Getting Your First Design Right on the First Try
Prestressed concrete works by putting the material under compression before it ever sees a real load. Concrete is strong in compression and weak in tension, so you force the tendon network to carry the tensile demand while the concrete takes the compressive side. That's the whole idea, but the execution is where most people trip up. I've seen engineers skip straight to the finite element model and waste two days chasing results that looked right but were built on wrong assumptions. Run the hand calcs first. A simple rectangular beam under uniform load: the required prestress force is roughly P equals M divided by e, where e is the eccentricity of the tendon from the centroid. If your eccentricity is 150 millimeters and your midspan moment is 450 kilonewton-meters, you need around 3000 kilonewtons of effective prestress. Check that against the allowable stress limits at transfer and at service. If the concrete stress at the bottom fiber during transfer exceeds 0.6 times the cylinder strength, you've already designed yourself into a corner. The loss estimation is where things get annoying. Short-term losses at transfer include elastic shortening, anchorage set, and friction. Long-term losses cover creep, shrinkage, and relaxation of the strands. American codes typically assume around 10 to 12 percent total loss for post-tensioned members, but that number is an average. When I was working on a bridge deck with curved post-tensioning ducts, the friction losses alone ate up 8 percent before I even accounted for anything else. I had to redo the jacking force calculations three times because the wobble coefficient and curvature coefficient from the duct manufacturer didn't match the default values in my software.
Here's a specific problem that cost me about a week. I was designing a double-tee floor system for a parking structure. The initial analysis showed adequate deflection under service loads, but after applying the time-dependent loss factors from the code, the camber at transfer was 35 millimeters upward and the deflection under full dead load plus 60 percent of live load came out to 18 millimeters downward. The total deflection was within limits on paper, but when I actually looked at the camber schedule, the tees were so pre-cambered that they wouldn't sit flat on the bearing seats. The workaround was to specify a lower initial prestress force and make up the difference with a secondary gravity load pattern that shifted the moment demand. It meant recalculating everything, but the tees finally behaved during installation. Segmental construction changes the game entirely. The tendon layout isn't just about flexure. Shear key performance, diaphragm action, and the continuity moments from negative regions over supports all compete for the same tendon space. You can't just stack tendons vertically and hope for the best. In practice, I usually separate the tendons into three families: the bottom tendons for positive moment resistance, the top tendons for negative moment at continuous supports, and the web tendons for shear. Each family has its own stressing sequence, and getting the order wrong can crack the web at a support before the member is even fully stressed. A counter-intuitive thing about prestressed concrete is that more prestress isn't always better. Over-prestressing a member at transfer can cause crushing at the anchorage zones, especially in regions with high stress concentrations around the deviators. The code allows higher compressive stresses at service because creep and shrinkage will redistribute them over time, but at transfer those stresses are fully active. I once saw a design where the bottom fiber stress at transfer hit 11 megapascals in a 50-megapascal concrete. The designer had optimized for span length without checking the transfer limit. The beam developed hairline cracks along the bottom within hours of release.
Another thing beginners miss is the difference between the transfer condition and the stressing condition. Transfer is when the tendon force is first released from the stressing jack into the concrete. Stressing is the initial application of force. Between those two moments, you lose force to elastic shortening of the concrete. For a member with multiple tendons stressed sequentially, each subsequent tendon causes additional elastic shortening in the already-stressed ones. The code gives you a simplified approach, but it's not exact. I use a step-by-step calculation where I track the shortening strain in the concrete at each stressing stage and adjust the force in the previously stressed tendons accordingly. It adds about twenty minutes to the calculation but catches errors that would otherwise show up as unexpected deflections later. Allowable stress limits are another area where people get sloppy. At transfer, the compressive stress limit is typically 0.60 f'ci and the tensile stress limit is around minus 7 point 5 root f'ci for pretensioned members. At service, the compressive limit drops to 0.45 f'c and the tensile limit becomes much smaller. The tensile limit at service is often the governing constraint, not the compression limit. If your member is in a corrosive environment, you may need to reduce the allowable tensile stress even further or provide additional cover. I learned this the hard way on a coastal structure where the design passed all the stress checks but failed durability inspection after five years because the cover over the tendons was insufficient for the chloride exposure. Shear design in prestressed concrete is fundamentally different from reinforced concrete. You don't need as much stirrup reinforcement because the axial compression from the prestress reduces the principal tensile stresses in the web. The code provides a modified shear stress formula that accounts for the prestress contribution. But here's the catch: the prestress contribution decreases as the member deflects. Under heavy load, the tendon line of action changes, and the vertical component of the prestress force that was helping with shear resistance diminishes. I always run a second shear check at the deflected configuration, not just the undeformed one. It adds negligible time and has prevented me from specifying inadequate stirrup spacing on a few occasions.
Get the Full Details
If you're doing this professionally, you'll need software. I use structural analysis packages for the global model and dedicated prestress design tools for the detailed checks. The big risk is that the software will give you an answer and you'll assume it's correct. It might be using different loss assumptions, different load combinations, or a different code edition than your project requires. I always spot-check a simple span with the software output against my hand calculations. If they agree within 5 percent, I trust the model. If they don't, I dig into the assumptions until I find the mismatch. Detailing is where theory meets reality. Development length for prestressing strands is different from rebar development. Strands develop through bond, but the bond characteristics are different because of the spiral deformations on the strand surface. The code gives you an explicit development length formula, but for high-strength strands in high-stress regions, you sometimes need additional straight length or mechanical anchorage. I've seen designers neglect this on short-span beams where the available embedment length was marginal. The beam worked fine under normal loading, but under overload conditions the strands slipped at the ends because the development length wasn't sufficient. Fire resistance is another overlooked area. Prestressed concrete members lose strength at elevated temperatures because the steel relaxes. The tendons are usually covered by concrete, but in thin members like double-tees or hollow-core slabs, the cover may not provide adequate protection for the fire rating required by code. I've had to add supplementary reinforcement or increase the member depth specifically for fire resistance on several projects. It's not always obvious from the structural analysis, so don't skip the fire check.
The economic sweet spot for prestressed concrete is typically spans between 6 and 20 meters for floor systems and 10 to 40 meters for bridges. Beyond that, you start competing with structural steel or steel-concrete composite sections. Below 6 meters, conventionally reinforced concrete is usually cheaper because the prestressing hardware and labor costs don't scale down linearly. I've seen firms specify prestressed concrete for short spans out of habit, and the unit cost was 30 to 40 percent higher than reinforced concrete for the same performance. For learning resources, the PCI Design Handbook is the standard reference. It covers both pretensioned and post-tensioned design with worked examples. ACI 318 has a dedicated chapter on prestressed concrete that's more concise but legally binding for code compliance. For advanced topics, the post-tensioning institute publishes technical bulletins that go beyond the code requirements. The Prestressed Concrete Analysis And Design Fundamentals aren't complicated, but they require attention to detail in every step from load estimation to final detailing.
