Using linear regression to analyze concrete strength data, the Corps of Engineers has developed two computer programs that help predict 90-day strengths from one-day accelerated cured-cylinder strengths. These programs use equations from ACI 214-77 and provide the user with standard deviations, predicted strengths, and graphs of the data for concrete that can be used for any design age.

The current trend of fast construction has necessitated the knowledge of the 28-day strength of the concrete cylinder at an early stage. This is achieved by using prediction relations to predict the 28-day strength based on the early test. This article presents a general method to develop the 28-day strength prediction models by analyzing the compression test data statistically and by using the linear regression analysis. A procedure is discussed to develop a computer program to perform the linear regression analysis. A method is also offered to obtain the nonlinear curves from the linear regression analysis program. Two applications of these prediction relations are discussed--prediction of 28-day strength and specification of early strength as an acceptance criterion. As an example, the prediction relations developed for the City of Phoenix are presented.

In-place testing offers the contractor an alternative to testing field-cured cylinders for assessing concrete strength during critical phases of construction. A relationship between cylinder compressive strength and the quantity measured by in-place tests is needed to use these methods; this requires statistical analysis of test data. The personal computer and "electronic spreadsheet" software provide a cost-effective means of performing such analyses. Available spreadsheet software can be used to perform the required calculations automatically, thus providing the user with the benefits of automated computation without requiring knowledge of a scientific computing language. Data from a study of the pullout test method are used to illustrate the use of spreadsheet software to perform linear regression and determine the required correlation relationship. In addition, a simplified technique is proposed for using the results of in-place tests to make a reliable estimate of the characteristic strength of structural concrete. The in-place characteristic strength permits an evaluation of structural safety during construction.

Pullout testing provides a measurement of in-place strength of concrete, and maturity can be used to estimate strength development. Test statistics are useful tools for interpreting the significance of the data. Maturity may correlate to pullout strength by regression analysis. The simplest correlation would be linear using the logarithm of maturity. Linear relationships may be compared using test statistics to determine if slopes and/or intercepts are significantly different. Statistical analysis techniques are used to evaluate pullout strength and maturity data. Some limitations associated with these techniques are discussed. A programmable calculator was used to develop the programs to illustrate the application of these analysis techniques.

This report presents a computer program for the computation of required average strength, f'cr, based on the field strength test data. Based on the recent Code (ACI 318-83), four different cases--available data more than or equal to 30, available data for two consecutive sets more than or equal to 30, available data more than or equal to 15 but less than 30, and no field data available for comparable concrete--are identified and solved. The necessary equations and flow chart are presented so the reader more clearly understands the sequence of operations. Input/output data are explained using sample problems that illustrate the capability of the program to handle the different cases. The steps presented are written for one particular brand of computer, but the specialized commands are kept to a minimum so that the user can easily modify the program to other types of similar widely used computers.