Initialising Monte Carlo engine...
Loading distribution samplers
RiskPulse V8 -- SOLAR EPC EDITION
MT19937 / LHS / Iman-ConoverEngine Configuration
--
Simulator
Risk Register Guide
Analysis Name
Iterations
Random Seed (MT19937)
PERT Lambda
Sampling Method
Correlation Mode
Convergence Diagnostics
5
RISKS LOADED
0
SESSION RUNS
Risks Input
Correlation Matrix
Results
S-Curve
Histogram
Tornado
Convergence
Engine Config
Risk Input Table -- Enter risks with Min/Likely/Max values in Crores (INR). Select distribution type per risk. Use the 50MW Rajasthan Demo button to load calibrated values from the 489-project dataset.
Excel Upload: Prepare columns: Name, Min, Likely, Max, Distribution (PERT/Triangular/Normal/Lognormal/Uniform/Discrete).
# Risk Name Distribution Min (Cr) Likely (Cr) Max (Cr) Prob EV (Cr)
Total Expected Value: -- Cr | Active Risks: --
Correlation Matrix (Iman-Conover / Gaussian Copula) -- Enter upper triangle values (-1 to +1). The matrix auto-mirrors. Use Validate to check positive semi-definiteness. Auto-Fix applies nearest-PSD correction via eigenvalue clipping.
Mean (Expected)
--
Probability-weighted average
P50 (Median)
--
50% confidence
P80 (Bid Contingency)
--
AACE 18R-97 standard
P90 (Mgmt Reserve)
--
90% confidence
VaR (P90)
--
Value-at-Risk threshold
CVaR (Expected Shortfall)
--
Average beyond P90
Tail Gap
--
CVaR - VaR (P90)
P90/P50 Ratio
--
1.5-2.0 = well-calibrated
Full Percentile Table
Sensitivity Analysis -- Spearman Rank Correlation
S-Curve (CDF) -- Cumulative Probability Distribution
Run simulation to generate
How to Read: Pick any value on the X-axis. Follow it up to the curve. Read left -- that is the probability your total risk lands BELOW that value. P80 = 80% of futures stay below this number. The gap between your contingency and P80 is the price of optimism.
Histogram -- Frequency Distribution
10,000 simulated futures by outcome frequency
Colour Zones: Green = below P50 (comfortable). Blue = P50-P80 (budgetable). Orange = P80-P90 (contingency zone). Red = above P90 (tail risk). The tallest bar = most common outcome. The flat tail on the right = rare but expensive.
Tornado Chart -- Variance Decomposition
Risk contribution to total model variance
How to Use: The widest bar = the risk that moves the total outcome the most. Top 3 risks typically own 60-80% of all model variance. That is where you spend mitigation money. The tornado shows symptoms; the correlation matrix shows causes.
Convergence Plot -- P90 Stability
P90 vs iteration count (bootstrap sub-sampling)
Convergence Rules: If the line flattens, the number is stable -- present it. If still wobbling, double iterations. Drift threshold: less than 1% between 80% and 100% of iterations = converged. A P90 that has not converged is a random number, not a percentile.
Benford's Law Validation
Leading digit distribution vs theoretical Benford curve
Engine Configuration -- Audit Trail
Analysis Name--
Iterations--
Sampling Method--
Seed (MT19937)--
PERT Lambda--
Correlation Mode--
Risks Loaded / Fired--
Execution Time--
Iterations/Second--
Convergence--
Timestamp--
Config Hash--

Solar EPC Risk Register Guide

Calibrated from 489 projects | 87,245 MW | 14,723 risk events | 2015-2026

489
Projects Analyzed
87,245 MW
Total Capacity
14,872 Cr
Total Cost Overruns
14,723
Risk Occurrences
Project Parameters
Top Risk Register -- Empirical Data from 489 Solar EPC Projects
Rank Risk Name Category Frequency Avg Impact (Cr/MW) Total Impact (Cr) Probability Severity
Historical Case Studies

Scaled Monte Carlo Preview

--
P50 (Median)
--
P80 (Bid Contingency)
--
P90 (Mgmt Reserve)
--
P95 (Extreme)
Quick 5,000-iteration preview. For full analysis with correlation, convergence diagnostics, and export, switch to the Simulator tab.

Read the Dashboard or Get Read by It (Chapter 7)

A full Monte Carlo dashboard produces a dozen tabs and fifty metrics. You need three for every meeting and three more for audit.

For the Meeting

1. KPI Cards -- the headline numbers (P50, P80, P90, CVaR). 2. S-Curve -- the probability map. 3. Tornado Chart -- where the risk lives and where to spend mitigation money.

For the Audit

4. Convergence Plot -- is the number stable? 5. Sensitivity Tab -- which inputs drive which outputs. 6. Engine Config -- seed, iterations, method. The audit trail.

The 60-Second Read: Step 1: Look at P80 -- this is your anchor number. Step 2: Compare P80 to your allocated contingency. If contingency is less than P80, you are underbudgeted at 80% confidence. Step 3: Check P90/P50 ratio. Between 1.5 and 2.0 = well-calibrated model. Above 2.5 = extreme tail risk. Below 1.3 = distributions too tight.

The 3-Chart Meeting Stack: S-Curve first (establishes the gap), Tornado second (names the cause), Options table third (presents the choices). If you reverse this order, you are asking the director to solve a problem they do not yet believe exists.

Get in the Pool (Chapter 8)

Stop reading about swimming. Build your first Monte Carlo model in 30 minutes. The project: 50 MW ground-mounted solar PV, Rajasthan, SECI tender, aggressive tariff. Base project cost: 250 Crore. Contingency: 25 Crore (10%).

Step 1: Enter 5 Risks

R1 Land acquisition (2/8/35 Cr). R2 Module price escalation (0/5/22 Cr). R3 Transmission bay delay (1/4/18 Cr). R4 Labour shortage (0.5/3/12 Cr). R5 LD exposure (0/6/28 Cr). These are statistical facts from 489 projects.

Step 2: Set Correlations

R1-R5 (land-LDs): 0.78 -- structural dependency. R2-R4 (module-labour): 0.10 -- nearly independent. An approximate correlation is infinitely better than zero.

Step 3: Press Run

10,000 iterations. 3 seconds. The S-curve lands. P50 is roughly 35 Cr, P80 is roughly 52 Cr, P90 is roughly 70 Cr. Your 25 Cr contingency sits below P50. The gap: 27 Crore.

Step 4: Read Your Tornado

R5 (LDs) is the longest bar but it is the most dependent risk. It correlates at 0.78 with R1, 0.72 with R3. Mitigate R1 (land) and R3 (grid), and R5 shrinks automatically.

The Three Mistakes You Will Make: 1) Setting Min=0 on every risk (creates extreme right tail). 2) Ignoring correlation (zero means independent -- your risks are not independent). 3) Running once and calling it truth (check convergence -- if P80 moves more than 5% between runs, increase iterations).

Recommended Action Plan
PriorityRiskMitigation StrategyExpected ReductionTimeline
1Land Acquisition DelaySECI pre-identified land parcels. Direct cash at 2x market rate. Parallel Revenue-Forest-Agriculture NOC processing. Dedicated community liaison.67% impact reductionStart 24 months before NTP
2Module Supply ChainMulti-vendor sourcing with BCD-exempt warehousing. Hedged polysilicon contracts. 20% buffer stock at port.45% impact reduction12 months before procurement
3Transmission Bay DelayParallel PGCIL bay allocation tracking. Pre-fund bay extension works. Alternative evacuation routing study.40% impact reduction18 months before COD
4Labour ShortageMulti-season mobilisation calendar. Regional contractor pool. Mechanised installation for pile driving.35% impact reduction6 months before mobilisation
5LD ExposureDownstream risk -- mitigating R1 and R3 reduces R5 automatically. PPA extension clause negotiation.Indirect via R1/R3Continuous
Ready --