These trends have inspired the research and engineering of robotic inspection technology for industry use.  Diverse organizations like the US DOE, NASA, aerospace technology companies Lockheed Martin and General Electric, petroleum giants BP Amoco, Standard Oil of New Jersey (Exxon/Mobil), an interested regulator Texas Natural Resource Conservation Commission and an advanced technology service provider Solex Robotics have been active participants and agents of this change.  Submergible robots have been developed and can now perform ultrasonic surveys of a tank floor bottom while the tank remains in-service and full of product.  The systems have been field proven in tanks fixed tanks, floating roofs and internal floating roofs.  The American Petroleum Institute has recognized these recent advances in robotic tank inspections in their amendment to Standard 653.¹  As the first official departure from the requirement for a manual inspection, tank owners and operators can now choose to use in-service robot submarines for their tank bottom inspections to determine corrosion rates calculations and floor settlement measurements.  The result of this change will be significant savings to the owner with increased safety for the work force and greater control for meeting more stringent environmental standards.  According to the ballot language justifying the modification to API 653, the cost of a tank outage ranges from $50,000 - $500,000.  The new robotic systems have demonstrated savings of more than 80% of these costs. 
 

SAFETY FIRST

Before the cost savings of not having to take a tank out of service can be realized, the question of safety has to be answered.  In order to operate in Class I Division 1 Group D environments with gasoline (flash point 536-880 ºC 280-471 ºF) or fuel oils (flash point 410-765 ºC 210-407 ºF), the equipment²  operating in that environment is required to be approved or listed by a nationally recognized independent testing laboratory³, such as Intertek Testing Services, Underwriters Laboratory, or Factory Mutual Research.  These organizations apply rigorous testing and ongoing auditing to support the certification process.  As part of the product safety testing process, these organizations will set materials on fire, cause parts to fail, ferret out the engineering deformities, and otherwise have the interesting job of making products explode.  The reasons are many.  Electronic equipment is capable of generating and releasing electrical and thermal energy during both normal and abnormal operating conditions.  The materials that equipment is constructed from can be capable of causing sparks.  In powered equipment, heat can build up to the point where it is capable of igniting flammable vapors.  Oxygen, which is a key element in creating an explosion, is also present inside most equipment, even if it is not present in the explosive environment.  A common sense example of one safety issue for tanks follows:  At the manway opening, oxygen and fumes are present in mixture.  The manway is usually made of carbon steel.  If any material that can spark scrapes against the manway enclosure, a spark can be generated.  Because of the size of robot it will scrape the sides of the manway at the time of insertion.  Therefore, the outer shell and attachments for in-service robots must be made of materials that do not spark.  For these reasons, safety standards and the certification process have been developed to insure safe operations of equipment in hazardous environments. 

In connection with a research contract with the U.S. DOE Office of Industrial Technologies⁴, Solex Robotics developed Maverick to meet safety certification requirements for operation in Class I Division 1 Group D hazardous environments.  This means that a robot machine is now capable of relieving the burden and hazard of sending and supervising large work crews into a tank to prepare it for inspection.  Normally, the work force will accumulate several thousand man-hours of work in confined and hazardous spaces during a tank outage for an inspection.⁵  The work is typically performed with respirators or fresh air masks to protect the preparation crew from the residual fumes.  Additionally, explosion- proof lighting, ventilation equipment, air monitors, and body harnesses are just a part of the tools necessary to perform the task of making a tank accessible by the inspection crew.  As a paradigm shift from the current manual inspection, in-service robotics eliminate or significantly reduce these requirements. 
 

THE ECONOMICS OF ROBOTICS

In addition to the manpower and personnel protection equipment, conventional inspections require an array of equipment including hydroblastors and washing equipment, incineration blowers, methane tanks, particulate filtration and collection systems, vacuum truck and waste containment systems, and inspection equipment.  All this has to be planned, managed and paid for by the owner or operator. 

Frequently, elements of outage costs are overlooked at the maintenance and inspection level since the items do not appear in their department’s budget.  Lost revenues will directly reduce sales income.  The cost of draining and refilling will end up in the utility budget.  The insurance premiums for hazardous and confined space work will be included in the insurance expense ledger.  The safety and education training for personnel will be in the training budgets or in payroll.   Actions like taking a tank out of service will usually necessitate locating and often renting alternative storage.  In a marketing terminal, the lack of a suitable alternative could mean sending the customers to competitors.  Performing a turnaround is lost opportunity cost for management to work on other projects.  Effectively locating the source of leaking product will reduce inventory losses and costs of environmental clean-up and fines.  These are just some of the cost factors that the terminal operator must weigh in order to keep from draining the bottom line.  The following table illustrates “typical” costs incurred due to a tank outage for an inspection.
 

ENVIRONMENTAL AND REGULATORY IMPACT

Leakage of petroleum and chemical products into the soils, ground waters, coastal environment and air can produce significant consequences at the federal, state and local levels.  The major federal legislative acts that are applicable follow: Clean Water Act; Spill Prevention, Control and Countermeasures; DOT Pipeline Tank Regulations; Clean Air Acts; Resource Conservation and Recovery Act; SARA Title III; and, Oil Pollution Act.  Demonstrating to the regulators that active leak detection, inspection, maintenance and repair programs are in place help in terms of keeping their goodwill.  In-service robotic inspections coupled with effective communication should help provide comfort with regulators in meeting compliance requirements.  An active program to locate and repair leaks is a prime goal for these pollution prevention measures.  Not having to degas, process or collect VOCs and ventilate tanks for manned entry is a derivative benefit from automated robotic inspections.  According to the Texas Natural Resource Conservation Commission, substantial greenhouse gas emissions occur during tank cleaning operations.  The use of in-service inspections will source reduce these greenhouse gas emissions by eliminating the need to drain, vent and clean tanks prior to inspection.  In order to avoid releasing greenhouse gases, such as benzene and other chemicals on the Toxic Release Inventory, operators will typically burn the evacuated fumes with methane and other gases to eliminate the VOC release.  Nevertheless, incineration generates significant CO2 emissions.
 

Resource Issues	Current  Technology	In-Service Tank  Inspection	In-Service  Savings
CO2 and VOCs (tons)	102.3005	.3003	102.0002
SO2 (tons)	0.0253	0.0028	0.0225
NOx (tons)	0.1720	0.0124	0.1596
Particulate (tons)	0.0328	.0025	.0302
Product Waste (Btu)	1,260,000,000	84,000,000	1,176,000,000
Man Hours in Confined Space	1,200	0	1,200

The above table represents typical pollution prevention and energy savings calculations that are associated with an in-service robotic inspection a 30 m. (100 ft.) diameter 13.8 m. (45 ft.) tall tank containing diesel.  These calculations assume best practices are being used to ventilate and thermally treat VOCs.

In-service robotic inspection is a useful advanced technological tool in the hands of a professional inspector.  Like all tools, it is primarily useful within the context of knowing its capabilities and limitations.   The system is designed to be able to change out the ultrasonic transducers. This means state-of-the-art UT is available.  According to third party experience, the UT data collected by means of in-service robotics is comparable to data taken by out-of-service techniques.⁶   Inside the tank environment, sludge, heavy top-side corrosion, heavy scale, irregular linings or coatings, drain hoses, mixers, pillars, piping and sumps represent obstacles or limitations to performing an in-service inspection.  As the cumulative wealth of inspection experience is continuing to increase and familiarity with operating the robotics tool is acquired, each of these issues is being addressed.  Fundamental to the use of robotics, as well as any other professional evaluation, the skill of the operator is paramount to acquiring useful data for analysis.  The main feature of a robotic inspection is the automation of data acquisition.  A traditional tank inspection will be made on the basis of a visual survey, MFL scan and UT prove out.  The UT prove out will typically consist of acquiring 1,000 - 2,000 discrete UT data points.  A robotic inspection report will consist of 50,000 - 5,000,000 discrete UT data points. In digital form, data can be accumulated and recorded in A, B and C scans.  Recording an A scan slows the rate of travel and data acquisition, but it provides the potential for post analysis by third party inspectors.  Data acquisition and analysis can be done in real time.  This means a curious inspector can go over and over an area by approaching it from different directions.

In a tank without major internal obstacles, the systems have been able to survey 15,000 data points per hour.  Experience has shown that 15% of the floor can be surveyed in three shifts, collecting 275,000 valid data points, in a tank with a floating roof that has no floor lining and containing refined product with moderate sludge. For statistical purposes, this provided an adequate sample size.  Thoroughness of the coverage is dependent on time allowed inside the tank.  The tank operator has the opportunity to expand the coverage by extending the inspection. 

In addition to UT, the robotics systems will carry video capability and depth sensors in the instrumentation payload.  In refined product, particulate is visibly evident.  The depth sensor will be able to perform an API 653 floor level survey.

Using statistics and extreme value analysis to predict the worst case bottom thinning, an estimated corrosion rate can be derived for extended service of the tank.  Prime candidates for in-service robotic inspection are those tanks that have had a base API 653 inspection and need an interval follow-up, ones that need an inspection to determine remaining useful life before being taken out of service, that have relatively clean product, tanks that are in jurisdictions requiring periodic inspections, and tanks that cannot otherwise come out of service due to the cost to inspect.

The major hurdle on acceptance of new technology is the question of reliability and its performance.  To address this issue, Hartford Steamboiler and Inspection Company has agreed to underwrite with a performance money back guarantee certain specific inspection services of Maverick.
 

CASE STUDY

The cost savings provided by in-service robotic inspections can be substantial.  As an example, one tank owner had a gathering tank for light crude production.  The large 206,000 bbl. floating roof tank had been in continuous use since its construction in 1979.  This tank is used for gathering light crude for pipeline transport to barge terminal facilities on the Mississippi River.  In order to complete a conventional inspection, it would have been necessary to hire barges for 35 days to provide alternate storage capacity.  The tank would have needed to be drained, the sludge removed, the bottom cleaned and the shell degassed prior to inspection.  The API 653 inspection consisted of: an engineering evaluation of the foundation and edge settlement; roof, roof vents and seal inspection; wall and roof UT examination; and, featuring an in-service robotic UT floor survey.

The robotic inspection was made difficult by several factors.  Based on samples, the bottom sludge was estimated to be less than 2 inches.  In the initial insertion, Maverick encountered sludge in excess of 25 inches deep and was completely submerged below the sludge line.  Countermeasures were used to dissipate the sludge and enabled the inspection to proceed.  Due to the beveled construction of the chine ring, Maverick was not able to survey the ring.  Collecting ultrasonic data through the bottom floor fiberglass coating was at first considered to be a major obstacle.  The technicians were able match the UT transducers with the substrates and gate the fiberglass readings.  This allowed the system to take back wall readings off the carbon steel floor.  An independent inspection firm was used to verify equipment and operating procedures, the UT data reliability and collection repeatability. 

The results were substantial.  The sonar mapping and data acquisition provided more than 100,000 discrete UT data points for analysis.  The estimates on the setup and take down time were accurate.  Even though there was a torrential downpour on the second and third days of operations, the Maverick inspection proceeded uninterrupted.  Based on prior bids, the in-service robotic inspection saved the tank owner more than $185,000.
 

SUMMARY COMPARISON
 

Conventional  Inspection	In-Service Robotic  API 653 Bottom Inspection
. Tank has to be drained and cleaned.  Waste has to be collected, treated and disposed.  Heavy equipment, such as hydroblasters, vacuum trucks, degassing systems, and waste treatment, is required. .	. Tank can remain operational and full of product.  Certified for operation in #2 fuel oil, diesel, jet fuel, kerosene, naptha, gasoline, and a full range of light and middle distillates. .
. Significant turnaround and tank outage planning required.  Project management for 40+ workers for one or more months is typical. .	. Scheduling is on as-needed basis.  Project management is limited to contractor’s daily work permits. . .
. Confined air space entry is between 600 and 10,000 manhours.  Continuous oxygen, LEL, toxicity monitoring and holewatch is required. .	. Confined air space entry is 0 hours for fixed roofs and 4-8 manhours for floating and internal floating roofs. . .
. Alternate storage is a factor. . .	. Need for alternate storage is eliminated. .
. Tank has to be degassed and vented.  In a tank with 100 ft. diameter, typically the equivalent of 80-120 tons of CO2 is released. .	. . Tank air emissions are –0-. . . .
. Spot inspections and limited surveys are not cost effective. .	. Spot inspections and routine limited surveys are possible. .
. . Typical tank survey consists of visual, MFL and a UT survey of 1,000-2,000 discrete UT data points. . . . .	. Ultrasonic survey is automated and able to record as many as 15,000+ discrete UT data points per hour.  Reports and analysis is based on a survey range from 50,000 – 500,000 data points to determine inspection interval. .
. Can save data in A, B, and C scan formats and differentiate and quantify top-side and bottom-side corrosion. . .	. Can save data in A, B, and C scan formats and, in most instances, differentiate and quatify top-side and bottom-side corrosion. .
. Extent of repairs are unknown until outage.  Vendors and supplies are maintained on standby rates while project management draws up repair plan and scope. .	. Based on survey, tank repairs can be projected with materials and services pre-planned, thus reducing turnaround. . .
. . Sludge and scale can be removed. . . .	. Sludge and heavy scale is a factor and can prevent an effective robotic survey.  Light sludge can be swept aside. .
. Commissioning and refill required. .	. No refill is required. .

Footnotes

1.  American Petroleum Institute Standard 653.  “Tank Inspections, Repair, Alteration and Reconstruction”, Jan. 1991, 4.4.1.2, Addendum 3, 1998.

2.  National Fire Protection Association Section 497 Chapter 2 and National Electrical Code Article 500.

3.  OSHA  29 CFR 1910.106.

4.   http://www.oit.doe.gov/nice3/projects/fctshts/robot.shtml

5.  OSHA  29 CFR 1910.146.

6.  Reynolds, J. and Bell, M., “Utilizing On-Stream Inspection Techniques in the  Petroleum/Petrochemical Industry in Lieu of Internal and Turnaround Inspections”, NPRA Maintenance Conference, New Orleans, May, 1999