3.1 Engineering Data Analysis and Statistical Applications

3.1 Engineering Data Analysis and Statistical Applications

Civil engineering projects demand rigorous statistical data analysis to manage inherent uncertainties in material properties, environmental loads, geotechnical parameters, and construction quality. Whether evaluating the compressive strength of structural concrete cylinder breaks, characterizing aggregate gradation from sieve analyses, predicting peak flood discharges, or establishing traffic arrival patterns, engineers must translate sample observations into reliable probabilistic decisions.

---

Measures of Central Tendency

Measures of central tendency locate the numerical center of an engineering dataset:

• Sample Mean (Arithmetic Mean, $\bar{x}$): The sum of all observed values divided by the total number of observations $n$: $\bar{x} = \frac{1}{n} \sum_{i=1}^{n} x_i$ For population data of size $N$, the true population mean is denoted by $\mu$.
• Median: The middle value of an ordered dataset arranged in ascending or descending sequence. For an odd sample size $n$, the median is the value at position $(n+1)/2$. For an even sample size $n$, it is the arithmetic average of the two central values at positions $n/2$ and $(n/2)+1$. Unlike the mean, the median is resistant to extreme outliers (such as a single rogue defective cylinder test caused by improper capping).
• Mode: The most frequently occurring data value in the sample. A dataset may be unimodal, bimodal, or multimodal. In highway engineering, modal traffic speeds represent the speed chosen by the greatest concentration of drivers.

Skewness and Distribution Symmetry

• Symmetric Distribution: $\text{Mean} = \text{Median} = \text{Mode}$.
• Positive (Right) Skewed: $\text{Mean} > \text{Median} > \text{Mode}$. Extreme high values pull the mean upward (e.g., annual peak flood river discharges, construction project cost overruns).
• Negative (Left) Skewed: $\text{Mode} > \text{Median} > \text{Mean}$. Extreme low values pull the mean downward (e.g., concrete strength results where strict quality control truncates high-end variability).

---

Measures of Dispersion and Variability

Central tendency alone cannot describe material uniformity. Two concrete batches may both average $28\text{ MPa}$, yet one batch may exhibit tight clustering while the other exhibits hazardous wild fluctuations.

Population vs. Sample Variance

• Population Variance ($\sigma^2$): Used when every entity in the population is measured: $\sigma^2 = \frac{1}{N} \sum_{i=1}^{N} (x_i - \mu)^2$
• Sample Variance ($s^2$): Used when estimating population variance from a representative sample. Crucially, dividing by $n$ produces a biased estimator that systematically underestimates the true population dispersion. Applying Bessel's correction with $n - 1$ degrees of freedom restores unbiasedness: $s^2 = \frac{1}{n - 1} \sum_{i=1}^{n} (x_i - \bar{x})^2 = \frac{\sum x_i^2 - \frac{(\sum x_i)^2}{n}}{n - 1}$
• Sample Standard Deviation ($s$): The positive square root of sample variance, expressed in the same physical units as the measured parameter (e.g., $\text{MPa}$, $\text{kN}$, $\text{mm}$): $s = \sqrt{s^2} = \sqrt{\frac{\sum_{i=1}^n (x_i - \bar{x})^2}{n - 1}}$

Coefficient of Variation ($CV$)

The Coefficient of Variation ($CV$, or $V$) is the ratio of the standard deviation to the mean, expressed as a dimensionless percentage: $CV = \frac{s}{\bar{x}} \times 100\%$ Because it eliminates dimensional units, $CV$ allows direct quality control comparisons across different structural materials and specified compressive strengths ($f'_c$).

ACI 214R-11 (Table 4.3) rates the overall variation of general construction testing for $f'_c \le 35\text{ MPa}$ by standard deviation. For higher strengths, it prefers the coefficient of variation because scatter grows roughly in proportion to strength.

---

Probability Fundamentals & Counting Rules

Civil engineering probabilistic safety margins rely on basic counting theorems and probability axioms:

• Fundamental Counting Principle: If an operation consists of $k$ consecutive steps where step 1 has $n_1$ outcomes, step 2 has $n_2$ outcomes, ..., the total number of composite outcomes is $n_1 \times n_2 \times \cdots \times n_k$.
• Permutations (Order Matters): The number of ordered arrangements of $r$ objects selected from a set of $n$ distinct objects: $P(n, r) = \frac{n!}{(n - r)!}$
• Combinations (Order Does Not Matter): The number of unordered subsets of $r$ objects selected from $n$ distinct objects: $C(n, r) = \binom{n}{r} = \frac{n!}{r!(n - r)!}$

Addition and Multiplication Laws

• Addition Rule (Union of Events): $P(A \cup B) = P(A) + P(B) - P(A \cap B)$ If events $A$ and $B$ are mutually exclusive (disjoint, meaning $A \cap B = \emptyset$ and $P(A \cap B) = 0$): $P(A \cup B) = P(A) + P(B)$
• Conditional Probability: The probability that event $A$ occurs given that event $B$ has already occurred: $P(A \mid B) = \frac{P(A \cap B)}{P(B)}, \quad \text{provided } P(B) > 0$
• Multiplication Rule (Intersection of Events): $P(A \cap B) = P(B) \cdot P(A \mid B) = P(A) \cdot P(B \mid A)$ If events $A$ and $B$ are statistically independent ($P(A \mid B) = P(A)$): $P(A \cap B) = P(A) \cdot P(B)$

---

Discrete Probability Distributions

1. Binomial Distribution

The Binomial distribution models the number of successes $k$ in a fixed series of $n$ independent trials, where each trial yields only two possible outcomes (Success/Failure, Defective/Non-Defective) with constant probability $p$ ($q = 1 - p$): $P(X = k) = \binom{n}{k} p^k (1 - p)^{n - k} = \frac{n!}{k!(n - k)!} p^k q^{n - k}$

• Expected Value (Mean): $\mu = n p$
• Variance: $\sigma^2 = n p q = n p (1 - p)$
• Standard Deviation: $\sigma = \sqrt{n p (1 - p)}$

Civil Engineering Application: Calculating the probability of experiencing a 50-year design flood ($p = 1/50 = 0.02$) over the temporary 3-year construction life of a river bridge pier cofferdam.

2. Poisson Distribution

The Poisson distribution models the number of discrete occurrences of a rare event within a specified continuous interval of time, space, area, or volume: $P(X = k) = \frac{\lambda^k e^{-\lambda}}{k!}, \quad k = 0, 1, 2, \dots$ where $\lambda$ is the average arrival rate or expected number of events in the interval, and $e \approx 2.71828$.

• Mean: $\mu = \lambda$
• Variance: $\sigma^2 = \lambda$
• Standard Deviation: $\sigma = \sqrt{\lambda}$

Civil Engineering Application: Modeling vehicle arrivals at an unsignalized highway intersection approach during peak hour traffic flow, or the frequency of severe wind gust occurrences per typhoon season.

---

Continuous Distributions: Normal (Gaussian) Distribution

The Normal Distribution is the cornerstone of structural safety, geotechnical soil property modeling, and quality assurance. The continuous probability density function is symmetric and bell-shaped: $f(x) = \frac{1}{\sigma \sqrt{2\pi}} e^{-\frac{1}{2}\left(\frac{x - \mu}{\sigma}\right)^2}$

The Standard Normal Variable ($z$-score)

Any arbitrary normal variable $X \sim N(\mu, \sigma^2)$ transforms into the standardized normal variable $Z \sim N(0, 1)$ via: $z = \frac{x - \mu}{\sigma}$ The $z$-score measures the number of standard deviations an observed value $x$ lies above or below the population mean $\mu$.

The Empirical (68–95–99.7) Rule

• $P(\mu - 1\sigma \le X \le \mu + 1\sigma) \approx 68.27\%$
• $P(\mu - 2\sigma \le X \le \mu + 2\sigma) \approx 95.45\%$
• $P(\mu - 3\sigma \le X \le \mu + 3\sigma) \approx 99.73\%$

Central Limit Theorem (CLT) & Standard Error

Let $X_1, X_2, \dots, X_n$ be an independent, identically distributed (i.i.d.) random sample from any population with mean $\mu$ and finite variance $\sigma^2$. As sample size $n$ increases ($n \ge 30$), the distribution of the sample mean $\bar{X}$ approaches a normal distribution: $\bar{X} \sim N\left(\mu, \frac{\sigma^2}{n}\right)$ The standard deviation of the sampling distribution of the mean is the Standard Error of the Mean ($SE$): $\sigma_{\bar{x}} = \frac{\sigma}{\sqrt{n}} \quad \text{or estimated as} \quad SE = \frac{s}{\sqrt{n}}$ When performing hypothesis tests or calculating confidence limits on sample means, the standardized statistic becomes: $z = \frac{\bar{x} - \mu}{\sigma / \sqrt{n}} \quad \text{or} \quad t = \frac{\bar{x} - \mu}{s / \sqrt{n}} \quad (\text{with } df = n - 1)$

---

Quality Control & Concrete Acceptance (NSCP / ACI 318)

In reinforced concrete design and construction monitoring, field cylinder specimens ($150\text{ mm} \times 300\text{ mm}$ or $100\text{ mm} \times 200\text{ mm}$) are tested in compression at 28 days in accordance with ASTM C39.

Target Required Average Compressive Strength ($f'_{cr}$)

To ensure that the specified design compressive strength ($f'_c$) is satisfied with negligible probability of understrength, batch plants must design the concrete mix for a higher target average strength $f'_{cr}$. ACI 318-14 (Section 26.4.3), which NSCP 2015 follows, refers proportioning to ACI 301. When a plant has a record of 30 or more consecutive tests, its standard deviation $s$ gives:

1. For $f'_c \le 35\text{ MPa}$: $f'_{cr}$ is the larger of: $f'_{cr} = f'_c + 1.34 s$ $f'_{cr} = f'_c + 2.33 s - 3.5$
2. For $f'_c > 35\text{ MPa}$: $f'_{cr}$ is the larger of: $f'_{cr} = f'_c + 1.34 s$ $f'_{cr} = 0.90 f'_c + 2.33 s$

Official Acceptance Criteria

A concrete batch meets structural design criteria if both of the following conditions are satisfied:

1. Every arithmetic average of any three consecutive strength tests equals or exceeds $f'_c$.
2. No individual strength test (the average of at least two $150 \times 300\text{ mm}$ or three $100 \times 200\text{ mm}$ cylinders) falls below $f'_c$ by more than $3.5\text{ MPa}$ when $f'_c \le 35\text{ MPa}$, or by more than $0.10 f'_c$ when $f'_c > 35\text{ MPa}$.

---

Confidence Intervals & Hypothesis Testing

Confidence Interval for Population Mean

A $(1 - \alpha) 100\%$ confidence interval represents the range within which the true population parameter lies with confidence $1 - \alpha$: $\bar{x} - z_{\alpha/2} \left(\frac{s}{\sqrt{n}}\right) \le \mu \le \bar{x} + z_{\alpha/2} \left(\frac{s}{\sqrt{n}}\right)$ For small sample sizes ($n < 30$) with unknown population variance, the critical Student's $t$-value ($t_{\alpha/2, n-1}$) replaces $z_{\alpha/2}$.

Hypothesis Testing Framework

1. State Hypotheses: Null Hypothesis $H_0$ (baseline claim, e.g., $\mu = f'_c$) versus Alternative Hypothesis $H_a$ (e.g., $\mu < f'_c$ for lower-tail test).
2. Select Significance Level ($\alpha$): Common standard is $\alpha = 0.05$ (5% risk of false rejection).
3. Calculate Test Statistic: $z_{\text{calc}} = \frac{\bar{x} - \mu_0}{s / \sqrt{n}}$.
4. Decision Rule:
   • Type I Error ($\alpha$, Producer's Risk): Rejecting $H_0$ when $H_0$ is actually true (e.g., rejecting an acceptable cement shipment).
   • Type II Error ($\beta$, Consumer's Risk): Failing to reject $H_0$ when $H_0$ is false (e.g., accepting substandard structural concrete).

---

Worked Engineering Examples

Example 1: Concrete Cylinder Test Evaluation

Eight ($n = 8$) standard 28-day cylinder compressive strength tests yielded the following results (in $\text{MPa}$): $29.4, \; 31.2, \; 28.6, \; 32.5, \; 30.0, \; 29.8, \; 33.1, \; 31.4$ The specified compressive design strength is $f'_c = 28.0\text{ MPa}$. Calculate the sample mean $\bar{x}$, sample standard deviation $s$, coefficient of variation $CV$, and evaluate whether the plant meets the required average strength $f'_{cr}$, assuming its long-term record of 30 or more tests shows this same standard deviation.

Solution:

1. Sample Mean: $\sum x_i = 29.4 + 31.2 + 28.6 + 32.5 + 30.0 + 29.8 + 33.1 + 31.4 = 246.0\text{ MPa}$ $\bar{x} = \frac{246.0}{8} = 30.75\text{ MPa}$
2. Sum of Squared Deviations: $\sum x_i^2 = 29.4^2 + 31.2^2 + 28.6^2 + 32.5^2 + 30.0^2 + 29.8^2 + 33.1^2 + 31.4^2 = 7581.62\text{ MPa}^2$ $\sum (x_i - \bar{x})^2 = \sum x_i^2 - \frac{(\sum x_i)^2}{n} = 7581.62 - \frac{(246.0)^2}{8} = 7581.62 - 7564.50 = 17.12\text{ MPa}^2$
3. Sample Standard Deviation ($n - 1 = 7$): $s = \sqrt{\frac{17.12}{7}} = \sqrt{2.4457} = 1.56\text{ MPa}$
4. Coefficient of Variation: $CV = \frac{1.56}{30.75} \times 100\% = 5.09\%$ A standard deviation of $1.56\text{ MPa}$ is below $2.8\text{ MPa}$, which ACI 214R-11 rates as excellent control.
5. Target Strength Check ($f'_c = 28.0 \le 35\text{ MPa}$): $f'_{cr} = f'_c + 1.34 s = 28.0 + 1.34(1.56) = 28.0 + 2.10 = 30.10\text{ MPa}$ $f'_{cr} = f'_c + 2.33 s - 3.5 = 28.0 + 2.33(1.56) - 3.5 = 28.0 + 3.64 - 3.5 = 28.14\text{ MPa}$ The governing required average strength is $30.10\text{ MPa}$. The plant's mean of $30.75\text{ MPa}$ exceeds it, so the mixture meets the target.

Example 2: Hydrologic Flood Exceedance Probability

A cofferdam is designed to withstand a 20-year flood ($T = 20\text{ years}$). The bridge foundation construction period will last $4\text{ years}$. Determine the probability that the cofferdam will be overtopped at least once during the construction duration.

Solution:

1. Annual exceedance probability of a 20-year flood: $p = \frac{1}{T} = \frac{1}{20} = 0.05$
2. Non-exceedance probability in any single year: $q = 1 - p = 1 - 0.05 = 0.95$
3. Probability of zero overtopping occurrences in $n = 4$ independent years: $P(X = 0) = \binom{4}{0} (0.05)^0 (0.95)^4 = 1 \times 1 \times (0.95)^4 = 0.8145$
4. Probability of at least one overtopping occurrence ($X \ge 1$): $P(X \ge 1) = 1 - P(X = 0) = 1 - 0.8145 = 0.1855 \; (18.55\%$

---

Board Exam Traps & Common Errors