous metal fork attached to one end, the shape and dimensions of the fork depending on the shape and dimensions of the insulators to be tested.

In Fig. 5 is shown a type of insulator which it is assumed for the purpose of illustration is to be treated by this method. The method followed in testing this insulator, which it is also assumed is on a 60,000-volt line, will be exactly similar, fundamentally, to the method followed in testing other insulators on other lines. There are two distinct steps in the operation, as in the case of testing suspension insulators, one known as the feeling process, done with the feeling stick (Fig. 3), and one known as the shorting process, done with the shorting stick (Fig. 4). In testing an insulator both the feeling process and the shorting process must be used. The feeling operation or process always precedes the shorting operation or process and determines and indicates the general condition of the insulator. From these indications it is easily judged whether or not the shorting operation can be applied without danger to the line or to the men making the test. Sometimes the feeling operation indicates that the shorting operation cannot be applied at all, as is always the case with very defective insulators. The feeling operation is very inaccurate, but its results will always clearly show if the application of the shorting operation would be dangerous. The shorting operation is very accurate in the locating of defects, but is prohibitively dangerous unless preceded by the feeling operation and its application governed according to the indications of that operation.

In this operation the point of the feeling stick is touched to the line conductor and drawn slowly away, when a distinct buzzing sound will be produced, which in the case of a 60,000-volt line will be audible until the feeling point is an inch or more away from the line conductor. The operation is repeated by touching the cement between porcelain parts No. 1 and No. 2 (Fig. 5). The sound will not follow the point nearly so far as it did from the line conductor if porcelain part No 1 is perfect. When porcelain part No. 1 has become defective the sound drawn from the cement between parts No. 1 and No. 2 will increase, approaching nearer and nearer the sound drawn from the line conductor as part No. 1 becomes more and more defective. From the above it is natural to conclude that when part No. 1 is totally defective the sound from the line and the sound from the cement between parts No. 1 and No. 2 will become equal. This natural conclusion is only seldom verified by what actually happens on the line. The sounds approach near enough to equality, under almost any conditions, to excite suspicion about part No. 1, while it is the very exceptional instance in which they become equal.

The operation is repeated by touching the feeling point to and drawing it slowly away from the cement between parts No. 2 and No. 3. If part No. 2 is perfect, the sound given off by the cement will be less than the sound given off by the cement between parts No. 1 and No. 2. The nearer to equality the sounds approach the more defective part No. 2 is. When part No. 2 is totally defective the sounds from the cement on each side of the part will in all cases be equal. This operation is repeated for each unit part of the insulator, and the difference (the greater the difference the nearer perfect the porcelain part) in the sounds given off by the cement on each side of any part indicates the degree of defectiveness of the part between the cement sections. When the last part is reached, which in the case of the insulator shown in Fig. 5 is part No. 4, the feeling operation is made on the cement on one side and the pin on the other side.

Every part of the multiple-pin-type insulator reads true to form during the feeling process except the top part of the skirt. That is, the sounds given off by the cement on each side of a part will be equal when the part between the cement sections is totally defective or dead, whereas the top skirt or part is very often dead and the sound given off by the line conductor on one side and the cement on the other side of the part will not be equal.

When the feeling process gives indications which would tend to show that an insulator is very defective, there is no need to use the shorting process. When the feeling process indicates that the insulator is entirely good, the shorting process should be used. This process locates parts which are only slightly defective and also parts which are totally defective.

SHORTING OUT A PIN INSULATOR

The shorting-out process should be applied to an insulator only after the feeling process has been applied first and he has clearly indicated that it would be entirely safe to apply it. Those parts of the insulator under test which seem to show indications of being defective during the feeling process should be shorted first. This is desirable because if one of the good parts were shorted first the insulator might flash, whereas by shorting the defective parts first the number of good parts remaining is definitely established and is also the degree of defectiveness of each defective part.

To short an insulator of more than two parts, two feeling sticks are required, such as are shown as Fig. 1. In this operation place the feeling point of a feeling stick firmly against the cement between parts No. 1 and No. 2 (Fig. 5). Place one horn of the fork of the shorting stick against the line conductor and hold it there; make and break contact with the feeling point with the other fork of the shorting stick, when a snappy sounding spark will be produced. If the top skirt or part is perfect, the sound of the spark will be that of the normal, good top skirt. If the top skirt is defective, the sound of the spark will be less than the normal spark of the perfect top skirt. As the top skirt becomes more and more defective the spark obtained in the way described will decrease in intensity, but it will seldom or never reach zero.

Then place another feeling point on the cement between parts No. 2 and No. 3 and short with the shorting stick between this feeling point and the feeling point on the cement between parts No. 1 and No. 2. As described above for part No. 1, a snappy spark will be produced the sound of which, when part No. 2 is good, will be less than the spark produced across part No. 1 when part No. 1 is good. The sound of this spark will decrease as part No. 2 becomes more and more defective and will be zero when part No. 2 is totally defective or dead.

The same conditions hold good for all the remaining insulator parts. When the last part, No. 4, is reached, the shorting is done between a feeling point on the cement parts No. 3 and No. 4 and the pin supporting the insulator, if the pin is metal. If the pin is wood, a feeling point is forced as far as possible up inside the insulator next to the wood pin and the shorting is done between this feeling point and the feeling point on the cement between parts No. 3 and No. 4. In the case of wood pins the bottom skirt acts very much like the top skirt.

On the system with which the writer is connected two men per crew are used in applying this method, and the defective insulators are marked conspicuously with red paint. These testing crews are followed by the maintenance crews, which replace all marked insulators. Both the testing and the replacing of the insulators are done with the lines alive. This method will not indicate cracks on the outer edges of the porcelain parts, nor will it indicate broken insulators. Defects of this nature can be seen by the human eye and need no method of test. It will, however, indicate all defects when they extend into the main body of the insulator, such as cracks, holes, porosity and otherwise defective porcelain. By the use of this method and the removal of the defective insulators all occurrences of punctured insulators on a system can be eliminated.

The methods described for locating defective insulators have been developed by the writer in connection with the maintenance work on high-voltage lines of the Georgia Railway & Power Company and have been patented in this country, Canada and most foreign countries.

·

·