The isothermal corrosion behavior of two stainless steels and carbon steel in mixtures of NaNO3 and KNO3 has been evaluated to determine if the impurities found in commodity grades of alkali nitrates aggravate corrosivity as applicable to an advanced solar thermal energy system. Corrosion tests were conducted for approximately 7000 hours with Types 304 and 316 stainless steels at 570°C and A36 carbon steel at 316°C in seven mixtures of NaNO3 and KNO3 containing variations in impurity concentrations. Corrosion tests were also conducted in a ternary mixture of NaNO3, KNO3, and Ca(NO3)2. Corrosion rates were determined by descaled weight losses while oxidation products were examined by scanning electron microscopy, electron microprobe analysis, and X-ray diffraction. The nitrate mixtures were periodically analyzed for changes in impurity concentrations and for soluble corrosion products.

Experiments have been conducted with a molten salt loop at Sandia National Laboratories in Albuquerque, NM to resolve issues associated with the operation of the 10MWe Solar Two Central Receiver Power Plant located near Barstow, CA. The salt loop contained two receiver panels, components such as flanges and a check valve, vortex shedding and ultrasonic flow meters, and aq impedance pressure transducer. Tests were conducted on procedures for filling and thawing a panel, and assessing components and instrumentation in a molten salt environment. Four categories of experiments were conducted 1) cold filling procedures, 2) freeze/thaw procedures, 3) component tests, and 4) instrumentation tests. Cold-panel and-piping fill experiments are described, in which the panels and piping were preheated to temperatures below the salt freezing point prior to initiating flow, to determine the feasibility of cold filling the receiver and piping. The transient thermaI response was measured, and heat transfer coefficients and transient stresses were calculated from the data. Analysis is presented which quantifies the thermal stresses in a pipe undergoing thermal shock. In addition, penetration depths were calculated to determine the distances salt could flow in cold pipes prior to freezing shut and validated with panel tests. Freeze/thaw experiments were conducted with the panels, in which the salt was intentionally allowed to freeze in the receiver tubes, then thawed with heliostat beams to assess permanent deformation in the tubes, and to develop procedures to thaw a panel so minimal damage occurs. Slow thermal cycling tests were conducted to measure both how well various designs of flanges (e.g., tapered flanges or clamp type flanges) hold a seal under thermal conditions typical of nightly shut down, and the practicality of using these flanges on high maintenance components. In addition, the flanges were thermally shocked to simulate cold starting the system. Instrumentation such as vortex shedding and ultrasonic flow meters were tested alongside each other, and compared with flow measuremeni from calibration tanks in the flow loop. 4JJ DISTRIBUTION OF THlS DOCUMENT I S UNUh4lTED ACKNOWLEDGMENT We would like to acknowledge the following for their contribution to the molten salt panel and component experiments :

No abstract

No abstract

During the summer and fall of 1987 in Almeria, Spain, a wire-pack receiver was tested by the International Energy Agency/Small Solar Power Systems (IEA/SSPS). The basic operation of the receiver is that: air is drawn through several layers of stainless steel wire screen; concentrated solar flux is directed on the face of the screen pack; the oxidized wires absorb the solar energy; and heat is transferred to the air flowing through the screen. Although the experiment goal was strictly proof-of-concept and was not receiver characterization, modeling efforts were initiated to help understand the experimental results. The steady-state performance of the receiver is modeled using the fact that the net solar and infrared radiative energy absorbed by each screen layer must be transferred to the air by convection. Basic performance trends and typical calculations of receiver efficiency are given. Model predictions and experimentally measured temperatures and flow rates are compared. Model predictions of receiver power and efficiency are generally higher than the test results (operational modifications of the receiver absorber as tested are believed to have produced nonideal conditions), but trends are consistent with experimental data.