SUBSCRIBE NOW!

Vol. 23 No. 10 — Nov/Dec 2004

This automated storage and retrieval system offers nine or ten times the storage space as standard library shelving.

Valparaiso University is located in northwest Indiana, about an hour from downtown Chicago. With about 4,000 students, Valparaiso (Valpo) is primarily an undergraduate institution with professional schools in business, engineering, law, and nursing. While the law school has its own library, all other disciplines were always served through one main library located near the center of campus. Moellering Library was built in 1959 and expanded in 1969 to hold 266,000 volumes in 54,000 square feet. By the 1990s, the collection had grown past 350,000 volumes, and the need to also house computers and other technologies made it clear that Valpo needed a new library.

Back in the mid-'90s, Donna took part in the conceptual planning for a new facility. And by the time Rick arrived on campus in the summer of 1999, Valpo was anxious to begin detailed planning. While there was a clear need for more book space, the evolution of library services over the previous 4 decades demanded a new kind of facility that would meet a wide variety of needs throughout the campus community. So we were going to build a Center for Library and Information Resources nearly twice the size of Moellering that would house many more books while also providing computing centers, group study space, the campus writing center, a lecture hall, training classrooms, and community gathering space. We wanted to have open shelving space for approximately 300,000 volumes and high-density storage for another 300,000 volumes.

Initially, compact shelving seemed to be a logical choice for high-density storage, but we had concerns about it taking over most of the library in the future. Also, while the movability of compact shelving nearly doubles the capacity of standard library shelving, it is not particularly inviting for users. We needed to find an alternative high-density storage system so that we could make more creative use of space throughout the rest of the building. We also wanted to make sure that any storage solution would be located within the building rather than off-site. Since Rick had worked at the University of Nevada Las Vegas during the planning for its new Lied Library, he was familiar with HK Systems and its Automated Storage and Retrieval System (ASRS). This offered an excellent alternative for on-site, high-density storage at Valparaiso University, while offering as much as nine or ten times the storage space per square foot as standard library shelving.

Rick Explains the ASRS

An automated storage and retrieval system is a means by which a large quantity of books (or other materials) can be stored in a single room and can be easily accessed within seconds. A large rack system holds many 2-foot-by-4-foot steel bins. A crane in each aisle of the system retrieves and replaces the bins on demand.

Initially, books are scanned as they are placed in the bins, and a database keeps track of the location of each volume. Later, when an item is requested, a crane pulls the appropriate bin and delivers it to the designated service point. A staff member then picks the requested book out of the bin and scans it to update the book's location in the database. Since order doesn't matter (unless you want it to), when a book is returned and re-scanned, it can go back into any bin where there is space.

In our installation, we have 1,872 bins in two aisles that are 18 rows high and 26 columns long on each side of the aisle. This is all located within a space that's 30 feet wide, 80 feet long, and 30 feet high. The vast majority of our bins (1,456) are 14 inches deep, with 312 at 12 inches deep and 104 at 18 inches deep. The capacity of each bin varies based on the size of the items you place inside, but the total capacity our installation is designed for is 300,000 volumes.

Donna Explains How the Collection Was Prepared for Storage in the Bins

As assistant university librarian for access services, my duties include overseeing circulation, stack maintenance, and our integrated library system (from Innovative Interfaces, Inc.). So Rick asked me to lead the planning for material that would be stored in the ASRS. First, we needed to decide what we would actually store there. I thought that the ASRS would be better received on campus if the collections in it were less likely to be accessed by browsing, and I knew that we'd have enough space for all of the circulating monographs on the open shelves in the new building. Using a committee that included me, the assistant university librarian for collection management services, and the cataloging services librarian, in April 2002 we recommended that the following items be placed into automated storage:

• Periodicals prior to 2000, bound or unbound

• Government publications

• Selected large runs of reference annual publications

• Low-use print indexes

• Artifacts and restricted papers from the university archives

• Unprocessed gifts

This is the most important thing I learned at this early stage: For the system to work with an online request link in the public catalog, every item in the bins had to have an item record with a bar code. This fact was significant, because most of what we wanted to store had never been bar coded. Fortunately, we had about 2 years to prepare these collections before the scheduled move to the new location, where the ASRS would be built.

Although we have been adding Marcive records for our U.S. documents to our online catalog for a number of years, we had never received bar codes with the records. Simply bar coding at the time of circulation had worked very well for the limited use this collection got. But to prepare for storage, in 2001 we began getting bar codes for all currently received documents. The government information librarian and I determined that a smart bar coding project would be most practical for retrospective conversion, saving us the step of manually entering a bar code into each record. In September 2002, using an output file of our existing records, Marcive added bar code numbers, printed smart bar codes with SuDocs numbers and titles, and then reloaded the records into our catalog. We trained student assistants to attach the bar codes to the correct items.

For this system to fulfill patron requests from the OPAC, every item in every bin had to have an item record with a bar code.

In the process of bar coding, we discovered records in our system without corresponding items on the shelves and items on the shelves without records in the system. Thus the project grew to include weeding and database cleanup. Although these side projects were not necessary to the ASRS storage project, they were important in generally preparing the collection for the move to the new building.

IT staff from Valpo, a tech specialist from Innovative, and programmers from HK collaborated to set up this complex system.

Once the document bar coding was underway, I worked with the assistant university librarian for collection management services on the periodical bar coding project. In January 2002, the periodicals assistant began by adding bar codes and creating item records for newly bound volumes. We assumed we would keep backfiles for current subscriptions, so as each volume came back from the bindery, the periodicals assistant forwarded a printout with the correct Innovative bibliographic record to the circulation workers so that they could bar code the earlier volumes. If the run was a straightforward sequence, we used the Innovative multiple-item record-generation feature to rapidly create item records with sequential bar codes. If the run had more complex or irregular numbering, we used double-dumb bar codes. Student workers then applied the bar codes to the volumes.

This work also became a weeding and retrospective conversion project. Once we completed the backfiles for our current periodicals subscriptions, we worked on the closed runs. First, bibliographers reviewed the closed runs for possible withdrawal. Then the remaining volumes were bar coded.

For boxes of uncataloged books, personal files, and archival collections, we developed models and templates of simple bibliographic records to enter into the Innovative system, beginning in October 2003. These records are suppressed from public view. Although we could have stored them directly in the ASRS without a corresponding record in the online catalog, keeping everything in the Innovative system makes it easier for staff to find what they need and to request it.

All items to be stored in the automated system had to have a record in the ASRS database. We needed a way to easily pull together lists of item records in the Innovative system to transfer to the ASRS database. We did this using a combination of item type and location codes for our newly created records. For the government documents and index collections, we had to carefully pull together lists using existing codes. Once the lists were generated, we used the rapid-update feature to change a code that indicated the items would go into the ASRS.

Fitting the ASRS into the New Building Design

Although much of the operation of the robotic system is automatic, staff members still need to be involved for the physical retrieval and return of items in the bins. We were both already concerned about staffing a building nearly twice the size of the old library, and it was imperative that the ASRS be located in the new building in such a way that an existing service point, most logically circulation, could staff it. We worked with our architects (Esherick, Homsey, Dodge, and Davis) to explore our location options. Because of the size of the ASRS rack-and-bin system, this was no small task. Also, because of the extreme weight of the ASRS and the materials it would hold, it had to be installed on a thick slab at grade level. This challenge was further complicated by the usual need for circulation to be near the main entrance, which would be located on the second floor in our new building.

Fortunately, the ASRS could be programmed to deliver the bins at any height within its rack system. So we first placed circulation where we wanted it to be and then worked through various possibilities to build the ASRS at the back of the department. The way it ended up, from the point of delivery behind the second-floor circulation desk, we now look down into the system.

Hardware and Software

In April 2003, we chose the ASRS from HK Systems because this vendor had already worked with Innovative Interfaces on two previous projects. In fact, the Innovative software had gone beyond customized programming to a purchasable product listed in its catalog. Donna initiated the purchase of the Innovative ASRS interface and had the Innovative specifications sent to the HK Systems' main office in Milwaukee. The actual contract with HK Systems was handled by Pepper Construction, the building contractor.

During software setup and testing, we began moving the bin-bound books to the new library.

We knew that we would need a new Innovative server, since our existing one was reaching the end of its life span. Donna placed the order for the new server after ordering the ASRS interface so that Innovative could size our new server appropriately.

As HK Systems developed its project proposal, we brought our campus IT staff into the planning process. Donna arranged for a conference call so that our IT staff, an Innovative technical specialist, and the HK Systems programmers could collaborate. Our IT staff's security concerns over using the Windows operating system for the HK Systems server were solved by setting up an additional network connection to the Innovative server and limiting that connection to only the HK server.

As part of the contract, HK supplied the main robotic system server, a backup server, and a dial-in diagnostic modem. It also supplied PCs for each operator workstation at the front of the bin aisles with attached receipt printers and a networked report printer. Two terminal servers are hidden at the back of the cranes. We supplied the uninterruptable power supplies for the main and terminal servers, along with two PCs to be housed at the back of the cranes for running diagnostic software. Our campus IT unit runs daily backups of the ASRS server through its Legato backup system.

Innovative advised us early on to make only one change at a time so that if something didn't work, we could assume it was caused by the thing we'd just done. In February 2004, Innovative installed the ASRS interface software on our old server so Donna could test the basic OPAC displays. In March, the new server was installed in the old library. In April, HK Systems was ready to test the data transfer component. Our IT unit opened a port in our firewall so that our Innovative server could talk to our ASRS server in Milwaukee. After working through connectivity issues, we worked on transferring records. HK and Innovative worked together to make sure the records were transferring correctly. All the while, Donna transferred test lists after every change and reported back to HK. By the last week in May, HK completed the final installation and testing of computing hardware and software.

Loading All the Bins

While the setup and testing were underway in spring 2004, we began moving boxes and volumes slated for the bins to the new library and Donna transferred their records to the HK database in preparation for loading. Immediately, we discovered that the Innovative system did not transfer the volume number over to the HK database. Since the unique identifier needed to pull each volume from a bin is the bar code, we had our student workers write the bar code numbers on the tops of all bound periodicals.

HK staff provided training and stayed with us for the first week of the physical loading in June. There was some downtime with cranes and software, so the first week went very slowly. We loaded our boxes of uncataloged books first and then began on the bound periodicals.

We wanted to transfer item records to the HK database as close as possible to the time we actually moved them. As groups of items were loaded, Donna used the Innovative rapid-update function to change the item location code to the one that would put the request link in the catalog. This way we were able to use the Innovative list function to track our progress, and the catalog reflected the true location of the item. Some location-code and scanning problems occurred, causing records not to transfer. Once we corrected the problems, the item records were transferred to the HK database and the volumes loaded.

Soon we were able to establish a rhythm. At the end of each day, the move coordinator estimated what would be moved next. Donna made a list before leaving work and transferred the records from home early the next morning, so loading could commence before she arrived on campus. The team used a whiteboard to track bin numbers so that we could load the bins in sequence from front to back and keep track of how many we were using. Circulation staff worked full time for 5 weeks to load 59,000 items while still staffing the circulation desk at the old library.

Solving Problems and Using the Robotic System

With the initial load completed, we started looking at problems. First, we dealt with "lost" items. Occasionally a loader would make a mistake and be forced into a bin audit. If the person did not do the bin audit correctly, some items were unintentionally declared lost. It was easy to generate a report of lost items, find the suspect bin, and rescan the items.

When a record is transferred to the HK database, HK assigns it the status "delivered." Once the item is scanned in the store function and assigned a bin, the HK status is changed to "available." When a patron requests an "available" item, the crane delivers the bin to the ASRS workstation, then a circulation worker pulls the correct item and scans it. At this point, the HK system status changes back to "delivered." These "delivered" and "available" statuses display the same in the public catalog, but when a patron requests a "delivered" item, the system assumes it is sitting somewhere in the circulation area. The OPAC then directs the patron to the circulation desk.

So we generated a long report for items "delivered" but not "available." Many of these ended up being invalid bar codes due to a faulty scanner. Since these bar codes are no longer in the OPAC, we created dummy Innovative records with the invalid bar codes and ran the transfer program in reverse to pull the bad records out of the HK database. For all remaining lost items, a circulation manager went to the new building every afternoon to process any picks and to look for items that may have been put into bins without being scanned. We searched the HK database for related volumes in the runs to identify likely bins in which to look. The most likely bins were pulled and searched for the missing volumes. The found volumes were then scanned into the system.

Our Final Steps and Our Future Steps

We opened for business in our new building on Aug. 10, just a month before this article was due. We continued loading boxes of archival materials and a few overlooked sets through August. And we're still loading older runs of bound periodicals as they come back from the binder. Once we are a bit more settled, student workers will be assigned to do bin audits just as they are assigned shelfreading. In the old building, our goal was to shelfread the entire collection over the course of each year. Now that a portion of the collection is in the bins, we will want to audit at least the bins containing periodical volumes on an annual basis.

Our future steps will include refining the public display in the WebPAC; training our student workers to do routine picks, stores, and bin audits; training our patrons to make requests from the OPAC; and working with our Physical Plant staff to establish a preventative maintenance schedule.

Further Reading

Eagan, Michael. "Project ARS: Story." http://mike.passwall.com/ars/story.html

Haslam, Michaelyn et al. "The Automated Storage and Retrieval System (ASRS) in Lied Library." Library Hi-Tech, v. 20 no.1 (2002): 71­89.

Parsley, Stephen L. "Automated Storage and Retrieval Systems" in Handbook of Industrial Automation, edited by Richard L. Shell, Ernest L. Hall. New York: Marcel Dekker, 2000. pp. 643­657.