Full Equations Utilities (FEQUTL) Model for the Approximation of Hydraulic Characteristics of Open Channels and Control Structures During Unsteady Flow

U.S. GEOLOGICAL SURVEY WATER-RESOURCES INVESTIGATIONS REPORT 97-4037

*WRN:nn * MESSAGES

*WRN:501 * SUBSECTION VALUE FOR FIRST POINT IS MISSING. SUBSECTION = 1 ASSUMED.

The subsection for the first point on the cross-section boundary is missing. It is assumed that subsection 1 is meant and the computations continue.

*WRN:502 * UNEXPECTED SLOPE AT LEFT END. SLOPE EXPECTED TO BE < 0 AT LEFT BOUNDARY

The left end of the cross section would normally have a downward slope as the channel is entered with the leftmost point higher than any nearby point. A cross section has been detected for which this is not true. If the cross section is a natural or constructed channel but not a closed conduit, the cross section should be checked to make sure that it is meaningful that the leftmost point of the cross-section boundary is lower than the second from the leftmost point. In most cases, such a condition is not correct.

*WRN:503 * UNEXPECTED SLOPE AT RIGHT END. SLOPE EXPECTED TO BE > 0 AT RIGHT BOUNDARY

The right end of the cross section would normally have an upward slope as the channel is exited with the rightmost point being higher than any nearby point. A cross section has been detected for which this is not true. If the cross section is a natural or constructed channel but not a closed conduit, the cross section should be checked to make sure that it is meaningful that the rightmost point of the cross-section boundary is lower than the second from the rightmost point. In most cases, such a condition is not correct.

*WRN:504 * LINE SEGMENT ENDING AT (ff, ff) IS HORIZONTAL AND NOT AT MINIMUM OR MAXIMUM ELEVATION. RIGHT HAND END INCREMENTED BY ff

Every function given in a function table must be strictly continuous. Therefore, it is not possible to have a horizontal line at any other than the minimum or maximum (the case for some closed conduits) elevation in the cross section. Every invalid horizontal line segment is incremented at its right-hand end by the given amount. If this action is not acceptable, the user must increment one end point so that the line is not horizontal.

*WRN:505 * DECREASE IN CONVEYANCE IN SUBSECTION nn AT ELEVATION = ff

The conveyance as computed in the given subsection at the specified elevation has been calculated to be smaller than the conveyance at the previous elevation in this subsection. This is not hydraulically correct unless the cross section has converging walls. Therefore, in a natural channel, the conveyance should always increase with an increase in the water-surface elevation. Additional subsections may need to be added to the cross-section specification.

*WRN:506 * CONVEYANCE NON-INCREASING AT DEPTH= ff

The conveyance for the entire cross section is nonincreasing at the given depth (water-surface height). This message may not be issued even though WRN:505 is issued because the increase of the conveyance in one subsection is larger than the corresponding decrease in conveyance of some other subsection. A decrease in conveyance with an increase of water-surface height is only possible if the walls of the section are converging. Additional subsections may be needed to properly represent the conveyance for the cross section.

*WRN:507 * BRIDGE SKEW < 0 OR > 45. RESET TO: ff DEGREES.

The bridge skew for the Bureau of Public Roads (Bradley, 1970) bridge method is outside the valid range. The value is reset in FEQUTL computations to be within a valid range. This bridge-loss method is no longer supported in FEQUTL, and it remains in the code for the benefit of users with stream-system models developed for versions of the FEQUTL code prior to version 4.0. These users are encouraged to revise the models to apply the WSPRO (Shearman, 1990) bridge routines (section 4.1).

*WRN:508 * PIER SKEW < 0. RESET TO: ff

The pier skew in the Bureau of Public Roads (Bradley, 1970) bridge method is invalid. It is set to a valid value in FEQUTL computations. This bridge-loss method is no longer supported in FEQUTL, and it remains in the code for the benefit of users with stream-system models developed for versions of the FEQUTL code prior to version 4.0. These users are encouraged to revise the models to apply the WSPRO (Shearman, 1990) bridge routines (section 4.1).

*WRN:509 * M > 1 IN MCOMP. K = ff KSUB= ff

The contraction ratio, M, has been detected to be larger than 1.0, which is not possible. This may result from computing the conveyance for a subset of a cross section. The value will not, in general, differ much from 1.0, and the value is reset to 1.0 in FEQUTL computations.

*WRN:510 * FROUDE NUMBER=ff > 1.0 AT XSID=aa

A Froude number at a point in a profile computed in WSPRO that exceeds 1.0 has been detected in WSPROT14 (section 5.22) computations.

*WRN:511 * NUMBER OF SIDES= nn TOO SMALL. NSIDES RESET TO 10

The number of sides requested for the polygon approximation to a pipe is too small in the MULPIPES (section 5.17) or SEWER (section 5.14) commands. The value is reset to 10 in FEQUTL computations.

*WRN:512 * NUMBER OF SIDES= nn TOO LARGE. NSIDES RESET TO nn

The number of sides requested for the polygon approximation to a pipe is too large in the MULPIPES (section 5.17) or SEWER (section 5.19) commands. The number of sides is reset to the given value. The number of sides is set to the maximum number of points in a cross section less four. This number is more than adequate to approximate a pipe. Application of 20 to 30 sides yields an approximation that is within a fraction of a percent of the true values computed from a perfect circle. Most commercial pipes are not manufactured to this accuracy.

*WRN:513 * NUMBER OF SIDES= nn TOO SMALL. NSIDES RESET TO 10

See WRN:511.

See WRN:512.

No file name has been given for the output of the CHANNEL command (section 5.2). The name CHANNEL is used and processing continues.

*WRN:516 * MORE THAN ONE LEFT BOUNDARY. ONLY FIRST ONE RETAINED.

In the elevation-defined floodway (section 5.12), more than one left boundary has been found at the given elevation. Only the leftmost boundary has been retained, and the others are taken to be the boundaries of islands.

In the elevation-defined floodway, more than one right boundary has been found at the given elevation. Only the rightmost boundary has been retained, and the others are taken to be the boundaries of islands.

If an error is detected in processing the current command, the computations in FEQUTL skip to the next command. When this results, the table computed in the current command was not computed. The table file specified will contain all the tables from the successfully completed commands, and no indication that the table file is incomplete will be given. The output must be searched for the string *ERR: to find all reported errors.

Free flow normally should not decrease with an increase in downstream head. This probably results from a
decreasing value of critical flow in one or both cross sections of the transition. Use of the NEWBETAE option can sometimes correct the problem.

In compact cross sections (cross sections in which the hydraulic radius and the hydraulic depth increase with the water-surface height), the celerity will never decrease as the water-surface height increases. The meaning of celerity and the related concepts of critical depth and critical flow in noncompact cross sections is unclear. A decrease in celerity with an increase in water-surface height may or may not result in subsequent computational problems. Therefore, it is important in the computation of critical-flow tables (section 5.4) and in the computation of the generalized-Ritter dam-break flood peak (section 5.14) to be aware of cross sections in which the celerity decreases with increasing water-surface height.

In compact cross sections (cross sections in which the hydraulic radius and the hydraulic depth increase with the water-surface height), the flow rate at critical flow will never decrease as the water-surface height increases. In noncompact cross sections, the critical flow as computed can decrease. It is not yet clear what this decrease means. Therefore, in the computation of a critical-flow table it is unwise to use a cross section for which the critical flow decreases with increasing water-surface height.

The depth (water-surface height) is less than the minimum depth available in the Escoffier table. This message should not appear and may indicate a bug in the GRITTER command (section 5.14).

The depth (water-surface height) is greater then the maximum depth in the Escoffier table. The cross section for the reservoir in the GRITTER command (section 5.14) must be recomputed so that it is somewhat higher than the maximum dam height to be simulated.

The offset for the left-hand end of a cross-section subset is off the cross section on the left end in the computation of a floodway, an explicit request for a subset of the cross section, or the contraction ratio in the Bureau of Public Roads (Bradley, 1970) bridge-loss method. The subset boundary is reset to the left end of the cross section and the computations continue.

The offset for the right-hand end of the subset is off the cross section on the right end in the computation of a floodway, an explicit request for a subset of the cross section, or the contraction ratio in the Bureau of Public Roads (Bradley, 1970) bridge-loss method. The subset boundary is reset to the right end of the cross section and the computations continue.

The valid responses to the USGSBETA (section 5.1) are YES or NO. All requests other than YES are taken to be NO. This warning is issued to alert the user to the possibility of an input mistake.

The span inferred from the given rise for a reinforced-concrete arch pipe differs by more than 2 percent from the span given in the input. This might indicate an input error, that the pipe is not a pipe arch, or that the pipe does not conform to the standard dimensions programmed in MULCON (section 5.16).

The equivalent diameter of a nominal-elliptical pipe computed from the given span and also computed from the given rise differ by more than 3 percent in MULCON (section 5.16).

The width of the slot in a closed-conduit cross section is checked in FEQUTL. This warning is issued if the slot width is greater than 0.07 and if the cross section is used in a context that requires a closed conduit.

The elevation of the total-energy line is checked in the CULVERT command (section 5.5) and any increases in the downstream direction are reported. If the error is small, it may be reasonable to ignore it. The check is sometimes invalid or misleading when flow over the roadway is large relative to the flow through the culvert. Further testing is needed to develop a more refined check.

The elevation of the total-energy line is checked in the CULVERT command (section 5.5) and any increases in the downstream direction are reported. If the error is small, it may be reasonable to ignore it. The check is sometimes invalid or misleading when flow over the roadway is large relative to the flow through the culvert. Further testing is needed to develop a more refined check.

The elevation of the total-energy line is checked in the CULVERT command (section 5.5) and any increases in the downstream direction are reported. If the error is small, it may be reasonable to ignore it. The check is sometimes invalid or misleading when flow over the roadway is large relative to the flow through the culvert. Further testing is needed to develop a more refined check.

The head-water ratio for a pipe culvert for culvert-flow types 1, 2, and 3 exceeds the maximum of about 1.6 in the figures developed by Bodhaine (1968).

The transition between culvert-flow type 1 and type 6 is smoothed in the CULVERT command (section 5.5) by selecting the coefficient of discharge for culvert-flow type 6 so that full-barrel flow can match the culvert-flow type 1 at its limit. This match was not possible because the matching coefficient of discharge could not be determined.

The elevation of the total-energy line is checked in the CULVERT command (section 5.5) and any increases in the downstream direction are reported. If the error is small, it may be reasonable to ignore it. The check is sometimes invalid or misleading when flow over the roadway is large relative to the flow through the culvert. Further testing is needed to develop a more refined check.

There is a contraction of flow in the departure reach. Current experience indicates that this is a frequent occurrence at low to moderate flow rates. However, this is a departure from the standard culvert tests used to determine exit losses. Therefore, special judgments have been made in computation of the losses in this case. The departure section may not represent the departure reach, an error in the cross sections may be present, or an error in an elevation may be present in the CULVERT command (section 5.5) input.

A solution for the approach reach could not be computed in the CULVERT command (section 5.5) even though the approach cross section has been overtopped. This sometimes results during the process of finding a solution. If the warning persists, the cross section should be extended to some high value or the CULVERT input should be reviewed for errors. Too much flow over the roadway could be a possible cause for the message.

A Froude number greater than 1 by an internally set tolerance has been detected when computing the loss for culvert-flow type 3 in a box culvert with the CULVERT command (section 5.5). This can result especially at the onset of submergence when the loss coefficients have small inconsistencies. The value is set to 1.0 and the discharge coefficient for free flow is applied.

A root for culvert-flow type 1 has not been computed in the CULVERT command (section 5.5), and computations for culvert-flow type 2 are being attempted. This message may indicate some special problems with the computations. However, the flows often appear to be reasonable when this message is issued.

Informative message showing some aspects of the computation process in the CULVERT command (section 5.5). This message may indicate special problems in some cases.

Informative message showing some aspects of the computation process in the CULVERT command (section 5.5). This message may indicate special problems in some cases.

It is possible that an unanticipated flow pattern results, and this pattern cannot be properly interpreted in the flow-classification scheme. In this case, the free-flow type may not be identified in FEQUTL. Sometimes the output to the point of failure can help the user in correcting the problem. In other cases, the program must be changed. If program changes appear to be required, Linsley, Kraeger Associates of Mountain View, Calif., should be contacted.

A search and inverse interpolation are done in the CULVERT command (section 5.5) for the depth (water-surface height) at which the critical slope for a conduit matches the slope of the conduit invert. This interpolation sometimes has insufficient accuracy and subsequent computations fail. Therefore, the interpolation result is refined with an approximation to Newton's method. Sometimes the interpolation result fails to converge to the requested tolerance, and the last result is used and computations continue in the CULVERT command. The poor convergence often results at small depths. The interpolation can be improved, in some cases, if additional points are forced into the conduit cross-section description. This can be done by applying the DZLIM option (section 5.1) in the computation of the table for the conduit. However, care is required because a small DZLIM may increase the number of depth values above the number allowed. Computation for too many depth values should only result for rather large culverts. The conduit tables should be computed with a separate FEQUTL run if DZLIM is applied because the DZLIM option is global, applying to all cross-section tables computed in FEQUTL.

The maximum rounding or beveling value specified for the figures used to determine if flow is type 5 or type 6 for rough pipe culverts is 0.03 (Bodhaine, 1968). Therefore, all values greater than 0.03 are set to 0.03 for table look-up calculations in the CULVERT command (section 5.5).

The head to embankment width ratio for an embankment-shaped weir exceeds the maximum ratio. This warning is issued only once even though the condition may occur more than once for a given embankment. This message indicates that the heads may be large enough to invalidate the computed flows because the nature of the flow over the embankment at high heads may not be properly reflected in the weir-coefficient tables. The maximum ratio of 0.32 for embankment-shaped weirs is the approximate limit of experimental verification of the weir-coefficient variation with head.

The figures specifying the criteria for classifying flow as type 5 or type 6 do not go beyond a length to diameter ratio of 35 (Bodhaine, 1968). Values greater than this are set to 35 for table look-up calculations in the CULVERT command (section 5.5). This will lead to some inaccuracy.

This warning may indicate an input error in certain cases.

A unique subsection number must be assigned to each portion of the cross-section periphery with a different roughness, a noncompact shape, or both. Thus, repeated subsection numbers can only appear in an unbroken sequence in the input. New subsection numbers are added in FEQUTL computations to meet this requirement. The roughness values for the new subsection numbers are retained from the old subsection. For example, a concrete-lined channel with a concrete-covered berm must contain at least two subsections to account for the noncompact shape caused by the berm. The roughness is the same for each subsection. A wide, rectangular channel with only part of the channel bottom paved also will require more than one subsection to account for the change of roughness without a change in shape.

NEWBETA can only be applied to a cross section with no decrease in the offsets.

Several assumptions are made about the variation of critical flow in the solution process for the EXPCON command (section 5.7). When overbank flow is simulated, critical flow and velocity may decrease. This may yield incorrect solutions in EXPCON. Unrealistic variation in the results must be checked. The root of the problem is that the concept of critical flow does not extend easily to compound channels. It has been observed in previous computations that one of the free flows may decrease to a completely invalid value. Sometimes a variation in the downstream head sequence can solve the problem. Changes in some of the loss values also may solve the problem.

Several assumptions are made about the variation of critical flow in the solution process for the CHANRAT command (section 5.3). When overbank flow is simulated, critical flow and velocity may decrease with increasing water-surface height. This may yield incorrect solutions in CHANRAT. Unrealistic variation in the results must be checked. The root of the problem is that the concept of critical flow does not extend easily to compound channels. Sometimes a variation in the upstream head sequence can solve the problem.

Computations for culvert-flow type 2 did not converge. Thus, another flow type will be computed in the CULVERT command (section 5.5).

In EMBANKQ (section 5.6), the weir coefficients are determined under the assumption that the weir height is large enough to have flow over a weir and not flow over a sill. The boundary between the two flow types is not distinct. A warning is issued at a ratio of 4 to indicate that the ratios are becoming large. The flow over the weir at each point in EMBANKQ computations is limited to critical flow in the approach section. Flow over a sill can be approximated with this assumption. However, if the flow is deep (greater than four times the height of the sill), the sill becomes more like a roughness element of the channel boundary and the assumption of critical depth may no longer be valid.

This message is issued whenever a cross-section execution is extended on the left end (the end with the smaller
values of offsets) to compute the table of cross-sectional hydraulic characteristics. This extension is only applied if the EXTEND option (section 5.1) of FEQUTL is selected.

This message is issued whenever a cross-section elevation is extended on the right end (the end with the larger values of offsets) to compute the table of cross-sectional hydraulic characteristics. This extension is only applied if the EXTEND option (section 5.1) of FEQUTL is selected.

A point other than the end points of the cross section has been detected as the highest. The table computations are limited to the minimum-end-point elevation because a high internal elevation is indicative of an error.

A subsection consisting of a single vertical line has no area. Its perimeter does not belong to any other subsection so that it is in effect a frictionless boundary. If this is not what is desired, the input must be changed so that the subsection includes an area.

This message is issued if the EXTEND = YES option is selected in the input-header block (section 5.1). Once the table is computed, there is no consideration of the extension of one side of the cross section in FEQ simulations. Therefore, any section with a large extension should be checked before application in FEQ. In FEQ simulation, it is assumed that the section contains water at all water-surface levels in the table, and warnings about unrealistic results cannot be issued.

The valid responses for the EXTEND option in the header block (section 5.1) are NO and YES. Responses other than these will result in a default value of NO.

A slot width wider than expected for most closed conduits has been detected in the QCLIMIT command (section 5.18). If the slot width is larger than expected, the data must be checked to determine the reason for the difference. The computations will continue as if the slot width is correct.

In HEC2X (section 5.15) the indirect mode has been selected, but no file name has been given. The file name INDIRECT is assigned by default. Computations continue and the output will appear in the file INDIRECT.

A slot height that is too small had been detected in the MULCON (section 5.16), MULPIPES (section 5.17), or SEWER (section 5.19) commands. It has been increased to allow the computations to continue.

A roughness factor is used to classify culvert-flow as type 5 or type 6 for rough pipes in the CULVERT command (section 5.5). The figures in Bodhaine (1968) for flow classification specify the range of this factor between 0.10 and 0.30. However, pipes with roughness values outside this range may be present in a stream network. The figures in Bodhaine (1968) were extrapolated both to larger and to smaller ranges. Outside of the given range, the accuracy of the computations is unknown.

Culvert-flow is classified as type 5 or type 6 for smooth pipes or box culverts with use of figure 13 in the CULVERT command (section 5.5). The maximum value for rounding/beveling in figure 13 is 0.06, and any larger values are set to 0.06 in table look-up calculations.

The specific force function is inverted in the INVTM subroutine to find conjugate depths for a hydraulic jump. Convergence is checked in two ways in the iterative solution process. In the first way, the relative residual value is computed. In the second way, the relative change in the argument to the residual function is computed. If the successive arguments converge before the residual function converges to zero, a convergence problem may be present. This warning is issued if such a situation arises.

The tables for the culvert-flow type 5 discharge coefficient have not been established above a ratio of head to vertical diameter of 5 in the CULVERT command (section 5.5). However, culverts may be present in the stream network for which this ratio is exceeded. The tables have been extended by extrapolation of a constant discharge ratio above the upper limit. This is inaccurate. If culverts that cause this message to be issued are present, the discharge coefficient tables should be reviewed. These are usually stored in a file named TYPE5.TAB. If necessary, the extrapolation may be changed to better reflect the conditions of a particular culvert.

Submerged flows in the culvert are computed in the subroutine QVSTW in the CULVERT command (section 5.5). Submerged-flow computations for culvert-flow type 1 profiles can be difficult in smooth, steep culverts. A flow adjustment procedure for conditions arising when culvert-flow type 1 is the free-flow type is applied to try to compute these profiles in the CULVERT command. These adjustments sometimes fail and then the downstream water-surface elevation (tail water) is adjusted to a higher level.

Convergence problems noted in WRN:567 could not be solved by flow adjustment. Therefore, downstream water-surface elevation (tail-water) adjustment is applied in the subroutine QVSTM in the CULVERT command (section 5.5). The old downstream water-surface elevation would have been used without adjustment.

The pressure at the vena contracta near the culvert entrance is estimated in the CULVERT command (section 5.5) when the pipe is flowing full. A table of allowable subatmospheric heads as a function of elevation is included in the CULVERT command. When the estimated pressure becomes less than allowable, WRN:569 is written. This should only result at high heads not often encountered for culverts.

The subatmospheric pressure computed at the vena contracta in the CULVERT command (section 5.5) has become physically impossible. Therefore, the results computed are invalid. The culvert configuration must be changed, usually by rounding the entrance, to prevent this message from appearing. See WRN:569.

A request for extrapolation of a slotted cross section has been detected and is ignored. Slotted sections are always assumed to extend as high as necessary when they are computed. The methods used for extrapolation of the cross-section function table may not work properly with slotted sections.

The watersurface profile for culvert-flow type 2 could not be computed to the entrance of the culvert in the CULVERT command (section 5.5). The profile starts at critical depth at section 3, given by Y3. The critical flow is given by Q3, and this is the flow in the barrel. IS and SFLAG are used to diagnose problems should the adjustment fail. Contact Linsley, Kraeger Associates of Mountain View, Calif., if this warning results consistently for a particular stream system.

The depth in the approach reach could not be computed in the CULVERT command (section 5.5) for culvert-flow type 2. The critical depth at section 3, Y3, and the critical flow at section 3, Q3, are increased in an attempt to compute the approach-reach results.

A culvert-flow type 6 coefficient of discharge for which full-pipe flow matches the culvert-flow type 61 at the type 61 upper limit is sought in the CULVERT (section 5.5) computations. The search may not succeed, and this message is issued when it does not. The transition between the flows may be abrupt enough to require some manual adjustment of the table.

A positive residual for culvert-flow type 0 cannot be computed in the CULVERT command (section 5.5). Under these conditions, it is assumed that culvert-flow type 0 cannot be present in subsequent CULVERT command computations.

The tail water at section 43 is higher than the submergence limit, indicating that the flow over the road must be submerged; but the computations in the CULVERT command (section 5.5) were done with the assumption of no flow over the road. It is assumed that culvert-flow type 0 cannot be present, and computations continue. However, these computations may not be successful. The culvert representation may need to be changed. See ERR:687.

See WRN:576. The culvert representation should probably be changed in the CULVERT command (section 5.5). The flow length assigned to the flow over the roadway may have to be defined more carefully. The flow length can be difficult to assign because flow may move down the road instead of across the road if curbs and gutters are present and the overflow width is poorly defined.

It is assumed in CULVERT (section 5.5) computations that culvert-flow type 0 cannot be present. This may be incorrect, and the representation for the culvert may have to be modified as described in ERR:687.

An attempt is made in CULVERT (section 5.5) computations to reflect culvert barrels with variable slope. However, the analysis rapidly becomes complex. Drowning of culvert-flow type 1 by type 2 flow usually indicates that the invert slope in the barrel decreases in the downstream direction. If the slope change is too large, it may not be possible to analyze the flows in the CULVERT command.

In CULVERT (section 5.5) computations, a match between full-barrel flow and culvert-flow type 2 at its upper limit is sought to provide a smooth transition between the two flow types. This may not be possible in all cases, and when it is not, manual adjustment of the flows may be required to avoid an abrupt change in flow at the boundary between the two flow types.

In CULVERT (section 5.5) computations, a match between culvert-flow type 61 and culvert-flow type 2 at its upper limit is sought to provide a smooth transition between the two flow types. This may not be possible, and the transition between the two flow types may be abrupt enough to require manual adjustment.

The ability to represent culverts with barrels with slope breaks is limited in the CULVERT command (section 5.5). If the breaks are small and never represent a control section, the CULVERT command may prove adequate. However, if the breaks are such that a control section forms, it is unlikely that the computations will be successful. It is sometimes possible to divide such a culvert into two sections with the slope break as the point of division. However, it may be difficult to simulate these culverts because the approach and departure reaches that must be formed at the break are fictitious.

The adjustment factor for momentum flux for the transition from culvert-flow type 5 to type 4 contains an invalid value. It is set to 0 in the CULVERT (section 5.5) computations. The transition may be abrupt or invalid. The flow table computed for the culvert should be checked by the user.

A value of KWING not equal to 1.0 has been found for a culvert that is not a box culvert. The wingwall adjustment factor in the CULVERT command (section 5.5) only applies to box culverts.

A value of KPROJ not equal to 1 has been found for a box culvert. This adjustment only applies to other types of culverts in the CULVERT command (section 5.5).

The initial value for Manning's n for the left overbank was found to be zero in the HEC2X command (section 5.15). It is set to 1.0 and computations continue.

The initial value for Manning's n for the channel was found to be zero in the HEC2X command (section 5.15). It is set to 1.0 and computations continue.

The initial value for Manning's n for the right overbank was found to be zero in the HEC2X command (section 5.15). It is set to 1.0 and computations continue.

The boundary between culvert-flow types 1 and 2 can be subject to small variations. Whenever the flows used to check for the validity of culvert-flow type 1 are in close agreement, subsequent computational problems are more likely. If computational problems result, the upstream head must be changed in the CULVERT command (section 5.5) to avoid computations close to the boundary between culvert-flow types 1 and 2.

It is possible for a culvert to be in flow type 2 at small and moderate water-surface heights and to transition to flow type 1 at higher water-surface heights and then return to flow type 2, or more likely to one of the high-head flows as flow water-surface heights approach the crown of the culvert. Thus, there can be a lower limit at section 1 for culvert-flow type 1 and an upper limit at section 1 for culvert-flow type 1. The upper limit for the lower region of culvert-flow type 2 does not always match the lower limit for the culvert-flow type 1 region. This results because of unavoidable approximations in the loss-coefficient curves and tables and in the treatment of losses. It is therefore possible for the head at section 1 to be at a level such that culvert-flow type 1 computations fail and culvert-flow type 2 computations fail. The correction for this problem is to increment the upstream head value in the CULVERT command (section 5.5) at the point where the problem results so that the upstream head is clearly in the culvert-flow type 1 region. If this value of head was added in CULVERT computations to match the minimum point on the roadway, then the minimum point of the roadway also must be adjusted. The increment usually is very small so that computational accuracy will not be greatly affected as a result of these changes.

This message is related to that of WRN:590. The computational failure in the CULVERT command (section 5.5) has occurred because the increment to the head at section 1 placed the flow in the culvert-flow type 1 region, but the computational decision process for culvert-flow type 1 resulted in simulation of culvert-flow type 2. This can result because of the inherent convergence differences in computing the limits for culvert-flow types 1 and 2. See message for WRN:589.

A case of energy gain from section 1 to section 4 has been detected in the UFGATE command (section 5.20). This could indicate that the convergence tolerances, EPSF and EPSARG, are too small. The values of these tolerances should be reduced by a factor of two or more and the program run again. If the warning persists, a problem may be present in the solution method for the particular case. Contact Linsley, Kraeger Associates of Mountain View, Calif., with the details of the gate being simulated.

Computation of a supercritical profile in the CULVERT command (section 5.5) resulted in a gap between the subcritical profile and the supercritical profile because the profile is so short or does not start at all. This implies that no profile is present in this gap. This cannot be correct because a profile of some sort must be present at all locations in the culvert. Therefore, the estimated losses for culvert-flow type 6 applied downstream from the vena contracta location are reduced until a supercritical-profile length is computed that closes the gap.

The gap described in WRN:593 cannot be closed in CULVERT (section 5.5) computations. The CULVERT command computations may succeed with small changes in barrel slope or section. However, the culvert-flow type 5 computations are especially dependent on the assumption of a prismatic barrel. All options available in the CULVERT command to close the gap have been attempted and none have succeeded in closing the gap.