(a) Steady-state emission testing. Make sure your engine's speeds and loads stay within ±2 percent of the set point during the sampling period.

(b) Transient emission testing performed by EPA. Emission tests must meet the specifications of this paragraph (b). Otherwise, they do not comply with the test cycle requirements of the standard-setting part, unless we determine the cause of the failure to meet these specifications is related to the engine rather than the test equipment.

(1) Shifting feedback signals. The time lag between the feedback and reference-cycle values may bias results. To reduce this effect, you may advance or delay the entire sequence of engine-speed and torque-feedback signals with respect to the reference sequence for speed and torque. If you shift the feedback signals, you must shift speed and torque the same amount in the same direction.

(2) Calculating brake kilowatt-hour emissions. Calculate brake kilowatt-hour emissions for each pair of feedback values recorded for engine speed and torque. Also calculate the reference brake kilowatt-hour for each pair of reference values for engine speed and torque. Calculate to five significant figures.

(3) Performing regression-line analysis. Perform regression analysis to calculate validation statistics as follows:

(i) Perform linear regressions of feedback value on reference value for speed, torque, and brake power on 1 Hz data after the feedback shift has occurred (see paragraph (b)(1) of this section). Use the method of least squares, with the best-fit equation having the form:

y = mx + b

Where:

y = The feedback (actual) value of speed (rpm), torque (ft-lbs.), or brake power.

m = Slope of the regression line.

x = The reference value (speed, torque, or brake power).

b = The y-intercept of the regression line.

(ii) Calculate the standard error of estimate (SE) of y on x and the coefficient of determination (r ²) for each regression line.

(iii) For a valid test, make sure the feedback cycle's integrated brake kilowatt-hour is within 5 percent of the reference cycle's integrated brake kilowatt-hour. Also, ensure that the slope, intercept, standard error, and coefficient of determination meet the criteria in the following table (you may delete individual points from the regression analyses, consistent with good engineering judgment):

                   Table 1 of § 1065.530_Statistical Criteria for Validating Test Cycles
----------------------------------------------------------------------------------------------------------------
                                                Speed                    Torque                   Power
----------------------------------------------------------------------------------------------------------------
1. Slope of the regression line (m)..  0.980 to 1.020.........  0.880 to 1.030.........  0.880 to 1.030.
----------------------------------------------------------------------------------------------------------------
2. Y intercept of the regression line  [verbarlm]b[verbarlm]    [verbarlm]b[verbarlm]    [verbarlm]b[verbarlm]
 (b).                                   [le] 40 rpm.             [le] 5.0 percent of      [le] 3.0 percent of
                                                                 maximum torque from      maximum torque from
                                                                 power map.               power map.
----------------------------------------------------------------------------------------------------------------
3. Standard error of the estimate of   100 rpm................  15 percent of maximum    10 percent of maximum
 Y on X (SE).                                                    torque from power map.   power from power map.
----------------------------------------------------------------------------------------------------------------
4. Coefficient of determination        r\2\ >= 0.970..........  r\2\ >= 0.900..........  r\2\ >= 0.900.
 (r\2\).
----------------------------------------------------------------------------------------------------------------

(c) Transient testing performed by manufacturers. Emission tests that meet the specifications of paragraph (b) of this section satisfy the standard-setting part's requirements for test cycles. You may ask to use a dynamometer that cannot meet those specifications, consistent with good engineering practice. We will approve your request as long as using the alternate dynamometer does not affect your ability to show that you comply with all applicable emission standards.