Design Exceptions

During each phase of the preconstruction process a team may flag discrepancies between UDOT standards and what can be accomplished within the project context. These discrepancies often become design exceptions to the UDOT standards. The safety impacts of design exceptions should be evaluated and discussed with the design team as they arise so that appropriate decisions can be made early in the project process. The preconstruction team prepares the Project Design Criteria (PDC) form for each project prior to advertising. The PDC identifies the FHWA Controlling Criteria and identifies if any project criteria do not meet UDOT standards. Substandard criteria require Design Exceptions (DE). The preconstruction team is required to explain how each DE impacts public safety. 

This webpage suggests how to quantify the safety cost for a DE analysis. The preconstruction team should seek additional funds to meet the design standards if their costs are less than the safety impacts calculated. In many cases the cost to implement the standard exceeds the safety impacts. In this case, the design team justifies the DE by stating that it is not economically viable to implement the mandated standards.

Information Required for DE Analysis

The project design team is best suited to perform the DE analysis because they have access to all information required to achieve a proper analysis. On July 6, 2022, the Division of Traffic and Safety provided a training session to introduce the concepts of predictive modeling and its use for the evaluation of DEs. A link to this training and the associated slides can be found at the bottom of this article. 

Some design teams may not feel qualified to perform a predictive crash analysis or did not include it within the scope of their contract. In that case, it is imperative that a 3rd party analyst with a good understanding of the design be used to complete the analysis. Typically the following information is required from the design team:

1. A copy of the PDC describing the DE that should be evaluated.

2. A design file that the analyst can use to become familiar with the project. This might be achieved with PDF drawings, KMZ files, or actual design documents.

3. Extents of the DE and dimensions/measures of design attributes (e.g. average lane/shoulder widths).

4. Typical cross sections of the roadway and where the DE is applicable, this helps to segment the project. 

5. If horizontal curves are part of the DE analysis provide layouts showing the proposed alignment. A list of locations, lengths, and radius of curves within the project is required.

Beside the items listed above, the analyst will need to have a good understanding of roadway assets along the roadway. UPlan has asset inventory information that can be used by the analyst. It is also good practice to meet with the design team and discuss the results of the analysis before finalizing it to ensure that the assumptions used in the analysis are valid and the results are understood.

Predictive Modeling

The Predictive Modeling procedure is used to quantify the safety costs of the DE. The Predictive Modeling webpage provides calibration numbers necessary for various SPFs used in Utah. The Historical Modeling webpage provides statewide average proportions by crash type, manner of collision, and crash severity for use in predictive modeling. 

Predictive modeling is used because the proposed alternatives do not exist yet. The following concepts should be applied when doing a predictive model for DE:

1. Expected vs. Predicted crashes:

   • The calibrated predicted crashes should be used for comparison if the alternatives considered use different SPFs from the Highway Safety Manual (HSM). 

   • The expected crashes should be used for comparison if the alternatives considered use the same SPFs from the HSM and the predicted model is only changed via the use of CMFs. Expected crashes for a future condition is achieved by calculating the ratio of expected to predicted crashes for the current condition and applying that ratio as a multiplier to the future predicted crashes.

2. SPF selection:

   • Some roadway types in Utah are not presented in the HSM. The calibration numbers used in Utah give calibration values for SPFs used in the HSM for other roadway types (e.g 3 lanes with a two-way-left-turn lane (4T) modeled as a 3T).

   • If a roadway type is not provided in the calibration list (e.g. one-way streets), the analyst might seek third party studies to model the proposed alternatives.

3. CMFs used:

   • Base Conditions: Each SPF in the HSM uses “base conditions” for the generation of its specific SPF. These base conditions are CMFs that should be used when comparing proposed DE.

   • Third party CMFs: Many DEs address issues that are not covered with the “base conditions”. Third party CMFs might be used from the CMF Clearinghouse, or other published sources that the analyst finds acceptable based on engineering judgment. It is imperative that the analyst try to find appropriate CMFs considering:

     1. Roadway type.

     2. Crash types being addressed. CMFs that do not address all crash types must be adjusted based on the proportion of the crash type that was investigated.

     3. Crash severities considered. CMFs that do not address all crash severities must be adjusted based on the proportion of the crash severity that was investigated. 

     4. Vehicle volume.

     5. Reliability of CMF (comparing the star ratings of various CMFs).

   • Unusual CMFs: Some studies have been structured such that their results do not agree with the majority of other research pertaining to that subject. These CMFs should be avoided until their results are better understood.

4. No CMFs available: Some DE address changes that have not been studied for safety impacts. The analyst should declare that there is insufficient evidence currently available to quantify the safety impacts of the specific DE. General guidance can be discussed with the design team but quantification of safety may not be obtainable. 

5. The predictive analysis should calculate the total crashes and then apply the severity distributions provided for Utah.

6. Crash costs should be applied based on the most current values. 

   • The unweighted crash costs are used when a severity distribution is available.

   • The weighted crash costs should be used when no severity distribution is available. 

7. Evaluations should use AADT values reported in the PDC for both the current and future conditions. Predicted/expected crashes will be extrapolated over the analysis period declared in the PDC. If no future AADT is provided the analyst should state that no growth in AADT was considered.

8. A discount rate of 3% is used.

9. Report the change in crashes between the standard condition and the DE, and the resulting present value in dollars over the analysis period.

10. Typically a PDC assumes a 20 year analysis period. If the design team knows that the roadway under consideration has a project horizon less than 20 years, it should be noted in the analysis and an agreed upon period should be applied.

Information to Provide in the Analysis

Each DE analysis memo should provide the following items to the preconstruction team. 

1. Statement of DE investigated and the methodology used (predicted/expected crashes).

2. AADT values used.

3. Summary of current crashes observed.

4. Crash frequency by severity level for each year of the analysis period (i.e. beginning and ending year frequency or an average frequency).

5. The safety cost in dollars for the implementation of the DE. This value is used by the preconstruction team to decide on an action.

6. Assumptions that were applied including copies of third party CMFs and the application methodology. 

The analyst may also provide other low cost countermeasures that address the crashes affected by the DE if the standard design cannot be met.

The following presentation and training recording was presented on July 6th, 2022.