5.2.1 Geometrical parameters determining the optical performance

5.2.1.1 Mirror shape

Geometric deviations can be distinguished in mirror shape deviations and misalignment of the receiver (eccentricity of absorber tube and focal line for PTC) [3]. Both parameters have a similar effect on the optical performance [4]. For the mirror geometry, either slope or shape deviations can be measured. For the optical performance, the slope is the relevant parameter as the deviation of the reflected ray in the receiver region is determined mainly by the direction of the reflection while the exact position has a minor effect. Slope deviations are commonly denoted according to the direction of curvature (SDx, SDy) according to the definitions given in Chapter 3, Eqs. (3.4)–(3.7).

The focus deviation as a system specific parameter also takes into account the distance from the point of reflection (surface element or SE) to the receiver (focal point or FP):

$F{D}_{x,\mathit{local}}=2\cdot \left({\left(\frac{d}{dx}z\right)}_{\mathit{meas}}-{\left(\frac{d}{dx}z\right)}_{\mathit{ideal}}\right)\cdot d\left(\mathrm{S}\mathrm{E},\mathrm{F}\mathrm{P}\right)$

Shape deviations imply the difference in 3D coordinates between measured and design points of interest. Slope deviations can be obtained from shape deviations via differentiation of point measurements with sufficient spatial resolution.

5.2.1.2 Receiver position

Possible misalignments of the absorber for line concentrating (mainly PTC) system in x- and z-direction strongly affect the intercept. Here, deviations have a similar effect on the optical performance and thus have relevance as the mirror shape [4].

For point-concentrating systems like solar power towers, misalignment of the receiver plays a minor role. In some case, wind-induced oscillations of the tower have been reported to have an effect on the IC.

5.2.1.3 Tracking deviation

Tracking deviation in PTCs refers to a deviation between the focal plane (defined by the optical axis and the parabola vertex) and the vector to the sun. It is caused by the accuracy and non-continuous tracking of the drive system and misalignment of collector modules. Tolerable tracking deviations for a EuroTrough-type PTC should be below 1 mrad [3]. Dynamic effects caused by external loads influencing the tracking (friction, wind, and static unbalance) are covered in the next paragraph.

For heliostats, the direction of the optical axis of the concentrator has a very large influence on the performance, as small angular deviations lead to distinctive focal deviations due to the relatively large distances between heliostats and receiver. Requirements on heliostat tracking accuracy depend on several factors, like heliostat size, receiver aperture, aim point strategy, and distance between heliostat and receiver. Typical values of the root-mean-square (RMS) tracking accuracy are between 0.8 and 1.6 mrad.

5.2.1.4 Deformation under operational loads

Collectors are exposed to variable loads during operation, affecting their shape and hence their optical performance [5,6]. Operational loads can be classified into wind forces and moments [7,8], bearing friction moments, and dead load, leading to sagging/bending and, if the center of mass of the collector and rotation axis do not match, static unbalanced moments. Possible effects are mirror deformation, lateral deviation of the absorber tube from the focal line (for PTCs), and tracking deviations caused by torsion.

5.2.2.1 Methods to measure concentrator slope

Local slope of reflective concentrator surfaces is directly assessed by exploiting the distortion or deviation of reflected patterns or light rays. Deflectometry or fringe reflection [10,11] uses known regular stripe patterns on a screen or target whose reflection in a specular surface is observed by a digital camera (see Section 3.4.2). From the deformations of the stripe pattern in the reflection, the local slopes of the mirror can be calculated. This method is applied to measure heliostats [12] (see example in Section 5.2.3.1), dishes [13,14], Fresnel mirror panels [15], or single mirror panels of parabolic troughs [16].

The application of a target-based method to entire parabolic troughs in the solar field is limited because of the considerable effort in placement and alignment of the screen or target relative to the concentrator surface. As PTCs are line-focusing, in each measurement position only a small stripe of the collector can be measured. An optimized target-based setup uses a multicamera system to characterize an entire parabolic trough module in an indoor production environment [12].

In order to overcome the difficulties of target-based deflectometry for parabolic troughs in outdoor conditions, a special type of deflectometric measurement was developed using the reflection of the absorber tube as pattern. This approach, called TARMES (trough absorber reflection measurement system), is based on the distant observer approach [17] and was developed to achieve high accuracy and high spatial resolution slope deviation maps for PTCs in curvature direction from a set of photos [18]. A high-resolution digital camera is required for the image acquisition. The position of the camera is calculated using a laser distance meter to measure the distance between camera and the parabola vertex and inclinometers to obtain the collector elevation. Normal vectors of the mirror surface are derived from the spatial coordinates of the absorber tube edges, the position of the absorber tube edge reflection on the mirror surface, and the nodal point of the camera (see Fig. 5.4). An example for the application of TARMES is presented in Section 5.2.3.2.

Another method for PTCs called TOP involves overlaying theoretical images of the absorber tube in the mirrors onto surveyed photographic pictures [19]. The video scanning Hartmann optical test system (VSHOT) is a laser ray-trace system to characterize both point- and line-focusing concentrators [20]; see Section 3.4.2.

VSHOT, TOP, and TARMES have in common that the collectors are easily measured when facing to the horizon, while the measurement of a collector module in its prevailing operating position (around the zenith) involves considerably increased effort like articulated man lifts. Airborne image acquisition (see Chapter 7) overcomes this limitation and enables a higher degree of automation and considerable larger measurement volumes [21].

5.2.2.2 Methods to measure concentrator shape

3D coordinates of reflective concentrator surfaces, concentrator support structure, receivers, assembling devices (jigs), and other components can be assessed by various means. The challenge is to achieve accuracy in the submillimeter range for objects with an extent of more than 10 m.

Close range photogrammetry (PG) [22] has been applied to various CSP collectors; see Shortis and Johnston [23], Pottler et al. [24], and Fernández-Reche and Valenzuela [25]. This technique derives the coordinates of points of interest, highlighted by special markers from a set of images taken from different positions. The application of these markers to large collector surfaces or structures means a high preparation effort. The basic characteristics of this methodology have been explained in Section 3.4.2. Photogrammetry can be applied to any collector orientation and with sufficient spatial resolution to detect characteristic shape deviations of mirrors and other points of interest like absorber tubes or axes of rotation. It is especially suitable for deformation analyses of prototypes and in cases where no reflective mirror surfaces are mounted. As the spatial resolution depends on the density of markers, it generally delivers a lower resolution than deflectometric approaches. Postprocessing of the raw images is a two-stage process. The first step is the calculation of properly scaled but arbitrarily orientated 3D coordinates from raw images using commercial image processing and bundle adjustment software [22]. The following task of calculating CSP-specific deviations and performance issues in general is carried out by custom software. The calculation of shape, slope, and absorber tube deviations consists in comparing measured and design coordinates. The application of PG for the qualification of heliostats is described in more detail in Section 5.2.3.1.

Recent developments in commercial laser scanners/radars allow their use to measure 3D coordinates of larger surfaces as parabolic trough modules or heliostats. More details can be found in Section 3.4.2. Specular reflecting surfaces need to be coated with an opaque paint before performing the measurement. Ulmer et al. [26] used a laser radar to validate a deflectometric measurement on an entire parabolic trough module.

Coordinate metrology using a coordinate measuring machine (CMM) has been applied to single mirror panels as validation against other approaches [27].

5.2.2.3 Methods to measure tracking deviation

A first impression of PTC tracking deviation can be obtained from any possible asymmetry visible in qualitative flux images in the focal region (see Fig. 5.3 right). Another option for PTC is to use the eccentricity between the shadow of the absorber and the parabola vertex [28]. This signal is exploited in photovoltaic (PV) cell-based sun sensors as a common method for closed loop tracking control in commercial plants (see Fig. 5.5). Two PV cells are aligned symmetrically on both sides of the parabola vertex, and the shadow of the absorber tube produces an electric current depending on the shadowed area of the PV cell. Tracking deviations are also caused by misalignment between the collector modules. In the construction phase of solar fields, the measurement of module alignment [3] offers great potential for performance enhancement. Relevant angle offsets between modules can be quantified with inclinometers if a reference axis is defined. In the absence of such a reference axis, alignment can be checked via the height level of the outer mirrors or with a water hose level.

For point-focusing systems, tracking deviations is best assessed by flux density measurements (see also Section 5.2.3.1).

5.2.2.4 Assessment of mechanical properties and deformations under operational loads

Effects of operational loads on the geometry may be assessed directly by geometric measures like photogrammetry (PG). As alternative, mechanical properties can be measured and effects of operational loads on the geometry can be calculated by structural analyses. In the design phase, finite element method (FEM) simulations to deduce the deformations from CAD models and assumptions on operational loads is highly recommended.

Relevant mechanical properties for PTCs are the lateral stiffness of the receiver supports, and torsional stiffness of collector structure position. To measure these quantities, the structure is carefully exposed to increasing force and/or torque far below the expected operational loads. Resulting deformations and torsion can be measured with high precision inclinometers and linear/angle measures.

5.2.2.5 Raytracing

In order to derive the intercept factor from the information gained by the described geometrical measurements, raytracing (RT) analyses have to be performed. In addition to the (measured) concentrator-receiver geometry, RT requires the assumption of a certain sunshape, and information about structural elements which shade incoming and block reflected radiation. For annual yield analysis with RT, deformation data must be considered for different elevation and incidence angles.

The statistical approach presented in Bendt et al. [1] consists of folding the sunshape with additional statistical concentrator errors, thus generating an “effective” source with a wider angle distribution compared to the initial sunshape. Together with an angular acceptance function, the intercept values for a given set of assumptions (e.g., collector design, incidence angle, and geometric accuracy) can be derived. This method works very well when the concentrator shape deviations can be described with a Gaussian distribution. However, as soon as deviations are distributed in a systematic way (asymmetric distribution with a mean value≠0), the resulting intercept factor is not correct.

Applications of this statistical approach to PTC fields are presented among others in Lüpfert et al. [2] and Pottler et al. [3]. A brief explanation on advantages and drawbacks can be found in Zhu and Lewandowski [29]. The statistical representation of optical concentrator errors by Gaussian distribution may be a valid approach for large samples like complete solar fields, but in general, systematic errors are very likely to happen, e.g., due to mass production of solar field components. NREL [29,30] has developed an analytical approach called FirstOPTIC, which preserves the spatial information of shape and absorber tube deviations while employing a probability approximation from Ref. [1] for the sunshape, reflector specularity, and tracking accuracy.

Numerical RT is nowadays the state-of-the-art approach to assess the optical performance of CSP systems. Computational resources allow the tracing of simulated light rays in large and complex optical systems, delivering accurate predictions of the flux density distribution in the receiver region. Ho [31] and Bode and Gauché [32] present a review on available RT software tools and their abilities. Most raytracing tools have an interface to define the concentrator by shape and absorber position deviations derived from either of the methods presented before. Analog to the statistical approach, concentrators may also be defined as ideal geometry adding statistically distributed geometrical errors. If the real mirror shape, tracking angle, and receiver position are known, RT may predict the optical performance correctly even for cases where systematic errors are present. Another advantage of raytracing is that so-called blocking and shading elements can be defined in order to determine their influence on the optical performance. Especially for non-zero incidence angles, parts of the concentrator structure interfering with incident and/or already reflected radiation, like bellow protections and receivers supports, have an influence on the intercept factor which can be significantly larger than the geometrical imperfections.

5.2.2.6 Flux density measurement

Raytracing delivers radiation maps and statistical values, which can be validated with measurements of the measured flux density in the focal plane. Flux density measurements are a straightforward and practicable approach for a point-focusing system (solar tower), but need considerable effort for a line-focusing system [33]. In most cases, indirect measurement of the flux is carried out with a digital camera and a Lambertian target (see Fig. 5.3 right for a PTC). Luminosity of the spot on the target can be directly translated to radiative flux in kW/m² when the system is carefully calibrated, which involves proper spatial transformation of image locations to receiver/target coordinates and radiometric calibration.

In the case of solar towers, the flux distribution integrated over the aperture is the solar power reflected by the entire heliostat field and entering the receiver. The result of such measurements can be used as acceptance criterion for heliostat fields at specific testing conditions (see Section 5.2.3.1).

5.2.3.2 Parabolic trough collectors

In order to compare several independent methods, selected PTCs at the PSA have been characterized with deflectometry, PG, and other methods.

Object preparation and image acquisition for the PG is similar to the photogrammetric measurement of a heliostat. It can be useful to apply the PG directly to the structure with dismantled mirrors, in order to distinguish effects arising from the metal structure from effects created by the mirrors.

For the measurement of the absorber position, either PG or manual measurement of the glass envelope tube can be applied, followed by estimation of the eccentricity of the steel absorber tube relative to the glass envelope tube. For this purpose, digital images are semiautomatically evaluated. These images are taken along the optical axis and perpendicular to the optical axis and vertex. The edges of the steel and glass tube are calculated from a cross section perpendicular to the focal line (see Fig. 5.15).

The TARMES approach (deflectometry) described in Section 5.2.2.1 is often applied to parabolic troughs to derive slope deviation data in high spatial resolution. The absorber tube position is an important input for the TARMES measurement and for a subsequent raytracing analysis. Typical mirror slope deviations obtained with the TARMES method and an intercept map are presented in Fig. 5.16.

5.3.1.1 Illustration of input energy variations by season and time of day

Consider a parabolic trough system oriented on a north–south direction, tracking east to west. The useful or effective energy input to a trough collector is the vector of the solar direct radiation perpendicular to the aperture, which is termed ANI and calculated from:

$ANI=DNI*\mathit{cos}\left(\theta \right)$

where ANI is the aperture normal insolation (W/m²), DNI is the direct normal insolation (W/m²), and θ is the incidence angle of the direct normal radiation relative to the collector aperture plane.

The incident angle depends on the location of the plant (longitude and latitude), time of the day and the year, and orientation of the collector. Several algorithms are available to calculate the sun position and angles (see, e.g., Blanc and Wald [45] and Reda and Andreas [46]).

Consequently, the characteristics of the ANI are dependent on the weather, location, day of year, time of day, and collector orientation. Fig. 5.17 shows the variation of ANI at a good solar site in a western United States location from the period 08:00 to 16:00 for Jun., Mar., and Dec. A similar plot for DNI would show a much stronger variation. It can be seen that the ANI is quite steady during midday in both Jun. and Mar., and quite high in Dec.

A more important characteristic is the magnitude of the change in ANI during a test period. Fig. 5.18 shows the percentage changes in ANI over the previous 30 minutes during the daytime hours for these 3 months. It can be seen clearly that the variations between 09:30 and 15:30 in Jun. and Mar. are reasonably small. Such results could be used to identify suitable test periods with small variations in ANI.

5.3.1.2 Impact of variations on solar thermal and electricity test results

Variations in the solar resource affect both the solar system thermal tests and the power block electrical output tests. For the former, the measured output quantity is the energy input to the HTF. The HTF then flows to the steam generator trains to produce input steam for the turbine. Variations in the input steam conditions, particularly flow rate, will have a significant influence on the steady-state condition of the electricity output. Therefore, standard test procedures for conventional power plants cannot be applied and new procedures have to be defined that take into account the variable nature of thermal energy and electricity supply.

However, there are two important exceptions to this “rule”: plants with thermal storage and plants with auxiliary fossil firing. Consider the most straightforward thermal storage configuration, namely a tower system with a molten salt flowing through the receiver and then into a large hot molten salt storage tank at high temperature. Hot molten salt is drawn from the hot storage tank and passes through the steam generator complex to generate turbine inlet steam. In this case, the variations in weather are fully compensated by the buffering of the hot storage tank, and the steam production is isolated from the solar resource variations. The case of a plant with auxiliary fossil-fired boiler are even simpler in this aspect, for when these boilers are used—for example, during a week of inclement weather—the electrical output is steady due to a steady fossil fuel input.

5.3.4 Description of tests

5.3.4.1 Short-term steady-state tests

The purpose of this test is to measure the thermal or electrical power output and/or efficiency of the solar system over a short period at steady-state conditions. Since perfect steady-state conditions are difficult to achieve even for a short period, acceptable stability criteria have to be defined for the following variables:

• minimum level and maximum allowable variation of DNI;

• maximum allowable variation of mass flow;

• maximum allowable wind speed;

• level and variation of inlet and outlet temperature and, if appropriate, pressure;

• reflector cleanliness at or higher than a specified minimum level; and

• availability of collectors/heliostats at a minimum level.

Uncertainty analyses on the test results (discussed later) will provide quantitative insights on the allowable variations in the individual measurement components. Prior to starting the tests, the plant must be operated for a sufficient period of time (stabilization period) under steady-state conditions as defined above (e.g., 0.5–1 hours). If measurements have to be interrupted because stability criteria were violated, the tests have to be repeated, including the stabilization period. Typically, test duration ranges from 15 minutes up to 2 hours or more.

Four types of tests should be considered for the short-term test needs:

1. Design/full power capacity test: During this test the thermal power output of the solar field and/or the electrical output of the total plant will be measured. The objective of this test is to verify that the plant and all related components and subsystems, e.g., HTF pumps and collector field or power block, are able to deliver the full design power output. Minimum solar conditions may be defined for this test.

2. Power test at available conditions: While the full power test requires high radiation conditions in order to achieve full output, this test can be performed at lower solar radiation, expanding the range of possible test days. During this test the thermal and/electrical power is measured. This test does not have to be conducted at full capacity; in fact, one objective of this test is to verify the power output at less-than-design capacity conditions. The measured output is to be compared with an expected output calculated using the solar system performance model.

3. Outlet temperature test: The objective of this test is to demonstrate that the solar system is able to achieve its design temperature at the outlet. For the solar power tower plant the solar system definition includes the receiver. Thermal power tests and outlet temperature tests should be performed at the same time.

4. Efficiency test: For the efficiency tests, the same test results as for the power tests can be used, but the solar conversion efficiency at the available test conditions shall be calculated from the measured result. Corresponding equations are provided in Sections 5.3.5.2 and 5.3.6.2. The measured value can then be compared with the expected value calculated using the performance model.

5.3.4.2 Multiday production test

The objective of the multiday production test is to measure and evaluate the solar system total thermal energy production and/or the plant total electricity production over a longer period. Thus, different operating conditions are evaluated such as start-up and shutdown, part load operation, and transient conditions, which all contribute to the overall energy production at different performance levels. Both clear sky and partly cloudy conditions are acceptable. The result of the tests is to be expressed as a direct measurement of total energy production during the test period. The test period should be long enough to contain sufficient representative days. A period of at least 10 days of continuous operation is recommended, though longer periods are preferable, e.g., 30 days. The measured results are to be compared to the same values as calculated by a performance model for the same period.

The primary objective of this test is to run the test continuously without any interruption for the complete period. However, if a relatively short period for the duration was selected, e.g., 10 days, minimum requirements for the solar radiation need to be defined. Days that do not meet this requirement have to be excluded and the test has to be extended until the required number of days with acceptable conditions has been accumulated. For some commercial plants, debt providers have required one full year of successful operation at design conditions.

5.3.5 Solar system performance

5.3.5.1 Description of boundaries of the solar system

As with any performance test, test boundaries need to be defined. The solar system is comprised of the solar reflectors, solar receiver, and heat exchangers needed to transfer the collected solar energy to the working fluid of the power block. An example of the solar system boundary is shown in Fig. 5.19 wherein the inputs are the solar resource, inlet working fluid, internal solar system parasitic power,¹ and, if required, auxiliary heating. The sole output is the increase in enthalpy of the working fluid.

In the case of a parabolic trough system, the solar system includes the PTCs, all piping that connects the individual collectors with each other, and the piping required to transport the working fluid from the system boundaries to the collectors and back to the system boundaries. Usually it also includes the pumping station for the working fluid and other equipment required to handle the working fluid, like the expansion vessel and fluid conditioning system.

In the case of tower systems, the solar systems includes the heliostat field, the power tower receiver and all piping and equipment that is required to transport the fluid from the system boundary to the receiver and back to the system boundary. Configurations with water/steam and air working fluids are direct systems in which heat is added directly in the receiver to the working fluid of the power cycle. In an indirect system such as a molten salt configuration, heat is added to a molten-salt HTF in the receiver and subsequently transferred to the power-cycle working fluid in a steam generator train. The steam generation train may be inside or outside the test boundaries.

The appropriate test boundary locations for the measurement of thermal performance should be chosen by the test parties based on the equipment configurations and test requirements. The scope of this chapter does not include a thermal energy storage system, nor co-firing with a natural gas boiler within the test boundary but focuses on the solar boiler only.

To be clear, the term solar field is used in the case of parabolic trough technology for the array of PTCs connected to each other. In the case of a tower plant, the solar field is comprised of an array of heliostats.

5.3.5.2 Solar system thermal calculations

The primary objectives of the thermal assessment are to determine the following thermal performance values:

• thermal power output;

• outlet temperature;

• thermal efficiency; and

• thermal energy output within a specific period.

5.3.5.3 Instruments

The primary measurements needed to calculate the performance values of Section 5.3.5.2 are the:

• temperature at the inlet and the outlet of the solar system as defined by the system boundary;

• mass flow at inlet and outlet of the solar system;

• pressure at inlet and outlet of the solar system; and

• direct normal irradiance (DNI).

In addition, the following measurements should be performed since they are used as an input to the performance model:

• wind speed and direction; and

• mirror reflectance.

It is not the purpose of this book to provide a complete treatise on standard measurement techniques. Those can be found in relevant codes and standards like the ASME performance test codes (PTCs). Rather, a comprehensive overview of applicable instruments will be given and some specific information on solar-unique measurements will be provided.

Accuracy and measurement uncertainty always play an important role in performance measurements. Due to the nature of variability of the solar recourse and the imperfections of the control system, this is of particular importance for solar thermal power plants. Therefore, Section 5.3.5.4 will specifically discuss measurement uncertainties.

5.3.5.4 Measurement uncertainty in solar system testing

Due to various influences, any test result will have an associated uncertainty. The uncertainty interval around a measured result describes our lack of knowledge about the true value of a measured quantity. Uncertainty can be reduced by many repeated measurements and in some instances can be further reduced through the use of redundant instruments.³ The uncertainty of an interval about a measured value is usually expressed with a probability or level of confidence. Uncertainties arise from possibilities for error arising from or classified as: systematic errors, random errors, and human errors. It is very important in the measurements discussed here to quantify the uncertainty intervals to make judgments on the validity of the test results.

Test uncertainty is a very important element within Performance Test Codes, such as those published by the ASME in the United States, which has placed critical importance on test uncertainty analyses of all measurements and calculations associated with PTCs. ASME PTC 19.1—Test Uncertainty is devoted to this topic, and plays an important role in all ASME PTC codes. The text in this section is extracted for illustration from a recent NREL report [48] on solar system acceptance test guidelines.

Because of the resource variability and imperfections in control systems, variation in all of the measured parameters is inevitable. The frequency and period of data collection directly impacts the test uncertainty, and it is highly recommended that a pretest uncertainty analysis be carried out prior to selection and subsequent installation of any instrumentation. A solar power test will typically consist of more than one test run (data collected during a period of time in which the measured parameters are relatively steady). Conducting more than one test in addition to a post-test uncertainty analysis is recommended to verify the repeatability of the test results and the validity of the pre-test uncertainty analysis.

The systematic error associated with a measurement of a single parameter can come from many sources, including the calibration process, instrument systematic errors, transducer errors, and fixed errors of method. The test engineer should be diligent in identifying all of these sources of error, although it is often the case that one or several will dominate within a particular measurement parameter.

Random errors can similarly be based on a manufacturer's specification. However, the random uncertainty for a given measurement can be reduced based on repeated measurements over the interval in which the system is considered to be in thermal equilibrium (defined by a minimal change in the DNI plus working fluid flow control over the test period, such that the effects of thermal exchange between the working fluid and the receiver hardware are negligible). For repeated measurements, the random standard uncertainty can be defined by:

$s_{\overline{X}}=S_X/\sqrt{N}$

where S_X is the standard deviation of a series of sampled data and N is the number of data points collected over the test interval (e.g., 180 data points for a 30-minute test with data collected at 10-second intervals).

The essence of an example calculation is described below using the ASME PTC 19.1 principles and notation (compare also Ref. [49]). The purpose is to describe how the uncertainties in each of the measured variables X associated with an acceptance test propagate into the value of a calculated resulting quantity R.

Calculated results, such as the delivered power and the solar thermal efficiency, are not typically measured directly, but rather are based on parameters measured during the course of one or multiple tests. For this case, the result, R, is a function of individual or average values of these independent parameters as described by

$R=f\left(\overline{X_1},\overline{X_2},\dots ,\overline{X_i}\right)$

where the subscript i describes the number of parameters used in the calculation of the result and $\overline{X}$ is either the value of a single measurement of the parameter or the average value of the parameter based on a number of N repeated measurements.

The expression for the combined standard measurement uncertainty of a calculated result based on multiple error sources can in many cases be calculated from the root sum square of the total uncertainty of the individual systematic and random error sources⁴:

$u_R={\left[{\left(b_R\right)}^2+{\left(S_R\right)}^2\right]}^{1/2}$

where b_R is the systematic standard uncertainty of a result and S_R is the standard random uncertainty of a result as calculated by

$b_R={\left[\sum _{i=1}^I{\left(\frac{\partial R}{\partial {\overline{X}}_i}{b}_{\overline{X_i}}\right)}^2\right]}^{1/2}$

$S_R={\left[\sum _{i=1}^I{\left(\frac{\partial R}{\partial {\overline{X}}_i}{s}_{\overline{X_i}}\right)}^2\right]}^{1/2}$

For the equations above, $b_{\overline{X}}$ is defined as the systematic standard uncertainty⁵ of a component and $s_{\overline{X}}$ is the random standard uncertainty of the mean of N measurements. Definitions for the standard systematic and random uncertainty, as well as the methodology for calculating these values, are described in detail in PTC 19.1 and Dieck [49].

The total standard measured uncertainty of the calculated result given by Eq. (5.13) can be calculated using the methodology described above. It is important to note that this “standard” uncertainty implies that the calculated result will capture the true result within a 68% confidence level (one standard deviation). Typically, a confidence level of 95% (two standard deviations) is desired by the testing parties. For this case, the expanded uncertainty in the result is given by

$U_{R,95}=2u_R$

A detailed application of this method for both the solar system thermal power output and solar system efficiency can be found in the NREL work [48] for the cases of a parabolic trough system and a molten salt power tower system, respectively. Using reasonable values for the test parameters, representative uncertainties were calculated. Although any such analysis must be undertaken with the specific system and instrumentation in mind, the systematic and random uncertainties of measurement parameters calculated in the treatment represent what may occur in a typical field installation. Actual uncertainties depend on specific field instrumentation and calibration procedures associated with an acceptance test.

For a 100 MWe solar tower power plant with 6 hours of thermal storage, the results in that case were in the order of 3.2% and 4.1% for the solar field thermal power and thermal efficiency, respectively. For the trough plant, the results were 3.9% and 4.7%, respectively.

5.3.5.5 Example of thermal efficiency results

Fig. 5.20 gives an example of results for the measured thermal efficiency at a clear day for a state-of-the-art PTC (for details, see Ref. [51]) operated in a commercial power plant. The values are calculated based on Eq. (5.8) using measured values for temperature, flow, and DNI. Each dot represents a 10-min average value. The graph shows typical values for the efficiency and a characteristic shape of the curve over the day. The values are rising steep in the morning and falling steep in the evening due to shading effects of adjacent collectors. The typical dip during noon is caused by the incident angle effects. The error bars representing the mean uncertainty of the results are also shown.

Fig. 5.21 illustrates the effect of the incident angle on the collector efficiency based on measurements performed within different periods of the year. The abscissa shows the absolute values for the incident angle. It can be clearly seen that thermal collector efficiency is worse during winter than during summer, which is mainly caused by the higher incident angles in the winter period.

5.3.6 Total plant performance

5.3.6.1 Description of system boundaries

The assessment of the performance of a utility scale solar plant is tied to the warranty/guarantee agreements between the technology suppliers and financing parties. A typical scenario is that the plant is designed and constructed by an EPC company under contract to the owners. The EPC provides the guarantees for the performance based on the successful completion of specified performance tests. A crucial element of the EPC contract is the definition of the elements, requirements, and timing of those tests. A number of test definitions are possible, usually a function of the perceived risks in solar system performance.

Section 5.3.5 discusses acceptance tests of the solar system, which usually fulfill one element of the total EPC performance test requirements. In practice, the EPC contract with the owners is usually based on the electrical output of the total solar power plant, while testing of the solar system performance itself is carried out to satisfy contractual agreements between the EPC contractor and the solar system supplier.

The total plant includes the solar system, power block, and balance of plant, as well as ancillary systems such as thermal storage and fossil fuel co-firing. A schematic for a parabolic trough system including all these subsystems is shown in Fig. 5.22 (compare also Ref. [54]). Test boundaries for the total plant performance will include all these subsystems. The inputs into the system are:

• solar radiation;

• auxiliary fossil heat; and

• parasitic power for in-plant needs, that is, the difference between gross and net electrical output,

while the output is:

• gross electrical power at the generator; and

• net electrical power to the electrical grid.

5.3.6.2 Calculations

Following the test approach presented in Section 5.3.4, the following values will be either directly measured or calculated from the measurement result:

• gross and net electrical power output P_el (MW)

• solar-to-electric efficiency

• electric energy output within a specific period E_el (MWh).

5.3.6.3 Conduction and evaluation of tests

The concept of thermal equilibrium for short-term testing is of equal importance in the measurement of total plant electrical output and solar system thermal output. Therefore, the performance test procedures laid out in Section 5.3.4 (short-term steady-state tests and multiday production tests) are also applicable for electrical performance.

As no solar-oriented test codes are available for overall plant performance at present, more general test codes must help to fill that gap.

ASME PTC 46, described below, is one of several good choices to fill that need. The test results from the use of PTC 46 would be compared to a projection of electrical output based on the test conditions using a sophisticated total plant performance model as discussed in Section 5.3.3. Other than the stability of steam flow and conditions into the steam turbine, there are in general terms no solar-specific aspects to electrical output testing.

Typically, ASME performance test codes address tests of short duration (measured in minutes, not hours, days, or longer) with constant energy sources. For a solar power plant with a Rankine cycle steam power block, the EPC contract will typically include a requirement to demonstrate the warranted electrical power output in MWe. But, in contrast to a fossil-fired plant, there will also be a requirement for the solar power plant to show the actual energy generation over time in kW-hours compared to that projected by the plant performance model based on actual solar radiation conditions and contractual parameters for power block and solar system characteristics. The time period may be for several months or even up to 1 year.

For the assessment of the total plant performance, the performance model has to consider, in addition to the inputs discussed in Section 5.3.3, the input of fossil fuel during the test and the loading status of the thermal storage system at the beginning and the end of the test.

For the long duration energy generation test, there will be no requirement related to test uncertainty because test conditions will normally not be steady and equipment conditions can vary widely, e.g., the mirror reflectance over a large solar field.

5.3.6.4 Example plant performance data

Typical electrical performance data for large utility-scale solar plants are shown in Table 5.1. These values are representative for plants in high solar resource regions (e.g., southwest United States, South Africa, Morocco) and thermal storage in the 6–10 full-load hour range.

5.3.6.5 Electrical test considerations

The crucial solar-related impact on a standard test of the steam turbine power block is the stability of inlet steam to the turbine. Other than that, the performance testing of the power block is well covered in available test codes.

For the short-term test, the steam turbine cycle output can be measured in a thermally stable condition if the solar steam input is acceptably steady, as mentioned above.

ASME PTC 46 (see Section 5.3.7.3), as an example, also stipulates a significant number of other code requirements to ensure valid results, providing useful and extensive details. In concert with that code, ASME PTC 19.1 gives tight requirements on the uncertainty measurements.

For illustration, the following paragraphs show a few general comments on the requirements of PTC 46 for understanding the electrical side of solar plant performance tests. Note that all aspects of procedures, equipment, measurement, uncertainty, and other factors are covered in that code. For actual testing with valid results, the full stipulations of that code must be adhered to rigorously and in depth.

The primary objective is to obtain test results with a very low measurement uncertainty for specific test requirements defined to achieve that objective.

Examples of those requirements include:

• complete records of the testing, including equipment, configuration, valve line-up, instrumentation, data acquisition, test procedures, test logs, and test data;

• for all instruments: type, model, purpose, location, redundancy, and calibration records;

• detailed test plan and test procedures, including stabilization criteria, number and duration of test runs, requirements for initial equipment conditions, including solar field reflector reflectivity;

• performance model for comparison to measurements agreed upon by test parties;

• uncertainty analysis using ASME PTC 19.1, both pretest and for all test results; and

• criteria for acceptance/rejection of individual test results.

5.3.7.1 NREL performance acceptance guidelines

Lacking published international standards or codes for performance testing of utility scale solar thermal systems, NREL⁶ undertook in 2009 and 2013 the publication of “Guidelines” to provide direction for conducting performance acceptance testing for large parabolic trough [48] and power tower [48] solar systems that could yield results of a high level of accuracy consistent with good engineering knowledge and practice. The recommendations in those Guidelines were reviewed by stakeholders representing concerned organizations and interests throughout the CSP community. The Guidelines recommend certain methods, instrumentation, equipment operating requirements, and calculation methods. The primary objective of the Guidelines was that, when tests are run in accordance with the methodologies, the test results would yield a valid indication of the actual performance of the tested equipment. It is important to note that the Guidelines are only recommendations—to be considered carefully by the contractual parties involved in the acceptance testing—and it is likely that modifications would be required to fit the particular characteristics of a specific project.

Notably, the Guidelines do not specify means to compare test results to contractual guarantees. A key presumption is that the parties to a commercial test will address that issue, as well as agree on other important issues with regard to testing protocol and evaluation of test results. The approach taken in the Guidelines is that the measured test results will be compared to projections from a performance model based on the measured meteorological conditions and agreed-upon solar system characteristics.

The scope of the solar system discussed in the NREL Guidelines does not include a thermal energy storage (TES) system. However, even if the scope did include a TES system, the methods of testing the solar field/HTF system itself would be similar, if not identical, to the guidance given in the Guidelines. A separate set of tests and objectives would be required for TES system acceptance.

5.3.7.2 AENOR UNE 206010

In Apr. 2015, AENOR published the Spanish standard on tests for verifying the performance of solar thermal power plants with cylindrical PTC technology (Spanish original title: “Ensayos para la determinación de las prestaciones de las centrales termosolares con tecnología de captadores cilindroparabólicos”).⁷

The standard is applicable to solar thermal power plants of any size with parabolic-trough collectors, with the sun as main source of energy, optional non-solar energy sources, and an optional thermal storage system. It does not include any other CSP technology, such as solar towers.

The code is intended to be applicable to acceptance testing in such power plants, or other scenario in which their performance must be known and thus will help in identifying possible operational problems that are affecting output.

The evaluation of the test requires:

1. use of a plant simulation model that can generate reference values when the input conditions measured during testing are entered; and

2. a defined verification procedure, a way in which measurements and uncertainties are to be compared to the reference values from the plant simulation model.

Similar to the NREL performance acceptance guidelines, two types of tests are proposed, a short- and a long-duration test:

1. Short-duration testing (at least 24 hours): the purpose is to verify the performance of the power plant over a short period of time during clear days, including a full charge-discharge cycle of the thermal storage if available, recommended to be performed three times. DNI>700 W/m² must be available for at least 4 hours, recommended without auxiliary heat source.

2. Long-duration testing (at least 365 days): the purpose is to verify or validate annual plant production and auxiliary consumptions (electricity and non-solar fuel).

The tests combine measurement and comparison to a predefined simulation model for the following operation parameters:

• available solar radiation (DNI);

• plant electric consumptions;

• net electric generation (high voltage side of transformer);

• non-solar energy, backup fuel; and

• net plant efficiency.

The test definitions are based on practical experience with commissioning and acceptance testing of many of the Spanish parabolic trough plants that came online between 2009 and 2013.

The English version of the text is also the basis for the corresponding IEC TC117 (see Section 5.3.7.5) international working group from 2016 onwards.

5.3.7.3 ASME PTC 52

In 2009, the American Society of Mechanical Engineers (ASME) formed a committee to develop a Performance Test Code (PTC) for concentration solar power systems. The code will provide procedures for conducting tests and analyses to determine the performance of CSP plants such as in parabolic trough, continuous linear Fresnel reflector and power tower technologies. ASME PTC-52 is intended to cover the solar-to-thermal conversion system only and not the complete power plant. The solar-to-thermal conversion system, i.e., the solar system or solar field, is treated as one component as a part of the whole power plant similar to a conventional boiler or to a steam turbine generator that are covered under different ASME performance test codes. As such, ASME PTC-52 will cover some aspects discussed in this chapter.

In accordance with the current practice that is described in the NREL performance acceptance guidelines and that is also discussed here, PTC-52 proposes to conduct short-term tests as well as multi-day tests. The code also elaborates the use of a performance model to evaluate the tests and to establish the expected output against which the actual test output will be compared. Details on test procedures, instruments and methods for measurements, calculation and reporting of measurement results, as well as on uncertainty analysis, will be given in the code. It will additionally contain a short digression on thermal storage.

When complete, PTC 52 is expected to supplant need for the NREL guidelines. The issue of the first version of PTC-52 is planned for 2017.

5.3.7.4 ASME PTC 46

The ASME Performance Test Code 46 (PTC 46) entitled “Overall Plant Performance” is internationally respected and often used for overall plant performance testing. As stated in the introduction to the Code, the objective of PTC 46 is to provide uniform test methods and procedures for the determination of the thermal performance and electrical output of heat-cycle electric power plants at a steady-state condition over a few hours, and provides explicit procedures for the determination of the following performance results: corrected net power, corrected heat rate or efficiency, and corrected heat input. Test results provide a measure of the performance of a power plant or thermal island at a specified cycle configuration, operating disposition and/or fixed power level, and at a unique set of base reference conditions.

Due to the thermal stability requirement of PTC 46, only the short-term solar system testing discussed in earlier sections strictly fits into this dictate, though many of the specific code requirements are applicable to the longer multi-day tests as well, as can be seen in the following paragraphs.

Test results can then be used as defined by a contract for the basis of determination of fulfillment of contract guarantees. Test results can also be used for comparison to a design number, to trend performance changes over time, to help evaluate possible modifications or to validate them, or for any application in which the overall plant performance is needed.

5.3.7.5 IEC TC 117

The International Electrotechnical Commission (IEC) is currently developing a standard for CSP plants (IEC TC 117)⁸ that is not restricted to the solar-to-heat conversion system like the ASME PTC 52, but rather intends to cover the complete solar power plant for the conversion of solar radiation into electrical energy. When completed, this standard would not only cover the plant in total but will discuss all elements (including all subsystems and components) of the entire power plant. It will include all of the current different types of systems, as parabolic trough, solar tower, linear Fresnel, dish, and thermal storage. The standard would not only address performance measurement and test methods, but also design, installation and safety requirements, as well as power quality issues. The section of the code that covers total plant performance testing will be based on the Spanish standard UNE 206010 (see Section 5.3.7.2) and will probably not be published before 2018.

5.3.7.6 Other codes

Several other codes are available that cover performance aspects of solar thermal collectors as a component, such as the International Standard ISO 9806, the European Standard UNE-EN 12975-2 and the American Standard ASHRAE 93 and ASTM E 905-87. These standards cover explicitly or can be transferred to concentrating and tracking solar thermal collectors and, therefore, can be applied in principle to parabolic trough and linear Fresnel collectors.

However, these standards do not cover complete large-scale systems and cannot be applied to the systems described in this chapter. Some of the tests and measurements described in the standards mentioned above can be used to derive collector parameters that can be used as an input into the performance model.