LSST Summit Facility
Construction Progress Report
Reacting to Design Refinements and Field Conditions
Jeffrey D. Barr
, William Gressler
, Jacques Sebag
, Jaime Seriche
, Eduardo Serrano
LSST, 950 N Cherry Tucson, AZ 85719
LSST, Casilla 603, Colina El Pino, La Serena, Chile
The civil work, site infrastructure and buildings for the summit facility of the Large Synoptic Survey Telescope (LSST)
are among the first major elements that need to be designed, bid and constructed to support the subsequent integration of
the dome, telescope, optics, camera and supporting systems. As the contracts for those other major subsystems now
move forward under the management of the LSST Telescope and Site (T&S) team, there has been inevitable and
beneficial evolution in their designs, which has resulted in significant modifications to the facility and infrastructure.
The earliest design requirements for the LSST summit facility were first documented in 2005, its contracted full design
was initiated in 2010, and construction began in January, 2015. During that entire development period, and extending
now roughly halfway through construction, there continue to be necessary modifications to the facility design resulting
from the refinement of interfaces to other major elements of the LSST project and now, during construction, due to
unanticipated field conditions. Changes from evolving interfaces have principally involved the telescope mount, the
dome and mirror handling/coating facilities which have included significant variations in mass, dimensions, heat loads
and anchorage conditions. Modifications related to field conditions have included specifying and testing alternative
methods of excavation and contending with the lack of competent rock substrate where it was predicted to be. While
these and other necessary changes are somewhat specific to the LSST project and site, they also exemplify inherent
challenges related to the typical timeline for the design and construction of astronomical observatory support facilities
relative to the overall development of the project.
facility construction, field conditions, excavation, geotechnical, mass estimates, heat loads, interface control,
design requirements modifications
is an ambitious project to build an astronomical observatory that will conduct a wide, deep and fast survey,
mapping the night sky every few nights and alerting scientists to changing transient objects
the way astronomical research is conducted. Identifying and reacting to changing conditions is also is an integral part of
the entire design and construction process for observatory facilities, requiring prompt, sensible and resourceful handling
at every stage.
To underscore the way in which the time frame for the development of the LSST summit facility has created both
opportunities and challenges for timely design decisions, this paper first presents a brief annotated chronology of the
early process, highlighting the major project milestones and the commensurate level of facility design required to
support them. That is followed by a fuller description of recent modifications instigated by changing requirements and
field conditions encountered during construction, concluding with the current status of the summit facility project.
2. DESIGN AND DEVELOPMENT OF THE LSST SUMMIT FACILITY
2.1.1 2004 to 2006
Conceptual Development Leading to Site Selection
During the earliest exploration stage the basic nature of the project was characterized and implications for an appropriate
site and facility were identified.
Basic ideas were formulated for building size, accessibility, response to climatic
conditions and other mainly site-related factors. Early concept development included: space requirement definition,
feasibility studies to support the site down-selection process, and initial definition of the major driving requirements
from the other main subsystems of the project
camera and data management. Other existing observatory facilities were
also evaluated as paradigms for the LSST facility concept.
The Summit Facility Design Requirements Document
(DRD) was drafted.
The draft DRD included some requirements that are common to many observatory facilities: the typical control room
and standard maintenance facilities, the need to accommodate a coating chamber, provide for instrument support, avoid
thermal turbulence (seeing), design for a seismically active site, and satisfy critical cooling needs. Other requirements
were more particular to the LSST facility, one of which was driven by the compact optical design of the LSST telescope
coupled with a very high stiffness requirement for fast telescope pointing. This dictated a wide pier inside of a relatively
compact enclosure, and for the facility meant that the transport and coating facilities for the 8.4-meter M1M3 mirror
would not fit within the telescope enclosure. A substantial service building adjacent to the telescope enclosure would be
necessary. Other special design factors stipulated that the only observing instrument, the camera,
would be serviced
exclusively on-site during the 10-year survey duration, requiring a dedicated clean room and preparation areas. More
generally, all support systems for LSST had to comply with a mission-critical requirement to minimize down time.
Concurrent with the DRD, the Facility to Camera Interface Control Document
(ICD) was initiated. The camera is a
major system dependent on extensive summit facility support and is not under control of the telescope and site team. As
such, this was one of the first ICDs for the LSST project, as early engagement with the camera team to develop facility
requirements was considered critical.
In close conjunction with emerging technical requirements, early conceptual design
including 3D modeling (Figure 1),
focused on site selection factors: comparative cost, topographical site space, wind studies and facility operations support.
- 3D models developed by the T&S team to explore building layout, massing and orientation
March 2006 -
The El Peñón site on the Cerro Pachón ridge in Chile is selected as the site for LSST.
Informed now by a specific site, a number of more in-depth development activities were undertaken. Early exploration
of environmental permitting requirements began, requiring definition of facility layout, ground coverage, extent of
terrain alteration and other basic parameters. Trade studies were conducted of alternative building arrangements, shapes,
and wind orientations. Geotechnical testing was initiated to determine basic characteristics of the rock strata and to
support finite element analysis of the telescope pier. Rock stability and seismic design studies were prepared by
contracted consultants, as well as alternative layout studies and independent cost estimates.
2.1.2 2007 to 2009
Proposal and Concept Design
The LSST Project Major Research Equipment and Facilities Construction (MREFC) Proposal is
submitted to the National Science Foundation (NSF).
To support the Proposal and in preparation for Concept Design Review a definitive site layout began to be developed for
the summit facility, satisfying the requirements defined in the DRD and now more comprehensively folding in the site-
specific criteria. An Environmental Conditions Document
was drafted. Further site characterization also included
geotechnical testing by IDIEM.
Multiple borings with extraction and testing of core samples were done in the rock
strata under the area that would support the telescope as depicted in Figure 2
right, while the remainder of the site was
characterized by inspection, collected samples and a topographical survey. Results of the geotechnical report helped
refine the optimal location for the telescope pier and confirmed the competent rock necessary for a stiff mount. The
geotechnical and topographical data was also used to determine appropriate site contours for design and specification of
initial leveling excavation.
One clear implication from the first documentation of environmental conditions was that the prevailing steady wind
would be a major factor in shaping the buildings and their relationship to the dome and telescope. Keeping the buildings
low and providing turbulence-suppressing treatment on any large structures adjacent to the telescope would help avoid
ground-heated air being pushed up into the observing path of the telescope. Also affecting building layout and massing
was the available area on the selected site. El Peñón peak is a relatively narrow ridge which is steeply sloped on the
approach side. These factors logically favored a multi-level facility stepping down into the saddle between the main peak
and a small adjacent hill, which was a convenient location for the smaller calibration telescope facility
, another unique
requirement for the LSST facility, depicted in Figure 2
left, adjacent to the concept design for the main LSST facility.
El Peñón with proposed platform & building outlines / Model of conceptual main facility and calibration telescope
While the buildings were being shaped, the telescope height and access for major assemblies was also being established.
Seeing conditions and structural studies were the principal rationale to define the 20.6-meter-above-ground height of the
elevation axis, thereby also establishing the basic height of the enclosure, dome and pier. ICDs for Facility to Dome
and other major subsystems were initiated. Early interactions with vendors and internal engineering
studies determined the configuration of the coating plant, a platform lift, and transport carts as well as other critical
maintenance and handling equipment for large observatory components such as the mirrors and camera.
LSST Concept Design Review is successfully conducted.
In the months following the Concept Design Review, development continued with additional, more focused study of
efficient building layout for the defined configuration. This fuller definition of the facility was driven in part by the
initiation of formal environmental permitting process in Chile, which, as in the U.S., requires baseline biological studies
as well as a definitive facility description adequate to evaluate the proposed use and its potential impact.
Environmental permission is granted by the Chilean environmental management agency.
By the beginning of 2009 the site for the LSST project had been established and investigated for three years,
environmental permission had been secured and the fundamental driving requirements were well defined for the
infrastructure and buildings that would occupy the site.
At this point, preparation began for contracting full design services for the summit facility. In that regard, the LSST
Project management indicated a preference for primary preparation of facility plans and specifications by Chilean-based
architectural and engineering firms. While this was recognized as a departure from the usual approach to facility design
for U.S. and European observatory projects in foreign countries, the familiarity of Chilean firms with the local
construction industry, codes, natural conditions and available materials coupled with the increasing capability of
professional services in Chile advocated in favor of in-country design of the site infrastructure and buildings.
August to December, 2009
Bidding is conducted for full design services of the LSST Summit Facility.
2.1.3 2010 to 2014
Contracted Design and Construction Procurement
Full architectural and engineering (AE) design of the LSST summit facility was kicked-off with the AE firm Arcadis
Chile in March, 2010. This early start, well ahead of construction funding which was anticipated to arrive in FY-2012,
was intended to optimize the use of design and development funding and to allow for a rapid start to the summit facility
once construction funding arrived. The timeline for the MREFC funding of LSST by the NSF would subsequently slip
three years, putting the design of the buildings and site infrastructure even more out in front of contracted technical
advancement of other major subsystems. Preliminary ICDs defining facility requirements related to the camera,
telescope mount, dome, coating facility, and platform lift were pressed into service immediately in the contracted design
of the site and facility.
The building design continued to evolve under the guidance of the Arcadis AE team and LSST engineering, most visibly
in exterior shape. The architects suggested continuous bands of angled wind deflector panels along the service building
walls to catch upwardly induced air flow and channel it around
rather than over
the building. This concept was
proven effective and optimized by computational fluid dynamics modeling undertaken by LSST
Building section and 3D model as developed by Arcadis / Excerpt of wind studies to test building shape variations
While full contracted design of the facility was advancing, the overall LSST project continued in design & development
activities, pending confirmation of funding. To continue timely advance of the summit facility, to better characterize
topographical and rock conditions and to further leverage private preconstruction funding
leveling of the site was
undertaken in 2011. The requisite area and general arrangement of the site had been well established by the emerging
site design. Technical specifications for the leveling excavation had previously been prepared by Arcadis.
Site leveling work: drilling, loading holes, geophone monitoring equipment, one of the ~30 blasts
Figure 4 depicts the basic process for controlled blasting accompanied by vibration monitoring to safely remove about
three-quarters of the total amount of rock necessary to create the requisite platforms and roads. This process protected
the competency of the rock substrate and also provided more precise, dynamic characterization of the rock properties.
LSST Preliminary Design Review is conducted
Summit facility final design package is prepared.
As the NSF-sponsored Preliminary Design Review of the LSST Project was being conducted, the summit facility design
was being carried to a 90% completion level, defined as a formal contractual deliverable. This final design review
package was submitted by Arcadis at the end of 2011. Following an extended period of in-depth review by the LSST
team and outside reviewers, a comprehensive summit facility design review was conducted in March of 2012. The
results of that review were conveyed to Arcadis, who subsequently advanced the plans and specifications to 100%
completion. To compensate for the somewhat early completion of the facility design, still well ahead of contracted
technical development of other subsystems, a subsequent design refinement phase was programmed and later carried out
as described in Section 3 below. Otherwise, the facility construction documents for the summit facility were essentially
shelved for more than a year while NSF funding for LSST construction was secured.
LSST Final Design Review is conducted
Bidding for Summit Facility Construction begins.
The NSF Final Design Review for the LSST project was successfully completed at the end of 2013 (despite a U.S.
government shutdown!). With that major project milestone passed, and the realistic prospect of NSF construction
funding on the near horizon, competitive procurement was initiated for the summit facility construction project. During
that bid process, including two extensions of the due date for proposals, over five-hundred questions were received from
the bidders and responded to by the AE design team and LSST. This extensive consultation between the bidders, the AE
firm and LSST served as a rather intensive
review in its own right of the constructability and completeness of the
design. After all the questions were answered and the bid proposals evaluated, Besalco Construcciones was selected as
the successful vendor and subsequently awarded the contract.
3. RESPONDING TO EVOLVING REQUIREMENTS
At the end of 2014 the summit facility design was complete, a contractor had been selected, and construction was set to
commence. Meanwhile the design-build contracts for the telescope mount assembly (TMA) and for the platform lift had
just been signed, and the contracted design of the dome and coating plant had yet to begin. As those and other major
subsystems began to take shape with the involvement of the respective vendors, a number of changes impacting the
facility would emerge.
This out-of-phase situation of the summit facility with respect to other large contracted elements of the LSST Project had
been anticipated and mitigating measures had been taken. Firstly, ICDs had been elaborated to the fullest extent possible
to define the facility-related requirements of those other major subsystems. The same ICDs that were utilized to design
the facility were subsequently incorporated into the bid packages for the TMA, dome and coating plant. Though the
prospective vendors identified some interface issues, in general the ICDs served the intended purpose of allowing
independent, non-concurrent design of two major subsystems based on a defined, common set of parameters. Secondly,
the initial stage of the summit facility construction contract incorporated a 3-month design refinement period, intended to
Final refinements by the AE firm, Arcadis, in part to address necessary clarifications or incomplete areas of the
design made evident during the extensive question-and-answer rounds of the summit facility construction bid
Suggestions from the selected construction contractor, Besalco, for alternate materials or methods affording
improved economy, maintainability or other advantages (the firm fixed price to be established post refinement);
Interface control revisions based on preliminary design modifications and initial vendor input on related systems.
Lastly, a defined early deliverable in each of the contracts for the TMA, dome, coating plant, platform lift was the
review and recommendation of any final modifications to the interface of their system to the facility.
These measures were carried out to beneficial effect in the optimization and coordination of the summit facility drawings
and specifications, which now effectively comprised the construction documents for the summit facility contract. Even
so, there were still some changes pending and yet to be revealed which would have significant impact on the project.
3.1 Facility Construction Starts and Changes Arrive
Hot and Heavy
In preparation for competitive bidding of the dome
, the LSST engineering team refined its preliminary design. One
notable outcome of that refinement was an increase in the total mass estimate of the dome. The previously defined 556
metric tons (t) plus a 10% contingency was increased to a new total weight supported by the lower enclosure of 691 t,
now including the non-rotating elements (drives and azimuth track) and also stipulating that an additional factor of safety
was to be defined by the AE structural designer. In total, this amounted to an effective ~17% increase in the vertical and
seismic loads imposed by the dome on the facility structure. The lower enclosure supporting the dome had been designed
several years before as a continuous concrete cylinder, primarily to dampen dome induced vibration and to minimize its
transmission to the telescope. This atypically massive structure and the natural stiffness of the curved form of the lower
enclosure now also proved beneficial in accommodating the increased dome mass with limited impact. The main
contributor to the lower enclosure vertical load was its own self weight and the heavy cylindrical structure was also
naturally stable against lateral loads and overturning. Arcadis’
structural engineer determined that
the dimensions of the
lower enclosure walls and foundation did not have to change and that only a minor increase in reinforcing steel and
additional foundation anchors was necessary. The dome contractor, European Industrial Engineering (EIE) has now
verified that the mass of the dome will not exceed the revised ICD-defined maximum mass of 691 t.
The initial work with the dome contractor concentrated specifically and intentionally on the facility interfaces. Early
requests by EIE for modifications to the defined interface focused mainly the base-mounting condition of the dome to
the lower enclosure. As informed by their previous projects, EIE preferred a wider baseplate than defined in the ICD,
which would effectively cap off the lower enclosure wall and provide an outward-projecting edge to help form the
required labyrinth seal. They also recommended the inclusion of an inflatable seal to be deployed when the dome is
parked and being air conditioned during the day, requiring the addition of facility compressed air lines to the dome base.
To ensure coordination of the embedded anchor bolts and base plates for mounting the dome azimuth rail and drives,
EIE preferred to expedite the design and fabrication of the plates and bolts to allow their arrival at the construction site
ahead of the lower enclosure concrete pour. The permanent mounting plates could thus also serve as the ideal templates
for the precise location of the bolts. These enhancements to the design, shown in Figure 5, were all accepted by LSST,
each requiring corresponding modifications to the design and construction of the summit facility.
Evolution of dome base mounting interface and similar previous detail for VISTA telescope (photo credit EIE)
Another significant facility modification was related to the cooling of the dome. According to EIE’s calculation of heat
transfer, the rate of cool air flow from the four air handling units (AHUs) serving the dome would have to be basically
from a specified flow of 14,756 m
/hr. per AHU to 31,000 m
/hr., in order to meet the relatively stringent
required temperature range of evening start-up ambient +0.5/-1⁰ C. That finding was substantiated via coincidental
concurrent review by the facility contractor’s supplier
of heating ventilation and air conditioning (HVAC) equipment.
Since the procurement of the AHUs was already in progress, the findings of EIE and the HVAC vendor were sent
to Arcadis’ engineer
who verified that
given the efficiency factors of the heat exchanger coils available
for the AHUs
a doubling in the air delivery rate would in fact be necessary to meet the defined cooling criteria. The
requisite modifications to accommodate much larger AHUs were immediately undertaken. The more complicated issue
turned out to be a corresponding increase in duct size, which was coupled with a request by EIE to relocate the ducts to
align with the parked positions of the dome arch girders. While this was also a sensible and justifiable change, duly
accepted by LSST, it resulted in necessary modifications to the structural and architectural elements in the two floors of
the lower enclosure that the ducts pass through, while shop drawings for those floor structures were already in progress.
Other dome-related changes to the facility, yet to be fully resolved, include electrical connection location, safety
guards/railings, as well as specification and procurement of a manlift adequate for servicing the dome and telescope.
This latter issue resulted from an optimization of the dome design which increased its swept volume and reduced the
available envelope for the manlift. These details are currently being resolved by the LSST engineering team in
consultation with the related vendors.
3.1.2 Telescope Mount Assembly (TMA)
The design-build contract with Empresarios Agrupados (EA) for the TMA
was initiated roughly concurrent with the
start-up of the construction contract for the summit facility. This allowed for immediate input regarding the interface,
even during the design refinement period for the facility. The specified maximum mass and center of gravity of the
telescope were verified by EA as being consistent with the Arcadis/LSST design of the basic geometry and stiffness of
the telescope pier. The means preferred by EA for mounting the TMA to the pier, however, presented a significant
departure from the ICD and a challenge for the facility designer and contractor. EA expressed a strong preference to not
have anchor bolts embedded in the pier prior to the arrival of the TMA, but rather for the facility contractor to form 400
recesses, 10cm in diameter by 61cm deep, in the top of the pier. (Figure 6, right) These voids would then receive the
bolts which would later arrive at the site as part of the TMA, and then, once the telescope was leveled and precisely
located, the 400 bolts would be grouted into the recessed holes by the TMA vendor. This methodology presented two
main challenges for the facility project: 1) verifying the pull-out resistance of this proposed anchorage methodology in a
highly seismically active zone, and 2) from a simply practical standpoint, how to form the voids. The solution to the first
problem required, enlarging the concrete cap at the top of the pier, confirming the grout adhesion characteristics and also
forming the voids in a conical shape (wider at the bottom) to significantly improve pullout resistance and comply with
Chilean seismic codes. The revised shape of the voids exacerbated the second problem of how to: form 400
recesses, accurately secure the forms in place during the pour and then remove the forms after the concrete had
set. The solution, as proposed by Besalco, was to use forms of expanded metal which would be left in place permanently
after the pour. (Figure 6, center) The mesh of the expanded metal would need to be tight enough to keep the viscous mix
of the concrete and aggregate from flowing into the formed recess during the pour, but would allow the grout, poured
later into the void around the bolts, to come into direct contact with the concrete and create an effective bond. The basic
methodology, the appropriate expanded metal mesh size, and grout bond/pull-out were later tested by Besalco to confirm
the viability of the solution. (Figure 6, left)
Revised anchorage of TMA to pier, incorporating conical voids for bolts, and field testing of methodology
The refinement of the TMA design led to other changes impacting the summit facility. The ring-shaped structural steel
access platform around the base of the telescope, included in the facility contract, was preferred by EA to not be in place
during the installation of the telescope, when that is undertaken about a year after completion of the summit facility
structure. This platform has therefore been removed from the facility contract and is to be fabricated and installed by
others. Other facility-related issues with the telescope have been successfully resolved, including:
The set of hydrostatic bearing equipment defined by EA was larger than anticipated, which required an expanded
new location to be provided in the first floor mechanical area;
The telescope azimuth cable drape required and was provided support from a facility floor below the telescope
inside the pier;
A method was devised to embed attachment pieces for the main beam of the telescope over-travel stop system in the
walls of the pier as they are being poured. The required pieces were expedited for fabrication and shipment by EA.
3.1.3 Coating Plant
The estimated mass of the coating chamber also increased during the bidding process. The defined maximum weight of
the upper vessel increased from 52 t to 64 t based on the advice of the prospective vendors
one of whom identified the
need for a thicker steel wall on the vessel to resist the vacuum pressure and other stresses. The coating chamber is to be
supported on concrete beams in an elevated floor of the summit facility service building, and the resultant ~23% increase
in its defined mass occurred well into the construction process. A request to Arcadis for an expeditious recalculation
resulted in an increase of the reinforcing steel in the supporting beams, which was incorporated into the ongoing
formwork for the concrete structure. Also incorporated into this recalculation were concerns about the transfer of weight
between the lower vessel/mirror cell assembly and the upper vessel during the sealing and vacuum pumping process,
adding an additional 25% for an indeterminate load-transfer factor. Another structural concern by the coating system
vendor, Von Ardenne (VA), is the seismic design of the coating chamber support system. This and other final definitions
of the interface are now being addressed in the early stages of the coating plant design-build contract with VA.
Since the coating plant contract was not executed until more than a year after the facility was already in construction, it
represents perhaps the clearest case of interface challenges resulting purely from out-of-sync design, including a cooling
issue described below in section 3.1.4.
3.1.4 Heat Loads
At the outset of contracted design of the summit facility, the observatory systems that would require cooling by the
facility chilled water system were identified by the LSST team. To serve these needs a dual cooling system was defined.
Cold-temperature ethylene glycol water (EGW), with a varying temperature set point to allow equilibration of equipment
with exterior ambient air throughout diurnal/seasonal variation, was to be provided for telescope and enclosure loads.
EGW of a constant temperature of about 7⁰C, more normal for building HVAC systems, was to be provided by a
general-purpose chiller for cooling control/office areas, computing equipment and other facility-related loads. This
general-purpose chiller was also specified to serve the mirror coating plant.
Arcadis’ HVAC designer devised a system
comprising three, identical, 50-ton facility chillers, wherein at any time two of the chillers (100 tons of cooling) would
be dedicated to serving the cold-temperature loads, and one chiller (50 tons) would serve the general-purpose loads.
Those general-purpose loads, except for the computing equipment, were considered non-critical (offices) or periodic
(coating chamber), so the general-purpose chiller was defined to be configurable to be switched over to back up the
critical cold-temperature chillers. In this case the computer room cooling would be served by a heat-exchanger fed off of
the cold-temperature system.
During the facility design, the specifications for the majority of the heat producing equipment were still only conceptual.
The LSST team, therefore, generated estimates based on the most reliable information available
mainly heat loads of
comparable types and sizes of equipment at existing observatories and early consultation with prospective system
vendors. As can be appreciated by the comparison in Table 1 of these 2010 estimates with more recent 2016 estimates,
now based primarily on contracted vendor data
the largest heat loads on the cold-temperature chiller have increased.
Other heat loads which were not previously considered have been added or transferred from the general-purpose chiller.
As the totals indicate, the projected cold-temperature heat load to be served has essentially doubled. Though not shown
in the table, similar post-facility-design increases occurred in the two biggest heat loads served by the general-purpose
chiller as well, with the computing equipment load going from 30 to 53 kW (8 to 15 tons) and the coating plant heat load
still pending but expected to increase by at least 50% over the currently defined 70 kW (20 tons).
Estimate of Heat Loads for Design of the Facility Cold-Temperature Chiller and Related Equipment
In the facility HVAC design with the initially estimated loads there was ample reserve capacity in both the cold-
temperature and general-purpose chillers as the sum total of all the maximum loads was less than two-thirds of the
available cooling. The total loads as now estimated exceed the available capacity, if assumed to be in maximum demand
simultaneously. Also to be considered, the chillers, as specified for a semi-interior protected installation and capable of
cooling to a wide range of set points down to a low of -8⁰C., are not available in a higher capacity, and a fourth chiller
will not fit within the available area in the building. Moreover, the flow rates and head loss of the EGW system had also
been underestimated, and the specified recirculation pumps were integral to the chillers.
The effective resolution of these cooling issues required a multifaceted approach. To begin with, the heat loads were
more closely scrutinized for concurrence and criticality. An obvious target in that regard was the dome air cooling
biggest load on the cold-temperature system (See section 3.1.1 regarding a related revision to the AHUs and ducts for
dome cooling). Dome cooling is almost exclusively a daytime critical load, while the majority of the other cold-
temperature loads are in greater and more critical demand at night when the telescope is working. This offers a very
helpful non-concurrence of loads which substantially relieves the apparent cooling deficit. Another relief measure being
actively explored is to remove the coating plant loads from the facility chillers and require the vendor VA to provide a
separate dedicated cooling system. The coating system will be in use at very predictable well-planned intervals, so a
dedicated chiller or refrigeration unit could be located in a protected outdoor location to be uncovered and prepped for
use prior to operation of the coating chamber. This also affords the coating system vendor more flexibility to optimize a
system for this very special need, instead of utilizing standard temperature coolant from the general-purpose chiller. VA
is now exploring the viable options.
The issue of insufficient head pressure in the EGW pumping system was similarly addressed via analysis and targeted
solutions. Firstly, the power of the recirculation pump was maximized within the range of pumps available for the
increasing it from ~20 to 25 meters of head. The required flow and head loss from that larger pump
was then analyzed system-wide to identify the zones of high head loss. Based on that analysis, piping sizes in both the
facility and on the telescope were strategically increased to reduce head loss and increase flow. These steps effectively
brought the EGW flow into compliance with requirements, except for the distribution loop in the top end of the telescope
serving the camera and M2 mirror. A secondary booster pump is now being developed for the top end as well as a
dedicated on-telescope refrigeration system for the camera, from which the heat will be removed by the EGW system.
These measures in concert effectively address the increased and highly critical demand for coolant flow.
In conclusion, it should be noted that, had these higher heat loads and coolant flow requirements been known at the time
of the facility design, the chillers and piping systems would likely have been specified with greater capacity to serve the
loads with a conservative level of reserve contingency. However, even given the lateness and scale of the increases,
there was adequate designed capacity and enough degrees of freedom to be able to effectively address the issues and
adequately serve all the systems.
Heat Source (cooling demand)
2010 Estimate 2016 Estimate
Dome Air Cooling
Recalculation by vendor
Dome azimuth drives
Recalculation by vendor
Telescope azimuth drives
To be verified by vendor
Telescope elevation drives
To be verified by vendor
TMA electronics cabinets
To be verified by vendor
Recalculation by vendor
M1M3 thermal control
To be verified
Top end assembly thermal
Fixed electronics cabinets
Load moved from general purpose chiller
M1M3 recirculation pump
Camera loads on TMA
Load moved from general purpose chiller
3.1.5 Platform Lift
Although the platform lift is a constituent part of the summit facility project, its design and fabrication are under a
separate contract directly managed by LSST and not part of the main summit facility design and construction. The lift
vendor, Pflow Industries, was selected early (October 2010) so that preliminary vendor definition of the lift could be
incorporated into the building design. When preparation for the full contract for design and fabrication of the lift began
in late 2014, as was the case with other systems, some unforeseen evolution of requirements needed to be addressed. The
heaviest defined payload of the lift, the M1M3 mirror plus its cart and cover, increased from 72 t to 78 t, as estimated by
the LSST engineering team, and the weight of the roof that the lift would raise along with the payload had also
increased, from ~9 t to ~18 t, according to Arcadis. In total this presented a significantly higher maximum load that
would have exceeded the capacity of the available basic lifting mechanisms (Figure 7
left) and the available space for
structural support elements of the lift. The solution required multiple approaches
light-weighting the design of the lift-
up roof structure (from 18 t down to 12 t), trimming the stipulated payload (from 78 t down to 73 t) and more carefully
defining lift cases (e.g. the lift-up roof would not be subject to a snow load when raised).
Drive base and lift test columns in factory assembly / complexity of steel structure around lift and in service building form
Design of the structural steel tower to support the lift also proved challenging in final coordination. The six columns that
support and guide the lift carriage were specified in preliminary design by Pflow to be laterally braced by the building
Arcadis’ structural designers found this
to be problematic, considering the very high seismic loads induced by
the lift in its fully loaded condition. Since the columns and shaft extend well above the firm concrete and steel structure
of the service building, Arcadis asserted that the lift tower should essentially be self-braced and in fact should provide
support for the building structure and cladding in the upper sections
including a significant wind load. It was
determined, however, that the requisite heavy bracing of the tower proposed by Arcadis interfered with the Pflow
mechanisms. The eventual solution was for the lower enclosure and foundation of the facility structure to laterally
support the lift tower at the base-level and in the north-south direction, while the lift tower would be adequately braced
to laterally support itself and the adjacent secondary building structures in the east-west direction above the floor levels.
Not surprisingly, design and verification of this complex, codependent structure required multiple iterations of finite-
element and load-case analysis between the two companies. This new integrated design of the lift-tower/building
structure (Figure 7
right) was not completed until more than six months after the construction contract was underway.
The incorporation of this late change and late-arriving drawings into the already complex structural geometry of the
service building has intensified difficulties with finalizing shop drawings and timely advance of fabrication/erection of
the steel structure. These issues are currently being mutually addressed by LSST, Besalco and Arcadis.
3.1.6 Summary and Impact of Changes
Changes to facility requirements are an inevitable part of the process. Many have been dealt with and others will
continue to emerge. Among those still emerging are a requirement for a laser propagation room and openings/shaft-ways
for the beam to safely pass though concrete and steel structures up to the calibration screen
as well as extra measures to
provide effective electrical grounding. To date, the potential negative impact of the changes in requirements has been
largely mitigated by prompt attention to the issues as they have arisen and an outside-the-box approach to seeking
solutions in cooperative consultation between the LSST team, AE designers, facility construction contractor and other
vendors. Only minor, non-critical-path schedule slip has so far resulted, and the related cost increases have been
considering the significant scope of the revised requirements in some cases. The more important risk
is to satisfactory long-term facility performance and adequate support for all potentially impacted observatory systems.
That risk remains to be fully retired, demanding further verification of the applied solutions as well as continued
vigilance and attention to all evolving requirements.
4. RESPONDING TO FIELD CONDITIONS
Wide, Deep, not so Fast!
The obvious first construction task for Besalco, following mobilization to the site, was picking up the excavation work
where the early site leveling contract had left off in 2011. Enhancing the continuity of that work,
subcontractor was Rocterra Ltda., who had done the previous leveling work. Arcadis had provided the engineering
design and specification of that earlier work as well. As previously described, the underground conditions and
appropriate excavation methods were considered to have been well characterized by geotechnical testing and field
exploration as well as through dynamic impact monitoring during the previous controlled blasting work.
“The Big Rock”
A pleasantly unsurprising finding was that the rock strata at the platform for the telescope foundation was, as predicted
by all previous investigation, competent and very hard
with a compressive strength of 230 MPa and elastic modulus of
more than twice as hard as the concrete for the foundation. The excavation for the 2m-deep by 18m-diameter
pier foundation was required to be executed exclusively by mechanical means (no blasting)
essentially by drilling the
rock to weaken it and then breaking it away gradually with a hydraulic hammer mounted to a backhoe. Following the
completion of the final meter of levelling of the telescope platform, the excavation for the pier foundation began
immediately. Nine months later
and after 12 hydraulic hammer picks had been destroyed in the effort
570 cubic meters of rock had been removed. The slowness of this work, for which Besalco had originally scheduled
about 2 months, led them and Rocterra to recommend alternative, hopefully more expeditious means. Table 2 compares
the mechanical method with the productivity, impact and cost of several alternative methods which were tested.
- Comparison of Excavation Methods Utilized (experiential data from LSST excavation
your results may vary
Avg. Productivity * (m
Impact PPV @ ~10 m Direct Cost ($USD/m
Mechanical (hydraulic hammer)
10 to 50 mm/sec
200 to 300 mm/sec
Controlled traditional blasting (ANFO)
500 to 1000 mm/sec
* For reference regarding productivity, the total amount of required rock excavation by Besalco was approximately 5,000 m
**Indirect cost adds dramatically to overall cost for mechanical means, due to project overhead for slow, long-duration work.
Expansive grout was also tried, but proved incapable of effectively fracturing the hard rock, except in very narrow
wedges with a free face exposed. Unfortunately, these were not the prevailing conditions in the relatively constrained
area of the pier hole. Plasma excavation was tested with better success. The rock was effectively fractured by the rapid
plasma expansion, and with acceptably low impact (verified by testing) on the adjacent rock strata. Induced peak particle
velocity (PPV), as measured by accelerometers and geophones, was found to be an order of magnitude lower than
traditional blasting. This option was also roughly ten times more expensive than normal blasting and more costly even
than the time-consuming mechanical means, so Besalco and Rocterra opted not to use plasma as the primary excavation
technique. They did, however, resort to plasma excavation in the finishing stages where the hydraulic hammer/pick
could not be effectively deployed. In the end, the pier foundation excavation was completed (shown in-progress in
Figure 8), and the rock substrate was pronounced by geotechnical inspection to be undamaged and competent.
For reasons of efficiency, other areas of the work also merited consideration of alternative means. The area of the
platform lift shaft was too close to the telescope pier foundation to allow normal blasting
even utilizing the controlled
methods (limited charge-per-delay, buffering and pre-shear separation) that had been applied to the blasting to create the
platforms and roads. It was also too large a volume, >1,000 cubic meters, to be removed in any reasonable timeframe by
mechanical means, and plasma was again determined to be a prohibitively expensive option. A commercial product
called Softron® was suggested by Rocterra and tested for induced impact and fracturing caused in the adjacent rock.
PPV measurements, post-test inspection by Arcadis geotechnical engineers and cost quotation determined that Softron
was indeed a viable option for excavating the lift shaft, and it was also subsequently used in other semi-sensitive areas of
the site. With regard to rock excavation methodology (Table 2), it would seem apropos to paraphrase the old adage:
choose any two”.
In-progress images of excavation of the telescope pier foundation
aerial drone fly-through
(photo credit J. Sebag)
“The Swimming Pool”
In contrast to the very hard rock encountered from the top surface downward at the telescope pier location, the area
where the LSST service building was to be located was known to have several meters of loose rock fill, which had to be
removed. The undisturbed strata below that, however, were expected to consist of relatively competent rock, as has been
almost uniformly encountered on Pachón sites. This was not the case. Over 5,000 cubic meters of decomposed rock and
clay were removed in search of material suitable for the foundations of the 3-floor service building. This impressive hole
a and shown on the right in Figure 9, began as test trenches and bore holes along the column axes
eventually resulting in an expansive continuous pit roughly 30 meters square in plan and up to 10 meters deep below the
defined floor level. Once excavated, the work (and consequences for budget and schedule) were not yet over, as the pit
had to be filled with structurally sound material.
Excavation to find competent rock support for service building & concrete pads for foundations (photo credit E. Serrano)
Re-compaction with granular fill material was determined to be time-consuming and logistically burdensome, so instead
unreinforced lean concrete (H10, 100 kg/cm
) was poured to create platforms under locations of column foundations.
Following that work and the placement of column footings, the remainder of the pit was filled up to floor slab levels
with an even leaner grade of concrete (H5, 50 kg/cm
) (Figure 9, left). In the final analysis, the additional excavation, fill
and related remediation resulted in a significant financial cost and 2 months of schedule slip. Some of that impact would
likely have been inevitable, even had more complete geotechnical testing been conducted under the area of the service
building prior to the structural design. It is possible, however, that with prior knowledge of the conditions structural
engineers could have designed pilings, raft foundations or other means to mitigate the lack of competent rock structure
close to the defined floor levels. These were no longer viable options once the over excavation was done.
Some after-the-fact mitigation did, however, prove possible. The base floor level of the service building was expanded
5m wider into the area that had been necessary to over excavate seeking sound bearing conditions (Figure 9, left). This
expansion added 134 m
(1,440 sq. ft.) of useful floor space at the entry/utility area of the building and reduced the
quantity of required fill, helping to hold down the cost and also gaining practical benefit from the additional work and
4.1.4 Other Unexpected Field Conditions
Other more minor suboptimal natural conditions have been encountered on the summit site. A 1-meter-wide vein of clay
soil running through the service building area required mitigation measures at one of the lift column foundations, but
otherwise presented no problem for support of the structures as designed. Topographical conditions and hard rocky
terrain over the proposed route for the main utility lines required field adjustment for feasible and sensible installation
and maintenance of the lines. The detail for protection of the utility lines in an above-ground installation condition
required a more durable, mortared encasement than had been specified. In several critical areas
at the base of the ramp
that will provide truck access to the service building and at the exterior side of the platform lift
the rock structure
proved unstable and became fragmented during excavation. In these areas the loose rock was cleared back as needed
and sufficient concrete structures were implemented to support the adjacent loads. Some field conditions were found to
be better than expected. For example, the inherent solidity of the rock along the side of the vehicle access ramp was
deemed by geotechnical inspection to be naturally stable enough to allow an extensive quantity of specified rock
anchors, reinforcing mesh and shotcrete to be eliminated.
Adding to the challenges, Mother Nature was not particularly helpful during the first year of the construction project.
The winter season of 2015 on Pachón began with a torrential rain storm in late March washing out roads and ended in
September with the strongest earthquake in the region in 30 years. In between those events there was more snow than
had fallen on Pachón since 2011. The service provider for the site, NOAO-South Facilities and Operations, thankfully
worked promptly and efficiently following every event and only about 20 days were lost to weather and closed road
conditions. Otherwise Besalco worked continuously through the winter, as their proposed schedule had anticipated.
4.1.5 Impact of Field Conditions to Date
The primary impact of unexpected conditions has been the financial cost of the additional work and the schedule slip.
These impacts have been mitigated to the extent possible through negotiation and creative solutions in consultation with
the engineering designers and the contractor. The silver lining is that the most mission-critical aspects of site conditions:
hard competent rock where it matters most; generally clear skies and steady predictable wind; and reliable logistical
support by the site management agency, have not varied from expectations and in fact have been proven out during the
first 18 months of work on site.
5. CURRENT STATUS OF THE SUMMIT FACILITY PROJECT
Excavation work for the project is substantially complete. This activity had a sizeable head start by accomplishing the
first phase of site leveling work prior to initiation of the main construction contract. The subsequent excavation,
however, to complete the leveling and excavation for foundations has been fraught with time-consuming difficulties and
additional unanticipated work.
Following and somewhat concurrent with excavation, the concrete structural work (as shown in Figure 10) has advanced
steadily. By August, 2016 it is expected to be ~90% complete, forming the primary structure of the three floors of the
service building as well as the lower enclosure and telescope pier. This work has also been subject to some delays,
mostly associated with the necessary additional underground structures and, to some extent, due to weather factors and
lack of full optimal staffing for the labor-intensive formwork.
Pier foundation continuous-pour night work, 2/20/2016
floor of service building, 5/24/2016 (photo credit E. Serrano)
In addition to the work at the main site, utility work is also in progress. Besalco has laid 1.4 kilometers of main utility
lines and the construction project will soon be connected to commercial power. The other main focus regarding utilities
has been long-lead procurement of equipment. All the primary electrical equipment (transformer, generator, switchgear,
main panels) are now on site and major mechanical equipment (compressors, chillers, air handling units) are in factory
The steel structural work is currently advancing in both factory fabrication and on-site erection. The steelwork has also
been subject to additional scope to include the erection of the structural support tower of the platform lift. The Pflow lift
is approaching completion in the factory in Wisconsin and is scheduled for testing and acceptance in mid-July 2016.
Overall, the on-site progress of the summit facility is behind schedule. The prime contractor, Besalco, had originally
projected a 24-month construction project, which has been extended to 30 months with an anticipated completion date
now in June of 2017. This is still, however, within the LSST-defined 30-month performance period of the contract. The
team is focusing the contractor’s attention on necessary completion milestones to allow initial
integration of related systems on schedule for the vendors of the dome, platform lift and telescope. In September the
concrete lower enclosure will be topped out and the EIE will have representatives on site to oversee the installation of
the dome base plates and bolts. Integration of the platform lift into the building is expected to start by the end of 2016.
On the very positive side, safety on the LSST summit facility construction site has been exceptionally well managed by
Besalco with oversight by LSST. There have been no construction-related injuries or serious accidents. Safety
regulations and culture have been firmly established and actively enforced. All minor incidents and infractions have been
investigated and effectively addressed. The emphasis on safety
above schedule, cost, or any other metric of the project
will be increasingly critical as the on-site construction project grows in complexity over the coming months.
In early 2017 and beyond the LSST summit site will be subject to expanding occupation by multiple contractors, starting
most intensively with the platform lift and dome erection and followed soon thereafter by the telescope itself. To support
those efforts, an expansion to the Pachón dorm will be provided as well as other on-site logistical infrastructure. By mid-
2017 the summit facility contract is expected to be reaching completion and demobilization by Besalco.
of the site from an adjacent hill
July, 2006 /the same viewpoint
May, 2016 (photo credits J. Barr and E. Serrano)
The early development of the summit site and facilities for LSST was driven by overall project advancement objectives,
such as site selection, environmental permissions, operational commitments, and other early project milestones
requiring a defined facility concept for effective progression. Development of the facility was also advanced to the
forefront due to the obvious need to have enclosed protected structures for virtually all subsequent on-site activities. As
is frequently the case for observatory projects, this put the commitment to formal design of the summit facility ahead of
similar final definition of the very elements of the project that the buildings are designed to serve. Yet even with a head
start on the other parts of the project, the LSST facility is still just off the critical path and faces challenging milestones
for timely integration of the other major subsystems to follow. The situation is further complicated by late-breaking
changes in the facility-related requirements for those subsystems.
This prevalent Catch-22 for observatory facility development has available remedies, some of which were implemented
for the LSST summit facility, and others that would have been advisable based on our experience to date:
Formal facility design should not be undertaken any earlier than necessary, avoiding the temptation to engage AE
building services and expedite the facility design to demonstrate early tangible progress on the project.
Once design is undertaken, the facility should only be carried to a level commensurate with overall project
advancement and as required to serve imminent objectives: site selection, environmental permitting, cost-basis
determination, and major reviews.
Although funding is generally limited during design and development, it is a wise investment in the long-run to
undertake full early site characterization, in particular determining critical underground conditions in all areas where
buildings may be located.
Interface control documentation should be treated as an ongoing process and not as a finished product to be serially
finalized and revised. A formal means should be established to utilize ICDs for early development while still
keeping them flexible, including design alternatives appropriate to the realistic resolution level of the interface.
Systematic periodic updates of estimates for critical requirements of other subsystems should be programmed into
the design process for the facility, including: mass and dimensional estimates, heat and electrical loads, utility
connections, and transport equipment.
At the point of 100% completion of AE drawings and specifications, and ideally after bidding, some form of design-
refinement phase is extremely beneficial
preferably also including input from the facility construction contractor
and, to the extent possible, the vendors designing the other major facility-user systems.
As in many endeavors, establishing an effective team approach is the most important factor for successful development
of observatory facilities. At the core of the team is a qualified AE design firm. Selecting an AE based near the project
location is a significant benefit during both design and construction. The construction contractor is the other key
contracted team member. From the initial groundbreaking to the final punch-list, the construction contractor should be
treated as a full partner in the process. As a critical information resource, the facility-development team also needs to
include start-to-finish involvement by key members from other related parts of the observatory project. For the LSST
summit facility project, maintaining this collegial partnership has been key to its successful realization to date. Working
together we will continue to face all challenges and build a suitable place for transformative astronomy to take place.
This material is based upon work supported in part by the National Science Foundation through Cooperative Agreement
Award No. AST-1227061 under Governing Cooperative Agreement 1258333 managed by the Association of
Universities for Research in Astronomy (AURA), and the Department of Energy under Contract No. DEAC02-
76SF00515 with the SLAC National Accelerator Laboratory. Additional LSST funding comes from private donations,
grants to universities, and in-kind support from LSSTC Institutional Members.
Architectural & Engineering Design of the facility and construction support
General Contractor for Construction of the LSST summit facility
Kahn, S., “Final design of the Large Synoptic Survey Telescope,” in
[Ground-Based and Airborne
Telescopes IV], Hall, H. J., Gilmozzi, R., and Marshall, H. K., eds., Proc. SPIE 9906, in press (2016).
Gressler, W., et.al.
“LSST Telescope and Site Update,”
in [Ground-Based and Airborne Telescopes IV], Proc. SPIE
9906, in press (2016)
Barr, J., Sebag, J and Krabbendam, V.,
Summit Facility Design Requirements Document”, LSST LTS-53
)", (2005 orig., 2009 version-control).
Araujo, C., et.al.
“Overview of LSST Mirror System”,
in [Ground-Based and Airborne Telescopes IV], Proc. SPIE
9906, in press (2016)
Sebag, J., Nordby, M.,
to Summit Facility Interface Control Document”,
), (2005 orig., update 2014).
Sebag, J, Barr, J and Krabbendam, V.,
Environmental Conditions Requirements”, LSST LTS-54
), (2005 orig., version-control 2009, update 2011)
IDIEM, “Geotechnical Survey & Foundation Study –
LSST Telescope at El Peñón”,
Sebag, J., et.al. “Dome
to Facility ICD”, LSST LTS-101
), (orig. 2010, update 2016)
Sebag, J., et.al. “TMA
to Facility ICD”, LSST LTS-77
), (orig. 2010, update 2016)
Phillips, D., Barr, J., “Poster
on site-leveling blasting program
presented at Preliminary Design Review”, LSST
DeVries, J., et.al.
“LSST dome status”,
in [Ground-Based and Airborne Telescopes IV], Proc. SPIE 9906, in press
Callahan, S., et.al.
“LSST telescope mount assembly”,
in [Ground-Based and Airborne Telescopes IV], Proc. SPIE
9906, in press (2016)
J. Sebag; P. Zimmer; J. Turner; J. McGraw; V. Krabbendam; A. Tokovinin; Edison Bustos; M. Warner; O. Wiecha,
“Surface Layer Turbulence measurements on the LSST site El Peñón
using microthermal sensors and the lunar
Proc. SPIE. 8444, Ground-based
and Airborne Telescopes IV, 844469. (September 17, 2012)
J. Sebag; K.
Vogiatzis; J. Barr; D. Neill, “LSST
summit enclosure-facility design optimization using aero-thermal
Proc. SPIE. 8449, Modeling, Systems Engineering, and Project Management for Astronomy V, 844904.
(September 25, 2012) doi: 10.1117/12.925446
J. Sebag; K. Vogiatzis, “Estimating Dome Seeing
, Proc. SPIE. 9150, Modeling, Systems Engineering, and Project
Management for Astronomy VI, 91500R. (August 04, 2014) doi: 10.1117/12.2054436
J. Sebag; Victor Krabbendam; Chuck Claver; Jeff Barr; Joshua Barr; Jeff Kantor; Abhijt Saha; D. André Erasmus,
LSST Site Selection”,
Proc. SPIE. 6267, Ground-based and Airborne Telescopes, 62671R. (June 14, 2006) doi:
Ingraham, P. et.al. “The LSST calibration hardware system design and development”,
in [Ground-Based and
Airborne Telescopes IV], Proc. SPIE 9906, in press (2016)
Kurita, N. et.al.
“Large Synoptic Survey Telescope camera design and construction”
in [Advances in Optical and
Mechanical Technologies for Telescopes and Instrumentation], Proc. SPIE 9912, in press (2016)