6 - OIL INDUSTRY PART 1
by BERNARD A HODSON
Imperial Oil is an affiliate of the huge American enterprise called EXXON which, in the earlier days, went under the names of Standard Oil of New Jersey and Humble Oil I was hired for the Calgary office which was establishing a Western computing facility for what was largely a technical operation. Oil had been discovered a short time before in Leduc, Alberta, and there was extensive exploration work taking place, as well as the creation of petroleum and gas processing plants, pipeline laying and all the associated activity such as dealing with roughnecks who operated on the drilling platforms, oil drilling companies, road construction to the well sites, and of course the financial dealings. In addition there was loads of legal activity as partnerships were arranged and royalties negotiated. I was very naive in those days, thinking that an oil reservoir was like a pool of black liquid that, if it was on the surface, you could float a boat on. When I was later shown a section of rock taken from an oil reservoir it was quite a revelation to see what looked like solid rock, and be told that the rock was porous with the oil being within the very narrow pore openings.
I was the first computer Indian for the company in Calgary and had three Chiefs for whom I was working. Of the four I was the most knowledgeable in computers, Between us we built what became the most formidable technical team in the entire Exxon empire, as I will explain later. Some were hired from outside, others were transferred from other groups within the company. I became Supervisor of this high powered Southern Alberta group. One of my staff later rose to be a Vice President of Imperial Oil, while others were given key and influential posts in the United States.
One of my tasks was to find applications that could usefully be placed on a computer. We did not have a computer of our own at the beginning but used one in the Toronto Head Office (a business oriented IBM 705) and one in New York at the IBM offices there. For a considerable period I would fly down to New York once a month on the British built Viscount Aircraft, a seven hour journey from Calgary to Toronto, followed by a change of planes. From time to time I would also work on a system in Toronto for scientific work, one in an IBM office there. Security in those days was not a problem but was tightened up slightly after a rather amusing incident at the IBM Toronto offices. We were busy conducting our work when a man walked in to the computer room. We assumed he was an IBM maintenance technician until he started pressing buttons on the tape drive to see what happened, which of course aborted the job that was in process.
Shortly after arriving in Calgary I visited the Research labs to determine whether there were any problems that mightbe better done by computer. The knowledge of computers was quite limited, as perhaps is best illustrated by an example. I interviewed a scientist whose specialisation was the mixing of what are known as "miscible liquids". Sometime oil has to be encouraged to flow from the rocks by sending down the well detergents similar to the well known Tide, other times they pump down a fluid to push out the oil. The problem this scientist was working on was what happens when different liquids push against oil, do they mix or does the oil get pushed out by the liquid. His experiment consisted of putting two liquids in a cylinder and observing over time what the interface looked like between the two fluids. He explained the calculations he was making and I assured him that this could be computerised. He then said "would this mean that he only needed to calibrate the instrument, make his observations and the computer would do the rest". I assured him this was so and asked him what was involved with the calibration of the instrument. After being told I said that the computer could do that for him also, at which point he asked what was left for him to do, other than take the measurements. Productivity of research was greatly improved.
Synthetic seismograms were another example for computerisation, but this involved the development of what are known as "digitisers". A seismogram is a reflection of waves received from underground surfaces when some sort of explosion is set off at ground level. Sometimes a "stomping machine" can be used, equivalent to a big giant tramping his feet on the surface. As the underground surfaces are separated by some distance you will get sound wave reflections from each surface, and a good geophysicist can interpret these "seismograms" as they are called and perhaps predict that oil or gas might be found at such and such a particular spot. In many respects they look like electrocardiograms or electroencephalograms such as are seen in hospitals or on one or other of the popular hospital television shows. It would make life much easier if somehow or other these seismograms could be put in a format that was amenable for computer processing. This led to the development of "digitisers" which could take the wavy lines and sample them every few fractions of a millimetre, converting these sampled values to a number. These numbers could then be transferred to a computer for processing. This digitisation speeded up the work of the geophysicist but it was then suggested that if we could synthesise these diagrams, based on "known" parameters, that it might lead to matching the actual with the synthetic to detect what was called an anomaly. Anomalies usually indicated that there might be oil or gas down below.
Ironically, years later, when I was working in the medical field, medical researchers developed something similar in the form of synthetic cardiograms. Had there been any inter-disciplinary contact the medical researchers could have saved themselves a considerable effort by using the work done in the oil industry. This was more difficult than with other disciplines as the oil industry does not like to divulge its innermost secrets. Additional examples of application development within the oil industry are given in later sections, including the potential use of atomic bombs and the setting on fire of underground oil pools.
At this time the company in the West was thinking of installing its own computer for scientific computation, although we had installed a small system for doing data processing related to financial applications, called the IBM 1401. You were only supposed to be able to run one application at a time on the 1401 but we found a way of actually printing the data from a magnetic tape at the same time as it was processing another application. This was an early application of what is known as time sharing. Instead of each application taking an hour we were able to run the two applications together in 1.2 hours, saving us time in which we could do otheruseful activity.
Having been asked to do an evaluation of what computers might best serve our needs I proceeded to do so. At that time some people in the industry had developed what they called "measures of efficiency", in which they calculated the time it might take to do a standard set of calculations, a methodology which still exists today, unfortunately. It is totally unrealistic for most businesses, as I have indicated in my book Modern Data Processing for Management and bears no relationship whatever to what might occur in real life situations, but it has contributed to some of the garbage that often comes out of University computer science programs, and has no doubt pushed some professor or other forward in his or her academic career. Rather than get involved with such nonsense I told the possible vendors (IBM, Burroughs and UNIVAC) that they had to be able to run a series of sample applications I would give them, with real data, and that would be my measuring stick. Furthermore they had to run with whatever operating system was being proposed for our use. Burroughs at this time had come up with an interesting machiner based on a computer language, the other two were more conventional von Neumann type machines.
All three vendors were able to do the benchmarks in Los Angeles so I travelled down there to do them all within a few days. One vendor was located near the airport, one in Anaheim and the other in downtown LA. I spent a small fortune on taxicabs travelling from one vendor to the other, along the freeways of LA. The obvious "best system" was the Burroughs B5000 with the UNIVAC 1107 and IBM 7040 somewhat behind. But, as often happens with large companies, the decision was dictated by the US Head Office which was, "we don't take the best, only IBM", an attitude which was prevalent around the world and which I experienced years later when working for the subsidiary in Canada of a British company. The company decision was made to instal an IBM 7040 which was later changed to a much inferior IBM 1410.
As the initial decision was the IBM 7040, I decided that I should give a course on programming this system and obtained permission from the company to give it at the University College of Calgary, affiliated with the University of Alberta in Edmonton (later the College become a University in its own right) The company were delighted as they said it enhanced their image within the City. Later they jokingly said I was like a "bad apple" as two or three of my staff also decided to give courses of one type or another, again the company being delighted and writing about it in their newsletters. As it turned out many years later, when I was looking for consulting work in Calgary, I had finished up teaching many of the people who later became senior executives in the oil industry.
I also became interested in giving some presentations on local television (described in a later chapter) and again this very progressive company was delighted, writing it up in their newsletters. Three of my staff joined in the fun and we ran for 19 programs given live over local television.
The company was strong on education and also at identifying potential senior managers. I was quite pleased when I was selected for their annual "management development program". This involved spending quite a few days touring the country, being briefed on the working of different departments, visiting the research and refining facilities, meeting and socialising with the very senior management. For some reason I received the wrong instructions, or else misread the instructions, and landed in the first meeting in Toronto some hour or two after the initial briefings had taken place. While it might easily have proven negative for my career it actually turned out quite the opposite, as now I was "known" by everyone on the course, including the organisers. We had an unlimited expense account and part of the course was to see how vulnerable we were to "betraying secrets" if under the influence of alcohol.
I was also sent on another course, this time a five week course on the subject of "reservoir engineering", so that I could become more engineering oriented (being a qualified aeronautical engineer was not much help in getting oil out of the ground). This was held in Tulsa, Oklahoma, under the auspices of Humble Oil. This was interesting but perhaps my criticism of the course explains it best, following which the course was completely revised. and not given in that form again. At the end of the course I stated quite plainly that three weeks of the course was completely wasted as all we were doing was a series of repetitive calculations on a hand cranked calculating machine. In doing so we were able only to do one or two examples of how oil is extracted from the ground and received no feel for how the different parameters of the rock in which the oil was found (permeability, or how easily the oil comes out, porosity, or how many little holes are in the rock containing oil etc.). I stated that a much better approach was to computerise the course and let us vary the different parameters, so that we could get a better feel for the process. The point hit home and the course was for ever changed.
We also had to educate some of the supervisors of operations, one example being with oil drilling. The longer it takes to drill a well the more costly it is, and many parameters affect the speed. As an example the pressure you put on the drill bit, how quickly the diamonds on the drill bit wear out (varying with the rocks you are penetrating), what types of rock you are drilling through (shale, limestone etc.), what lubricants you are using to ease the cutting edge, the speed of rotation of the drill bit, all affect the drilling time. It is possible to develop mathematical equations that help evaluate the optimum drilling procedure but the user of the computer programs has to understand the limits of the mathematics. One solution to these equations says that if you put no weight on the drill bit and rotate it at infinite speed then that is the best way to do it. Obvious nonsense but it was necessary to give the operators some feel for how to interpret the equations.
The same sort of thing occurred with refinery operators. For refineries we were using a recently developed mathematical procedure called "linear programming", a complex series of equations which took considerable time to process on the computers of the day. A typical refinery would receive a variety of crude oils with varying viscosity (some thick, some thin), with varying density, and would have to process this crude oil to produce a mix of gasoline, tar, naphtha, benzene or whatever. Originally linear programming experts from Head office would be responsible for the equation handling but this took away the responsibility of the refinery operators, and the technical experts had other important work, so the company had to devise courses for non technical management that would enable them to run their refineries economically, but would also give them an insight to the correct interpretation of mathematical formulae.
At this time we were also designing several gas processing plants to extract sulphur and other undesirable items. The computer optimisations that were developed enabled many trials to take place, to arrive at the best design solution, rather than the previous situation where only two or three designs could be evaluated.
One of the amusing and instructive events about that time was the arrival in Canada, following the 1956 revolution in Hungary, of two mathematicians, of the same name but unrelated. One joined Imperial, the other IBM. Both were brilliant. Our Hungarian was able to predict, almost exactly, what the solution to the refinery optimisation might be. A few hours later, after the computer had churned away for a considerable time, it showed that he he was invariably correct.
Communications within a multi national company are never easy. We had a need for a program to do a mathematical procedure called "stepwise multiple regression" which would enable us to start the analysis of a problem with as many variables as we thought would influence the result and then, based on established criteria, eliminate those that were unimportant. We developed an excellent program for this. At the same time an almost identical program was developed by colleagues at Standard Oil. It was only when both programs were operational and we simultaneously advised our affiliated companies of our endeavours, that the duplication was discovered. Such events were almost certain to occur with the communication tools available at that time (no Internet or networks).
As a systems analyst as well as a technically knowledgeable person I was able, in a short study I conducted, to show that there was a considerable overlap in the works of several committees. Also, when the company had first discovered oil, a fast communication system had been established, which over the years had become an expensive and hardly useful historical rather than decision making system. Accepting my recommendation that committees be cut and the manpower heavy historical activity be abolished, the company saved far more than the cost of the study. This type of duplicated committee work and obsolescent reporting I found to be typical of several large organisations.
There were some amusing aspects to the "secrecy" within the oil industry. Imperial had developed a technique for deciding whether or not to invest in a project, so confidential that we were not allowed to discuss it openly. After I left Imperial I discovered that several other oil companies had also developed similar techniques, which were also considered "hush-hush".
"Scouting", which has nothing to do with Baden Powell, the founder of the Boy Scout movement, was a clandestine operation in which persons tried to reach, unobserved if possible, new wells which were being drilled by competitors, so that they could report on what they saw happen. In particular they were looking for evidence that oil or gas had been found. The problem in their reporting was to find a secure communication method back to their offices, one that would be free of being tapped. Such activity today has been made much easier by the use of sophisticated monitoring and encryption equipment but in the earlier days it was something of a "seat of the pants" operation.