The Deployment Challenge: Technology vis-a-vis the Business Case
Raymond A. Boy& Mark E. Thornton National Fuel Gas Distribution Corporation 10 Lafayette Square Buffalo, New York 14203 and Richard M. Jahoda AGRA Baymont 14100 58th Street North ClearWater, Florida 33760-3798
Identifying the Opportunity
The initial scope and functionality of National Fuel’s automated mapping/facilities management/geographic information system (AM/FM/GIS) was focused on its mapping and work order systems. There was general agreement that the existing maps, while of a high quality, were labor-intensive in terms of maintenance. Ad hoc map product preparation also was labor-intensive, involving the timeless clip-and-tape method. Furthermore, the map products were updated for users on an annual basis; during the period between updates, it was a common practice to post changes to the maps at the respective service centers in order to keep the maps “current”. Additionally, because the maps were hardcopy, they did not lend themselves to spatial queries and analysis which is possible with vector data attribution. The maps of National Fuel’s facilities were maintained by the respective Engineering Departments in New York and Pennsylvania. Prior to the inception of the project, the maps were routinely re-issued on an annual cycle basis, with mapping changes captured via the work order system. Hardcopy expenditure requests and completion reports (as-builts) as well as a plant accounting front end developed in-house (Capital Expenditure Request Tracking System -- CERTS) were used to provide the information for the mapping updates. Products of the mapping process include system maps, operating maps and age records (also referred to as “strip” maps). It was a primary objective to eliminate the redundant processing of information necessary to support the company’s various applications in its information infrastructure. This elimination of redundancy was the basis for the productivity improvement envisioned. Because the new system entailed the single-event processing of as-built drawings for updating graphical and other plant records, map updates were available to the Engineering, Gas Dispatch, and Operations Departments on a more frequent basis--- daily, if desired, if viewed electronically, or quarterly while in the transition mode moving away from hardcopy. From the inception of the project, it was agreed upon, that the maps depicting National Fuel’s facilities, would serve as an excellent graphic reference source once digitally converted and that these maps, once in digital format, could serve as the foundation upon which a host of second phase applications could be built and deployed. Global optimization of replacement, address look-ups, and network flow analyses were some of the applications at the top of the list of potential developments. Deployment Strategy National Fuel’s AM/FM/GIS team quickly came to the realization that a fast way to bring the project’s deployment to a screeching halt was to overburden the team’s limited resources by trying to bring up too much too soon. Deployments never occur without at least some degree of unexpected problems. Therefore, incremental deployment helps narrow (somewhat) the problems that can surface. Furthermore, and even more critical, incremental deployment lends itself to justifying the business case for each of the applications planned to be rolled out. The criteria used in these assessments of the individual applications included costhenefit ratio, quality of service, data integrity and reliability, and timeliness/accessibility of system data. These criteria were the driving forces behind the prioritization methodology. In order to expedite the system’s usability as well as to minimize cost outlays, National Fuel’s AM/FM/GIS team decided to take a unique approach to its deployment by adopting a mixed raster-vector hybrid system of maps. This hybrid approach was accomplished by scanning, warping and edge-matching all of its operating maps into a seamless, raster operating map within which pockets of vector would be populated over time as replacement work and new mainline was installed. An enhanced ETAK landbase was used as the reference landbase to which National Fuel’s maps would be warped. This overall approach offered National Fuel the distinct advantages of bringing a digital system “up” in a relatively short time period, and gaining the benefits more quickly by eliminating redundancies sooner. By shortening the implementation period, the backlog of jobs to be posted to the master model during the transition window was also kept more manageable. Transitioning Technologies Every effort was made to minimize internal cultural shock in terms of the system’s functionality. In as much as National Fuel’s hardcopy system was workable, albeit technologically based on “pen and ink,” much of the system was replicated digitally. Intergraph’s FRAMME system was the base system being implemented, although it was heavily customized to meet NF’s requirements defined in a detailed specification. The customization was delivered in modular fashion due to its extensive nature. Although originally intended to be deployed in Unix, and while integration testing of the modules was underway, a business case was made for migrating the system to NT; this was performed prior to the prototype phase. The shift from Unix to NT saved maintenance hard dollars as well as simplified the skill set requirements necessary to perform system administration. It was decided that migration to NT was inevitable and that the earlier it was performed, the impact to the system’s user group would be minimized in as much as the user group was still in the pre-AM/FM/GIS mode of operation. It was early in the project that everyone came to realize and appreciate the value of employing detailed specifications in dealing with outside parties (vendors) and the phased approach. Detailed, yet understandable specifications helped reduce the element of surprise for both the vendors and the NFG team, helping provide the needed mutual agreement on what was the desired deliverable and what the critical criteria were for acceptance. Given detailed specifications, the vendor could then intelligently quote a price with known deliverables and quality expectations, as well as a reasonable timeline. An early lesson we learned was that if concerns were openly communicated and looked upon as opportunities for solutions rather than problems to be endured, seemingly ugly issues were worked around. Simply put, in working through the specifications, get agreement to what is “black and white” and then explore a mutually acceptable and livable answer for the “gray”. Given enough creativity and open-mindedness, both parties may just end up in a win-win situation. An example of this is where the conversion effort involved product of highly variable complexity. A sampling of jobs from the spectrum of complexity was worked to build some experience and helped refine the estimate for doing the work. Job-by-job authorization based on individual scopes quotes also provided a manageable approach to control workflow and spending, while offering the opportunity to not contract out a particular potential project segment if the cost just seemed too steep and not cost justified. The phased approach entailed the gradual “easing in” of new functionality; the phases utilized by National Fuel were development, prototype, pilot and full implementation. The benefits are simple--one can test and tweak the newly delivered customized rulebase sofiware or conversion worksets on a limited basis, and if successful, expand the rollout. If the deliveries are “functionally-challenged,” the problems are limited in scope and are not spread or propagated throughout the entire system. To further compress the timeframe in which the newly designed system would become operational, National Fuel also used outside expertise to jump-start the conversion process. The Challenges of Conversion While the newly-developed system was still in the later stages of development and incremental delivery, and in order to minimize the total elapsed time to bring the system “up,” National Fuel set about the challenge of loading its several thousand hardcopy operating maps into its new state-of-the-art system. These operating maps were to be digitally converted with the goal of creating a single, seamless, positionally-representative operating map from which sections or subsets could be extracted and used for office and field reference, look-ups, as well as creating Expenditure Requests (ER’s). After a thorough multi-faceted review including actual conversion of a mini-area, AGRA Baymont was selected as the conversion vendor. This selection process was conducted in parallel with the tail end of the modular software deliveries and system integration testing and debugging. Recognizing the magnitude and importance of the conversion effort, these respective teams worked closely and in detail to define what the end users were doing with the data. This included how the final data looked, and what could be done as a team to get all the users trained to be productive. While at the same time, maintaining the quality mapping product that the field and other internal customers had come to expect over the years, in the least amount of time. Although the conversion costs were also significant in terms of total project spending, the team decided to proceed cautiously and perform a prototype test of a very small portion of the service territory in order to test the mutual clarity and completeness of the agreed-upon conversion specification. It was of primary interest to both parties to ensure that the conversion deliverables were in fact what was envisioned and that they are usable. This prototype was looked upon as a final chance to iron out any differences in interpretation of the specification, and to minimize any potential for after-the-fact rework. The scanning-warping-edge matching conversion proceeded largely to plan and with few surprises. The maps were seamless and maintained via the work order process, and also could be kept on a current basis; i.e., maps could be printed out and disseminated via hardcopy, downloaded and sent out via CD or tape or viewed via network connections and viewed “virtually real time”. What was created, then, by NFG and AGRA Baymont was a hybrid raster-vector set of maps at a fraction of the cost and time to create a fully-vector set of operating maps. Additionally, due to the shortened timeframe of conversion, the work order backlog of jobs requiring processing and posting to the master model was also kept to a minimum. Some “opportunities” did arise, however, after the initial digital conversion was complete. The service area needed to be re-grided in order to make all of the map sheets usable fill sheets. Pockets of non-legible documents also were vectorized since the scanning process, while not causing a degradation in map usability, could not improve the readability of a marginally readable original, even at high dpi scanner settings. Where overlapping maps carried different pressure lines, it was decided that confusion could be greatly reduced by consolidating the maps onto a single layer, using the best available Iandbase as the new master landbase, and having all information not on the master maps vectorized over to the new masters. The resulting seamless, nonoverlapping nature of the maps greatly enhanced readability and simplified use of the maps. Selective processing of backlogged work orders was also addressed through contract services. This business case decision was based on the need to get updated maps out to the field as soon as possible given the technology change to fully electronic mapping. The approach to attacking the backlog was a conscious decision to best utilize the vendor’s digital mapping expertise and National Fuel’s in-house institutional knowledge. As a cooperative effort, the backlog was whittled down in a timely fashion so that periodic updates could be issued to the user community as was targeted. As a retrospective lesson learned, it was found that productivity could be improved significantly if software changes were made even though the delivered software was built as specified. For example, it was found that enhancement and streamlining of the rulebase to allow users more privileges (e.g., to change and edit different states) was a definite enhancement. The addition of default values to eliminate extra key-ins was also helpful. Work procedures were also reviewed whenever they were identified as being unclear, or too cumbersome or inefficient. Going Forward Now that the base functionality of the proposed system is operational, National Fuel is exploring and moving forward with second-phase applications which could only become fictional after the core was built and running. Leak plotting is one such application, and interfacing with the Land Department and Continuing Property Records are two other such applications. Leak plotting (where leaks are plotted by latitude/longitude) allows Engineering and Operations personnel the capability to perform cluster analysis of leak locations. The clusters can then be examined for their viability of being repaired and/or replaced. The interfacing with the non-Engineering Department applications allows for the realization of productivity benefits achieved by the elimination of redundant data entry, as well as the ease of visual referencing of data which is less than user friendly when viewed in tabular format. Other applications such as spatial queries, network analysis, and new customer application processing are in various stages of review for potential deployment. While National Fuel feels confident that its design and specification of its system could support all of the applications cited, National Fuel instinctively first questions the business case before proceeding. Since, even though an application is technologically feasible, does it in fact help keep hold down costs, provide better service, or make for enhanced system reliability. Concluding Thoughts What National Fuel accomplished was only possible because of the team efforts of not only its own internal AM/FM/GIS team members, but also because of the team efforts of its vendors. In terms of a bottom-line, it took all of the team players working in concert to make the project a reality. National Fuel acknowledges the uniqueness of its approach and is convinced that a significant portion of its success was achieved due to its incremental and phased implementation approach, while also keeping in the forefront the business case being addressed. | ||
|
|