A new version (1.2) of PV_LIB Toolbox for Matlab has been released. It includes a number of new functions and important bug fixes. It is recommended that all current users download this new version.
New Functions Include:
- pvl_louche – Estimate Direct Normal Irradiance from Global Horizontal Irradiance using the LOUCHE model
- pvl_erbs – Estimate Direct Normal Irradiance from Global Horizontal Irradiance using the Erbs model
- pvl_orgill_hollands – Estimate Direct Normal Irradiance from Global Horizontal Irradiance using the Orgill and Hollands model
- pvl_reindl_1 – Estimate Direct Normal Irradiance from Global Horizontal Irradiance using the Reindl 1 model
- pvl_reindl_2 – Estimate Direct Normal Irradiance from Global Horizontal Irradiance using the Reindl 2 model
- pvl_martinruiziam – Determine the incidence angle modifier using the Martin and Ruiz incident angle model
- pvl_adrinverter – Converts DC power and voltage to AC power using Anton Driesse’s Grid-Connected PV Inverter efficiency model
Bug Fixes Include:
- Corrected pvl_klucher1979. Statement GHI(GHI<DHI) = DHI did not work for vectors, changed to GHI(GHI<DHI) = DHI(GHI<DHI)
- Corrected pvl_calcparams_desoto.m as fillows: a. Removed line M = max(M, 0); which set values of M which were less than 0 to 0 b. Removed line S(S==0) = 1E-10; c. Added line IL(isnan(M) | M<0 | S <=0) = 0; This sets IL to 0 when i. Irradiance is <0 or ii. Airmass modifier is <0 (most likely due to evaluating the polynomial at very high airmass or iii. Airmass modifier is NaN, which is returned by pvl_relativeairmass for sun zenith angles > 90 degrees d. Added line I0(IL==0) = 0; According the circuit diagram, IL is the only source, and therefore, there can be no reverse saturation current (I0) when IL is 0 e. Added line Rsh(S <= 0) = inf; This is due to the fact that Rsh is determined by dividing by S, negative values would give a negative Rsh, and at S=0 Rsh is undefined
- 2. Modified pvl_singlediode.m a. Pre-allocate memory to Imax, negPmp, Vmax, Ix, Ixx, Voc, and Isc. This is necessary in order to use the filter (see below) b. Added a filter u = IL > 0; Thus we only compute IV points and IV curves when there is a photocurrent (IL) c. Changed the Isc and Voc generation lines to only generate Isc and Voc under conditions which satisfy filter u. d. Changed Imax and negPmp finding line to only find Imax and negPmp when conditions satisfy filter u e. Changed Vmax, Ixx, and Ix generation lines to only generate the values under conditions which satisfy filter u f. Pre-allocate memory for Result.I and Result.V (necessary to implement filter u). g. Changed Result.I finding line in order to only find I-V curve currents under conditions which satisfy filter u h. Added line Result.I(:,end) = 0; in order to ensure that the current at Voc is exactly 0 (prevents numerical error which may result in a negative current at Voc).
- Corrected errors in wapr_vec.m. From line 134 through line 154 (inclusive), the Lambert W function is evaluated piecewise over the domain (-exp(-1), inf) using four approximations. The domain is partitioned into four disjoint sets and each approximation is applied to one (and only one) set. Filters were implemented incorrectly which allowed later approximations to overwrite values obtained from early approximations.
- Corrected pvl_calcparams_desoto.m documentation. The documentation used to specify that alpha_isc should be in units of 1/C, it now specifies that alpha_isc should be in units of A/C or A/K. Documentation was changed rather than code due to the fact that the SAM CEC module database lists alpha_isc in units of A/C. a. Thanks to Martin Herrerias Azcue for the bug find.
- Corrected pvl_calcparams_desoto.m. Prior input checking for EgRef was p.addRequired(‘EgRef’, @(x) (isnumeric(x) & isvector(x) & x>0)); this has been changed to p.addRequired(‘EgRef’, @(x) (isnumeric(x) & isvector(x) & all(x>0))); in order to ensure that all EgRef values are positive. a. Thanks to John McKeen at DOW Solar Solutions for the bug find
- Corrected pvl_clearsky_ineichen to use system-dependent file separators when using the default Linke Turbidity index. Prior code was: load(‘Required Data\LinkeTurbidities.mat’); this has been changed to load([‘Required Data’ filesep ‘LinkeTurbidities.mat’]); in order to allow for system-specific file separators. a. Thanks to Mark Campanelli at NREL for finding this bug
- Modified pvl_singlediode to not require pvl_fminbnd in order to find the maximum power point. Prior to this change, maximum power point was found by finding the maximum value of the power-current curve using a modified version of MATLAB’s fminbnd function. pvl_singlediode now finds where dP/dV as a function of current is equal to 0. Uses bisection techniques. a. New subfunctions “calc_phi_exact”, “calc_Imp_bisect”, “g”, and “calc_Pmp_bisect” were added. b. Prior code was left in the function and commented out to allow for reversion later if necessary.
- 1. Modified pvl_getaoi to avoid complex values. Changed the line: AOI = acosd(cosd(SunZen).*cosd(SurfTilt)+sind(SurfTilt).*sind(SunZen).*cosd(SunAz-SurfAz)); to the following: AOI = acosd(max(min(cosd(SunZen).*cosd(SurfTilt)+sind(SurfTilt).*sind(SunZen).*cosd(SunAz-SurfAz), 1),-1)); a. Roundoff errors previously could cause the argument of the arcos function to be greater than 1 or less than -1 (resulting in a complex output). The min and max functions prevent such an occurrence.
- Modified pvl_getaoi to avoid complex values. Changed the line: AOI = acosd(max(min(cosd(SunZen).*cosd(SurfTilt)+sind(SurfTilt).*sind(SunZen).*cosd(SunAz-SurfAz), 1),-1)); to the following: temp = cosd(SunZen).*cosd(SurfTilt)+sind(SurfTilt).*sind(SunZen).*cosd(SunAz-SurfAz); temp(temp>1) = 1; temp(temp<-1) = -1; AOI = acosd(temp); a. Roundoff errors previously could cause the argument of the arcos function to be greater than 1 or less than -1 (resulting in a complex output). The min and max functions did not prevent all occurrences.
About the author
Joshua S. Stein is a Senior Scientist at Sandia National Laboratories.
